L'anguille européenne, indicateurs d'abondance et de ... - ifremer

L’anguille européenne (Anguilla anguilla) est une espèce menacée.
Pendant trois ans, les principaux acteurs concernés par cette situation
– pêcheurs, chercheurs, institutionnels et financiers – ont mis en commun
leurs expériences, leurs savoirs institutionnels et vernaculaires et les ont
adaptés au contexte du réseau de bassins versants. Cet ouvrage fait
la synthèse du projet lancé par l’Ifremer et intitulé « InterregIII Espace
Atlantique Indicang » .
Ce guide méthodologique précise la biologie générale de l’espèce et met
en lumière des caractéristiques physiologiques, comportementales,
populationnelles et écologiques qui aideront les gestionnaires à mieux
évaluer son abondance, la nature des modifications que son
environnement naturel a subi ces dernières décennies et les pressions
exercées par l’homme sur un bassin versant déterminé. Il définit
des descripteurs, puis des indicateurs pour chaque stade du cycle
biologique : civelle, anguille jaune et argentée. La partie concernant
l’environnement est fortement développée compte-tenu de l’importance
de la qualité des milieux dans la reconstitution des stocks d’anguille.
Dès 2009, la réglementation européenne sur l’anguille prévoit de mettre
en œuvre des plans de gestion et de restauration de cette espèce.
Ce guide permettra de disposer d’une base pratique et théorique pour
installer et, au besoin, faire évoluer les indicateurs nécessaires à l’évaluation
de l’efficacité de ces plans.
Il intéressera un public averti, les naturalistes et les gestionnaires soucieux
de la préservation de cette espèce. En rappelant les principales notions
sur la dynamique des populations, il sera également un support didactique
pour l’enseignement universitaire.
Les coordinateurs sont des spécialistes reconnus de l’anguille européenne.
Patrick Prouzet, coordinateur du projet Indicang est directeur de programme
à l’Ifremer. Gilles Adam, animateur du comité de communication du projet,
est hydrobiologiste et chargé de mission à la direction régionale de l’environnement
Aquitaine. Éric Feunteun, responsable du comité scientifique et technique du projet,
est professeur d’ichtyologie et d’écologie marine au Muséum national d’histoire
naturelle. Christian Rigaud, co-animateur de la thématique « anguille jaune » du projet
est ingénieur de recherche au Cemagref et animateur du groupement d’intérêt
scientifique sur les espèces amphihalines (Grisam).
faire
Savoir
L'anguille européenne
Savoir
faire
L’anguille
européenne
Indicateurs d’abondance et de colonisation
G. Adam, É. Feunteun, P. Prouzet et C. Rigaud,
coord.
En couverture : L’arbre à anguilles (dessin S. Gros, Ifremer) ; civelle (photo G. Choubert, Inra) ;
anguille jaune (photo G. Adam, Diren Aquitaine) ; anguille argentée (photo H. Farrugio, Ifremer).
Prix : 55 €
Éditions Cemagref, Cirad, Ifremer, Inra
www.quae.com
ISBN 978-2-7592-0085-6
-:HSMHPJ=WUU]Z[:
ISSN : 1952-1251
Réf. : 02091

General introduction
1

Background to the general objectives of the InterregIIIB “Atlantic area”
INDICANG project
Initiated within the framework of the INTERREGIIIB - Atlantic area programme, the INDICANG
project federated 40 partners from 4 Atlantic Arc countries 1 . The objective of the project was to
develop and disseminate knowledge concerning the exploitation, habitat and evolution of the
European eel in order to restore eel stocks which are currently at risk.
The river basin is the relevant scale for the management of the European eel. This scale allows
eel production to be optimized by limiting the constraints related to various anthropogenic factors (one
of which is fishing). Furthermore, it allows commercial fishermen to be fully involved in the project: by
integrating their observations (the commercial fishermen is, in this context, a practitioner who also
undertakes environmental monitoring) and by involving them in the comparison of the results obtained
from scientific and technical monitoring with those from the monitoring of exploitation indicators (total
catches, fishing effort, catch per unit of effort, climatic variability….). The scale of the river basin also
makes it possible to undertake a systemic type of analysis.
Eel population abundance and habitat quality were therefore assessed in 13 river basins (figure
1). The principal disturbance factors were identified.
From these observations and their comparison, the project partners have defined relevant and
inexpensive indicators relating to the quality of the environment, the abundance of elvers migrating up
the estuaries, the intensity of upstream colonisation by young eels and the abundance of silver eels
migrating towards the Sargasso sea.
The project does not fulfil explicitly the request from the European Union concerning the state of
the population and the level of escapement in relation to the defined reference target i.e. “40% of
silver eel biomass produced in a pristine environment” because those involved in the project do not
currently have the elements to provide an accurate answer.
The project partners emphasised the need to implement tools and methodologies that would
enable managers to assess and compare the efficacy of species restoration plans which have to be
defined and implemented from the 1st of January 2009 2 .
1
EU Community Programme > 2000 – 2006, annex 1 of the Indicang Report,
http://www.ifremer.fr/indicang.
2
Article (CE) No 1100/2007 of the Council of 18 September 2007, Official Journal of the European Union, annex 2 of the
Indicang report, http://ifremer.fr/indicang
2

Tamar, Camel
and Slapton Ley
La Vilaine
La Loire et la
Sèvre Niortaise
Minho
L’Adour
La Gironde, Garonne
Dordogne
Nalon
Esva
L’Oria
Figure 1 -
Geographical location of river basins covered by the INDICANG project.
Some preliminary definitions to clarify the objectives of the handbook
The notions of descriptor and indicator are defined in the newsletter number 2 of the
INDICANG project 3 . Once an object is described according to various criteria, its state can then be
assessed using these two elements in comparison with various reference frameworks.
• Indicator: Analytical information summarizing the state of a system or its evolution in relation to
specific objectives. This second aspect is most notably used in ISO standard 8042.
• Descriptor: Qualitative or quantitative element, observed, measured or calculated, used to
describe an object, an individual or a system. For example, the taxonomy of the eel entity uses the
following descriptors: the class (Osteichthyes); the order (Anguilliforms); the family (Anguillidae);
the genus (Anguilla); the species (anguilla) and a vernacular name (European eel).
• Criteria: Element or information allowing a judgment or a choice to be made concerning an
object, an individual or a system. It implies the existence of a norm against which to evaluate and
judge.
3
Indicang newsletter, 2/2006, 3, annex 3 of the Indicang report, http://www.ifremer.fr/indicang.
3

A good indicator must:
be relevant and reliable
be sensitive
be synoptic
be shared and interpretable
be available over a significant period
of time
validated scientifically and statistically
reveals the evolution being monitored
summarises complex phenomena, it generally
results from the combination of several
descriptors.
including by non-specialists where it aids
management
takes into account the notion of technical and
economic feasibility
It is useful to collate these different indicators and their associated reference frameworks into
tools, usually called “Management charts”, which offer a synoptic view of the system under study as a
function of a certain number of evaluation criteria 4 .
The project partners considered that there was a minimal management chart as opposed to
an optimal management chart.
At a workshop held in Porto in 2006, it was agreed that a minimal management chart must
include indicators concerning the importance of available production areas taking into account the
difficulty of access to yellow and silver eel production sites and indicators of 2 types of anthropogenic
mortality: fishing and hydroelectric production. Anything less than this means that the eel production
capacity of the hydrographic system cannot be evaluated.
The optimal management chart is defined by the inclusion of all the indicators described in this
handbook. However, an improved management chart can be established by adding to the minimal
management chart some descriptors concerning the chemical quality of the aquatic environment and
the health status of the individuals observed or caught 5 .
Finally, the main technical terms used in this handbook and more broadly within the INDICANG
project framework are defined in the project glossary 6 .
4
For an illustration, please refer to the web site of the > http://www.anguille-loire.com.
5 Girard P., Elie P., 2007. Manuel d’identification des principales lésions anatomo-morphologiques et des principaux parasites
externes des anguilles, Cemagref, report 110, annex 4 of the Indicang report, http://www.ifremer.fr/indicang.
6
Collective, 2007. Glossaire Indicang – Langue francaise, annex 5 of the Indicang report, http://www.ifremer.fr/indicang.
4

Handbook contents and instructions
The biological objective of the European regulation relating to eel management 7 is to improve
very significantly the global flux of potential spawners leaving their growth area to return to their
reproductive area. This global flux includes the production from all river basins and coastal zones
colonised by the species. First it must be noted that the Indicang programme only took into account
inland areas (from the salt-water estuary to the upper reaches of rivers) and consequently so does this
handbook. This increased flux, which will inevitably be progressive, requires in particular a marked
increase in the survival rate from the glass eel stage, in all lower reaches of river basins, to the silver
eel stage over the complete distribution area of the species.
Starting from a very poor situation, monitoring via relative indices could show abundance
returning to 1970s-1980s levels, at least in a first intermediate phase. These types of monitoring could
also concern anthropogenic pressures exerted more or less directly on the species. Quantified
objectives of flux, stock and/or survival are also proposed based on the correlation between the
degree of colonisation at the entrance to a river basin (and more particularly to the estuary) and the
minimal resulting yield of spawners a few years later. Of course, these two complementary
approaches correspond to different estimation methods including in terms of cost and time. Within this
handbook, the review of currently available methodologies specifies the type of approach relating to
different methods.
Within a river basin, total spawner production includes that of tidal compartments (accessible
without having to pass any barriers and without having to swim against the current) and that of
upstream non-tidal compartments. Here also, the methodological review takes into account the
existence of these compartments and the impact of their characteristics on the monitoring procedures
that have to be implemented.
Finally, within the framework of a local eel management plan at the scale of the river basin,
two supplementary analyses are fully justified in these compartments:
• Identify the species local characteristics - mainly the distribution and the abundance levels by
sex, age or size class; but also the health condition and the reproductive quality of produced
spawners;
• Identify human pressures, their localisation, their intensity and the magnitude of their impact on
global survival between glass eels entering the system and silver eels returning to the sea and on
the quality of individuals which are to migrate to the Sargasso sea;
For each of these approaches, the actions undertaken aim to collect the elements required to:
• assess the initial local situation, in the form of indices or absolute quantification;
7 Article (CE) No 1100/2007 of the Council of 18 September 2007, Official Journal of the European Union, annex 2 of the
Indicang report, http://ifremer.fr/indicang
5

• judge the state of the species and/or its habitat by comparing findings with a reference situation;
• set an objective to be met: a biological objective concerning the density of the species or a
management objective concerned with controlling anthropogenic mortalities.
• monitor on a regular basis the impact of the implemented management measures, taking into
account that, due to eel dynamics, several decades will be required for significant recovery.
This general background having been set out, what of the handbook’s contents and
instructions?
First, in Part I 8 >, the reader will find a synopsis of the
knowledge concerning eel biology and the course of its inland growth phase together with a
refresher on sampling basics, an essential part of data collection in any kind of monitoring.
The reader will then find in Parts II 9 > and III 10 >, a review of existing methods for each of the
analyses mentioned above. This review was undertaken using available publications, reports or
notes. In some cases, the lack of reliable methods had to be acknowledged.
Tables 1 and 2 give the reader an overview of possible approaches and the methods currently
available to identify and then monitor:
• quantitative and qualitative eel characteristics in the river basin by broad type of compartment;
• the nature and/or impact of human pressures observed in the basin.
Each method identified is linked to a chapter reference in the handbook, where the reader can
find detailed information about it or the identity of the teams currently designing or finalizing it.
So far as possible, each method must address the following questions:
• What are the objectives? Quantify, estimate the pressures exerted on the species, highlight
evolution trends, etc.;
• What are the indicators required to address these questions?
• What are the descriptors required to build these indicators? What are the data required for these
descriptors?
• What are the methods used to calculate these descriptors and the protocols to collect these data?
What problems may be met in the implementation of these methods and these protocols?
• Once the indicators have been built, how to interpret them?
Finally, almost all of the methods mentioned and their implementation are illustrated with
examples taken from the work carried out in the river basins involved in the Indicang programme.
The currently limited nature of acquired knowledge and validated methods will become clearly
apparent when reading tables 1 and 2 and this handbook. This is hardly surprising given the extremely
8
See Chapters 1, 2 and 3.
9
See Chapters 4,5 and 6.
10
See Chapters 7,8 and 9.
6

apid development in the status of this species. Hence, in France, for example, the status of eels
evolved, over a 20-year period, from that of a pest species in salmonid rivers to that of a species that
should be under controlled management (1984 French law on fishing). For the moment, this
development has not been followed by a substantial effort to establish methodologies and acquire
knowledge comparable to that observed for salmonids although over the past 5 years, there has been
some progress in France and in other European countries. The review and the comparison of
methods undertaken within the Indicang framework have highlighted this situation and the effort
required if the quantified objectives established by the European regulation are to be achieved.
However, this work does mean that henceforth it is possible to harmonize descriptive
procedures of the state of the species and the pressures to which it is subjected in the basins. This is
a vital stage in sharing the diagnosis between all relevant stakeholders and implementing diversified
and coordinated management measures, which are well understood by all stakeholders.
Diagnosis and monitoring of the local characteristics of the species
Monitoring species abundance
The analysis of table 1 shows that when monitoring the species’ abundance, the great
difficulty resides in quantifying the phenomena at the scale of the river basin (total glass or silver eel
flux, existing stock).
As regards the existing stock, it should be noted that no deep environment (estuary or river)
has yet been estimated in any way. In shallow environments, electrofishing estimates are of course
possible, but the variability of the collected data in relation to the prospected habitat makes any
extrapolation to the whole hydrographic network extremely risky.
As regards the fluxes of glass eels and silver eels observed respectively at the entry and exit
of an axis or a basin, some methods do exist, mostly associated with a particular context (the
existence of a fishery, a trap, etc.). Although relatively demanding in terms of monitoring time and
specific equipment, they provide interesting signals at exploited sites. But these are mostly located
upstream of the estuarine zone, and even in the fluvial zone. Currently, the fluxes of glass and silver
eels have not been quantified at an estuary mouth, which means that no interpretation is possible at
the scale of the whole basin. It seems that the limiting factors relate to the significant, even colossal,
resources required rather than to methodological issues.
However, in the great majority of cases, it is possible to monitor the relative evolution of the
species status through abundance indices linked to fisheries monitoring, passability device
monitoring or permanent electrofishing networks. More information can be drawn from the analysis of
these data if the interpretation is in terms of size class and if the environmental context of the
sampling sites is adequately analysed. By comparing current indices to historical data, the extent of
the restoration required can be measured.
7

Table 1 -
Synopsis of the methods presented in the methodological handbook concerning the observation of eel life stages by river basin
compartments
Tidal
compartments
Upstream non-tidal
compartments
Individuals under observation
and processes being analysed
Type of
approach
Methods
Handbook chapter
and any Obs.
Methods
Handbook chapter
and any Obs.
GLASS EELS
Abundance
indices
CPUE of different fisheries (push
net, “pibalour” type push-net, etc.)
Chapter 6
Historical references
and
total basin recruitment observed in
the estuary
Absolute
quantification
Flux quantification in the estuary
Mark-recapture
Chapter 7
Difficult on very large
and stratified estuaries
YELLOW EELS
30 cm long
and
migratory potential
Abundance
indices
Absolute
quantification
CPUE passive fishing gears Chapter 6
No method available and/or
significant tools to be developed
Deep environments
(> 1,50 m)
Shallow environments
Deep environments
(> 1,50 m)
Shallow environments
CPUE passive fishing
gears
Analysis on the
permanent
hydrobiological and
piscicultural network
(RHP: réseau
hydrobiologique and
piscicole)
No method available
and/or significant tools
to be developed
Electrofishing but great
variability according to
the prospected habitat
Chapter 6 and 8
Chapter 8
Chapter 8 and 9
DOWNSTREAM MIGRATORY SILVER EELS
and
effective escapement
Abundance
indices
Absolute
quantification
CPUE but no specific fisheries in
tidal zones
No method available and/or
significant tools to be developed
Chapter 6
CPUE of silver eel fisheries (Historical
References)
Mark-recapture along the downstream migration
axis
Trapping on constructions at the exit of the axis or
the river basin
Chapter9
Raised water level
phases and
temporal variability
difficult to integrate

Monitoring the quality of produced spawners
As regards monitoring the quality of spawners produced by the basin compartments, two
monitoring angles are recommended:
• Sex ratio evaluation 11 (analysis by size class, sexing during sacrifices for contamination
analyses);
• Health condition 12 from external observation of pathological signs (recognition handbook and
data entry sheet) during field monitoring (trap, fisheries, specific monitoring) and internal analysis
(pathologies, level of chemical contamination); optimizing the sacrifice of individuals to collect
further information (sex, age, …).
Characterising human pressures and evaluating the magnitude of their impact
In a given compartment, eels are subjected to a range of more or less significant human
pressures, which affect their survival and/or quality.
In this approach, the following must be distinguished:
• Characterizing these pressures by compartment within a river basin. The initial observations
can be repeated at intervals in order to monitor the evolution of the context in which the eels
develop, particularly after management decisions;
• Evaluating the impact of these pressures in terms of species survival or distribution. This
evaluation, carried out on one axis, one compartment or even one site should ultimately of course
be considered in terms of the whole river basin.
The great majority of the methods currently available make it possible to standardize the
characterisation of human pressures in a river basin or in one of its compartments and to produce an
initial ranking. This stage is important as it allows basin zones to be characterized according to the
major pressures to which they are subjected. This analysis can contribute significantly to the choice of
priority actions in each of these zones.
On the other hand, evaluating the real impact of each pressure at the appropriate scale
(construction, compartment or axis) is rarely well-mastered compromising accurate evaluation at the
river basin level.
Some methods are being developed and would benefit from the creation of referentials
combining the context (environment + pressures) with the observation of the state of the species
(abundance index, size structure, etc.). These referentials would enable the relationships between
local context and local state of the population to be defined and could thereby lead to impact
estimations based solely on a description of the magnitude of the pressure observed (table 2).
11 See Chapter 9.
12
See Chapter 5.
9

Table 2 –
Synopsis of the methods presented in the methodological handbook concerning the inventory, the characterization and the
evaluation of pressures and their impacts on eels according to river basin compartments
Types of pressure Type of approach Methods
Tidal
compartments
Handbook chapter
and any Obs.
Methods
Upstream non-tidal
compartments
Handbook chapter
and any Obs.
Barriers to river
basin colonisation
Inventory and
characterization
Impact evaluation
Not applicable
Not applicable
Identification, description and
evaluation of barrier passability
Impact on distribution at axis level
Chapter 4 Evaluation by axis within the basin in
order to determine intervention priorities
English method being developed (Reference
Condition Model).
Impact on survival at axis level
Calculation in % SPR - being designed (Grisam)
Fisheries
Water abstraction
Inventory and
characterization
Impact evaluation
Inventory and
characterization
Impact evaluation
Inventory of different fisheries Chapter 6 Inventory of different fisheries Chapter 6
Glass eel survival
Yellow eel survival
Identification and
characterization of water
abstraction
Impact on glass eel survival Chapter 4
Impact on yellow eels
No data
Impact on silver eels
No data
Chapter 7
Fisheries monitoring and flux
quantification
Evaluation by filtered volume/total
volume (Glass Eel Model to Assess
Compliance – being designed –
Grisam)
Evaluation by size structure analysis
(Eel Length Structure Analysis - being
designed - Grisam)
Chapter 4
Yellow eel survival
Evaluation by size structure analysis (Eel Length
Structure Analysis - being designed - Grisam)
Silver eel survival Chapter 9
Fisheries monitoring and flux quantification
Identification and characterization
of water abstraction
Chapter 4
No estimation carried out currently
Turbines on
downstream
migration axes
Inventory and
characterization
Identification and characterization
of turbines
Impact evaluation Impact on downstream migratory
silver eel survival
Chapter 4 Evaluation by axis within the basin
Chapter 4 Analysis by construction
10

Scale of the study: the river basin
The entry of young individuals into a river basin and their “restitution” a few years later as
potential spawners, leads to a conception of the hydrosystem as a unit producing a fraction of the adult
stock returning to the Sargasso sea in order to ensure the long term survival of the species (Dekker,
2000). This global approach by river basin is a relatively new conceptualisation. It was totally absent
from the two synopses of Deelder (1970) and Tesch (1977). More recently, biological and fisheries data
have been progressively collected and analysed at this scale of observation (Elie and Rigaud, 1984;
Moriarty, 1987; Legault, 1987; Vollestad and Jonsson, 1988; Barak and Masson, 1992; Naismith and
Knights,1993; Smogor et al., 1995; Castelnaud, 2000; Feunteun et al., 2000; etc.).
Each basin presents a diversity of aquatic compartments, each with a functionality and a
particular influence on the species’ dynamics, resulting from both its particular local characteristics (or
habitat quality: shelter density, trophic level, etc.) and its position within the system (or habitat
accessibility: distance from the tidal limit, slope, degree of colonization, etc.). Within all these habitats,
the species’ distribution seems to depend on a large number of nested factors (Laffaille et al., 2004).
Only the inland area is considered here although the coastal maritime zone should be integrated
where management is concerned.
Having said this, for practical reasons, a structured division into compartments and/or
biological stages is often unavoidable, particularly in large hydrosystems. For example, the estuarine
section (affected by the dynamic tide) and the fluvial section (upstream of the dynamic tide limit) are
generally defined by environmental characteristics and administrative contexts (fishing or navigation
regulations) that can differ.
Figure 2 summarises the division of a French basin and specifies the main biogeographical,
hydrodynamic, administrative and regulatory limits of the fluvio-estuarine system.
11

Limite
(transversale)
Limite de
Dynamic Limite (de (ascending) remont ée tide ) limit =
(Transversal) limit
der
la mer =
salure Sea water des eaux limit
= LSE (old) de mar boundary é e dynamique between = maritime
of the sea
LMD
& fluvial navigation
=(ancienne) limite
Mar Tide é only
de l ’ inscription maritime
Mar Dynamic é e dynamique tide and (descending) et courant (descendant) river current
fluvial
(descending)
courant (descendant)
river fluvial current seulement only
Salt l Eau water
sal é e Brackish m Eau water
âe Fresh Eau douce water
Public Domaine maritime
public
maritime domain
Domaine Private and public public fluvial fluvial et domaine priv é fluvial
Sea Mer including
dont
Eaux Inland int waters
érieures (fleuves (rivers avec with avec estuaire an d estuary et affluents, et and tributaries, plans plans d water d bodies, ’ eau, nes) ’ eau, lagunes) lagoons
river embouchure mouth and
et c ô coast
te
P ê che Fishing (sous under r é glementation) maritime regulation
maritime
P êche Fishing (sous under r églementation) fluvial regulation
fluviale
Sea (salt Mer water).
Tidal. (eau) Maritime
sal é e
regulation
Mar é e
r é glementation
maritime
Brackish Partie du water fleuvepart of
(eau) the saum river
âtre
à mar ée
r é glementation
maritime
Fresh Partie water du fleuve part of
the (eau) river. douce Tidal.
Fluvial regulation
à mar ée
réglementation
rétation
fluviale
Fresh Partie water du fleuve part of
(eau) the douce river.
sans
Non-tidal.
mar ée
Fluvial regulation
r é glementation
fluviale
Sea
mer
Estuary
Estuaire (salt water)
(Estuaire sal é)
zone
Mixed mixte
fluvial fluviale
zone
(fresh water)
(Estuaire doux)
e
Strictly fluvial zone
Tidal Zone area à à mar r = Fluvio-estuarine é é e e= = Syst ème system fluvio = -estuarien Lower reach = Partie partie of river
basse du fleuve
Figure 2 -
Fishing-related administrative and regulatory divisions on a French fluvio-estuarine
system (modified by Castelnaud et al., 2006).
Figure 3 shows two maps, which allow a sufficiently accurate description of the geographical and
administrative framework of a French estuary (example of the Adour estuary).
12

The Adour catchment
Location of the catchment
‘Commune’ boundary‘
‘Cantons‘
‘Cantons‘
Adour catchment
‘Commune’ boundary
‘Canton’ boundary
‘Department’ boundary
Adour catchment
‘Commune’ boundary
‘Canton’ boundary
‘Department’ boundary
3a. General characterisation of the Adour catchment and estuary: a) geographic and hydrographic
context of the Adour catchment; b) administrative and hydrographic context of the Adour
13

Atlantic Ocean
Transversal
limit of the
sea
kilometers
Administrative division of the estuary
Maritime zone (22km) under the
control of Maritime Affairs
Mixed zone (21.9km) under the
Agriculture and Forestry Department
Directorate (DDAF)
Fluvial zone, under DDAF control
Boundary between each zone
Adour catchment
Mean high water mark
Hydrology
‘Communes’ of the Pyrenees
Atlantiques
‘Communes’ of the Landes
3b. Administrative and hydrodynamic context of the Adour estuary
Figure 3 -
General features of the Adour basin and its estuary.
14

Having established this general context, the different environmental parameters affecting the
various life stages of the eel must be identified and characterised 13 . The river basin characteristics
(upstream of the estuary) where eels grow from their estuarine recruitment until their silver
metamorphosis constitute the basin context. The determinant influence of the environment on many
population characteristics, for example the sex-ratio 14 means that it is of course the decisive factor
determining the population structure of glass, yellow, and silver eels produced in the basin and hence
the reproductive potential of the basin.
In this book, we have chosen a division by biological and ecological stage (estuarine
recruitment, fluvial recruitment, sedentary phase and downstream migrations) and by environmental
compartment of the basin (habitat quality and quantity, fisheries monitoring). But these two divisions are
often linked.
Intermediate indicators are therefore useful to help understand the consequences of the actions
taken with respect to each theme and basin compartment and to guide future efforts. These
intermediate elements must contribute to assessing the operational quality of the river basin as a silver
eel production line. Their comparative analysis is always required. For example, finding an
increasingly abundant stock in a basin with 5 to 10% of yellow eels silvering each year is of little interest
if they are going to be affected by significant mortality as they return to the sea.
For practical reasons, measures must be tailored and implemented for each large river basin or
each homogeneous group of small river basins, because these constitute coherent operational units for
the population fractions they recruit as well as representing a relevant scale for actions by managers in
charge of these aquatic territories (Feunteun, 2002; Baisez and Laffaille, 2005).
The fact that every local action must be integrated into the larger European framework must not
be overlooked. The Indicang network chose the “Atlantic Arc” as the first Ievel.
The way in which an eel population functions may be compared to a tree. This “eel tree” (figure 4)
can only work if its roots, anchored in the Sargasso sea, are rich in spawners, i.e. silver eels. It can only
flourish if sap rises or falls along its trunk, this represents the oceanic circulation. This circulation cannot
stop or even slow down, otherwise the “leptocephali” larvae (ascending sap) will not be carried
eastwards at least with the same celerity, and nor will the silver eels (descending sap) be carried back
to their spawning grounds. Hence the unanswered question: what will be the effect of climate change on
oceanic circulation and hence on the functioning of this population?
13 See Chapters 2, 7 and 9.
14
See Chapter 2.
15

Finally, the tree can only prosper if glass eels, originating from larvae, colonise the different parts
of its foliage (representing the river basins) and of course, if continuously thinned, the tree will eventually
die.
This arborescent structure set the context for the Indicang project, which took it into account by
recommending local action to take care of the leaf (action at the level of the hydrographic unit) together
with coordinated actions undertaken on a significant number of leaves (Atlantic Arc) so that foliage
restoration would be sufficient to have a significant impact on the future of the “eel tree”.
Figure 4 -
The eel tree (from : S.Gros, Ifremer, Indicang Project).
16

Part I
Biological and methodological bases
17

Chapter 1
The life of the eel
Gilles Adam, Gérard Castelnaud, François-Xavier Cuende,
Estibaliz Diaz, Eric Feunteun, Patrick Girard,
Pascal Laffaille, Vanessa Lauronce, Stéphanie Muchiut,
Iñaki Oroz-Urrizalki, Patrick Prouzet, Christian Rigaud,
Laurent Soulier, Nicolas Susperregui
18

The life cycle of the eel, an amphihalin Silver catadromous eel species, is complex and, unlike that of the
Atlantic salmon, remains shrouded in obscurity, especially in the marine environment. For example,
reproduction has never been observed in the natural environment and no egg or adult has ever been
harvested in the presumed spawning ground (Nilo and Fortin, 2001). The species’ taxonomic status
remains very imprecise and hybridations between the European eel (Anguilla anguilla) and the
American eel (Anguilla rostrata) are commonly observed (Boëtius, 1980, Avise et al., 1986 and 1990;
Okamura et al., 2004). However, recent work concerning the genetic diversity of European and
American eels (Wirth and Bernatchez, 2001, 2003) has shown a well established segregation between
the two Atlantic species.
Figure 1.1. Diagram of the European eel biological cycle (adapted from Schmidt 1922, Klechner
and Mac Cleave, 1985).
1.1. Reproduction
Among the 19 species and sub-species of Anguilla identified globally, two species, Anguilla
anguilla and Anguilla rostrata, inhabit the Atlantic ocean (Tesch, 1977). These two species’ reproduction
grounds most probably overlap in the Sargasso Sea. The location of spawning grounds seems to be
19

influenced by the subtropical convergence ((Kleckner et al., 1983), along a density front (MacCleave,
1993). Schmidt in 1925 (in Nilo and Fortin, 2001) observed, for Anguilla anguilla, a spawner peak in
March, a result which was confirmed in 1979 (Tesch and Wegner, 1990). More recent results suggest
that spawning occurs from January to July and that hybridation exists between the two species,
although Anguilla anguilla and Anguilla rostrata spawning grounds seem to be separated, which
constitutes a reproductive isolation mechanism (MacCleave et al., 1987; Albert et al., 2006).
Schmidt (1925) concluded that reproduction occurred at depths from 400 to 700 metres.
However; Robins et al. (1979) observed an eel at a depth of 2,000 metres, off the coast of the
Bahamas.
Given the morphology of spawners (thick skin, dilated pupils, retinal transformation, marked
lateral line) and the need for high pressure on their flanks in order to initiate the production of gametes
experimentally, it is indeed plausible that reproduction could occur at several hundred metres in the
epipelagic zone (Kleckner et al., 1983). The smallest larvae are caught at depths of 200-300 metres
(Schoth and Tesch, 1982).
Boëtius and Harding (1985) thought that Schmidt’s conclusion that this species does not
reproduce in the Mediterranean was somewhat unfounded but, to date, no observation supports their
statement.
According to Boëtius and Boëtius (1980), the fecundity of European eels is between 0.7 and 2.6
million eggs for individuals between 630 and 790 mm long; that is 1 million eggs per kg of female on
average.
1.2. Embryonic development and larval phase.
We only have limited information on the embryonic development of eels. To date, no egg has
been harvested in the natural environment. The most comprehensive observations available are those
made in an artificial environment on Anguilla japonica (Yamamoto and Yamauchi, 1974).
Figure 1.2. Photo of eel leptocephalus Anguilla anguilla (photo : Y. Desaunay, Ifremer).
20

The larva, called leptocephalus, is better known in the natural environment. The smallest ones,
recovered on the presumed spawning ground, measure about 5 mm. These larvae, also known as
“willow-leaf like” or etymologically as “slender- or thin-headed” (figure 1.2) feed on plankton. They are
toothed (with 3 to 20 teeth, depending on their size) (Bertin, 1951). Carried passively by oceanic
currents, they, nonetheless, undertake vertical migrations of several hundred metres, between 35 and
600 m deep (Tesch, 1982). The nyctemeral rhythm markedly increases the catch at shallower depth at
night (Tesch, 1980; Kracht, 1982).
As they reach the continental slope, aged about 1 year, the larvae metamorphose into glass eels
(Lecomte, 1991). At the time of this metamorphosis, leptocephali are some 70 to 80 mm in length. The
duration of this oceanic migration continues to be debated because recently published physical models
of larval transport indicate that inert particles take up to three years to be carried by the Gulf Stream
between the Sargasso Sea and the European coastlines (Kettle and Haines, 2006; Bonhommeau et al.
in press).
1.3. Phase of entry into inland water: the glass eel stage
The French term “civelle” (including both glass eel and elver) covers the whole phase of
metamorphosis of the leptocephalus larva ending with complete pigmentation (when viscera are no
longer visible and yellow pigmentation is beginning). This pigmentation process has been codified
(Strubberg, 1913; Elie et al., 1982), its kinetics depending in particular on temperature and salinity
(Briand et al., 2004).
Figure 1.3. Estuary entry phase: glass eels (photo: G.Choubert, Inra).
21

As it enters the estuary, the glass-eel is transparent, with little pigmentation and not feeding. It is
said to be at the stage V A (Elie et al., 1982). Gradually, pigmentation develops, and it reaches stage V B
when rostral and caudal spots are clearly distinct. It is mainly at these stages that elvers are the most
abundant in salt- and fresh-water estuaries (Lefebvre et al., 2003; Briand et al., 2005; Lafaille et al.,
2007) and that they are fished, even in the lower reaches of rivers (Prouzet et al., 2002; De Casamajor
et al., 2003). The pigmentation process continues and elvers become fully pigmented at the VI B stage.
This marks the end of the elver stage. The following stage VII (yellow eel) is characterised by the
appearance of yellow pigments and an increasingly benthic behaviour.
Feeding begins again during the pigmentation stage, most frequently as soon as the VI A2 stage is
reached. Very quickly afterwards, there is a cessation of the weight and size loss that characterises the
first stages in all observation sites (Tesch, 1977; Charlon and Blanc, 1982; De Casamajor et al., 2000
and 2001; Lafaille et al., 2007). Of course, as feeding starts again, individuals acquire new physical
capacities. The following stages, particularly the VI A4 stage, can last several months until complete
pigmentation (White and Knights, 1997).
Elver migration has been studied in some detail on the Adour (Prouzet et al., 2003) where it was
confirmed that the upstream migration of elvers is passive, following the tidal front, with their position in
the water column dependant on light intensity (De Casamajor et al., 1999). The entry of glass eels into
the estuary is not a continuous process but occurs in “waves” (Rochard and Elie, 1994; De Casamajor
et al., 2000). During the upstream migration season (generally from October to March on the Adour), a
reduction in the size and weight of harvested elvers has been observed (De Casamajor et al., 2000 and
2001 ; Charlon et Blanc, 1982). Glass eels generally average between 68 and 76 mm in length (De
Casamajor et al., 2003) close to the shores of the Bay of Biscay.
1.4. The colonisation phase of estuarine and inland waters: yellow eels
The end of passive migration is linked to changes in the behaviour of elvers, which from being
passive and pelagic-like, become increasingly active and autonomous. Not all elvers migrate upstream;
some remain in the lower reaches of rivers and esturaries, and even in the coastal transitional waters:
lagoons, salt marshes (Daverat et al., 2005 and 2006). Some eels, having inhabited fresh water, will
return to brackish or salt waters a few months or years later. The dispersion of those that colonise areas
upstream of the tidal limit is determined by factors that remain poorly understood. For this reason, it is
difficult to establish a link between estuarine recruitment, which depends on the seasonal waves of
glass eels in the estuary, and river recruitement that comprises the fraction that colonises, apparently
progressively, the inland waters upstream of the tidal limit (chapter 8, yellow eel indicator). It appears
that individuals who cross this limit have the highest condition (or weight) coefficients (Edeline et al.,
2006). Feunteun et al., (2003) proposed a four-category classification for migratory behaviour. The
“founders” who settle as soon as they find a favourable habitat; the “pioneers” who migrate to the
22

highest reaches of the river; the residents or "home range dwellers" who settle in a given area for
several years and the “nomads” who move from one habitat to another in order to feed or to settle
temporarily. Work undertaken on the Vilaine (Briand et al., 2000) shows that the catchment basin is
colonised through migratory waves and suggests that density-dependency mechanisms might exist. On
the Severn, Ibbotson et al. (2002) suggest instead that the catchment basin is colonized by downstream
to upstream dispersion. Recent work shows that, in fact, there is a synergy between these two
colonization strategies which are conditioned by environmental conditions (particularly the quality and
accessibility of habitats), allowing eels to colonise all the available inland habitats, hence improving the
reproductive success of the species (Lasne and Laffaille, 2008).
Figure 1.4. Sedentarisation phase: yellow eels (photo : G. Adam, Diren).
Generally speaking, males predominate where densities are highest, usually the lower reaches of
river basins, whilst the females, which are older and fatter, predominate in sparsely populated areas,
generally upstream of the river basins (Parsons et al., 1977; Aprahamian, 1988; Vollestad and Jonsson,
1988; Acou et al., in press). However, this is not always the case, especially in smaller river basins
(Laffaille et al., 2003). It is however important to note that this theoretical spatial structuring does not
determine the relative contribution of each of the large compartments of a river basin to the production
of females in that basin. In fact, females seem to be in a very small minority in heavily-populated
downstream zones and to predominate in upstream zones but with very low densities, similar to
downstream female densities.
Ovogenesis and spermatogenesis occur at the same time in the gonad, the former starts at
around 14 cm and the latter at around 18 cm. This hermaphroditic stage preceeds a permanent
masculinisation or feminization phase (Bertin, 1951). The factors that influence sex determination, after
the so-called non-functional hermaphroditic phase, are essentially unknown (Tesch, 1977). Some
23

Downstream migration occurs all year round but varies in intensity from summer to spring. The
seasons during which migratory intensity peaks vary with latitudes and the presence of barriers to
downstream migration (Feunteun et al., 2000; Acou et al., in press). Generally, in the central distribution
zone (Bay of Biscay), it occurs in the autumn (Langon and Dartiguelongue, 1997; Gosset et al., 2000).
Variations in some environmental parameters (temperature, flow, conductivity, atmospheric pressure,
etc) and lunar cycles play an important role in triggering downstream migration (Smith and Saunders,
1955; Winn et al., 1975; Gosset et al., 2000; Durif, 2003; Acou et al., in press). On average, the males
migrate when smaller and younger than the females, whose size generally exceeds 40 cm (Acou et al.,
2003).
Other noticeable anatomical transformations observed in the eel include modifications to the wall
of the gas bladder which allow the eel to reach depths of at least 2,000 metres (Robins et al., 2000) and
changes in the retinal cells which develop characteristics similar to those of abyssal fish. Deep migration
would enable eels to use the deep Gulf Stream countercurrents in order to reach their spawning ground
(Tucker, 1959). The direction in which the water is moving might be determined with the help of induced
electric fields (Rommel and Stasko, 1973; Rommel and Mac Cleave, 1973) which can be detected by
eels. Other authors suggest that the magnetite cells located in the jaws of silver eels may play a role in
their orientation. Once at sea, it appears that silver eels travel in a generally westerly direction until
reaching the Gulf Stream currents. Their chemical senses (importance of olfaction) and tactile senses
(importance of the lateral line) would then enable them to find the spawing areas. In any event, the
mechanisms guiding the return to the Sagasso Sea depend on complex physiological mechanisms
which can only be fully functional in healthy individuals. The physiological functions used in guidance
mechanisms may be impaired through contamination by metals and some organic compounds and/or
the poor health state caused by various pathogenic organisms.
1.6. Mechanism for dispersion from the spawning area towards the
productive inland zones
The panmixia hypothesis means that European eels make up a single population mating at
random in the Sargasso Sea. This implies that because of the random dispersion of larvae by oceanic
currents, all the European eels scattered in the river basins of the colonization area, from Mauritania to
the Arctic polar circle, belong to the same reproductive population. However, this generally-accepted
hypothesis has been called into question by recent work on the genetic diversity of the population (Wirth
and Bernatchez, 2001, 2003; Maes and Volckaert, 2002). From nuclear DNA analysis, these authors
have established the following facts:
• The genetic markers used show that Anguilla rostrata and Anguilla anguilla are two welldifferentiated
species and the non-arborescent structure of Anguilla rostrata seems to indicate its
panmictic nature.
25

• On the other hand, the arborescent structure of Anguilla anguilla shows that several groups exist : a
Mediterranean group, a North Sea group and an Atlantic group, leading to the hypothesis that the
species is not panmictic;
• The sample of Icelandic origin occupies an intermediate position and confirms work by Avise et al.
(1990) which found a significant ratio of hybrids from the 2 species.
According to these authors, the possible existence of several, genetically-distinct units may be
related to the distance separating the river basins from the Sargasso Sea breeding grounds. However,
the position of the Adour sample (Bay of Biscay coast) and the Minho sample (North of Portugal) whose
arborescent structures are close to those of samples from the North Sea, show that the respective
location of river basins in relation to the Sargasso Sea does not explain everything. The complexity of
the oceanic circulation from the Sargasso Sea to the European coastlines may, through the respective
durations of migrations, furnish supplementary factors explaining this genetic proximity, which is not
explained by the geographical proximity of the stocks under consideration.
However, more recent work from Dannewitz et al. (2005) has shown a strong temporal variability
within the sample in excess of the variability related to the geographical origin. This strong genetic
variability had been observed by Cagnon et al. (2004) on samples collected from the Adour, and
originating from different migratory waves during a given migratory season. These latter results suggest
that the panmixia hypothesis for European eels remains valid and emphasise the need to implement a
more stratified sampling programme for this kind of study. This may also show whether the spatial
structuring identified by the work of Wirth and Bernatchez (2003) is an artefact of the temporal structure
which characterizes metapopulation dynamics (Albert et al., 2006 ; Maes et al., 2006 ; Pujolar et al.,
2006).
Even in the absence of genetic structuring, three distinct stocks or sub-populations, which
produce silver eel populations with different characteristics, can be identified through geographical,
ecological, fisheries, and biological specificities (relating in particular to recruitment intensity and the
diversity of oceanic migratory channels).
The first group is in the “North of the European distribution zone” (North Sea, Baltic). Glass eel
recruitment appears to be very limited, biological cycles are long and densities low, eel exploitation is
essentially focused on the silver and yellow stages. This group mainly produces females.
The second group is in the "Centre of the European distribution zone" (Atlantic, English Channel),
a reference location for Indicang. The stock is characterised by higher glass eel recruitment as the river
basins are almost completely colonized, biological cycles last between 5 and 15 years but are shorter
than for group 1 and the sex ratio varies according to the sub-population parameters and the physical
and trophic characteristics of the habitats. Exploitation mainly concerns glass eels (in the South),
although yellow and silver eel fisheries are well developed on some rivers (Somme, Loire, Gironde) and
on the coastal marshes of the Atlantic coast.
26

The third group, the "Mediterranean zone", is characterized by glass eel recruitments that are
dispersed but more important than in the North of the colonisation zone. Biological cycles are often short
and populations are mainly confined downstream in coastal lagoons, especially in North Africa.
Exploitation focuses essentially on yellow and silver eels.
Cold currents
Warm currents
Figure 1.6. Map of the Northern Atlantic Ocean outlining the oceanic circulation, a vector for
leptocephali larvae.
Given the weak swimming capacity of the larvae, it is thought that the main branch of the Gulf
Stream, followed by the North Atlantic Drift, provide the transport for the major part of their migration
eastwards. The northern component of the Subtropical Convergence, the Azores current, carries the
larvae towards the Mediterranean whilst the northern branch of the North Atlantic Drift transports them
towards the northern part of the distribution area. The southern branch of the North Atlantic Drift, which
is the most important, carries the larvae toward the central part of Europe. This dispersion phenomenon
is very important as it explains why the main glass eel fisheries have developed in the central part (Bay
of Biscay, south of the British Isles) which is the first to receive the highest concentrations of glass eels.
Elvers migrating upstream in the rivers of the northern Iberian Peninsula, earlier than in the rivers in the
north of the Bay, may be transported by the Azores current rather than the southern branch of the North
Atlantic Drift.
Colonization of the boundaries of the distribution area follows later, through the dispersal of
individuals which have not been drawn into the inland waters of the central zone. Conversely, the
recruitment into the Irish Sea or the south of the North Sea and also the northern branch of the North
27

Atlantic current is naturally lower or low. This explains the low level of eel densities in Scandinavian
regions, although these are compensated by a high proportion of females (Knight, 2001).
1.7. Status of the species and assessment of the resource
1.7.1. Status of the population and of its exploitation by fishing
Studies compiled by the joint ICES/EIFAC working group (Anonymous, 2003; Dekker, 2003;
Dekker et al., 2003) show that the number of individuals in the population of, and the catches from
fisheries based on, the European eel (Anguilla anguilla) have declined considerably in the great majority
of river basins in the distribution area. Currently, the population is considered to be at risk, “outside safe
biological limits”, and the fisheries cannot sustain their production level in most of the river basins in
question.
However, the French and Spanish glass eel fisheries continue to be of very great socio-economic
importance for small-scale coastal fisheries (Leauté, coordinator 2002,; Prouzet, coordinator 2002) as,
despite falling catches, Asian demand (China in particular) keeps the first-sale prices high, with such
prices reaching 600 to 700 euros per kilogramme landed value for live glass eel.
The situation is similar for the American eel (Anguilla rostrata) (Nilo and Fortin, 2001; Dekker et
al., 2003). There are few data on elver upstream migration as it is only monitored intermittently in the
United States and in Canada but, on numerous American watercourses (Casselman et al., 1997;
Chaput et al., 1997), a significant drop in the number of juvenile or yellow eels has been observed in
migratory pass counts or in the densities on different rivers estimated by electrofishing. Both
populations, European and American, showed a sizeable drop in abundance at least towards the end of
the 1970s.
However, it must be noted that situations differ markedly between river basins and are often
related to their location with respect to the main currents in the recruitment of this species. The situation
seems to be more critical in the northern part of the distribution area than in the southern part and also
in accordance with the degree of anthropisation of the river basins in question, with a very significant
drop in yellow eel density in extensively-modified river basins (Lobon-Cervia et al., 1995). This negative
trend is less noticeable river basins which are less anthropised, even those with significant fisheries (for
example the Adour, Prouzet et al., 2002), or are more open to the entry of glass eels due to the absence
of fishing even if there is a high level of degradation of the river basin (for example the Frémur, Lafaille
et al., 2005).
Globally and as with many marine fish stocks, a diagnosis can in theory be established on the
basis of state indicators for stocks and the pressure levels to which they are subjected in comparison to
reference points. These indicators and reference points are often expressed in terms of biomass
(generally spawning biomass) and mortality. In the case of marine stocks, fishing accounts for most of
28

the mortality of human origin. In the case of eels, there are multiple causes of anthropogenic mortality
and they are not due exclusively to fishing.
Comparing the level of observed biomass and the corresponding mortality level of anthropogenic
origin (figure 1.7) shows three possible stock states in relation to four reference points. Limit reference
points (1 and 3) are the values beyond which there is a high risk of stock collapse (red area). The
precautionary reference points (2 and 4) make it highly probable that reaching these limit reference
points can be avoided. The more uncertain these limit values, the further the precautionary reference
points are from them. The yellow area corresponds to this safety margin, the uncertain state of the stock
then requiring close monitoring. Finally, the blue area corresponds to a stock considered to be
sustainably exploited.
In the case of eels, although the reference points are not clearly identified, we are probably below
the limit biomass as recruitment has decreased by a factor 10 to 15 over the last 25 to 30 years.
According to the great majority of expert scientists, it is highly probable that we are above the limit value
for mortality of human origin. Thus, the stock is most certainly today in the red zone implying a
significant risk of collapse. Management measures must therefore aim to reduce mortality of
anthropogenic origin rapidly, in order to restore the spawning biomass in the long run.
Biomass Limit
lim Biomass it (1(1)
Biomasse
Precautionary
pBiomass écautio n(2)
(2
Leveveau
Level of
mortality Mortaliof
é
anthropogenic anthrLeveopiqu
origin
(pollution,
(pollutions
abstraction, pompages
turbines, turbinages
fisheries etc
pêches et .
Valeur Acceptable limite admissible value for mortaility de of é
d’origine anthropogenic anthropique origin (3)
Valeu Preca
d
p éc
Precautionary value for mortality of
anthropogenic de
é d’origine
origin (4)
-(4
anthropique
Biomass
Figure 1.7. Diagram showing stock diagnosis as a function of mortality and biomass targets
(source: Grisam, 2006).
1.7.2. Millennial, centennial and current evolutions: climatic
influence
Work by Wirth and Bernatchez (2003) analysing the genetic diversity of eels from nuclear DNA,
using the method developed by Beaumont (1999), appears to show that the eel population has declined
29

over a period of time corresponding to some 766 to 5,132 generations. With female sexual maturity
being reached between 10 and 15 years, this would indicate that the period of decline has lasted at
least 8,000 years and at most 75,000 years. The average values chosen by these authors, 2,000
generations and 10 years for the duration of a generation, date the origin of this decline towards the end
of the Wisconsin glaciation. These climate changes have certainly affected oceanic circulation
(Duplessy, 1999) and larval transport by the Gulf Stream. The results of this work also showed that the
decline started earlier in North America than in Europe and that the rate of decline has been more
pronounced in European eels.
The NAO (North Atlantic Oscillation) is a phenomenon of fluctuations in the pressure difference
between the Azores subtropical high and the Icelandic low. NAO variations have also been noted on a
centennial scale. The strong influence of these atmospheric pressure variations on oceanic circulation is
well established. They have beneficial effects on some amphihaline fish species such as salmonids
(Beaugrand and Reid, 2003) as they increase the trophic chain productivity. However, by slowing down
oceanic circulation, they may have a negative impact on species such as eels which, during their marine
ecophases, feed very little (leptocephalus) or not at all (silver eel and glass eel) and which use the
current as a means to migrate.
1.7.3. Anthropogenic factors: recent evolution
The analysis of the evolution of recruitment in the northern part of the distribution area,
particularly in Sweden and in the Baltic, shows that recruitment and escapement indicators began to fall
well before the 1970s (Anonymous, 2002).
Many anthropogenic factors act in synergy on the various river basins. The principal ones, not in
order of importance, are discussed below. They vary between river basins.
1.7.3.1. Reduction in habitat accessibility
Across the whole distribution area, one of the major causes of the decline in the species is the
fragmentation of river basins and the non-accessibility of many habitats due to the building of an
increasing number of barriers to migration.
During the 20th century, more than 25,000 large dams were built to meet multiple needs
throughout the world (drinking water, irrigation, industry, electricity, etc.). European Union member
states regulate 60 to 65% of their river flows and the building of large dams accelerated after the second
World War (In Anonymous, 2003). These barriers have greatly restricted access to the habitats of the
middle and upstream reaches, which are far less degraded than downstream ones. The immediate
impact has been a reduction in the inland distribution area of the eel. This impact is certainly more
noticeable in the peripheral zones of the colonization area, particularly in Scandinavian zones which are
markedly less irrigated by glass eels than the central zone. Sweden, for example, a pioneer in
30

hydroelectric development in Europe, began to develop this energy source from the end of the 19th
century. At the beginning of the 20th century, as energy transport techniques developed, it became
possible to equip large basins. This enabled Sweden, which has considerable water reserves (more
than 100,000 lakes and 13 rivers with an average flow of 100 m 3 /s) to rely on hydroelectric energy to
cover its total energy requirements until 1967. It is highly likely that the impact on silver eel production in
this zone was very negative. From the end of the 1940s (although eel exploitation did not increase
during the Second World War or immediately after), there was a very clear fall in glass eel catches in
this region which was followed, ten years later, by a marked drop in eel catches in the Baltic.
In France also, such fragmentation continued throughout the 20th century, although to a lesser
degree. Some dams, such as the Arzal on the Vilaine, and those built on many estuaries opening into
dyked marshes (Couesnon, Sèvre Niortaise, Seudre, etc.), were built in the river mouth in order to
prevent the tidal flow upstream. The effect of this was to block the influx of glass eels and to increase
the impact of fishing downstream of these constructions (with exploitation rates over 90%, ICES -
Working group on eels, 2002; Briand et al., 2003). Although important efforts have been aimed at
facilitating the passage of migratory salmonids or alewives through these barriers as they migrate
upstream, far fewer fish passes (Legault, 1988; Legault, 1992; Knights and White, 1998; Briand et al.,
2005) or manoeuvres of the construction (Legault, 1990; Laffaille et al., 2007) have been dedicated
specifically to eels. It is only recently that large dams have begun to be equipped. The Tuilières dam on
the Dordogne is a perfect example (Teyssier et al., 2002). In 2002, a bristle pass was installed (this
being a device adapted to the passage of juvenile and yellow eels). It allowed 28,329 individuals to pass
over it, whereas the pool-type fish pass (designed for migratory salmonids) had only let 3,521 juvenile
eels through (Lauronce and Garcia, 2007).
The negative impact of these dams on the production of future spawners in a given river basin
can be seen during the downstream migration (Behrmann-Godel and Eckmann, 2003) when many eels
die or are mutilated as they pass through hydroelectric turbines (Boubée et al., 2001; Watene and
Boubée, 2005; Winter et al., 2006) or in reserved flow pipes (Legault et al., 2003). Numerous
experiments have shown that the survival rate of silver eels after their passage through these turbines
may be very low. The mortality generated depends on the type and characteristics of the turbine that is
used (figure 1.8), on the position of the water inlets in relation to the river axis, on the presence or
absence of protective grates and on the hydrostatic pressure created by the height gradient and the
speed of the currents which can pin the eels to the grates. The problem becomes particularly acute
when several hydroelectric dams are located along the same downstream migration axis. On the
Meuse, Prignon et al. (1998) estimated the direct mortality to be between 34 and 45% in the case of
males and between 40 and 63% in the case of the females, which are of larger size. On the Rhine,
Dönni et al. (2001) showed that between Schaffhouse and Bâle, the cumulative silver eel mortality after
the passage through 13 hydroelectric power stations was some 92.7%, very close to the mortality rate
estimated by Behrmann-Godel and Eckmann (2003) after the passage through 14 successive
31

constructions on the Moselle. On the Rhône, the eel mortality rate from Lyons to the sea is some 90%
(Feunteun et al., (1999).
Figure 1.8. Eel migrating downstream killed by the blades of a turbine (photo : Migado).
Thus, the future of this species depends on improving, as quickly as possible, the liberty to
migrate both upstream (juvenile eels and yellow eels) and downstream. Otherwise, the reduction in
silver eel production in the middle and upper reaches of the river basins will accelerate. In this respect,
the ICES working group is greatly concerned about the possible negative impact on this species of the
European guideline (2001/77/CE) that seeks to increase “green” energy in member countries. This
guideline aims to increase the proportion of energy obtained from renewable resources from 13.9 to
22%. It will result in a multiplication of hydroelectric dams on rivers.
1.7.3.2. Degradation of habitat quality
The degradation of habitat quality (Laffaille et al., 2004) and in particular, the reduction in wetland
surface area (Feunteun, 1994 ; Baisez, 2001) have also been identified as having a particularly negative
effect on the species. From the 1960s onwards, changes in farming practices in the central and
southern part of the colonization zone led to more water being abstracted or diverted for irrigation
purposes and many wetlands situated on lower basins have disappeared through reclamation or
drainage. Of the 268 million hectares irrigated in the world, some 30 to 40% rely on water reservoirs.
This causes increasingly lower watercourse levels and, as a result, a marked reduction in their
piscicultural production. France, where agriculture represents a major activity, is obviously no exception
32

to this rule. For example, in the Garonne-Charente-Dordogne river basin, the irrigated cereal area has
increased five-fold: from 100,000 hectares in 1970 to 500,000 hectares in 2000 (Teyssier et al., 2002).
In the Adour river basin, the situation is similar, with irrigated crops having increased four-fold (Prouzet
et al., 2002).
The farming of these wetlands and the increase in this type of agricultural production also go
hand-in-hand with the increasing use of pesticides. At the same time, numerous industrial and urban
poles have developed and resulted in PCBs and flame retardants being found in the aquatic
environment. But these various types of pollutants (chlorobiphenyls, heavy metals, organochlorinated
pesticides, etc.) quickly build-up in the adipose tissue of eels. A recent study of eel PCB contamination
in Belgium (Goemans and Belpaire, 2002) showed that 80% of samples from 244 sites exceeded the 75
µg/kg acceptable threshold. High concentrations have also been found in this species in the basins of
the Adour (GIS-ECOBAG data) and of the Gironde (Tapie et al., 2006). The impact of this contamination
on eel physiology and in particular on its reproduction remains to be determined but it could trigger an
early silvering process and an early downstream migration before the energetic resources required for
the transoceanic migration for reproduction have been stored (Robinet and Feunteun, 2003; Thuillard et
al., (2005). Likewise, it would seem that this degree of organic contamination reduces eel fertility, and
may even cause sterility (Robinet and Feunteun, 2003).
Figure 1.9. Water primrose treated with weed killer in a recalibrated canal in the Adour river
basin. (photo : P. Prouzet, Ifremer).
To these various anthropogenic factors and their probably sizeable negative impact on this
population’s future must be added the consequences of the introduction of exotic plant or animal
species such as the Florida shrimp (Procambarus clarkii), extremely abundant in the lower parts of river
33

asins, and plants bred for aquaria such as water primrose or parrot’s feather which invade irrigation
canals and can only be eradicated through the intensive use of weed killers (figure 1.9).
1.7.3.3. Development of Anguillicola crassus
Multiple pathogenic agents can interfere with the species dynamics (Bruslé , 1994; Vigier, 1990)
and it seemed useful, within the Indicang programme framework, to increase the awareness of various
partners to the collection of data related to the health status of individuals observed during monitoring
operations (fisheries, passes, inventories) (Girard and Elie, 2007). Anguillicolosis has been one of the
most problematic parasitoses over the past few years. At the beginning of the 1980s, an eel-specific
parasite, Anguillicola crassus (figure 1.10), responsible for anguillicolosis, was introduced into our
environment (table 1.1) following the importation of Japanese eels (Anguilla japonica) into the
Mediterranean (Peters and Hartmann, 1986).
Figure 1.10. Abdominal cavity of an eel infested by the parasite Anguillicola crassus.
This nematode worm multiplies in the swim bladder and induces lesions in its wall. The lesions
are caused either by the haematophagous adults who feed on the blood of the capillaries irrigating the
bladder wall, or by the larvae L3 which cross the wall and, after ingestion by eels of intermediate
infested hosts (zooplankton, small fishes) migrate towards the swim bladder where they may develop
into a cyst before transforming into adults (Lefebvre et al., 2004). After several successive infestations,
the swim bladder wall progressively loses its elasticity and its suppleness. This probably reduces its
ability to ensure the eel’s hydrostatic equilibrium during its marine migration towards the Sargasso Sea
(Möller et al., 1991). Boury et al., (2005) estimated that 15 to 20% of silver eels leaving the Loire river
basin have suffered alterations to their gas bladder that will make their migration towards the Sargasso
sea practically impossible.
34

Diagnosing infestation is not always easy as the absence of adults or evolved larvae in the
bladder cavity does not necessarily mean that there are no young larvae within the bladder wall.
Moreover, the infestation can also be temporary. Thus, Lefèbvre et al. (2002) developed an observation
method and a bladder deterioration index in order to assess current and/or past intensity of the parasitic
aggression.
Table 1.1. Prevalence and intensity of the infestation by Anguillicola crassus for various subpopulations
of the European eel in its distribution area.
References Country River basin Year Prevalence (%)
ICES-EIFAC, 2006 Italy Tiber 1996 66,3
Infestation
intensity (number
of parasites/eel)
ICES-EIFAC, 2006 Italy Comaccio 1997 11,9 -
ICES-EIFAC, 2006 Italy Figheri 1997 9,1
ICES-EIFAC, 2006 Italy Burano 1997 37,4
Evans et al., 1999 Ireland Erne river 1998 3,2-22,2 2,3-5,6
Gallastegui et al., 2002
Basque
Country
Butrón 2000 6,7-8,9 8-10
Aguilar et al., 2005 Spain Ulla 2000 0 0
Aguilar et al., 2005 Spain Tea 2000 55,5 5,5 5,82
Maíllo et al., 2005 Catalonia Coastal lagoons 21,3-30,8
Audenaert et al., 2003 Belgium All the Flemish basins 2000 88,1 5,5
Genc et al., 2005 Turkey Río Ceyhan 2002 72,41-82,86 3,20-3,31
Adam, 1997
Lauronce and Garcia,
2007
France
Grand-Lieu lake
(Loire basin)
1991-1995 64-31,7
France Garonne Dordogne 2002 45 2,02
Marty, 2004 France Adour 2004 52,74 (0-100) 3,36 (0-23)
ICES-EIFAC, 2006 Sweden North of the Baltic Sea 2002-2005 62
ICES-EIFAC, 2006
Sweden
South of the Baltic
Sea
2002-2005 10
ICES-EIFAC, 2006 Germany Rhine 2006 60
ICES-EIFAC, 2006 Germany Dosel 2006 80
ICES-EIFAC, 2006 Denmark Fresh water < 12ppm 2006 >50
ICES-EIFAC, 2006
Denmark
Brackish water >
12ppm
2006

The introduction of this parasite into the United States appears to be more recent, with Fries et al.
(1996) reporting its presence in a fish-farm in Texas. It then spread quickly as it was reported in the
Hudson River (New York state) (Barse and Secor, 1999).
The fact that this nematode does not cause any specific pathological disorder in Anguilla japonica
(Ooi et al., 1996) should be noted.
1.7.3.4. A fishery exploiting every biological stage
Fishing can be added to the multiple oceanic or inland factors mentioned previously. The
historical analysis of the species’ abundance seems to show that this fishery, which has developed on
all inland stages (Dekker, 2003) is not the factor triggering the drop in numbers (Prouzet, 2003), but it
appears to be an aggravating factor as, in many cases, fishing pressure has not adapted to the changes
in abundance. Having said this, its intensity varies greatly depending on the river basins. Thus,
exploitation rates on migratory eel glasses vary from 0% (in Northern Europe or in the Mediterranean
where, in many countries, fishing this stage is prohibited) to more than 90% below estuarine dams
which block all migration (Anonymous, 2002).
In an open-type estuary such as the Adour, the Garonne or the Loire, estimations of glass eel
abundance and exploitation rate showed that the latter varied between 5 and 40% depending on the
year and the estuary (Bru et al., 2004; Bru et al., 2007).
Figure 1.11. Fishing for elvers, using push nets on the Isle, a tributary of the Dordogne (photo :
Cemagref).
36

As regards yellow eels, the dataset on exploited systems also shows the diversity of exploitation
intensity. According to Dekker (2000), it was 85% for males and practically 100% for females in the
Ijsselmeer lake in Holland for the 1989-1996 period. The exploitation rate is practically the same on the
western coast of Sweden resulting in a 15% escapement (Svedäng, 1999), and according to Adam
(1997) it is about 45 and 50% in the Lac de Grand-Lieu in France.
Yellow eel exploitation rates vary according to the fishing gear used and depend on the size (and
therefore on the age) of the eels that are targeted. Fisheries can overwhelmingly target small-size eels
as in the Monaci lagoon in Italy (Ardizzone and Corsi, 1985) or focus on larger size silver eels as in the
Porto Pino lagoon in Sardinia (Rossi and Cannas, 1984) or on the Lough Neagh in Northern Ireland
(Allen et al. 2006) or even modulate their fishing effort on a wider range of sizes as is the case in the
Lac de Grand-Lieu in France (Adam, 1997).
Figure 1.12. Fishing for yellow eel with a fyke net on the Lac de Grand-Lieu (photo : P.Prouzet P,
Ifremer).
As regards silver eels, their exploitation also varies greatly in intensity. Exploitation rates are low
on the French Atlantic coast and many rivers such as the Adour or the estuarine-river system Gironde–
Garonne–Dordogne do not have fisheries targeting this phase. This is not the case on the
Mediterranean coast where this ecophase is targeted along with yellow eels. Through tag–recapture,
Feunteun and Boisneau (anonymous, 2003) estimate that the escapement rate is at least 80% from the
Loire “guideau” fisheries (fishing traps equipped with leaders). It is lower in the Baltic where silver eel
37

escapement is estimated at 60% (Moriarty, 1997). This is also the case on the Erne River (Matthews et
al., 2001) or on the river Shannon in Ireland (McCarthy and Cullen, 2000) where, on average,
escapement exceeds 60%.
Figure 1.13. Fishing silver eels with a “guideau” on the Loire. (photo : Ph. Boisneau).
In many French rivers, exploitation has tended to decrease markedly because of increasing
resource scarcity, and also because this metier has become much less attractive to young professional
fishers: this is the case, for instance, of the Adour (Lissardy et al., 2004) or of the Garonne and
Dordogne (Girardin et al., 2006).
However, both the global exploitation rate and fishing mortality remain unknown, whatever the
ecophase. Available data only concern river basin stocks where professional fisheries are well
established. Very little is known about sport and illegal fishing. The first estimates (Changeux, 2003;
Baisez et al., 2006) show that their production is far from negligeable. But a comprehensive evaluation
has yet to be undertaken on all the basins concerned with this activity.
38

Chapter 2
Further information on biology
Christian Rigaud, Pascal Laffaille, Patrick Prouzet, Eric Feunteun,
Estibaliz Diaz, Jaime Castellano
39

2.1. Variability of estuarine recruitment characteristics
2.1.1. Intra- and inter-seasonal variability
Oceanic and climatic factors explain the monthly variations in the length and weight of glass eels
arriving in the estuaries. Fluctuations in the strength of marine currents, temperature and food
availability seem to be the most important parameters. Oceanic conditions are more favourable for
individuals migrating at the beginning of the season than at the end (Lecomte-Finiger, 1978). Thus, in
many estuaries, the length and weight of European eel glass eels decrease during the year (Lecomte-
Finiger, 1976; Boetius and Boetius, 1989; Desaunay and Guerault, 1997; De Casamajor et al., 2000 and
2001; Lambert et al., 2003; Lefebvre et al., 2003; Laffaille et al., 2007). This trend during the recruitment
period has also been observed in other eel species (Jellyman, 1977; Chisnall et al., 2002 for New
Zealand eels; Jessop, 1998 for American eels).
Some authors argue that the inter-annual variation in biometric characteristics may be related to
estuarine recruitment fluctuations. Glass eels tend to be rather thin in years of poor recruitment (Haro
and Krueger, 1988; Castonguay et al., 1994; Désaunay and Guérault, 1997).
2.1.2. Geographical variability
Morphometric variations for a given year and across the whole distribution area may also be
explained by energetic losses during the continental shelf crossing when leptocephali metamorphose
into glass eels. Losses are more or less significant depending on the shelf width (Haro and Krueger,
1988). The most recent work undertaken on the Adour by Cagnon et al. (2004) also showed an interannual
genetic variability in glass eel fluxes. This can complicate the studies on genotypic heterogeneity
of the European eel population carried out in recent years and which, for the time being, have neither
rejected nor confirmed definitely the panmixia hypothesis 1 .
2.1.3. Evolution of the degree of pigmentation during entry into the
estuary and inland waters
During anadromous migration, several stages can be identified from the extent of tegument
pigmentation 2 . These life stages correspond to progressive physiological and behavioural changes:
functioning of the digestive tract, intensification of swimming activity, functioning of the gas bladder,
development of teeth, etc. The pigmentation stage is generally determined on the basis of the
classification established by Elie et al. (1982). This standardised classification allows a comparison to
be made of the various studies that have been undertaken. But a few experiments are also based on
1 See Chapter 1.
2 Id.
40

the work of Boëtius and Boëtius (1989). A catalogue based on the classification of Elie et al. (1982) and
developed by Grellier et al. (1991) can be found in annex 6 of the Indicang report 3 . The pigmentation
stages are used to define glass eel development stages because size and weight are not relevant and
determinant variables (as already mentioned, both of the latter decrease during the season for a given
pigmentation stage).
The time taken for glass eels to become fully pigmented is observable, but the impact of
environmental factors on this process makes it difficult to establish a linear relationship between the
time spent in the estuary or inland waters and the degree of pigmentation (Gascuel, 1987; Ciccotti et al.,
1995). Work by Briand et al. (2005a) allows predictions to be made concerning the effect of temperature
and salinity on the pigmentation process and shows that, for a given increase in temperature, when
salinity is high the process is much slower than in fresh water. This is the major environmental factor
affecting the speed of the pigmentation process.
The V B pigmentation stage predominates in estuaries (De Casamajor et al., 2003; Lefebvre et al.,
2003; Briand et al., 2005a; Laffaille et al., 2007), but it is also frequently found in glass eels sampled at
sea (De Casamajor et al., 2003). The same authors also show that this pigmentation stage comprises
elvers with otolithometric structures belonging to the 3 types found in the classification of Lecomte-
Finiger and Yahyaoui (1989). Type 1 has no estuary transition ring, type 2 has one on the edge of the
otolith whilst type 3 shows growth beginning after this transition ring (De Casamajor et al., 2003).
Antunes (1994) found that the otoliths of 40% of Minho glass eels had 1 or 2 rings.
The pigmentation stage V B cannot therefore be used as an indicator of the time spent in the
estuary. Furthermore, the V B pigmentation stage can be found up to the natural tidal limit (Lecomte-
Finiger, 1978; Jorge et al., 1990; Desaunay et al., 1993; Prouzet et al., 2003) or the artificial one
(Laffaille et al., 2007). On the other hand, the V A stage can be considered to characterise glass eels that
have just entered estuarine waters and therefore to correspond to estuarine recruitment.
Growth and feeding recommence around pigmentation stages VI A2 to VI A4 . Swimming then
becomes active, with a tendency to switch from a pelagic to a more benthic lifestyle (Sorensen and
Bianchini, 1986; Belpaire et al., 1992). From stage VII, individuals are referred to as juvenile eels (Elie et
al., 1982).
2.1.4. Behaviour and environmental factors during estuarine
passage
Glass eels arrive in the estuary all year round with an intensity that varies according to one or
more normal distribution (Gauss) curves with migration peak(s) appearing earlier or later depending on:
3 Grellier P., Huet J., Desaunay Y., 1991. Stades pigmentaires de la civelle Anguilla anguilla (L.) dans les estuaires de la Loire et
de la Vilaine, Ifremer, annex 6 of the Indicang report, http://www.ifremer.fr/indicang.
41

• The latitude of the estuary: main arrivals occur later in the north and south of the distribution area
than in the centre.
• The variability of oceanic factors that affect successive leptocephalus waves (Boetius and Boetius,
1989) originating from a spawning period spread over several months, from January to July,
according to the most recent observations (Albert et al., 2006). This generates different glass eel
arrivals over the season whether on the Atlantic coast (Cantrelle, 1984; Guérault et al., 1992;
Desaunay et al., 1993; De Casamajor et al., 2000), in the English Channel (Laffaille et al., 2007) or
on the Mediterranean coast (Ciccotti et al., 1995; Lefebvre et al., 2003). This rhythm is further
modulated in the estuary by the effect of various continental hydrodynamic factors 4 .
Work undertaken on the Adour since 1995 (De Casamajor, 1998; Bru, 1998; Prouzet et al., 2002
and 2003) and during the INDICANG project has shown the very strong influence of hydroclimatic
variables, such as turbidity, flow or tidal current, on the vertical or horizontal migratory behaviour of
glass eels in an open estuary (with no estuarine dam) and the resulting catch variation. Recent work
undertaken on the Oria (Castellanos and Diaz, 2006) in the proximal zone of the estuary, where
thermoclines and haloclines are still marked, showed that glass eels were entering mainly along the
bottom of the salt water section until they became accustomed to desalinated waters. In a closed
estuary (the Couesnon), Laffaille et al. (2007) have recently shown that water levels (linked to tidal
amplitude) and the estuarine water temperature are the main factors explaining temporal variations in
glass eel colonisation dynamics.
2.1.4.1. Hydrodynamism and lunar cycle
Flows and tide
It is often difficult to dissociate the impact of tides and flow which jointly affect the progression of
the dynamic tide inside an open estuary. It has been shown, both experimentally (Barbin and Krueger,
1994) and the real world (McCleave, 1980; Legault, 1987; Prouzet et al., 2003), that glass eels cannot
battle for very long against oncoming currents faster than 0.3 m/s.
Creutzberg (1961) showed in the laboratory that glass eels exhibit positive rheotaxis in currents of
0.2 m/s. It is in fact at these speeds that glass eels can be first observed in the water column (figure
2.1). When ebb tide currents are faster than 0.36 m/s, they swim close to the bottom or else burrow. The
former tends to occur when the bed is sandy whereas burrowing is preferred when the substrate
consists of gravel.
4 See § .
42

Density densité g/100m³
m3
7,00
6,00
5,00
4,00
3,00
2,00
1,00
0,00
débi t : 440
Flow
coef f : 97
Coeff
BM : 21h18
HT PM : 3h46
LT
1,20
1,00
0,80
0,60
0,40
0,20
0,00
-0,20
-0,40
00:37:03
00:59:55
01:23:00
01:47:15
02:08:10
02:27:45
02:51:00
Speed vitesse m/s m/s
03:10:40
03:32:50
03:58:10
04:19:00
04:38:20
densité fond
Bottom density
vitesse courant
Current speed
Figure 2.1-
Eel glass densities in the water column on the Adour during the rising tide of
22/12/1999. Speeds below zero indicate a downstream current (from Prouzet et al.,
2003)
When flow rates are high, it may happen that the tide no longer spreads up the estuary. A
hydrodynamic blockage is produced with the current in the estuary still flowing downstream, even during
the incoming tide. Under these conditions, large numbers of glass eels can be seen pressed to the
banks of the lower part of the estuary, seeking calmer waters and this can lead to significant catches
with a simple scoop net during the flood. This phenomenon is very well known to fishers using this type
of gear. Often following a flood period, daily catches increase as soon as the water level begins to fall
(figure 2.2).
CPUE (kg/fishing trip)
Flow in m3.s¯¹
CPUE Push-nets ------- CPUE hand scoop-nets
Figure 2.2 -
Catch trends, by trip, with push nets and scoop nets, on the Adour during the
November 2001- March 2002 fishing period, from Prouzet et al. (2001).
43

These conditions of hydrodynamic blockage often cause arriving glass eels to accumulate at the
mouth of the estuary or at sea (Cantrelle, 1981; Elie and Rochard, 1994) a situation very similar to the
blockage effect caused by physical barriers such as estuarine dams (Laffaille et al., 2007).
The first glass eels can be observed at oncoming speeds of 0.2 m/s. A progressive increase in
density occurs with increasing speed, something which was generally observed during Indicang
fieldwork undertaken on the Adour, the Isle or the Loire.
Finally, the increase in the river flow seems to generate a significant intrusion of fresh water into
the marine environment which, in turn, seems to attract glass eels on the continental shelf to the estuary
mouth (Lecomte-Finiger, 1978; Cantrelle, 1981).
Tidal cycle and lunar cycle
Very simplistically, tidal and lunar cycles are related through Newton’s law of “universal
attraction”. According to Newton’s theory, the tidal cycle is explained by the gravitational pull of the
moon and the difference in oceanic water attraction between the visible and dark sides of the Earth. The
attraction force of the sun also plays a role, reinforcing or opposing the moon depending on the position
of the latter compared to the sun. When the two planetary bodies are in conjunction, i.e. during the new
moon, or in opposition, i.e. during the full moon, the attraction forces are added together and the
amplitude of tides is significant. During the quadratures, i.e. the quarter moon phases, when the two
planets are at right angles, the amplitude of the tides is very low. It should be noted that high and low
sea levels are out of phase from one day to the next by about 24 hours and 50 minutes and that the
lunar cycle lasts 27.7 days.
It is therefore difficult to dissociate these two cycles. Yet, their effects on glass eel catch
variations appear to be different (figure 2.3). Average catch abundance indices recorded on the Adour
for the month of January over the 1985-1993 period (Prouzet et al., 2001) in relation to the different
lunar phases show a very significant increase during the new moon but not during the full moon, despite
the fact that both phases have tides of very significant amplitude. These observations confirm those of
Fernandez and Vazquez (1978), Tzeng (1985) and those made on the Isle during the Indicang
programme 5 (Lauronce and Susperregui, 2006). Furthermore, cyclical rings on glass eel otoliths every 7
and 14 days suggest that the lunar rhythm may affect the species’ metabolism (Lee and Lee, 1989).
As mentioned previously, the effect of the tide is modulated by the intensity of the river flow. The
propagation speed of the tidal movement and the depth of its penetration into the estuary can only be
related to the tide coefficient at constant flow. It is not, therefore, a simple effect and it has to be
considered in the context of hydrological events.
5 Lauronce V., Susperregui N, 2006. Newsletter No2 – Bassin Gironde Garonne Dordogne, Migado/AADPPEDG, annex 7 of the
Indicang report, http://www.ifremer.fr/indicang
44

The lunar cycle has a completely different impact related to moonlight intensity, which is itself
modulated by cloud cover and by water turbidity (especially during the quadrature periods). These three
factors play a dominant role in the quantity of light that penetrates the water column and therefore on
the vertical migratory behaviour of glass eels, as we will see below. But one of the consequences is that
this lunar effect is not observed in highly turbid estuaries (Laffaille et al., 2007).
Mean CPUE
Lunar phase
Figure 2.3 -
Variation of an average index of catch abundance centred on average January
catches each year between 1985 and 1993 on the Adour as a function of the lunar
phase. NL: new moon; PQ: first quarter; PL: full moon; DQ: last quarter.
The simple effect of tidal movement at constant flow
The many observations made on the Adour since 1995, and also on the Isle and the Loire (during
work carried out in the Indicang project), show that glass eels migrating passively do not accumulate at,
but slightly downstream from, the tidal front (which corresponds to a hydrological boundary separating
upstream and downstream flows).
In fact, during the incoming tide when the tidal movement spreads, glass eels progressively
migrate upstream just behind the tidal front, a finding that has been confirmed in other estuaries
(Cantrelle, 1981; Sheldon and McCleave, 1985).
Active movements along the longitudinal axis seem to be acquired progressively as pigmentation
develops (Gascuel, 1987). This author differentiates “carried” migration in the lower part of the estuary
from “swam” migration in the upper part and in the river.
Various works highlight the absence of glass eels in the water column during ebb tide. They adopt
a benthic or demersal behaviour pattern and remain burrowed in the sediment or move up the estuary
45

close to the bottom (Creutzberg, 1961; Gascuel, 1987), which prevents them from being pushed back
downstream. They are then very difficult, if not impossible, to catch with fishing gear. Elie (1979) noted
some glass eel concentrations close to the surface during the ebb tide on the Loire and on the Vilaine.
This was not observed on the Adour during fieldwork undertaken from 1995 to 2002 (Prouzet et al.,
2003), nor in Sèvre Niortaise (Gascuel, 1985). For some authors (Gandolfi et al., 1984; Gascuel, 1987;
Sheldon and McCleave, 1985), the low-water slack, the high-water slack and the flood-ebb turning tide
are favourable periods for catches, particularly the high-water slack. However, the observations made in
recent years on watercourses in the Indicang network have not confirmed these findings.
Wind action
The wind causes a circulation of superficial waters and influences water-mixing, particularly fresh
and salt water. In the Mediterranean, when blowing onshore, the wind is the main contributor to the
penetration of marine waters into the coastal lagoons or the estuaries. It influences the water
temperature and thus can affect the behaviour of migratory individuals 6 .
The wind speed must exceed 10 m/s in order to activate the mixing of fresh and marine waters
(Lecomte-Finiger, 1978; Elie, 1979; Weber, 1986). Although professional fishers state that wind action
affects the fishing outcome, they did not identify it as a major factor in the Atlantic when asked about it
(DeCasamajor, 1998).
2.1.4.2. Estuary physico-chemistry
The estuary morphology affects tidal spread upstream and the river flow. The way in which the
two water masses mix is specific to each estuary and determines the length of the zones where haline
and thermal stratification of the waters occurs.
Climatic conditions (rainfall, air temperature, wind, cloud cover, etc.) also modify the physicochemical
characteristics of the waters. These fluctuations have a direct impact on glass eel migratory
behaviour and in particular on their presence near the surface or in the water column. These factors are
important as they can significantly affect sampling when the fishing gear used can only explore the
superficial part of the water column.
Influence of salinity variations
Glass eels moving from sea water to fresh water present physiological adjustments at several
levels: increase in tissue water content during migration (Charlon and Blanc, 1983); increase in
dissolved oxygen consumption (Fontaine and Callamand, 1943); thyroid hyperactivity resulting in
positive rheotropism, which varies according to the season (Fontaine, 1976); development of the
olfactory epithelium (Sorensen, 1984; Crnjar et al., 1992); modifications in the function and structure of
6 See § .
46

tissues involved in osmoregulatory mechanisms: branchiae and branchial epithelium, oesophageal and
intestinal walls (Heinz and Hansen, 1979).
Salinity is identified as a factor affecting the speed of the pigmentation process of individuals
(Briand et al., 2005a). In the laboratory, Crean et al. (2005) showed a clear change between nonpigmented
glass eels that prefer salt water and pigmented elvers that prefer fresh water. Despite their
strong osmoregulatory ability tested in the laboratory (Tosi et al., 1988; Ciccotti et al., 1993), many
authors have noted that individuals remain in the estuarine haline zone for a certain period of time,
which may last for 3 or 4 days. According to McGovern and McCarthy (1992), this explains the
discrepancy between highest catches and highest tide coefficients. The salinity front can constitute a
barrier below which glass eels accumulate on arrival in the estuary (Ciccotti et al., 1995).
Work undertaken by Castellanos and Diaz (2006) on the Oria within the Indicang project
framework provided interesting observations on glass eel catches in the estuary below the salinity front
(figure 2.4). They showed a significant increase in density below the halocline (between 3 and 5 metres
during the observation period) whilst above, in very desalinated water around 2 metres deep, density
was very low and even nil. It may also be noted that maximum glass eel density occurs when salinity is
at a maximum at depth, and that the density then decreases rapidly as soon as this maximum is
reached, an indication that the group of glass eels is moving rapidly upstream of the observation station.
Salinity in ppm
35
30
5.00
4.50
4.00
25
3.50
20
3.00
2.50
15
2.00
10
1.50
1.00
5
0.50
0
0.00
295 245 195 145 95
45
-5
Time in minutes before high tide
CPUE near the CPUE at depth Salinity 2m Salinity 3m Salinity 5m
g³/100m filtered
Figure 2.4 -
Relationship between the variation in glass eel density on the surface and at
depths of about 4 metres and the degree of salinity on the Oria. (modified
according to Castellanos and Diaz, 2006).
47

This illustrates the progressive adaptation of glass eels to fresh water and their attraction to this
non-salinity (and not especially to preferred currents) because the analysis of surface and bottom
currents shows that these both run upstream and are of similar force.
Effect of water temperature
Although it is recognised that temperature affects glass eel behaviour, the relationship remains
poorly understood. As for all poikilotherms, temperature significantly affects their metabolism and
therefore their energy expenditure. Its action depends on the biological development stage and on the
impact of other environmental factors (Tongiorgi et al., 1986; Ciccotti et al., 1993).
Observations made on the Adour during the 2001-2002 fishing season (Prouzet et al., 2003)
(figure 2.5) showed that there were no catches in the estuary for the 15 days when water temperature
fell below 6°C with a minimum of 2°C. The same result can be observed in other estuaries such as the
Couesnon (Lafaille et al., 2007). It seems that glass eels become inactive below 4-6°C (Deelder, 1958;
Elie and Rochard, 1994). On the other hand, it appears that above 10 - 12°C, glass eels start to migrate
actively upstream (Gascuel, 1986; White and Knights, 1997).
CPUE (kg/fishing trip)
CPUE push-nets (2) CPUE hand scoop net (1) T⁰ estuary (⁰C)
Figure 2.5 - Evolution of daily catch (in kg per fishing trip) and of water temperature on the
Adour during the 2001-2002 fishing season (from Prouzet et al., 2003).
However, the absolute thermal threshold has to be considered relative to sea water temperature.
It is more the thermal differential that matters (figure 2.6), because catches are made below 6°C in
estuaries further north. Thermal differentials greater than 5°C seem to inhibit upstream migration as
noted by McGovern and McCarthy (1992). But this inhibition seems to start at a difference of 3°C.
48

Temperature (⁰C)
Thermic differential (⁰C)
Sea temp Estuarine temp Thermic differential
Figure 2.6 - Evolution of the thermic differential between fresh and sea water on the Adour
during the 2001-2002 fishing season.
The speed of upstream migration is a function of the temperature and the pigmentation stage
(Ciccotti et al., 1993). For Lambert et al., (1995), glass eel sedentarisation in estuaries occurs when
temperatures are low, whilst upstream migration accelerates with high temperatures. Hence, estuarine
acclimatisation time also depends on temperature (Jellyman, 1979; Gascuel, 1985).
Effects of turbidity and quantity of light in the water column.
Due to the lucifugous behaviour of glass eels, the light intensity penetrating the water column
modifies the behaviour of individuals and hence their vulnerability to fishing gear and sampling. As
Bardonnet et al. (2005a) showed, these young individuals are particularly sensitive to low light
intensities. These authors also estimated that light avoidance thresholds are around 10 -11 W.cm -2 for
non-pigmented glass eels (stage V B ) and around 10 -10 to 10 -8 W.cm -2 for later stages (VI A0 to VI A3 ). Thus,
in the majority of estuaries, anadromous migration occurs at night and near the surface, although it can
occur by day particularly in very turbid waters such as those of the Couesnon estuary (Laffaille et al.,
2007).
The impact of light intensity on the vertical migratory behaviour of glass eels has been studied
under experimental conditions or in situ in the Adour and Nivelle catchments (De Casamajor et al.,
1999; Prouzet et al,. 2003. Bardonnet et al., 2005a). In situ analysis of the glass eel density distribution
in the water column shows that during the period of the new moon (masked moon), glass eels are
distributed throughout the water column and are to be found in the layer covered by the surface net
around 1.5 metre deep, regardless of turbidity. When waters are clear (30 to 35 NTU) and in the new
49

moon, this corresponds to a light intensity less than 10 -11 W.cm -2 . Graphs relating light intensity to depth
for various hydroclimatic conditions and lunar phases are supplied by Bardonnet et al., (2005a). During
the quadratures (first and last moon quarters), turbidity and cloud cover are the major determinants of
the presence of glass eels near the surface. When turbidity is high (over 100 NTU), glass eels are found
throughout the water column regardless of the lunar phase, whilst when waters are clear, glass eels
avoid waters between 0 and 2 to 3 metres deep.
Chemical action of flows
The factor(s) attracting glass eels into the estuary continue(s) to be debated: decreased salinity
and temperature, presence of dissolved substances in the fresh water, combination of amino-acids
originating from estuarine vegetal populations (Lecomte-Finiger, 1978; Cantrelle, 1981; Gascuel, 1987;
Tosi and Sola, 1993).
2.2. Fluvial recruitment and river catchment colonisation
2.2.1. Characteristics of the yellow eel stage
The term “Yellow eel” (table 2.1) describes the linking stage between the two migratory marine
stages i.e. glass eels and silver eels. However, whether in terms of morphological or behavioural
characteristics, the limits between these three stages are less clear cut than they appear initially.
Table 2.1 -
Synoptic limits of class sizes of yellow eels that can be sampled in inland waters.
50-150 mm) Individuals in their first or second year of inland life.
150-300 mm) Growing individuals (2 to 6 inland summers of growth depending on sites and
individuals).
300-450 mm) Male individuals possibly silvering or female individuals in period of growth
450-600 mm) Female individuals possibly silvering
Small sizes (150 – 400 g) generally associated with shallow environments.
600-750 mm) Female individuals possibly silvering
Average sizes (400 - 800 g).
Over 750 mm
Female individuals possibly silvering
Large sizes (more than 800 g) generally associated with deep environments.
Various observations confirm the diversity of yellow eel behaviour. Monitoring of pass devices
located in the downstream zone of river catchments records mainly yellow migrants smaller than 300
mm (Opuszynski, 1965; Tesch, 1977; Naismith and Knights, 1988; Legault, 1994; White and Knights,
1997b; Laffaille et al., 2000; Legault et al., 2004). Sizes up to 650 mm were recorded on devices located
further upstream in the catchments (Legault and Feunteun, 1992; Baras et al., 1994; Baras et al., 1996;
Logrami, pers. com.).
50

Vollestad and Jonsson (1988) caught yellow eels behaving differently in that they were migrating
down the river Imsa. Adam (1997) observed the same phenomenon at the exit of the lake of Grand-
Lieu, as did Chadwick et al., (2007) on a small Scottish river. Knights et al. (1996) concluded, moreover,
that yellow eels could cover significant distances and even change environment because they found
that animals tagged in fresh water could be recaptured in the coastal zone, over 20 km away. Daverat
and Tomas (2006) identified this behaviour pattern from their observations on eel runs in the Gironde
based on the microchemical analysis of the otoliths.
Finally, tag-monitoring shows the limited movement of some individuals over a period of several
months (LaBar et al.,, 1983; LaBar et al., 1987; Dutil et al., 1988; Adam and Elie; 1994; Parker, 1995;
Baras et al., 1998; Laffaille et al., 2005a) whilst, at the same time, the majority of individuals have
disappeared from the initial observation site. Leaving aside the implausible hypothesis that high
mortality affected all of these monitoring efforts, it seems logical to think that these individuals changed
their whereabouts within these habitats (Rodriguez, 2002), which, incidentally, could resemble renewed
migration when these phenomena involve a significant number of fishes at the same time.
According to some authors, these different but co-existing behavioural patterns could be linked to
sedentary and nomadic individuals (Feunteun et al.,2003). For others, these two patterns may exist and
alternate in all individuals, expressing themselves in times of environmental crisis and/or in well defined
temporal slots. For example, Baisez (2001) in marshes and Ximenes (1986) in lagoons showed that an
individual can, from time to time, change its activity zone under the influence of various factors such as
the lack of food, aggression, water quality deterioration, water movements, etc. This hypothesis is
probably not specific to eels as in many animal species (Van Baalen and Hochberg, 2001), individual
histories seem to result from a succession of key moments triggering rapid behavioural responses
(usually “stay or leave”) related to the status of the individual at that precise moment (energetic state,
physiological state, etc.).
2.2.2. The various recruitment levels of a catchment
Estuarine recruitment or total catchment recruitment consists of the glass eel influx into the
estuarine zone. Such recruitment is highly dependent on the general state of the European eel stock
(Dekker, 2000) and/or on oceanic conditions for reproduction and larval migration (Knights, 2003). The
geographical location of a river catchment in the inland distribution zone and its main characteristics
(flow and water quality in particular) also have an important impact on the size of this estuarine
recruitment 7 .
Natural mortality, the sedentarisation of some individuals in tidal zones and in coastal wetlands,
fishing and other anthropogenic mortalities (Briand et al., 2003, 2005b and 2005c) then determine the
7
See § , p.43
51

fraction of this estuarine recruitment which leaves the zone under the influence of the dynamic tide and
comprises the fluvial recruitment.
2.2.3. Passage into fresh water
As mentioned previously, glass eel progression in the estuarine zone relies mainly on using the
flood tide currents (McCleave and Klechner, 1982; Gascuel, 1986; McCleave and Wippelhauser, 1987;
Elie and Rochard, 1994; Beaulaton and Castelnaud, 2005). Older descriptions also attest to their
capacity to progress in shoals or in huge columns against the current in the freshwater tidal zone during
some or all of the ebb tide (in Deelder, 1970; in Tesch, 1977). Tidal transport ceases in the zones
demarcating the limits of the dynamic tide. Swimming against the current then becomes imperative in
order to continue the progression upstream. Therefore, these particular zones are the site of
behavioural changes, with most notably the development of a more benthic behaviour and the
appearance of density-dependent phenomena (Bardonnet et al., 2005a; Bardonnet et al., 2005b; Briand
et al., 2005c; Edeline, 2005).
The triggering of migration beyond this limit seems to be significantly related to temperature, river
flow (particularly attraction flow) and tide coefficients (Jellyman and Ryan, 1983; Tongiorgi et al., 1986;
White and Knights; 1997b). No significant migration occurs until the temperature reaches 6 to 9°C;
estimates of this temperature threshold vary by site. White and Knights (1997b) highlight the major role
played by significant temperature variations in triggering migration. Authors observing migration on
dams equipped with bristle passes report higher triggering temperatures (between 12 and 15°C). This
discrepancy is most probably linked to the fact that the animal comes out of the water in order to
progress. The difference between water and air temperature must therefore also be taken into
consideration.
2.2.4. Upstream progression
2.2.4.1. Main characteristics of the individuals concerned.
Swimming against the current over long distances or climbing fish ramps mainly concerns
individuals with significant thyroidal activity and high condition (or body condition) coefficients (Edeline
et al., 2006).
Whether in the tidal zone or further upstream, their concentration along the banks or their
formation into columns (Cairn, 1941 in Cantrelle, 1984) reflects the identical response of individuals
(glass eels/elvers or yellow eels of various sizes) to the same constraint. This shoaling behaviour allows
progression whilst minimizing energy expenditure. The massive nature of these columns has
disappeared with the heavy fall in recruitment, which must have some impact on the colonisation
dynamics of river catchments.
52

The size of individuals observed transiting through fish passes ranges from 80 to 650 mm. The
absence of larger sizes may be related to their lower abundance but also to the selective character of
some passes (Legault, 1992).
Observations made on estuarine passes (Legault, 1994; Briand and Boussion, 1997; Anonymous,
2003) or during estuarine dam operations (Laffaille et al., 2007) reveal the huge numerical dominance of
elvers (95 to 98% of the total count).
Practically all available results from monitoring over the past 20 years highlight the rapid
disappearance of these elvers at a short distance upstream of the tidal limit (Moriarty, 1986a;
Aprahamian, 1988; Naismith and Knights, 1993; Laffaille et al., 2000; Feunteun et al.,2003; Legault et
al., 2004; Briand et al., 2005b). However, in 1982, in the Vilaine river catchment, despite the presence of
the Arzal dam in the estuary, Elie and Rigaud (1984) identified VI A4 stages in September-October
immediately downstream from Rennes, 130 km from the estuary and upstream of 7 barriers. Their
upstream limit had regularly progressed over the course of the summer. Some documents also attest to
the past presence of small individuals (less than 150 mm) far upstream in the river catchments, for
example at 240 km from the tidal limit on the Vienne (Legault, 1996). Oral accounts (Steinbach, pers.
com.) even confirm the presence of a large quantity of these small individuals 25-30 years ago, around
Orléans (about 300 km from the tidal limit of the Loire). And Feunteun et al. (2000a) note the anecdotal
presence of small individuals over 80 km from the sea on the Rhône.
The evolution of the average sizes currently observed on fish passes further and further away
from the tidal limit confirms these findings. The average size is, for example, 130 mm on the Frémur, 6
km away from the tidal limit (Laffaille et al., 2000), 220 to 270 mm on the Dordogne 70 km away
(Legault, 1992; Pallo and Travade, 2001), 280 mm on the Garonne 80 km away (Legault and Feunteun,
1992) and 300 mm on the Meuse 250 km away (Baras et al., 1994).
In conclusion, today the very great majority of small individuals quickly disappear upstream of the
tidal limit, most probably due to the fall in fluvial recruitment.
Migratory eel monitoring at a given site reveals that their average size usually decreases over the
course of the season. This is probably related to the early passage of older individuals waiting
immediately downstream of the barrier, followed by that of eels coming from further downstream. This is
the case for example for the smallest individuals on stations which are close to the sea (Moriarty, 1986b;
Baras et al., 1994; Pallo and Travade, 2001; White and Knights, 1997a).
2.2.4.2. Progression kinetics
As with estuarine dams, (Laffaille et al., 2007), significant passages over barriers located further
upstream may be observed over short periods of time, generally 3 to 6 weeks, especially concerning
young individuals (Deelder, 1958; Moriarty, 1986b; Legault and Feunteun, 1992; Baras et al., 1994;
53

Carry and Delpeyroux, 2003). The passages of older individuals, in much smaller numbers, are spread
between May and October (Moriarty, 1986b).
On a daily scale, upstream progression is essentially nocturnal (Tesch, 1977; Baras et al., 1996)
except, once again, for the youngest individuals (Baras et al., 1996) with less lucifugous behaviour
(Sörensen, 1950 in Deelder, 1984).
In the majority of cases, estimates of active swimming progression speed vary between 10 and
45 km per year (Hussein, 1981; Moriarty, 1986b; Aprahamian, 1988; Mann and Blackburn, 1991; Baras
et al., 1996). Briand et al. (2005b) observed a slightly higher annual progression (50 km) for migrants
transiting through the Vilaine estuarine pass. For the youngest individuals, Baras et al. (1996)
estimated that the migration speed could reach 75 km per year due to migration being both diurnal and
nocturnal. However, these averages hide strong inter-individual variability in progression speeds.
Feunteun et al. (2003) noted the existence of very fast individuals that could travel up to 200 km per
year whilst slower ones migrated at around 50 km per year, even with no major barrier to their migration.
2.2.4.3. Impact of barriers on this progression
All these estimates must be set within their respective physical context taking into account in
particular river slopes and barrier density.
Thus, on very steep rivers in the north of Spain (1.4% average slope), Lobon-Cervía et al. (1995)
noted much lower propagation speeds than those stated previously. When comparing migration
kinetics on the Dee and the Severn, Aprahamian (1988) also mentions that this parameter may strongly
contribute to the differences observed.
When they exist, estuarine constructions are of course the primary elements disrupting
colonisation as they create a physical barrier to the free upstream movement of individuals and modify
interactions between fresh and salt water masses (Gascuel, 1986; Laffaille et al.,2007). When rivers are
at flood stage in particular, only the highest tides will allow individuals being carried by the flow to
position themselves immediately downstream of these constructions. In the case of an estuarine barrier,
inappropriate management (the total absence of downstream-upstream water passage) and/or illadapted
equipment (no pass or no dam operations) create the conditions to block potential migrants,
leading to significant mortality from predation (birds, carnivorous fishes) and the possible development
of very efficient fisheries downstream. For example, monitoring downstream of the Arzal dam in the
Vilaine estuary (a 10,000 km² river catchment) showed that, during each night's fishing, the push-nets in
use could jointly filtrate the total fishing zone volume 5 to 7 times. These studies estimated the
harvesting rate of the annual glass eel flux at 97% (Briand et al., 2003) in this type of developed
estuarine context, notwithstanding the existence of an eel pass. In fact, this pass only becomes effective
from April onwards (Briand et al., 2005c).
54

Attempts to increase the permeability of these estuarine constructions have shown the
complementarity over time of two strategies. The first strategy consists of opening the sluice gates for
thirty minutes to an hour during the nocturnal slack water of spring tides (Legault, 1990) when tidal and
river flow conditions mean that glass eels are present immediately beneath the dams. This strategy can
be very effective for glass eels in the passive migratory phase (Laffaille et al., 2007). The second
strategy consists of equipping the construction with an eel and glass eel pass (Legault, 1992). This type
of device generally only becomes operational from April-May when water is released and when glass
eels begin to swim actively (Briand et al., 2005c). This strategy must therefore replace the former during
the spring and summer periods. Finally, when operating the dam is impossible during the winter and
spring, it is also possible to plan fishing or trapping operations below the construction followed by
releasing operations directly upstream of it (Rosell et al., 2005). This trap and transport procedure must
of course use methods which guarantee maximum survival of individuals (fishing techniques, season,
release conditions) and is only recommended as a last resort.
Over the rest of the network, further upstream, barriers contribute to restricting the areas where
eels can be found. White and Knights (1997a) noted the strong impact of successive barriers over a
short distance along an axis. Briand et al. (2005b) also showed that eel density in the Vilaine catchment
was more closely correlated with the number of dams downstream than with the distance to the tidal
limit. They lead to total blockages or successive delays to upstream migration along an axis (Elie and
Rigaud, 1984; Feunteun et al., 1998; Briand et al., 2005c) and ultimately to the declining presence of the
species in the upstream reaches. Furthermore, during this more or less lengthy blockage at the base of
a barrier, natural and anthropogenic mortality increases (predation and/or fishing) particularly though the
emergence of competitive and cannibalistic behaviours that are well documented in the context of high
density populations (Sinha and Jones, 1967; Knights, 1987; Lamothe et al., 2000).
Using fish-pass traps set up on the various barriers along an observation axis, White and Knights
(1997a) concluded that migratory individuals, during the short favourable spring period when there is an
attraction flow, pass though a variable number of dams depending on their passability level then stop.
They only recommence their progression the following year after a growth period. This upstream
progression, spread over several years, naturally leads to an observed average age of the animals
which increases with the distance to the tidal limit (Deelder, 1984; Moriarty, 1986b; Aprahamian, 1988;
Vollestad, 1988; Barak and Masson, 1992; Tzeng et al.,1995; Ibboston et al., 2002).
Finally, it is often noted that, in the past, the presence of many non-equipped barriers along river
corridors was not incompatible with the maintenance of a very large distribution area. Two points should
be made about this observation:
• First, although barriers of limited size have long existed in the catchment meanders, in the 1960s
large barriers often appeared in downstream areas and on major axes with an impact that is easy to
imagine, given their position and the absence of specific equipment or management.
55

• Second, the characteristics of old dams and mills have developed significantly. Over at least the
past 20-25 years, they have been modernised (for example, waterproof metallic doors have
replaced wooden sluice gates) and the management of river flows has changed greatly (little or no
summer flows, massive and brutal winter releases).
These changes have greatly affected the roughness of many river axes as regards colonisation
migration by reducing favourable summer periods with attraction flows and by affecting the passability of
barriers especially as recruitment and escapement levels in estuarine zones have declined to low levels.
2.2.4.4. Intensity of the phenomenon
Recruitment indices can be obtained from fish-trap facilities located on barriers at different levels
of the river catchment (Legault, 1994). While the fish passes located at the estuary exit can reveal all or
part of the fluvial recruitment in the catchment (Briand et al., 2005c), those located along the axis show
the evolution in migrant abundance with increasing distance from the tidal limit (White and Knights,
1997a).
On a fish pass very close to the Shannon estuary (Ireland), 11 to 35 individuals per km² of
upstream catchment were recorded between 1973 and 1983 (Moriarty, 1986b). Around the same time, it
varied between 29 and 382 individuals per km² on the Imsa in Norway (Vollestad and Jonsson, 1988).
More recently in 1992, on a small Breton river (the Arguenon), Legault (1994) estimated the annual
number of passages at 531 individuals per km² a few kilometres away from the sea. On the Vilaine
estuarine dam, Briand et al. (2005b) observed that, between 1998 and 2003, the numbers of individuals
passing through varied between 20 and 240 individuals per km². Similarly, the range was between 50
and more than 500 individuals per km² on the Frémur between 1997 and 2003 (Legault et al., 2004).
The range is narrower on the Enfreneaux dam in the Sèvre Niortaise estuary where the number of
passages recorded between 2000 and 2002 fluctuated from 60 to 120 individuals per km² (Anonymous,
2003). These indices, calculated from fish-traps very close to the estuarine zone, give some idea of the
range of ratios that can be found in the most downstream reaches but also reveal a high inter-annual
and inter-site variability.
Further upstream, the ratios drop significantly. Hence, Pallo and Travade (2001) on the Dordogne
and Carry and Delpeyroux (2003) on the Garonne observed 1 to 3 individuals per km² of upstream
catchment on the first barriers located about 250 km from the entry of the estuary and 80-90 km from
the tidal limits of these two rivers. Several hypotheses have been put forward to explain this
phenomenon. The first hypothesis concerns a progressive colonisation of the catchment as downstream
sites become increasingly saturated. The speed of progression of yellow eels in the river catchment
then depends on the trophic and physical conditions encountered (Briand et al., 2005b). Hence, fluvial
recruitment intensity may have a sizeable impact on migrant distribution within the river catchment
through behavioural patterns adapted to the density of individuals in situ (Ibboston et al., 2002; Lasne
56

and Laffaille, 2008). The second hypothesis suggests a distribution of individuals according to their
physiological profile and their physical capacities (Feunteun et al., 2003; Edeline et al., 2004; Edeline,
2005). The final hypothesis proposes that the distribution of individuals in the different catchment
compartments is quite homogeneous, and explains falling density (number of individuals per m²) simply
in terms of the “dilution” of the colonizing flux as the watercourse increases in importance once the
upstream meanders are reached.
2.3. Sedentarisation phases of variable onset and duration
2.3.1. Coastal or estuarine sedentarisation
Extensive eel production from fish ditches (marshes or catchment) has often been described
(Labourg, 1976; Rossi and Colombo; 1979; Arias and Drake, 1985; Massé and Rigaud, 1998; Feunteun
et al., 1999) but it was often interpreted to be the result of individuals being trapped. Yet, the highly
significant presence of the species had also long been noted in open saline environments such as
saline estuaries or lagoons (Deelder, 1958) but here also, for a long time it was suggested that this
concerned above all individuals in transit. It was only from observing the particular growth rhythms on
otoliths (Mounaix and Fontenelle, 1994), and the unique microchemistry of their bony structures (Tzeng
et al., 1997; Tsukamoto et al., 1998; Daverat et al., 2006) and through tag-based monitoring (Parker,
1995; Morrison and Secor, 2004) that it was possible to clearly identify individuals which grow in the
estuary and even in saline zones.
However, Daverat et al. (2006) showed that for Anguilla anguilla, as for A. japonica or A. rostrata,
the complete absence of contact with fresh water only concerns a relatively small number of individuals.
The large majority of eels observed in salt or brackish water are therefore probably individuals that have
resided in fresh water for the first 3 to 5 years of their inland life or that periodically move between
environments of different salinities.
2.3.2. Sedentarisation and territory prospection
Studies in deep environments were undertaken on a limited number of individuals (LaBar and
Facey, 1983; Bozeman et al., 1985; LaBar et al., 1987; Dutil et al., 1988; Parker, 1995, Baras et al.,
1998). They showed a significant predilection for a certain type of context (salt, tidal fresh-water zone,
etc.) even for a particular burrow. To the best of our knowledge, tag analysis and subsequent monitoring
of a large number of individuals has never been undertaken in these deep environments (estuaries or
rivers), certainly because of the very high cost involved in operations of this type in such systems.
This kind of monitoring in the canals of the Breton marshes (Baisez, 2001) showed that the
position of summer burrows depended strongly on the size of individuals (very close relationship
between the water level in the summer burrow and the size of the animal). Similar results were found on
57

a small watercourse in North Brittany where the 20% of animals that were recaptured at least once
were, in 90% of cases, recaptured on the site of their initial tagging during the 8-year-monitoring period
(Laffaille et al., 2005a). Such observations agree with the findings of Lobon-Cervía et al. (1990) in
Spain concerning the very slow recolonisation process of stations that follows a period of withdrawal.
Estimates of the extent of the “territory” explored around its burrow by a yellow eel that is
considered to be sedentarised vary significantly according to the environment. Mark-recapture
experiments led Mann (1965) to estimate it to be around 40 km in rivers. Telemetric monitoring over 20
to 30 days gave much lower estimates in lakes or rivers. Hence, LaBar et al. (1987) assess it to be
between 1,300 and 2,700 m² in a small Spanish lake. Baras et al. (1998), in a small river, found the
average explored territory to be from 100 to 150 m². Through PIT-tagging and monitoring over 2 years in
small size canals, Baisez (2001) estimated the average territory of an individual to be 300m of river
bank and around 1,000m² with the size of the area covered increasing with the size of the individuals.
2.3.3. Activity rhythms
Yellow eel activity follows essentially a seasonal cycle with a day-night rhythm (Neveu, 1981a;
Jellyman and Sykes, 2003).
The intensity of young yellow eel movement in the river is also strongly correlated with
temperature and particularly with rising spring temperatures which act as a trigger (Moriarty, 1986a;
Naismith and Knights, (1988). They are not very active when temperatures fall below 12-13°C or in high
summer temperatures which are often associated with a nocturnal fall in oxygen concentration (Baras et
al., 1998; Baisez, 2001). During these slower or interrupted activity phases, eels shelter in hiding places
or burrows. Winter hiding places are mostly located in deep zones and in muddy areas. Activity peaks
observed, in particular during passive gear monitoring, are practically synchronous with those recorded
during the monitoring of dam fish passes (Naismith and Knights, 1988) or by telemetry (Baras et
al.,1998).
Yellow eel activity is also very dependant on the day-night cycle, with a predominantly nocturnal
character. This finding is supported by the majority of studies, particularly those on feeding rhythms
(Neveu, 1981a; De Nie, 1987; Belpaire et al., 1992; Baras et al., 1998; Baisez, 2001). However some
minor variations do occur. Small yellow eels (less than 150 mm long), particularly in their active
colonisation phase, may undertake very significant diurnal activity (Tesch, 1977). Baisez (2001) also
noted that environmental conditions could induce a temporary modification in the nychthemeral rhythm
of eels over 250mm long. She also noted that in marshes, during summer nocturnal phases when
oxygen availability is problematical, dominant activity seems to shift to the day time and even to the
middle of the afternoon. Finally, these bespoke monitoring programmes also showed the existence of
individuals with strong diurnal behaviour although these were not in the majority over the whole year.
58

Some authors also highlight the joint influence of the nychthemeral rhythm, the lunar cycle and
water turbidity suggesting that yellow eels are especially sensitive to the amount of light they actually
detect in the water mass (Adam and Elie, 1994; Baisez, 2001), as also seems to be the case for glass
eels 8 (Bardonnet et al., 2005a; Laffaille et al., 2007).
2.3.4. Feeding behaviour
Eels are characterised by their great ability to adapt to the environmental trophic supply, with their
diet spectrum evolving as they grow, from plankton to crustaceans and insects then to fish (Bergersen
and Klemetsen, 1988; Costa et al., 1992; Michel and Oberdorff, 1995; Radke and Eckmann, 1996;
Schulze et al., 2004). The size of their prey seems to depend primarily on the size of their mouth
opening (Lammens and Visser, 1989; Costa et al., 1992).
This ability to profit from the various trophic levels of the environment explains why this fish,
despite its status of being both carnivorous and piscivorous, constituted, in years of high abundance,
around 50% of the total biomass in sites close to the sea, whether in lentic waters (250 to 300 kg of
eels/ha in brackish marshes according to Feunteun (1994)) or in flowing waters (around 110 kg/ha in
the flowing waters of Basse Nivelle in the Pyrénées Atlantiques according to Neveu (1981b)). Fish begin
to appear episodically as a prey when eels are between 250 and 300 mm. Ichtyophagy generally
becomes dominant in the inland environment between 400 and 500 mm (Bergersen and Klemetsen,
1988; Barak and Masson, 1992; de Nie, 1987; Schulze et al., 2004). In saline or brackish zones, the
significant abundance of crustaceans minimises the importance of this ichtyophagy even in large
individuals (Rigaud, unpublished data). It is important to differentiate this natural ichtyophagy which
appears during the animal’s development from the aggressive and cannibalistic behaviours often
observed in high-density conditions (Knights, 1987) or in environments which are poor in trophic
resources (de Nie, 1987).
2.3.5. Abundance levels
2.3.5.1. Macro-distribution
In addition to the monitoring of fish passes, electrofishing undertaken across the hydrographic
network generates interesting information on the evolution of the species abundance along a
watercourse.
Thus, on a given axis, the probability of observing an eel on a station increases with the distance
from the source. It is also influenced by the number and the passability of barriers downstream of the
8 See § , p.43
59

station (Feunteun et al., 1998), by its altitude and by the surface area of the catchment upstream
(Oberdorff et al., 2001).
In France, various authors (Elie and Rigaud, 1984; Legault, 1987) have noted this decrease in
abundance and the increase in average size when moving up floodplain river catchments (respectively
the Vilaine and the Sèvre Niortaise). These observations were associated with a highly disrupted
context, due in particular to estuarine hydraulic constructions. Lobon-Cervía et al. (1995) on a short (50
to 60 km) and steep (1.4%) river in Galicia and Naismith and Knights (1993) on the Thames observed
the same phenomenon.
Smogor et al. (1995), comparing the migration of American eels to a dispersion phenomenon
along catchment watercourses, suggested that abundance changed as a function of the distance to the
tidal limit, following the declining half of a normal distribution. They validated this hypothesis for
individuals between 61 and 374 mm long in four river catchments of Virginia (United States). Ibboston et
al. (2002) studying the European eel in 2 British river catchments reached similar conclusions, namely
that abundance by age group declined following a negative exponential. However, Laffaille et al. (2003)
showed that this was not the case on small watercourses, particularly on the Frémur (North Brittany)
which is 17 km long and highly populated with eels. On the other hand, they did confirm it along the
barrier-free Loire, with a rapid decrease in abundance indices relating to individuals smaller than 300
mm and a near total absence of these individuals 100 km from the tidal limit (Lasne and Laffaille, 2008).
Overall, the great majority of “recent” data shows rapidly decreasing eel abundance beyond the
first 80-100 kilometres upstream of the tidal zone. However, it is important to note that all these studies
were carried out after 1985, during a period when the collapse in recruitment had already been apparent
for at least 5 years. Similar descriptions of this phenomenon are not available from 50 years ago, at a
time when the species’ abundance and recruitment levels were much higher.
2.3.5.2. Impact of more local factors (meso- and microdistributions)
It should first be noted that there is an almost complete absence of qualitative and quantitative
data on population fractions in the main and deep axes and in deep unexploited waterbodies. This point
was highlighted in France by the Eel Group as early as 1984. Yet these deep zones seem to be
interesting compartments likely to produce the large individuals observed during downstream migratory
phases. At least, this is what is noted by Oliveira et al. (2001) as regards the American eel.
In shallow meanders, monitored by electrofishing, the factors mentioned previously (distance to
the tidal limit, altitude, etc.) are unable to explain all the observations, especially downstream of the
catchments. The effects of local factors such as the water level, the current intensity, the substrate
granulometry, the type and density of aquatic vegetation on the station have therefore been tested. In
the majority of cases, the results obtained in rivers are inconclusive (Smogor et al., 1995), doubtless
60

ecause the approaches used do not take into account the nesting of the spatial structures mentioned
previously, with abundance depending on the distance to the tidal limit and the passability of barriers as
well as, at a given location along the river, a heterogeneous distribution of individuals within the various
available habitats (Laffaille et al., 2003).
Feunteun et al. (1998), examining the results obtained over short stretches of the river, also noted
this heterogeneous eel distribution, even within each successive stretch. This finding tends to confirm
the impact of local factors and complicates of course the determination of a density or a biomass in a
given stretch of the river. It also shows the importance of station choice.
Work undertaken on zones located relatively close to the sea also shows the significant influence
of these local factors and the relationships between the size of individuals and their preferred habitat
(Baisez, 2001; Laffaille et al., 2003; Laffaille et al., 2004). Hence, the level of water or silt (in wetlands),
vegetation abundance and the number of shelters affect size distribution within the environment. Small
individuals (less than 150 mm, or sometimes less than 300 mm) seek shallow zones (riffles, runs,
banks, etc.) offering lots of shelter (stones, vegetation, branches, etc.). Some authors have also noted
the marked preference of large eels for deeper waters (Sloane, 1984; Aprahamian, 1988; Baisez; 2001;
Laffaille et al., 2003; Laffaille et al., 2004). This change in preference seems to occur around 300-350
mm.
These observations suggest the need for a data analysis by size class taking into account the
nature of the habitats concerned.
Finally, it is important to remember that the population fraction observed in a given station is the
result of several cumulative years of colonisation, each of which occurs in a particular context
(recruitment level, river flow, opening of dams, fishing pressure, etc.). This temporal dimension must
always be integrated into the data analysis, increasing its complexity (see for example Briand et al.,
2005b).
2.3.6. Quality of observed individuals
In addition to abundance analyses, it is important to take into account the quality of individuals
produced by each of the catchment compartments. Size structures, eel sizes, sex ratios, growth
rhythms, and levels of parasitic or chemical contamination (heavy metals, PCBs, pesticides, brominated
flame retardants, etc.) are all indicative of this quality in the various areas under observation.
2.3.6.1. Size-weight relationship
Analysis of the size-weight relationship was first developed mainly so that, after a calibration phase in a
given area, systematic collection of weight data in the field could be avoided, size data being easier to
61

collect 9 . Table 2.2 shows the variability of observations collected in various contexts in Europe, based
on 13 references covering 17 sites (11 coastal and 6 fluvial).
Authors Year Country Environment N r 2
Table 2.2 - Size-weight relationship (weight = a x size b ) observed in various European sites.
(cm) A b
Size range
Vollestad 1986 Norway Bay 274 0,98 30-80 0,50 10 -3 3,29
Lee 1979 France Dyked salt 55 30-75 1,06 10 -3 3,14
marshes
Melia et al. 2006 France Lagoon 18293 0,98 6-78 0,24 10 -3 3,36
Svedäng 1999 Sweden Western coastline 1924 0,93 29-90 0,36 10 -3 3,21
Mounaix 1992 France Vilaine - Atlantic 289 0,99 22-84 0,43 10 -3 3,22
salt estuary
Mounaix 1992 France Loire - Atlantic salt 91 0,98 21-72 0,16 10 -3 3,38
estuary
Rossi and Villani 1980 Italy Lesina Lagoon 444 7-65 0,71 10 -3 3,21
Neveu 1981 France Basque river 35 0,96 6-70 2,00 10 -3 3,00
Carss et al. 1999 Scotland River and lake 332 10-60 0,86 10 -3 3,18
Mounaix 1992 France Vilaine rivers and 280 0,99 13-82 0,83 10 -3 3,21
streams
Mounaix 1992 France Loire river 130 0,98 17-84 0,23 10 -3 3,04
Arias and Drake 1985 Spain Salt marshes 1489 15-100 1,45 10 -3 3,06
Fernandez- 1989 Spain Atlantic salt 51 0,98 24-70 1,23 10 -3 3,08
Delgado et al.
estuary
Lobon-Cervía 1995 Spain Cantabrian river 435 0,97 24-40 0,54 10 -3 3,33
et al.
Rasmussen and 1979 Denmark River 358 7-60 1,59 10 -3 3,02
Therkildsen
Rossi and 1976 Italy Comacchio 680 0,99 35-70 0,60 10 -3 3,28
Colombo
lagoons
Rossi and 1976 Italy Valle nuova 418 0,99 35-70 0,50 10 -3 3,32
Colombo
lagoons
Vollestad and
Jonsson
1988 Norway River and lake 1283 0,95 7-90 1,07 10 -3 3,07
In the great majority of cases the coefficient of determination (r²) is reported and is always very
high (> 0,93), indicating a strong relationship of type P= aL b between weight and size in a given site.
On the basis of these references, the average weight can be graphed against each length increment,
together with the corresponding standard deviations (figure 2.7). The two extreme relationships are also
included, with the maximum relationship found in a Mediterranean lagoon and the minimum relationship
in the western coastal area of Sweden. The observed weight differences increase with size and
become significant from 450-500 mm. However, the reproductive potential of females produced in these
different contexts must be taken into consideration.
9 See.Chapter 5.
62

Maxi
Weight (in g)
Mean
Length (in cm)
Figure 2.7 - Relationships between size (total length in cm) and weight (in g) observed in
Europe in saline environments, rivers or streams (13 references, see table 2.2).
Various authors (Ximenes, 1986; Lobon-Cervía et al., 1995) have also shown that there is a
seasonal evolution in the weight-size relationship observed on their study sites. The body weight of
individuals falls from October to April then increases before generally levelling off at the beginning of the
summer.
2.3.6.2. Condition factors
A relative condition factor (Kr) can be calculated for each individual, or each group of individuals,
under observation using the size-weight relationship observed in an area (distribution scale, river
catchment, sub-catchment or compartment). This factor Kr is equal to 1 if the observation fits perfectly
with the expected mean in the given area and is lower (or higher) than 1 if it is below (or above) this
mean. Eels observed by Melia et al. (2006) in Mediterranean lagoons have a Kr around 1.30 compared
with the mean performance observed across the 17 European sites (figure 2.7). Swedish eels observed
by Svëdang (1999) have a Kr around 0.76.
At another scale, in the Gironde-Garonne-Dordogne catchment (table 2.3), Lamaison (2005)
found that the relative condition factor Kr varied significantly according to the compartment with the
highest values in the salt estuaries.
63

Table 2.3 -
Relative condition factors observed in different contexts within the Gironde-
Garonne-Dordogne catchments (Lamaison, 2005).
Relevant compartment
Number
of
individual
s
Mean Kr
Standard
deviation
Salt estuary 313 1. 18 0.13
Tidal fluvial zones
Non-tidal fluvial zones
Observations on the first fluvial passes
Observations on the rivers
Garonne 231 1. 14 0.13
Dordogne 141 1. 12 0.13
Garonne 117 1. 13 0.12
Dordogne
Garonne 4259 1. 01 0.29
Dordogne 1009 0. 88 0.19
Garonne
tributaries
Dordogne
tributaries
128 1. 09 0.20
129 0. 95 0.23
Fulton’s condition index (K= 10 3 x P (in g) / L 3 (in cm)) (Ricker, 1980) must also be mentioned. It is
based on the hypothesis of a coefficient b equal to 3 in the size-weight relationship whilst the analysis of
the 17 sites mentioned previously gave a mean b of 3.23. Fulton’s index is therefore of limited interest,
this rough estimate giving significant differences with real life, especially for large sizes. However, it is
possible to take another value of b equal to 3 in this relationship using the method suggested by Bolger
and Connoly (1989) where b represents the slope of the linear regression between log weight and log
length taking into account all individuals.
2.3.6.3. Growth levels
The source of the wide diversity in eels’ growth performance is a complex and controversial topic.
The diversity in colonised environments (differences in salinities, in annual temperature profiles, in
environmental quality, etc.) and in individual growth potential are often highlighted. However, there are
also methodological issues related to the preparation and reading of otoliths, bony structures inside the
inner ear, where winter interruptions must be identified in order to estimate the age of individuals
(Fontenelle, 1991; Panfili et al., 1994).
General observations
The inland growth phase of the European yellow eel lasts between 3 and 15 years in the very
large majority of European sites.
64

There is usually a greater size gain in the first year of inland life than in the following ones.
Hence, Panfili and Ximenes (1994), collating observations from 29 sites in Europe, noted an average
increase of 9.4 cm (± 4.4 cm) during this first year. The results observed in purely extensive fattening of
elvers in small waterbodies (Belpaire, 1987; Belpaire et al., 1992; Klein-Bretteler, 1992) confirm these
figures, with a mean growth of 7-8 cm and maximum performance of 12-13 cm over a 6 month period
(from March to October). Similar results were obtained by Meunier (1994) who, over a period of 68
months, monitored glass eels marked by immersion then released into the Alsatian Rhine, and who
observed a mean growth of 10 cm during the first year and 5.5 cm per annum over the first 5 years.
Monitoring usually reveals, for a given site, practically linear mean growth at least over the first 6
to 7 years of inland life (Panfili and Ximenes, 1994; Adam, 1997), with an overall mean of around 5 cm
per annum. All these results are obtained either by comparing observed size with estimated age from
otolith reading or by retro-calculation techniques. This retro-calculation is based on the observation of a
linear relationship between the size of an individual and the size of its otolith (Rossi and Vilani, 1980;
Vollestad and Jonsson, 1988; Fernandez-Delgado et al., 1989; Mounaix, 1992; Panfili, 1993). Once
winter interruptions have been identified, this gives an estimate of the size an individual must have been
at each stage.
PIT-tags (Passive Integrated Transponder tags) used on a substantial number of eels have
furnished new information sources concerning individual growth in situ. Hence, in the Breton marshes,
considered to be fresh water, Baisez (2001) noted annual average size gains of 8cm (± 5cm) for 7
months of effective growth in the year and for individuals between 28 and 65cm (44cm on average).
Impact of individual factors
Regardless of the site studied, the growth pattern of eels placed or observed in the same living
conditions is very heterogeneous and this seems to be a characteristic of the species (Wickins, 1985;
Berg, 1990; Vollestad, 1992; Panfili et al., 1994). Meunier (1994), when monitoring the fate of marked
elvers released into the Rhine noted a size disparity between individuals (or ΔMax-Min) which increased
in a quasi-linear fashion with time (ΔMax-Min = 1.5 cm when released, ΔMax-Min = 10 cm after 20
months, ΔMax-Min = 22 cm after 55 months).
Sex is often put forward to explain in part this disparity in a given environment. Davey and
Jellyman (2005) produced a very comprehensive synopsis of knowledge concerning this issue. It seems
that in Anguilla anguilla, the initial growth rate is significantly related to the future sex of the individual
(Kuhlmann, 1976; Holmgren and Mosegaard, 1996; Holmgren et al., 1997). Hence future males have a
higher initial growth rate. But the situation quickly reverses itself with females then having a higher
specific growth rate (ΔL/L) (Davey and Jellyman, 2005) and greater absolute growth (Vollestad and
Jonsson, 1988). This may be linked to sexual differentiation which seems to appear quite early on
65

etween 15 and 30cm (Bienarz et al., 1981; Colombo et al., 1984; Colombo and Grandi, 1996; Melia et
al., 2006).
Purely to establish orders of magnitude, the data compiled by Adam (1997) in his synopsis of
annual mean growth rates of the two sexes at 48 European sites (table 2.4) was used. For males, the
mean growth rate recorded across the sites ranged from 5.3 cm per annum during the first 3 years to
1.7 cm per annum between the 7th and 9th years. Over the same period and for the same intervals,
female rates were higher (7.3 and 5.0 cm per annum respectively). This agrees with observations
showing male silver eels of smaller size than their female counterparts at the same age (Sinha and
Jones, 1967; Tesch, 1977; Vollestad and Jonsson, 1988; Fernandez-Delgado et al., 1989; Parker,
1992). This earlier and marked decrease in the male growth rate is probably linked to the departure of
fast-growing individuals which have become silver eels during the 3 to 9 summers spent in inland
waters. Vollestad (1992), in his synopsis on the age and size of silver eels in Europe, gave a mean male
age of 6 years (± 3 years) for 40.6cm (± 3.3cm) and a mean female age of nearly 9 years (± 4.5 years)
for 62.3cm (± 10.6cm).
Table 2.4 -
Analysis of mean growth performance observed at 48 European sites (references
from Adam (1997) with maximum, minimum and mean values from the sites after 3,
6 and 9 years of inland growth.
Yellow eels
Males
Mean size (MS)Mean growth (MG)
Females
Mean size (MS)Mean growth (MG)
3 years
inland
MS = 22 cm
MG = 5.3cm.year -1
Max. MS = 35cmMax.
MG = 9.3cm.year -1 MS = 28cm
Min. MS = 10cmMin. MG
MG = 7.3cm.year-1
= 1.5cm.year -1
Max. MS = 40cmMax MG
= 11.0cm.year -1
Min. MS = 14 cmMin. MG
= 2.3 cm
6 years
inland
MS = 34 cm
MG = 4.0 cm.year-1
Max. MS = 44 cmMax.
MG = 3.0 cm.year -1 MS = 45 cm
Min. MS = 20 cmMin. MG MG = 5.7 cm.year-1
= 3.3 cm.year -1
Max. MS = 60 cmMax.
MG = 6.7 cm.year -1
Min. MS = 30 cmMin. MG
= 5.3 cm.year -1
9 years
inland
MS = 39 cm
MG = 1.7 cm.year-1
Max. MS = 46 cmMax.
MG = 0.7 cm.year -1 MS = 60 cm
MG = 5.0 cm.year-1
Min. MS = 26 cmMin. MG
= 2.0 cm.year -1
Max. MS = 80 cmMax.
MG = 6.7 cm.year -1
Min. MS = 45 cmMin. MG
= 5.0 cm.year -1
The eel therefore displays sexual dimorphism based on size. Currently, a silver female is always
longer than 40 cm and any individual more than 45 cm is highly likely to be a female in practically all
European sites. Smaller individuals are generally males whose size varies between 25 and 45cm
(Lobon-Cervia and Carrascal, 1992; Vollestad, 1992; Kushnirov and Degani, 1995; Poole and Reynolds,
1996; Laffaille et al., 2006).
66

Impact of environmental factors
Growth performance also varies greatly with geographical sites (Sinha and Jones, 1975; Berg,
1989; Fontenelle, 1991; Panfili and Ximenes, 1994; Meunier, 1994). The analysis of mean, minimum
and maximum performance at European level (table 2.4) corroborates this. The annual temperature
profile of the site, the trophic and chemical quality of the environment and fish abundance (eel
abundance but also total multispecific abundance) are often among the factors cited (Fontenelle, 1991;
Vollestad, 1992; Aprahamian, 2000).
Within the same catchment, marked differences in growth performance are also observed
between the salt estuary, the fresh estuary, the river and the stream. Hence, Panfili and Ximenes (1994)
noted a mean increase of 9.4cm (± 4.4cm) during the first year at 29 European sites, but with a distinct
difference between saltwater sites (10.9 ± 4.8cm) and freshwater sites (6.8 ± 1.6cm). Likewise, Mounaix
(1992) detected noticeable ecotypes in the Vilaine (Brittany) from examining growth interruptions on
otoliths. Panfili and Ximenes (1994) also noted this in Mediterranean environments. And Lamaison
(2005) observed this growth heterogeneity between downstream compartments of the Gironde
catchment with mean growth rates of 5.0 to 6.8cm per annum in tidal zones and of 4.0 to 5.0cm per
annum in the river and the fluvial zone.
Finally, the temporal dimension of environmental factors must also be considered. The ecological
factors (Fernandez-Delgado et al., 1989; Panfili et al., 1994; Melia et al., 2006) and demographic factors
(Moriarty, 1973; Tesch, 1977; Parsons et al., 1977) mentioned previously can vary very significantly
from one year to the next at a given site.
Overall, this growth variability of multiple origin leads to significant overlaps in the size distribution
of successive cohorts in a given catchment or site, which excludes a simple polymodal size distribution
(Adam, 1997) beyond the first 15 to 20 centimetres.
2.3.6.4. Size and/or weight structures
Distribution variability and heterogeneity within habitats.
In shallow environments, numerous studies have identified the high spatial variability of signals
(abundance, size and weight structures) mostly collected by electrofishing:
(1) Between the sub-catchments of the same river catchment (Aprahamian, 1988; Barak and
Mason, 1992; Lambert and Rigaud, 1999; Lasne and Laffaille, 2008) with the characteristics of each
sub-catchment being important (density of constructions, slope, hydrology, water quality, exploitation
context, etc.).
(2) Between upstream and downstream reaches of the same corridor (Naismith and Knights,
1990a; Lobon-Cervía et al., 1995), with the slope and construction density impeding to a greater or
67

lesser extent individuals’ upstream migration (reduced abundance, colonisation spread over several
years). The migration itself depends on the number of potential migrants observed downstream.
(3) Within a river section at a given level of the corridor. Size classes are distributed differently in
pools and riffles (Glova et al., 2001) and the structures observed by electrofishing vary greatly within the
same reach between 2 successive dams (Feunteun et al., 1998). Identical results were found on canal
networks (Baisez et al., 2000; Laffaille et al., 2004) with a significant structuring effect due to the water
level.
(4) Finally, at the scale of a fishing station, there is a preference for the banks (Broad et al., 2001)
and a heterogeneous class-size distribution (Glova, 2001; Glova, 2002) which depends on the substrate
and the density of aquatic vegetation (Laffaille et al., 2003; 2004).
Moreover, in addition to the spatial variability at different scales, regular station visits during the
year demonstrated the temporal variability of most of the characteristics collected (Larsen, 1972; Barak
and Mason, 1992; Lobon-Cervia et al., 1995; Carss et al., 1999; Baisez et al., 2000) . This finding holds
for all the sampling strategies typically used (electrofishing, pots, fyke-nets, etc.).
In deep zones (estuaries and rivers), fewer studies have been undertaken, but the little
information available (Peňáz and Tesch, 1970; Elie and Rigaud, 1984; Fontenelle et al., 1990; Naismith
and Knights, 1990b) reveals the same type of variability, in particular sex ratio and density differences
between the coastal zone, the estuary (external and internal) and the river. But no information is
currently available on the density levels in these environments nor on eel distribution within these deep
zones. The question is whether the distribution favours the riparian zone (in which case, the length of
banks must be taken into account in density extrapolation) or whether the distribution is more
homogeneous (in which case, the total water surface area must be taken into account).
This “depth” effect is important when analyzing the results observed in the different parts of the
river catchment. Very high water levels seem to be linked to the significant presence of large individuals
(over 60cm long), which are rare in meanders and shallow environments. Hence, data collected by Lee
(1979) and Baisez (2001) in the shallow environments of dyked coastal marshes (maximum depth of
2m) indicate the presence of sizes between 6 and 60cm but very rarely that of larger individuals. And
Oliveira et al (2001), observing American eels, noted that the production of large females was favoured
by the large waterbodies and deep lakes of river catchments.
All other sizes have been observed in all environments, although small and young individuals
gradually disappear during the upstream journey towards the catchment headwaters (Laffaille et al.,
2003). This most probably relates to the growth and aging of migratory individuals. This phenomenon
has always existed, even when the species was very abundant, and used to be expressed in terms of
increasing average size of individuals with the distance to the tidal limit (Barak and Mason, 1992;
68

Lobon-Cervia et al., 1995). The fall in fluvial recruitment during the last 25 years has, however, changed
the nature and especially the intensity of these phenomena.
Analysis of observed size structures
In coastal zones, the size and weight structures of population fractions are influenced neither by
the distance to the tidal limit nor by the presence of barriers, which are typically considered to be
structuring factors for population fractions situated further upstream in the river catchment (see previous
sections). The characteristics of population fractions observed in these downstream zones result
essentially from anthropogenic (fishing, abstractions, etc.) and natural (sedentarisation, migration,
mortality, etc.) pressures exerted on the elver and yellow eel stages.
In such zones, Svedäng (1999) compared size structures in neighbouring sectors which were
comparable in terms of habitat quality, and where some sectors were exploited and others not. Size
structures differ significantly in these two cases, with very few large individuals in the exploited sectors.
In the case of Svedäng (1999), the total mortality of exploited sectors is 2 to 3 times higher than in nonexploited
sectors, despite an annual exploitation rate which at first sight seems low. This result
highlights the significant impact of moderate but consistent exploitation over a period of some ten years
on the same cohort, especially when the stock is in poor condition as is the case for the European eel.
This conclusion supports that drawn from limit sizes estimated with a von Bertalanffy model. These vary
between 42 and 70cm for males and between 76 and 151cm for females for lightly-exploited population
fractions at the yellow stage (Leo and Gatto, 1995; Poole and Reynolds, 1996) whilst these limit sizes
drop, for example, to 39 and 58cm (respectively male and female) in the exploited population fraction in
the Camargue (Melia et al., 2006).
It is therefore potentially of interest to use this analysis of the size structure observed in a territory
in order to develop a rule for the total mortality level (natural, fishing and other mortalities) applicable to
the existing population fraction. However, this requires a coherent and reliable size structure in a given
territory and reference data (reference sites or periods). Scientific teams continue to work on its design
and it must therefore be used with caution. It would be of interest to establish a data collection network
covering the various contexts, which would help to calibrate and validate the approach (by associating
“size structure / pressure level / mortality rate").
2.3.6.5. Health condition and level of contamination by
Anguillicola crassus
Multiple pathogenic agents can affect the species dynamics (Bruslé, 1994; Vigier, 1990) and it
seemed interesting, within the Indicang programme framework, to increase the awareness of the
69

various partners to the collection of data relating to the health condition of individuals observed during
monitoring operations 10 (fisheries, passes, inventories).
Anguillicolosis is one of the most problematic recent parasitoses. The presence of Anguillicola
crassus, a parasitic haematophagous nematode that lives in the swimbladders of eels, first described in
France in 1984, has been detected in practically all environments (including brackish) and all regions
(Blanc, 1989). The impacts of this contamination are multiple, although debate continues as to their real
influence on an individual’s resistance capacities during difficult episodes (high temperatures, anoxic
phases, etc.) as well as their ability to grow normally and/or to migrate at great depths towards
spawning grounds (Lefèbvre et al., 2004; Eel Rep, 2005). However, it should be noted that in France,
this infestation occurred after the significant drop in glass eel arrivals (1980).
Diagnosing infestation is not always easy as the absence of adults or evolved larvae in the
bladder cavity does not necessarily mean that there are no young larvae within the bladder wall.
Moreover, the infestation may be temporary. Lefèbvre et al. (2002) therefore developed an observation
method and a bladder deterioration index in order to assess current and/or past intensity of the parasitic
aggression.
2.3.6.6. Chemical contaminations
An increasing number of studies focus on the presence and the impact of xenobiotics, molecules
created by the chemical industry and other industrial activities (pesticides, herbicides, brominated flame
retardants, PCBs, etc.) to which can be added heavy metals present at abnormally high concentrations
in certain zones (for example: Moreau and Barbeau, 1982; de Boer and Hagel, 1994; de Boer et al.,
1994; Knights, 1997; Robinet and Feunteun, 2002; Bordajandi et al., 2003; Branchi et al., 2003;
Yamaguchi et al., 2003; Morris et al., 2004; Santillo et al 2006; Tapie et al., 2006). Unfortunately, eels
are highly exposed to, and likely to heavily concentrate, these products because of the duration of their
inland growth phase, their carnivorous status, their benthic behaviour and the fact that they build up
significant amounts of fat before silvering (Robinet and Feunteun, 2002).
It is strongly suspected that there is a reduction in reproduction and/or migration and/or resistance
capacities due to the impact of these various pollutants, which are present in the majority of inland
aquatic environments (Bruslé, 1994; Hodson et al., 1994; Couillard et al., 1997). The ‘Eel Rep’
programme (2005) highlights the low probability of a contribution to the reproductive phase for a very
significant proportion of silver eels from various European catchments and tested under experimental
conditions (resistance over long distances, deep diving, maturation, gamete and larval quality). The
harmful effect of PCBs is particularly highlighted.
10 See Chapter 5.
70

The presence of these molecules in the water is also frequently cited as a factor reducing the
attractiveness of inland waters to juvenile eels (consequently reducing colonisation), but this aspect has
yet to be clearly demonstrated.
The levels of contamination by xenobiotics and heavy metals which have started to be found are
worrisome. It is important to undertake a contamination diagnosis on a large scale by standardising
sampling and analysis methods (season, environment, sizes, etc.). Work undertaken on the Gironde
fluvio-estuarine system (Tapie et al., 2006) shows significant differences in the contamination levels
between the zones of a same catchment and between size classes.
2.4. Metamorphosis into silver eel and downstream migration
2.4.1. Metamorphosis into silver eel
After several years of growth, within or close to inland or insular waters, yellow eels, immature up
to that stage, begin their second metamorphosis to prepare for marine migration. This is called the
“silvering metamorphosis”. Yellow eels then turn into silver eels (Fontaine, 1975; Thompson and
Sargent, 1978; Kleckner and Krueger, 1981; Pankhurst, 1982abc; Pankhurst and Lythgoe, 1982;
Fontaine, 1989; Fontaine, 1994; Durif et al., 2000; 2005; Acou et al., 2005; Ginneken et al., 2007). They
adapt in order to return to the oceanic spawning grounds.
Besides anatomical and physiological transformations at the time of silvering, eels begin a period
of fasting which follows the accumulation of fat during their sedentary life (Ancona, 1921; Fricke and
Kaese, 1995). This fast, as described by Fontaine and Olivereau (1962) “combines the catabolism and
the use of certain substances not only to provide energy but also to transfer, transform and re-use some
metabolites in other organs, especially the gonads". Boëtius and Boëtius (1980) analysed the
distribution of energy derived from using accumulated fat (table 2.5).
Table 2.5 -
Energy distribution of fat catabolism in silver eels (from Boëtius and Boëtius,
1980).
Energy attribution % attributed
Development of gonads 18 %
Routine metabolism 27 % Reproduction (75%)
Migration 30 %
Post-reproduction activities 25 %
The main energetic reserve of silver eels consists of stored lipids, mainly in the muscles as
triglycerides (80% of total energetic reserve) (Boëtius and Boëtius, 1985). The liver, which also
accumulates fat, is the second energetic reserve but only represents 1 to 2% of the total body weight
71

(Dave et al., 1979). Lipid reserves play a major role in the migratory phase because, as already
mentioned, eels stop feeding during silvering metamorphosis and reproduction. The lipids available at
the time of departure for the Sargasso sea should a priori be sufficient for eels to cover 5,000 to 6,000
km without having to feed (van Ginneken and van den Thillart, 2000; van Ginneken et al., 2005a), even
if a 6,000 km journey means that eels have to use up to 60% of their lipid reserves, depending their
starting point and the currents they follow (van den Thillart et al., 2004).
Silvering metamorphosis gives rise to a change in the chloride cells (Cl - /Na + /K + ) in the branchiae,
comparable to the salmonid smoltification (but more flexible in time) and under the influence of cortisol
(Epstein et al., 1971; Fontaine et al., 1995). Despite being an archetypal amphihaline species, eels
seem to be able to maintain the same internal osmotic equilibrium whether they are in fresh or sea
water.
Despite all this knowledge, the most recent major study (EELREP, 2005) concluded that no
internal or external factor has yet been clearly identified as the trigger for silvering metamorphosis.
When studying geographical variations in the age and size at silvering of eels from numerous
European and North African locations, Vøllestad (1992) noted a relationship between the latitude of the
growth environment and the age at silvering. The latitude of the growth zone, together with the longitude
of this zone within the watercourse (Krueger and Oliveira, 1997; 1999) are determining factors for the
time that eels remain in inland waters and therefore for the turnover of each population fraction. The
underlying biological process probably combines the mortality rate in fresh water and the distance to the
reproduction zone (within each river catchment and from these inland zones).
For the same latitude, recent studies show that the age at silvering depends on the context of the
river catchment since neighbouring and/or similar-sized watercourses can produce young and small
silver eels in one river catchment and older and larger ones in neighbouring catchments. For example,
silver eels from the Loire catchment upstream of Ancenis (a river catchment of about 100,000 km²) are
essentially females of 800 to 1,000g and 8 to 9 years old on average (Boury et al., unpublished data),
whilst in the Frémur, a small catchment of 60 km², females are small (350 to 400g for 600mm on
average) and young (4 to 6 years old) (Acou, 2006; Laffaille et al., 2006). In the Oir, a small river
catchment very close to the Frémur but where there is far less human activity, silver females are of
intermediate size (500 to 600g) and age (6 to 7 years old) (Acou et al., in press).
These studies therefore indicate the river catchment as the study unit of choice. They also
highlight the need to include not only the diversity of the typology and the latitude of watercourses but
also the chosen survey methodology (on which part of the river catchment was sampling done, or
should it have been done?) when extrapolating local information, particularly on yellow eels, to estimate
the number of spawners at the regional scale.
72

2.4.2. Unpredictability of the downstream migration pattern
2.4.2.1. Factors involved
In temperate regions, it is currently impossible to identify precisely the factors triggering
downstream migration in Anguilla species. However, it is generally agreed that a massive silver eel
migration is usually closely related to hydrological factors which seem to act in synergy, such as flood
flow, and water temperature (Vøllestad et al., 1994; EELREP, 2005; Acou et al., 2008). These latter are
inevitably related to other parameters such as changes in atmospheric pressure, rainfall and the lunar
cycle (Tesch, 1977; Todd, 1981; Boubée et al., 2001; Acou et al., 2008) even though the role of this final
factor is often questioned.
Water temperature
Significant relationships can be found between water temperature and the number of migrants.
Generally, silver eels from all continents migrate downstream when temperatures fall (Todd, 1981;
Haraldstad et al., 1985; Hvidsten, 1985; Vøllestad et al., 1986; Haro, 1991; Boubée et al., 2001) even
though the variance explained by this factor can be very low in catchments where barriers impede
migration (Acou et al., 2008). However, silver eels tolerate a very broad range of water temperatures.
For example, in Scandinavia, the minimum and maximum of this factor that can interrupt migration are
respectively 4°C (Vøllestad et al., 1994) and 14°C (Hvidsten, 1985). In Spain, this interval varies
between 10 and 16°C (Lobon-Cervia and Carrascal, 1992). American eels also tolerate this kind of
range, between 9°C and 13°C (Euston et al., 1997). In New Zealand, the migration of A. australis and A.
dieffenbacchi stops below 11°C (Boubée et al., 2001), whilst they seem to tolerate temperatures
between 7°C and 19°C according to Todd (1981). Other studies have shown 6°C to be the threshold
above which eel activity increases and 10°C to be the threshold temperature triggering the migration of
eels > 260 mm (Naismith and Knights, 1988; White and Knights, 1997; Chadwick et al., 2007). The
behaviour of silver eels from the Frémur river catchment seems to be more flexible, as they can migrate
downstream at temperatures ranging between 4 and 23°C, although the majority migrate at
temperatures ranging between 6 and 10°C (Acou et al., 2008. These observations support the
hypothesis that there is no threshold temperature as was previously suggested by Haraldstad et al
(1985). However, as proposed by Vøllestad et al (1986), it is possible that water temperature is a long
term factor involved more in the control of the physiological preparation phase prior to migration than in
the actual departure.
Lunar cycle and luminosity
Boëtius (1967) suggested the existence of an endogenous rhythm in silver eels linked to the lunar
cycle. Numerous studies confirm that the migration activity of this ecophase is at its maximum between
the last quarter and the new moon, and at its minimum during the full moon (Frost, 1950; Lowe, 1952;
73

Deelder, 1984; Todd, 1981). However, other studies do not show a clear migration pattern linked to the
lunar cycle (Acou et al., 2008). Furthermore, in a recent study carried out under experimental conditions,
Durif (2003) concluded that the lunar cycle probably affected silver eel migration through luminosity
rather than through an endogenous rhythm. It seems that silver eels are highly sensitive to the
luminosity within the water column; this luminosity depends on the lunar cycle but also on cloud cover
and turbidity.
The importance of luminosity for migration is also illustrated by the fact that silver eels are
essentially active during the night (Deelder, 1984; Haraldstad et al., 1985). Hence, strong luminosity
(during the day or during full moon nights) clearly inhibits migration, whilst factors likely to decrease the
luminosity (night, cloud cover, turbidity, new moon) facilitate migration (Durif, 2003).
Rainfall and flow
Rainfall and flow have a significant impact on eels’ downstream migration (Frost, 1950; Lowe,
1952; Deelder, 1984; Vøllestad et al., 1986; Jonsson, 1991; Chadwick et al., 2007; Acou et al., 2008). A
significant increase in the flow in autumn triggers the start of silver eel migration (Vøllestad et al., 1986;
1994; Behrmann-Godell and Eckmann, 2003) although Chadwick et al., (2007) have observed these
autumn migratory peaks at low flows. But the flow factor itself seems to be strongly related to
atmospheric pressure variations and in the end to significant rainfall (Acou et al., 2008). The significance
of the flow seems to be confirmed by the fact that silver eels generally migrate actively in the reaches of
rivers and streams where flows are highest (Jonsson, 1991; Tesch, 2003).
2.4.2.2. Migration periods
Peak migratory periods depend on latitude, with the migratory peak usually occurring when the
temperature falls and the water level rises significantly (figure 2.8). At a latitude of 45-50°N, this peak
corresponds to the end of autumn or the beginning of winter (Deelder, 1970; Durif, 2003). A significant
rise in water temperature can stop the migratory process and even reverse it (EELREP, 2005).
74

cpue
flow (m3/s)
CPUE (mean catch by fishery)
1400
1200
1000
800
600
400
200
0
01/10/2001
05/10/2001
09/10/2001
13/10/2001
17/10/2001
21/10/2001
25/10/2001
29/10/2001
02/11/2001
06/11/2001
10/11/2001
14/11/2001
18/11/2001
22/11/2001
26/11/2001
30/11/2001
04/12/2001
08/12/2001
12/12/2001
16/12/2001
20/12/2001
24/12/2001
28/12/2001
01/01/2002
05/01/2002
09/01/2002
13/01/2002
1400
1200
1000
800
600
400
200
0
Daily mean flow (m3/s)
Catch date
Figure 2.8 - Daily catches of silver eels from professional fisheries in the Loire (histogram) and
river flows (dotted line) in 2001 (Boury and Feunteun, unpublished data).
Downstream migratory peaks are often related to river floods.
The main downstream migratory period for silver eels is usually between August and December
on the Atlantic coast (Tesch, 2003; Behrmann-Godel and Eckmann, 2003; Chadwick et al., 2007).
However, migratory peaks may appear from July until the spring, in particular when environmental
conditions are unfavourable in the autumn, in river catchments such as the Frémur with significant
barriers to migration (Acou et al., 2008), although the same may also be true in river catchments where
migration is unhindered by hydraulic constructions (Frost, 1950; Deelder, 1984; Hvidsten, 1985; Lobon-
Cervia and Carrascal, 1992; Wickström et al., 1996). In Sweden, Holmgren et al (1997) noted that this
migration can occur both in autumn and in spring in the Fardime träsk Lake. Similar results were
obtained by Chadwick et al (2007) on a small Scottish river, although these authors indicated that the
spring peak related mainly to sexually non-differentiated individuals.
Silver males appear to migrate downstream earlier than their female counterparts (Acou, 2006).
This is probably due to their geographical location, further downstream in river catchments, assuming
that the beginning of downstream migration is synchronous for both sexes. Males are more abundant
than females near the estuary and are caught first in downstream fisheries (Boury and Feunteun,
unpublished data).
Silver eel downstream migration seems to be related therefore to particular hydrological
conditions (Tesch, 1977; Haro, 2003; Acou et al., 2008). On several occasions on the Loire, marked
silver eels have been caught up to 60km downstream from the site where they were marked the
previous day (Boury and Feunteun, unpublished data). Individual tracking experiments have shown that
75

when, in large rivers, silver eels were unable to reach the sea in a single trip, they migrated downstream
an average of 380km and could stop their migration for the year (EELREP, 2005). Hence, in very large
catchments, silver eels that have lived far from the sea can take several years to complete their
downstream journey towards the sea. Furthermore, if conditions are unfavourable for downstream
migration during a whole winter, potential migrants are probably forced to stay in the catchment and wait
until the following year for conditions to become favourable. Among them, a small proportion may even
regress to the yellow stage (Feunteun et al., 2000b). When the autumn and/or winter is too dry, and
conditions are therefore unfavourable for the migration process, a small downstream migration of silver
eels does not necessarily equate to a low potential production of the catchment.
2.4.3. Impact of the catchment context on the reproductive
potential
The eel is a fish with high phenotypic plasticity and this makes it particularly sensitive to its
environment. One of the main factors of this high plasticity is sexual differentiation: in eels, it appears
not to be inscribed in the chromosomes but determined by environmental characteristics and the density
of the population fraction in situ. But this density is often deeply modified by the physical context of the
catchment (see previous sections on sedentarisation). The catchment context therefore plays a role in
controlling eel sexual differentiation and consequently in the reproductive potential of each river
catchment.
2.4.3.1. Implications of the environmental determination
of sex
An excessive local density of eels causes stress in young individuals. This stress results in a
particularly long period of undifferentiated sex, leading systematically to undifferentiated gonads
(ovotestis) developing into male gonads (testis). The consequence is the numerical dominance of males
in high density environments. Conversely, low population density leads to female predominance. This
phenomenon can be observed both in the natural environment (Parsons et al., 1977; Vollestad and
Jonsson, 1988; Acou et al., in press) and in aquaculture (Egusa, 1979; Holmgren, 1996). Hence,
typically, in estuaries, small coastal watercourses, coastal marshes and the downstream reaches of
river catchments, the production of male spawners always greatly outnumbers females and can
sometimes be exclusive. On the other hand, females greatly outnumber males in sparsely-populated
zones further upstream in the larger inland hydrosystems. It is however important to note that this
theoretical spatial structuring does not determine the relative contribution of each of the large
compartments of a river catchment to the production of females in that catchment. Although females
predominate in upstream zones, the densities are very low and are in fact similar to densities observed
in heavily-populated downstream zones where females are in a very small minority.
76

Currently, the decline in total and fluvial recruitment, together with the fall in density or in
abundance indices including in the zones that are close to the sea, has resulted in an increasingly
significant number of females (Laffaille et al., 2006 ; Evans, unpublished data). If such findings were
confirmed, they would strengthen the highly likely hypothesis of a relationship between the predominant
sex and the population density.
On both sides of the Atlantic, large sex-ratio discrepancies from one catchment to the next have
been observed for several years running (Krueger and Oliveira, 1999). In a global population where
recruitment is panmictic (Avise et al., 1986), these observations suggest that sex is determined by the
environment (Krueger and Oliveira, 1999). Environmental sex determination has been confirmed by a
number of laboratory and field studies for the European and American species (Tesch, 1977; Wilberg,
1983; Helfman et al., 1987; Holmgren, 1996; Holmgren and Mosegaard, 1996; Krueger and Oliveira,
1997).
Because of the strong sexual dimorphism and the significant difference in abundance,
particularly along watercourses, this environmental sex determination is crucial for the reproductive
potential. In European eels, as in other eels from temperate regions, growth strategies are
fundamentally different between the sexes. The size (between 25 and 45cm in our regions) of males
migrating downstream and the duration of their maturation (2 to 6 years in an inland environment) are
respectively smaller and shorter than those of their female counterparts (from 35 cm to 1 m and
≥ 4 years) (Dekker et al., 1998; Vollestad, 1992; Laffaille et al., 2006). This is probably because the
fertility of males does not depend on their size and it is in their interest to migrate as soon as possible
whereas the opposite is true for females (Colombo et al., 1984; Helfman et al., 1987; Barbin and
McCleave, 1997; Oliveira 1999; Haro, 2003). Male and female size distributions commonly overlap in
many hydrosystems (Haraldstad et al., 1985; Vollestad and Jonsson, 1988; Rosell et al., 2005; Laffaille
et al., 2006) and in aquaculture (Holmgren, 1996) making it impossible to use size structures between
35 and 45cm to determine the sex ratio. Sexual differentiation generally begins between 150 and
250mm, i.e. in juvenile eels over a year old which, at that age, are still in the downstream reaches of
watercourses, although larger individuals (over 35cm) of undifferentiated sex can still be found (Bienarz
et al., 1981; Colombo et al., 1984; Colombo and Grandi, 1996; Laffaille et al., 2006; Melia et al., 2006).
Hence, the sex ratio of the silver eel fraction produced by the catchment may reflect the population
structure from which it is derived (Lobòn-Cervia et al., 1995; Feunteun et al., 2000b; Boury and
Feunteun, unpublished data; figure 2.9) and consequently it may indicate a kind of “hosting potential”
(Krueger and Oliveira, 1999; Lambert, 2005).
77

Frequency
Males
Females
Distance to the sea (km)
Figure 2.9 - Sex ratio of silver eels caught with a large push-net ("guideau") in the Loire,
between 80 to 155km from the sea (Boury and Feunteun, unpublished data). The
further from the sea, the higher becomes the proportion of female silver eels (up to
100%) and the larger becomes the average size. And inversely, the closer to the
sea, the higher is the proportion of males (up to 60% of the sex ratio) and the
smaller is the average size.
However, other studies, over longer periods of time, show that it is not possible to estimate
accurately the sex ratio of eels migrating downstream from the stock in situ, particularly on the Frémur
where the sex ratio of migrating eels has undergone significant modifications over recent years, going
from a practically exclusive male predominance to a slight female predominance (Laffaille et al., 2006),
notwithstanding the fact that the stock in situ has hardly varied (Laffaille et al., 2005b). Similar results
were noted on Lough Neagh in Northern Ireland (Evans, unpublished data). Likewise, Krueger and
Oliveira (1997; 1999) did not find the same sex ratio between sexually-differentiated yellow eels (size
> 30cm, = population inside the river catchment) and silver eels in their downstream migration trap.
They always underestimated the percentage of migratory males.
Of course, the sex ratio of silver eels is affected by available habitats in the growth zone with the
presence of lakes or large ponds promoting female predominance in the downstream migratory fraction
(Parsons et al., 1977; Oliveira et al., 2001). Hence, density alone does not determine the sex ratio;
physical characteristics such as the presence of different types of habitat also play a role.
The sex ratio of sexually-differentiated eels in the river catchment is not therefore the best
estimator of the sex ratio of the fraction of silver eels actually migrating downstream. On the other hand,
predictions as to whether the sex ratio of the migratory fraction will be biased towards male or female
78

can be made on the basis of the size structure of the eel population in the river catchments. This
information, even if is not precise, can be useful when judging the reproductive potential of the
catchment.
2.4.3.2. Density and sex ratio thresholds
In rivers where eel biomass is around 100-150 kg/ha of water (Frémur : 110-170 kg/ha, Acou et al
in press a; Esva: 90-159 kg/ha, Lobòn-Cervia et al., 1995), silver eel production is predominantly male
(between 40 and 90% depending on the year on the Frémur, Laffaille et al., 2006; > 99 % for the Esva,
Lobon-Cervia et al., 1995). On the other hand, when the biomass of the population fraction is low (3.5
kg/ha on the river Imsa, Vøllestad and Jonsson, 1988; 35-45 kg/ha on the Oir, Acou et al., in press), eel
production is predominantly female (> 90 % and ~ 80 % for the Imsa and the Oir respectively).
2.4.4. Impact on the reproductive potential
Given that (i) the “downstream” context of a river catchment seems to be responsible for the sex
ratio bias towards males or females and that (ii) males are smaller than females at silvering, the
"downstream" context of a catchment should be responsible for significant differences in the silver
biomass produced.
For this amphihaline species which starts catchment colonisation downstream, the “downstream”
context (in terms of eel density) comprises two factors 11 :
• Fluvial recruitment;
• Upstream zone accessibility.
Barriers to upstream migration, especially when they are very difficult or impossible to pass,
cause local accumulations of eels and, consequently, large disparities in population structures
downstream and upstream of these constructions (Aprahamian, 1988; Lobòn-Cervia et al., 1995; White
and Knights, 1997; Ibbotson et al., 2002; Domingos et al., 2006; Lasne and Laffaille, 2008). For this
reason, these must affect the characteristics of the pre-migratory fraction of silver eels (= eels in the
process of silvering and preparing to migrate).
Consider the example of two coastal catchments of similar size and at the same latitude but with
very different situations concerning barriers to migration. One (the Frémur), obstructed by multiple
barriers which are difficult to pass for juvenile eels, produces essentially males. The other (the Oir), in
which upstream circulation is much less hindered, produces essentially females (Acou et al., in press).
However, whilst it takes 2 to 6 years to produce a male, it can take much longer to produce a female.
The difference in maturation times required to produce each sex (= turnover in inland water), together
with the greater total natural mortality rate in long turnovers, leads to a difference in the biomass of
11 See Chapter 8.
79

spawners which is produced and in the sex ratio (Acou et al., 2008). This independence of biomass and
sex ratio means that when estimating the reproductive potential both the biomass of silver eels
produced and their sex ratio must be taken into consideration. A male can of course fertilise several
eggs; and group spawning behaviour has recently been described in captivity (Dou, pers. comm. on A.
japonica ; van Ginneken et al., 2005b on A. anguilla). Only the number of ovules laid by the females
restricts the number of fertilised eggs and therefore the future number of larvae produced and finally, in
part, the recruitment. Future recruitment seems to be limited far more by the number of female
spawners than by the number of males even though the latter are indispensable for reproduction.
Estimations of spawner biomass must therefore be given together with silver eel sex ratio estimations as
a female predominant sex ratio adds more value to the reproductive potential of the catchment than a
male predominant sex ratio.
Using a numerical simulation model which calculates the number of females produced in the
Maine river catchment in the United States of America (McCleave, 2001) and taking into account
parameters such as the location of barriers, their upstream passability or the mortality rate of silver eels
passing downstream during their migration, it was found that in order to improve the reproductive
potential of the river catchment, the first factor to take into consideration, other than fisheries, was the
arrangement of barriers, i.e. the surface area of accessible habitats and the succession of high and low
densities in the catchment. This latter parameter affects both the sex ratio, and therefore the number of
silver females potentially produced, and the natural mortality rate. Taken together, these observations
imply that the eel density profile along the catchment is the predominant factor characterising the
reproductive potential of the catchment.
80

Chapter 3
The foundations of sampling
Noëlle Bru and Simplice Dossou-Gbete
81

The objective of this part is not to reproduce basic textbooks on sampling but to draw the
attention of readers to, and remind them of, some of the concepts used to define the indicators
described previously and to interpret their variations.
Various books that are well known to biologists and ecologists can be consulted:
- Sokal R.R., Rohlf F.J., 2001. Biometry, W.H.Freeman 3 rd edition.;
- Scherrer B., 1984. Biostatistics (1984), edited by Gaëtan Morin;
- Venables W., Ripley B.D., 2002. Modern Applied Statistics with S, Springer Verlag, 4 th edition.
We are interested in drawing inferences about the unknown characteristics of a population using
a sample taken from it. We will present an overview of some of the fundamentals which make it possible
to take a sample and to deduce from it, with a certain margin of error, the characteristics of the
population being studied (abundance, size and weight structures, pigmentation characteristics…).
3.1. Sampling in fisheries ecology: basic background
3.1.1. Introduction
This part outlines the main principles underlying sampling in fisheries ecology for the
establishment of protocols (choice of sites, periods, duration and measurements) or for the analysis of
statistical data (sampling error and sample collection).
An essential characteristic of any living system is that it exists in space and in time. This important
characteristic must be included in the sampling plans used in fisheries ecology. Therefore, the means to
record the relevant descriptors and the mathematical analytical methods specific to this type of
information must be put into place. No data processing can ever provide information a posteriori which
has not been included in an appropriate sampling plan at the very start. Setting up a “sound" data
collection protocol is therefore essential and it must comply with certain rules.
Stage 1: Define the specific issues related to the objectives of the study requiring data collection;
sampling objectives can vary greatly. The following examples are illustrative:
• Aid to decision-making within the framework of natural resource management (fisheries in
particular). For example, identifying the abundance level of a resource at a given date before
management measures concerning this resource are defined and before the uses affecting its
abundance are regulated;
82

• Comparison of a population’s sub-groups with one another, using one or several characteristics
of the elements that comprise them. These sub-groups may result from partitioning the
individuals comprising the population using some discriminatory criterion (temporal, spatial,...)
• Assessment of trends in some indicators (types of temporal or spatial variation);
• Assessment of the distribution of a characteristic within the population;
• Validation of a hypothesis about one or several characteristics of a population’s elements
Stage 2: Establish the statistical analysis methods to be used in order to ensure that the data
collected can be processed in an effective way and address the issues identified in stage 1.
3.1.2. Basic statistical concepts
3.1.2.1. Population, statistical individual and sample
Target population = exhaustive set of elements to be studied and identifiable without ambiguity.
Two population types can be distinguished: first, finite or infinite discrete populations (elements can be
counted) and second, continuous populations (for example, a surface area or a well defined volume).
Sampled population = part of the target population which is identifiable without ambiguity and
whose elements are accessible for data collection.
When the sampled population is distinct from the target population, it is important to verify that
conclusions reached from data collected on the sampled population are valid for the whole target
population.
Example: suppose the population is a population of eels, in a given zone, to be sampled with
electrofishing. The sampled population might be the population of eels within zones that are accessible
on foot.
Statistical unit (or individual) = An element of the target population or of the sampled
population that can be observed.
Examples: Each fish in a specific population; a surface of specified size, shape and orientation whose
localization is unambiguously defined; ….
Measurement support = When the constraints of an observation protocol require the processing
of a portion of the statistical units, this portion is called the measurement support.
Example: using 100 grammes of eel flesh to measure the concentration of a pollutant.
Sample = a finite set of statistical units of a population selected according to an appropriate and
explicit method. The size of the sample is the number of statistical units that comprise it.
83

3.1.2.2. Observation scales
The choice of sampling strategy depends on the objective of the study (question/hypothesis)
which is determined according to temporal and spatial variations of the studied phenomenon. The
objectives of a study can be of various kinds:
• Describe a situation at a given time (issue fixed in time) and/or in a given place (issue fixed in
space);
• Clarify the trends (temporal and/or spatial variations);
In order to define spatial observation scales, the following information must be collated when available:
• The amplitude (or size) of the area to be sampled; correlate this size with the feasibility of the
sampling. There may be a variety of meaningful scales based on the practitioner’s knowledge
and experience and on the working hypotheses;
• Available information on the spatial distribution of the population in the spatial area selected for
this study: zones of concentration or not (spatial heterogeneity), spatial discontinuities in the
distribution;
• Scale dependency: going from local to global (river basin ...).
In order to define temporal observation scales, the following must be identified:
• The possible existence of different functional scales;
• Discontinuities within these scales.
In this way, it may be possible to establish a measurement frequency and the regularity (or not) of
sampling.
• Temporal constraints: period, duration, frequency, synchronisation ….;
Example: suppose the behaviour of a fish changes according to night or day and to rising or descending
tides. If this is the case, the observation scale might be fixed at night during a rising tide.
Example: biomass of glass eels entering the Adour estuary during a fishing season.
• Target population = the water circulating through the estuary from the 1 st of November to the
31 st of March from the mouth to a given zone;
• Sampled population = the water circulating in the estuary during the nocturnal rising tide
• Statistical unit = the volume of water filtered by scoop nets (in m 3 ) in fixed locations in the water
section. In this case, the sampling support is the volume filtered by the scoop net.
The choice of sampling strategy must take into account the nature of the objectives of the study
and this requires a sound understanding of the scales of temporal and spatial variation of the
84

phenomenon being studied. The major difficulty is that, generally, the main significant phenomena occur
simultaneously but at different temporal and spatial scales. And technical and financial constraints have
also to be taken into account: available time, measurement feasibility ….
3.1.2.3. Statistical attribute
An attribute is a characteristic of the statistical individuals of a population capable of describing
their variability (the differences between them)
When selecting attributes in the study of a population, the 2 issues below must be taken into
account:
• Relevance: they must have a logical link with the phenomenon;
• Information loss: is it possible to have more information? If the answer is yes, variables should
be hierarchised so as to clarify the information loss if only some of them are to be measured
because of the “cost" involved ….
3.1.2.4. Classification scale
This is the set of possible and discernible modalities of an attribute associated with the mode of
assessment of this attribute in the statistical units. The elements that make up a classification scale
must be collectively exhaustive and mutually exclusive.
It is also necessary to examine some of the notions relating to the classification scale:
• Exhaustivity : all the possible values of an attribute that can be observed must be present;
• Assessment methods of an attribute: physical measurement, counting or assessment of a
modality or a state;
• Measurement unit and nature of the measurement scale: quantitative information (usually
requires powerful and expensive measuring equipment) or sufficient qualitative information.
3.1.2.5. Definition of statistical variables or descriptors
When an attribute is associated with a classification scale, it is called a variable or a descriptor.
Four broad descriptor types can be defined: qualitative descriptors; ordinal descriptors; semiquantitative
or discrete descriptors and quantitative descriptors. A descriptor can be obtained by the
direct observation of an attribute or by combining other descriptors, they are then called synoptic
descriptors.
They are usually classified by 2 main themes: environmental descriptors and population
descriptors (spatial occupation; biometric; demographic; structural; ecotoxicology of individuals; …).
85

Table 3.1. Examples of population descriptors
Population
Spatial occupation descriptors
Statistical
unit
Attribute Scale Types
Presence / absence
qualitative
Geographical space
Spatial size unit
Presence of a
species
Type of spatial
occupation Low,
average, high
Number
ordinal
quantitative
Types of habitat
Brackish water, fresh
water, salt water,
field…
qualitative
Biometric descriptors
A species in a given
geographical area
An individual
Age
Alevin, juvenile, adult
In years
qualitative
quantitative
Pigmentation stage 5A qualitative
Size In cm quantitative
Condition index
(ratio, size, weight)
synoptic
3.1.2.6. Definition of the “datum”
A datum is a variable or several variables observed jointly on a statistical individual.
3.1.3. Collection of information
3.1.3.1. General background on the data recording
mechanism
The following assessments are important:
• As regards qualitative descriptors: assess discernible states and identify selected states;
• As regards quantitative descriptors: assess measurement accuracy, which depends on
instrument characteristics;
Example: turbidity measurement: an instrument must be chosen which is accurate for very low turbidity
measurements as there is a threshold for glass eels. The same is true for the speed of filtration of a
gear and of the current which must be measured accurately given their importance when calculating fish
densities and estimating biomass for example.
86

Measurement systems whether in terms of people or equipment must therefore be "faithful": in
the case of replications or measurements repeated under similar conditions, the results provided must
be identical, allowing for the degree of accuracy of the equipment.
For all descriptors, it is important to assess information variability whether this concerns the
magnitude of variability over a homogeneous temporal and spatial area or whether it relates to future
data processing needs….
Example: on the Adour, turbidity varies little during the rising tide, therefore a single measurement is
sufficient at the tidal scale. This is not the case on the Isle where turbidity peaks have been observed
before and after high water during the rising tide.
3.1.3.2. Sampling protocols
There are 2 main families of sampling protocol: probabilistic sampling and non-probabilistic
sampling. Probabilistic protocols are based on the random selection of sample statistical units. Each
statistical unit included in the sample is therefore selected with its own, known inclusion probability.
Unlike probabilistic sampling, the selection of statistical units comprising the sample using a nonprobabilistic
protocol is based on expert opinion.
Table 3.2. Probabilistic protocols versus non-probabilistic protocols
Advantages
Probabilistic
• Makes it possible to calculate the
statistical precision of calculated
indicators
• Provides reproducible results give or
take statistical uncertainties.
• Makes statistical inference possible
• Makes it possible to assess the
probability of erroneous decisions
Drawbacks • Statistical units where spatial
locations have been chosen by a
probabilistic method can be difficult
to reach
• The accuracy of uncertainty
calculations depends on the
adequacy of the probabilistic model
describing the population and
sampling variability
• Inexpensive
Non-probabilistic
• Easy to implement
• Depends on the knowledge of the
expert.
• Does not make it possible to assess
uncertainty on the basis of statistical
indicators.
• It is not possible to use statistical
data processing to extend the
conclusions to the whole target
population.
• Expert appraisal plays a prominent
part in the analysis of results
concerning the study’s objectives.
87

The cost of collecting sample data may make it difficult to implement sampling with the desired
degree of accuracy. In fact, the choice of a plan is a compromise between:
• Parsimony (notion linked to cost constraints);
• Sampling accuracy (notion linked to minimizing the sampling variance when estimating a
relevant quantity, see below).
The sampling plan must also be realistic. Information must be sought and collected according to a
well defined mechanism, which is called the sampling protocol.
In order to be comprehensive, a sampling plan must also justify the sample size and establish
criteria making it possible to localise observation units and perhaps also determine the time at which
information is to be collected. In ecology, these are defined by geographical locations and/or collection
dates.
Examples:
• In order to study the various types of habitat in a river basin, the statistical population is the
surface area of the river basin and the observation units are surface units (generally called
sampling stations) which must be defined precisely by their location: longitude, latitude, depth,
altitude ….;
• In order to study salinity in an estuary, the statistical population is the estuary and the
observation units are units of volume defined by their localization, and by the date and duration
of the observation (which might be, say, a period of 48 hours starting at 8.00 in the morning …).
Classic probabilistic sampling plans are the following:
Simple random sampling plan:
The statistical units are selected randomly. For a given number of units, all possible choices have
the same probability of being selected.
Examples: In order to study the various types of habitat in a river basin, the statistical population is the
surface area of the river basin and the observation units are surface units (generally called sampling
stations) defined as 10 m². If 30 are chosen, their geographical localization is selected randomly.
Systematic sampling plan:
Samples are chosen at regular “intervals” in space or time. The initial location or starting time is
chosen randomly and thereafter other points are defined at regular intervals over an area (grid) or in
time (systematic). The latter is called a chronological series.
88

Specific sampling plans are used to assess the size of an animal population (transects,
quadrats..).
The principle is to make observations along a series of lines or points in an exhaustive search for
statistical units.
Line sampling consists of counting the number of statistical units observed whilst following a line
of pre-defined length. The distance between each statistical unit and the line must also be noted.
Point sampling consists of selecting a number of well-identified geographical points and counting
the number of statistical units seen from these points. The distance between each statistical unit and the
observation point must also be noted.
Stratified and proportional allocation sampling plans.
In this case, the target population is divided into distinct strata or sub-populations defined in such
a way that there is less heterogeneity between statistical units belonging to the same stratum than
between statistical units belonging to different strata. Strata must be chosen as a function of the spatial
and temporal proximity of the statistical units or on the basis of prior knowledge concerning the
phenomenon.
Adaptive and combined protocols can also be mentioned.
These different sampling protocols are governed by implementation rules, which, if respected,
lead to representative samples and unbiased results.
It is always essential to verify that the sampling approach works well before full-scale
implementation in order to check that the design meets the objective sought.
3.1.3.3. Choosing the sample size
The choice of sample size is crucial in a probabilistic sampling framework as it has a large impact
on the statistical accuracy of the calculated indicators. The chosen size generally depends on the
desired level of accuracy and on the protocol selected.
The survey rate and the sample size are therefore both linked to the variability of the
phenomenon being studied.
89

3.1.4. Statistical estimators and indicators
The aim of most sample-based statistical techniques is to generalise the sample results in order
to describe the population. Sampling aims to provide sufficient information so that inference at the
population level is possible. The choice between different statistical techniques is related to:
• The observation mode of the information (measurement, counting, qualitative
evaluation);
• The types of decision to be made (understand, plan, classify ….);
• The desired degree of generalization of the techniques and results: spatial or temporal
extrapolation or interpolation, adaptability to other times and places...
3.1.4.1. Definitions
An “estimator” is a mathematical formula that combines data of varying complexity (linear or
non-linear estimator), enabling the estimation of a population parameter from a sample. Because it is
calculated on the basis of a sample taken from the population and because different samples give
different values for the same estimator, the latter exhibits a variability that can be determined using the
laws of probability.
The estimator is in fact a random variable which follows a probability distribution or statistical law
characterised by statistical parameters such as the mean or variance, used to assess its qualities: no
bias, variance ….
An estimate is a particular value calculated by applying the estimator formula to data from a
given sample.
As a general rule, each estimate is associated with a confidence interval (calculated from the
standard error, i.e. the standard deviation of the estimator which varies with the sample size) for a given
risk so as to take into account the fact that the estimate is only one possible vision of reality, as a
consequence of the calculation process described above. The confidence interval gives a measure of
the estimation accuracy of an estimator derived from sample data. In order to evaluate this accuracy, a
sufficiently high probability level 1 − α is set such that it can be considered certain that the difference
between the true unknown value of the relevant parameter and its estimation is below a certain
threshold. If the variations in parameter estimates as a function of possible samples follow a normal
distribution and if there is also no bias, then the confidence interval can be determined solely from the
mean squared error which in this case is equal to the variance of the estimator.
An indicator is a parameter which describes a population and is related to a variable.
Examples: the average size of eels, the proportion of population presenting a particular modality of an
attribute, the occupation rate ….
90

3.1.4.2. Bias and its sources
In general, a bias is a systematic error in the results of a study. Nevertheless, it is important to
differentiate between various notions of bias in statistics. The different notions discussed here are:
estimation bias (mentioned above), sampling bias and measurement biases. A sample is biased when it
is not representative of the population from which it is drawn (selection bias):
• Sampling bias: might be due to the selectivity of the fishing gear. It is often due to the fishing
technique which is itself linked to fish behaviour leading to non-constant vulnerability. This
poses a problem of missing or poor data;
• Measurement bias: sometimes related to human factors, for example differences in efficacy, or
to poor staff training. Attention must be paid therefore to data-entry and measurement errors,
and ways must be found to control and correct them;
• Estimation bias: systematic error introduced by the estimation method selected to approximate
the value of a descriptive population parameter from the sample data.
Example: This refers specifically to the following issues:
• Staff training …;
• Conservation of sample material ….;
• Power of the fishing gear ….;
• Verification of measurement instruments … .
3.1.4.3. Quality of estimators
The sampling error is the error made when estimating a relevant descriptive population parameter
using sample data. Of course, this error varies from one sample to the next (for a given protocol and a
fixed sample size). Therefore, the objective is to control this sampling error.
• An estimator is said to be unbiased when the mean of sampling errors calculated over the whole
range of possible samples is equal to 0 (for a given protocol and a fixed size sample).
Otherwise, this mean sampling error is called a bias.
• The mean squared error is the mean of the squares of sampling errors, calculated over the
whole range of possible samples (for a given protocol and a fixed size sample). This mean
squared error defines the efficiency of the estimator. Therefore, it must be as low as possible.
• Finally, sampling error should decrease as the sample size increases. This requirement may be
expressed as the fact that the probability of the sampling error exceeding a given threshold
tends towards zero as the sample size increases. When this property is true, the estimator is
said to be convergent.
91

3.1.5. The special case of fisheries
It is therefore recommended to choose several spatial and temporal scales. If this is not possible,
the smallest scale should be used that will reveal the variability of a very small part of the global
space/time sample. Mathematical techniques can then be used to aggregate information over larger
scales.
It is recommended to ensure that samples are comparable, which requires that:
• They are collected either with the same gear or with different gears with identical characteristics
used in the same way;
• They result from identical fishing effort:
• They are of sufficient size so that all the individuals present and catchable by the relevant gear
can be harvested.
92

Part II
Evaluation of the main pressures
93

Chapter 5
Inland environmental indicators
Stéphanie Muchiut, Nicolas Susperregui, Iñaki Oroz-Urrizalki
94

This chapter covers all the descriptors and indicators identified in order to characterise the
species’ environment and to monitor its condition adequately.
A collection and monitoring method is suggested for each indicator, examples are developed and
a table summarises the current state of knowledge in each river basin of the Indicang programme. Any
information which is difficult or impossible to obtain may have been entered as non-existent by the basin
coordinator in these summary tables. In fact, any information that could not be retrieved easily by a
manager was considered to be non-existent.
This chapter does not deal with socio-economic factors even though they may have an impact on
the evolution of some indicators. For example, the growing demand for energy can affect the quantities
of water abstracted and the quality of habitats.
Appendix 8 contains the field sheets for every indicator together with their completion
instructions 1 .
4.1. Optimum and minimum frameworks
This refers to the accuracy of the optimum and minimum management charts described
previously 2 .
They consist of indicators which are easily available and give a general picture of the
environmental evolution by defining the minimum framework required to get an idea of the
environmental condition and its evolution and the optimum framework required to assess the true
productivity of the area that is actually or potentially colonised.
The minimum framework covers the indicators required for an initial diagnosis of the
environmental situation and of the potential disruptions likely to affect the healthy growth of eels.
The optimum framework defines a sound reference basis as regards eel production but does
not represent the maximum possible. It comprises all the indicators covered in this chapter.
1 Soulier L., Muchiot S., Susperregui N., Oroz-Urrizalki I., 2007. Guide de remplissage des fiches terrain,
Ima/IKOLUR, http://www.ifremer.fr/indicang.
2 See >.
95

Table 4.1. Principal indicators developed in this chapter. Red arrows precede indicators belonging to the minimum framework.
Type of indicator Indicators Descriptors Required parameters
Barriers to anadromous migration
Construction passability
Production
Habitat available
Potential habitat surface area
Water surface area of the 3 compartments:
estuarine zone, fluvial zone and related
zones (wetlands)
potential of the
environment
Functionality and quality
of habitats
General habitat quality
Water abstraction intensity
Habitat quality, based on physical and
ecological parameters
Inventory of water abstraction across the
whole basin
Fishing mortality
Professional and amateur fishing
Number of fishermen, type of gear,
catch sites and weight
Antropogenic
mortality
Downstream migration
mortality
Turbines and water abstraction
Mortality characteristics according to the type
of turbine
Accidental mortality
Occasional mortality events
Records of high mortalities and
relevant causes
Mortality from water
abstraction
(official) abstraction stations in the lower
reaches of the axes
Power stations, agricultural and industrial
water abstraction
Species
Restocking or transfer of
individuals
Population enhancement
Basin policies and transfer
characterisitcs
96

4.2. Indicators concerning the productive potential of the environment
4.2.1. Habitat availability
The North Atlantic, the Mediterranean Basin, and the (marine, coastal, and inland) growth zones
situated between Mauritania and the Arctic Polar Circle comprise the global habitat required for the
European eel biological cycle (ontological niche). This ontological niche can be divided into larval and
adult migratory areas (North Atlantic) and into growth areas (river basins, lagoons, marshes, etc.). Only
so-called “inland” growth zones are dealt with here, i.e. the river basins, each of which theoretically
functions as an autonomous system, supplied with glass eels and, depending on habitat quality and
accessibility, producing the future spawners.
A river basin, with its complex hydrological system (river, meanders, ponds, wetlands, etc.), is a
considerable potential habitat for eels. However, as with all migratory fish, eels are confronted with
numerous pressures which restrict or make their progression inland impossible. It is therefore
imperative to know, in a basin, what area is really available for their growth and future development.
Two descriptors are necessary to determine this “habitat availability” indicator: passability of
barriers to anadromous migration and potential habitat area.
4.2.1.1. Barriers to anadromous migration
Context and objective
The quasi-totality of hydrographic networks without barriers can be colonised from the estuaries
to the heads of the basins if their natural slope is not excessive. However, the colonisation mechanisms
of hydrosystems are poorly understood and continue to be debated 3 (Briand 2002; Ibbotson et al., 2002;
Feunteun et al., 2003; Lambert 2005; Lasne and Laffaille, in press).
Be that as it may, the natural distribution of eels can however be significantly hampered by the
erection of barriers on migratory pathways.*
During the 20th century, more than 25,000 large dams were constructed for multiple purposes
throughout the world (Prouzet, 2003). In 1997, Moriarty and Dekker reported that European Member
States were regulating 60 to 65% of their river flows. Similarly, 40% of freshwater habitats are
inaccessible or inappropriate, partly due to natural constraints (high altitudes) but largely due to
anthropogenic factors, one of the most important being the physical obstruction of river basins.
97

Barriers to free circulation are in fact one of the main factors restricting the inland distribution
area. They block migration or delay colonisation, hence decreasing densities in upper reaches 4 . The
impact of a construction depends on its physical characteristics but also on its hydraulic management
and this is particularly true for barriers situated in the zone of the dynamic tide.
Currently, the European Union within the framework of its “eel regulations” requires Member
States to take the necessary steps to avoid, to the greatest extent possible, hindering the natural
migration of the species, as one of the measures likely to increase eel production significantly in many
river basins.
The objective of this descriptor is to identify, chart and characterise barrier passability to eels'
anadromous migration in the river basins. The mapping of barriers, and the evaluation of their density
and passability on an axis are tools that provide guidance to management and habitat restoration
policies.
Study scale of the descriptor
The scale covers the migratory axis and not just one barrier. The most important step is to select
priority axes on the basis of the most coherent geographical and biological criteria possible.
The hydrological basin is considered in terms of 3 large units:
• Estuaries: transversal limit of the sea – dynamic tide limit;
• Rivers: dynamic tide limit – first totally impassable barrier ;
• Wetlands: wetlands must be considered as a separate habitat. However, as these environments
are complex, the methodologies suggested in this document have not been validated for wetlands
and require further development.
Studies should start at the barrier which is located furthest downstream and then move upstream
taking into account the following criteria: proximity of the estuarine zone; zones most likely to be subject
to drought and the level of water quality.
Priority sectors are those that host the different life stages of the eel (biological criteria) and are
consistent with the geographical approach described above.
• Priority 1: active zone hosting individuals less than 30cm long which will renew the existing stock.
The upstream limit of this zone reached by these young individuals is called the “colonisation
front” 5 .
• Priority 2: colonised zone with eel populations of various sizes and ages.
• Priority 3: zone suitable for colonisation where eel presence has been reported.
3 See chapter 2.
4 See chapters 2 and 8.
5 See chapter 8.
98

The upstream limit of the zone under consideration is the first totally impassable barrier. A
definition of such a barrier is given in the passability grid for the descriptor entitled "barrier to
anadromous migration". No passability improvement is possible for this barrier (no eel pass) given the
present state of technical knowledge. All territories located above such a construction are considered to
be permanently lost for eels.
A natural limit to colonisation is also set by altitude as eels have never significantly colonised
territories above 1,000m.
Once a study zone has been defined, the relevant watercourses have to be selected. In order to
limit the study size, only the following are included:
• coastal watercourses with sources less than 100km from the coast;
• watercourses ranking above 5 in the Strahler classification;
• watercourses listed for eels under the regulation to ensure free circulation of fish (in the case of
France);
• lower reaches of watercourses where abundance is being assessed.
99

Selection of important axes:
Sélection des axes importants :
Strahler stream order > 5
+ rank 4 at less than 100km from the dynamic tide
or on the main axis
+ rang 4 à moins de 100 km
de la marée dynamique ou sur
l’axe principal
ME_20_06_05_csp par O_str
4 (532)
5 (137)
6 (40)
7 (15)
8 (6)
Figure 4.1. Example of watercourse classification in the Loire river basin (source : P.Steinbach).
Precise accurate field measurement is not necessary for watercourse ordering. The order can be
calculated automatically using a GIS programme for the whole hydrographic network In France,
“BDCarthage” is used for this kind of work.
The Strahler system includes the notion of hydrographic network segmentation. At its source the
watercourse is ranked 1. When it merges with another watercourse, two possibilities arise: if the 2
watercourses have different ranks, then the downstream watercourse is given the highest rank; if both
watercourses have the same rank n, then the downstream watercourse is ranked n+1 (figure 4.2). Maps
used for the Strahler classification must be 1:50,000.
100

Figure 4.2. Principle underlying the Strahler stream order classification system.
Data acquisition
Field information must be collected if possible during peak migratory periods in average flow
conditions i.e. outside floods and low water levels, usually between the months of May and July. It must
be as reliable as possible in order to achieve a coherent classification of barriers based on their
passability.
Mathematical models are being developed by French GRISAM scientists (a Scientific Interest
Group for Amphihaline Migratory fish) in co-operation with INDICANG participants. Using information
collected at each dam, their objective is to give a passability rating on a consistent basis across all
basins, independently of the observer.
One of the objectives of these models is to incorporate both the colonisation indicator, in terms of
a construction density percentage compared to a barrier-free reference situation, and the downstreammigration
mortality indicator, expressed in terms of the proportion of spawners reaching the sea as silver
eels compared to a situation with no mortality. However, an initial analysis of the impact of these
constructions can be made at an intermediate stage.
Whilst waiting for these models to be validated, the information required to ensure that they
become operational can be collected, based on a method developed by Steinbach (Onema, France) in
the Loire river basin. The principle relies on collecting information for each dam (head-drop, slope,
rugosity, etc.) and then giving it a passability rating. However, as the objective of the descriptor is to be
applicable to, and representative of, all river basins, this method could not be used in its initial version.
101

It is necessary, therefore, to construct an inventory of all dams and to collect certain information
which is then recorded on the appropriate field sheet (figure 4.3). To help complete this sheet,
instructions can be found in appendix 8.1 6 at the following address: http://www.ifremer.fr/indicang.
It is also imperative to take photos of constructions, of passes and of the upstream and
downstream banks.
Figure 4.3. Field sheet for the > descriptor.
6 Muchiot S., Susperregui N., Oroz-Urrizalki I., 2007. Guide de remplissage des fiches terrain – Obstacles à la
migration, Ima/IKOLUR, appendix 8.1 of the Indicang report, http://www.ifremer.fr/indicang.
102

The information collected for each dam is entered into a local database. However, the fields and
their format presented in table 4.2 must be scrupulously respected so that the work across the whole
range of basins is coherent and information can be exchanged.
Data exploitation
Using this mathematical model, a homogeneous passability rating can be given to all basins,
independently of the observer. This rating is based in particular on the height of the construction and not
the “height class” which would tend to minimise the impact of dams lower than 2 metres.
For the time being, some idea of passability can be gauged from a construction ranking based on
5 criteria: head-drop; dam profile; rugosity; edge effect and the diversity of passes.
The head-drop gives a rating to the construction which is then weighted by other criteria. In the
end, constructions can be divided into 6 classes: class 0 (unobtrusive and/or no barrier); class 1
(passable without apparent difficulty); class 2 (passable with some risk of delay); class 3 (difficult to
pass); class 4 (very difficult to pass); class 5 (impassable). Further information on the criteria in figure
4.4 can be found in appendix 8.1 7 in the Indicang report.
7 Muchiot S., Susperregui N., Oroz-Urrizalki I., 2007. Guide de remplissage des fiches terrain – Obstacles à la
migration, Ima/IKOLUR, appendix 8.1 of the Indicang report, http://www.ifremer.fr/indicang.
103

Watercourse N° :
Passability diagnosis (based on expert opinion)
Date :
Class :
Observer:
Criteria Contribution / impact reduction Rating
≤ 0.5 m + 1
Head-drop
Profile
≤ 1m + 2
≤ 2m + 3
> 2m + 4
Vertical part ≥ 5H(rise)/1L(run) and/or very marked break in the
slope
Very steep part 5H/1L to 3H/2L and/or marked break in the slope + 0,5
+ 1
Slope of downstream side 2H/3L to 1H/5L - 0,5
Very mild slope of downstream side ≤ 1H/5L - 1
Waterproof and smooth materials + 1
Rugosity
Very rough downstream facing (rocks, vegetation or mixed) - 1
Rough downstream facing (irregularities, mosses) - 0,5
Edge effect Favourable lateral dip - 0,5
Diversity
Observations :
Existence of a far easier pathway - 1
Existence of an easier pathway - 0,5
TOTAL
Figure 4.4. Sheet used to establish the passability diagnosis (according to Steinbach, Onema).
104

Table 4.2. Definition of database fields and their characteristics.
Name of the field Format / Type of data Expected response
Expert organisation Text Ekolur, AZTI, MIGRADOUR, MOGADO, CIIMAR, WRT, LOGRAMI, …
Name of the expert Text DIAS, DIAZ, MARTY, LAURONCE, BAISEZ, …
Country Text Portugal, Spain, France, England
River basin Text Specify the code of the river basin
Zoning Text Enter the letter corresponding to the typology
Name of the zone Text Specify in full the name of the estuary, of the watercourse, …
Date of the survey Date Format dd/mm/yyyy
Longitude X Numerical Decimal degree
Latitude Y Numerical Decimal degree
Construction code Text Allocate a construction code = basin code + primary key number
Left bank ‘”Commune” Text Specify in full the name of the “commune” on the left bank
Right bank “Commune” Text Specify in full the name of the “commune” on the right bank
Department Text Specify in full the department or “province”
Owner Text Specify in full the name of the owner
Construction type Numerical Specify the construction type code number
Other construction type Text If the answer to the previous question is 11, specify the type in full
Construction use Numerical Specify the use code number
Other type of use Text If the answer to the previous question is 11, specify the type in full
Construction state Text Specify the construction state code number
Year modernised Numerical Year its height was raised, turbines were changed, etc.
Length of the dam Numerical Specify the length of the dam in metres
Width of the watercourse Numerical Specify the width of the watercourse in metres
Height gradient at low water level Numerical Specify the height gradient at low water level in metres
Threshold slope Numerical Specify the threshold slope in metres
Distance to the sea Numerical Specify the distance to the sea in kilometres
Distance to the dynamic tide Numerical Specify the distance to the dynamic tide in kilometres. Give the relevant symbol
Position of the main dam Numerical Specify the code number
If on a river loop, length of the by-passed river
branch.
Numerical
Specify the length of the by-passed river branch in kilometres
105

Name of the field Format / Type of data Expected response
If on a river loop, reserved flow in the by-passed
river branch.
Numerical
Specify the flow in the by-passed branch in cubic metres / second
Number of passes Numerical Specify the number of existing passes on the dam, if no pass score 0 and move on to the height of the construction
Type of pass 1 Text Enter the relevant letter: C/A/P/K/O
Position of pass 1 Text Enter the relevant letter: M/D/G
Presence of traps in pass 1 Yes/No Tick the box if answer is yes
Is pass 1 working? Yes/No Tick the box if, according to you, eels can pass.
Year when pass 1 became operational Numerical Format yyyy
Type of pass 2 Text Enter the relevant letter: C/A/P/K/O
Position of pass 2 Text Enter the relevant letter: M/D/G
Presence of traps in pass 2 Yes/No Tick the box if answer is yes
Is pass 2 working? Yes/No Tick the box if, according to you, eels can pass.
Year when pass 2 became operational Numerical Format aaaa
Type of pass 3 Text Enter the relevant letter: C/A/P/K/O
Position of pass 3 Text Enter the relevant letter: M/D/G
Presence of traps in pass 3 Yes/No Tick the box if answer is yes
Is pass 3 working? Yes/No Tick the box if, according to you, eels can pass.
Year when pass 3 became operational Numerical Format yyyy
Construction height Numerical Specify the construction height in metres
Head-drop Numerical Specify the head-drop in metres
Profile (slope) Text Enter the relevant code: P1/P2/P3/P4
Rugosity Text Enter the relevant code: R1/R2/R3
Edge effect (favourable lateral dip) Yes/No Tick the box if answer is yes
Diversity (presence of easier pathways) Text Enter the relevant code: V0/V1/V2
Median flow (May/July) Numerical Specify the median flow over the period May to July in cubic metres / second
Upstream migration observations Hypertext link If they are available, add the profile of daily flows over the period.
Remarks about the pass Text If the pass is not working, specify why.
106

Various representations are available to users and managers depending on their requirements.
However, in order to facilitate a rapid comparison between river basins, information is presented in a
summary table as shown below:
Rating
Number of
Percentage compared to the
Density compared to the total area
constructions
total number of dams
of the river basin
0 x XX % …
1 Y YY % …
Barrier mapping also gives an overview of the issue of “construction and upstream migration” for
each river basin. It provides in particular the density of barriers on each of the main axes and sub-river
basins and their passability which then helps in the implementation of the management measures
required to ensure free circulation.
Finally, as soon as this level of exploitation is reached, the construction should be surveyed for
upstream and downstream migrations. It is suggested, therefore, that the “barriers to anadromous
migration" descriptor be dealt with at the same time as the "barriers to catadromous migration" indicator
because this also requires a field survey.
As indicated previously, barrier and potential habitat area maps must be compared in order to
determine which areas can be accessed easily, with some difficulty or not at all by eels.
The barrier map must also be compared to the map of the presence or absence of yellow eels
and the limit of the colonisation front 8 .
Applied example in the Loire river basin
The Loire project operations cell of ONEMA (National Office for Water and the Aquatic
Environment) has established a Geographic Information System and a database where more than 1,800
constructions are recorded. Barrier passability (figures 4.5 and 4.6) is analysed by experts, specifically
for eels, using criteria such as the head-drop, the slope of the downstream facing, the lateral dip, the
rugosity of materials and the diversity of passes. The diagnosis takes into account the impact of
constructions on migrations that take place at the end of spring and during the summer, when flow
regimes are low to moderate.
8 These indicators are defined in chapter 8.
107

C4
19%
C5
2%
C0
12%
C1
23%
C3
22%
C2
22%
Barriers: Classement des obstacles par niveau en fonction de leur
franchissabilité
Unobtrusive Class 0
Passable Obstacles : without obvious difficulty
Effacé Class 1
classe 0
Passable Franchissable with sans blockage difficulté or apparente seasonal delay Class classe 2 1
Passable Franchissable but avec difficult blocage ou retard saisonnier Class classe 3 2
Passable Difficilement but franchissable very difficult Class classe 4 3
Impassable Très difficilement franchissable Class classe 5 4
Infranchissable classe 5
Figure 4.5. Distribution of barrier passability in the Loire basin (Source: ONEMA).
Unobtrusive
Passable without obvious difficulty
Passable with blockage or seasonal delay
Passable but difficult
Passable but very difficult
Impassable
To be specified
Figure 4.6. Distribution of barrier passability for anadromous migration in the Loire (Source:
ONEMA).
The barrier density on the main axes (figure 4.7) can be established from the inventory of the
main constructions and the analysis of their passability. This shows here the need to concentrate efforts
on developing the Maine basin and the Sèvre Niortaise where an important number of constructions
108

curb the colonisation of highly favourable areas for the species (estuary proximity and appropriate
habitats).
Legend
Impassable barrier
Significant barrier
Nuclear power station
Barrier density
No barriers
< 5 barriers per 100km
5 to 10 barriers / 100km
10 to 25 barriers / 100km
>25 barriers / 100km
Figure 4.7. Map of obstacle density on the Loire (Source: ONEMA).
Applied example in the Garonne and Dordogne basin
The survey of constructions on the Garonne and Dordogne tributaries together with their
passability characteristics, presented as a graph, shows the number of passable barriers and those
which need to be equipped with passes on the different axes of the basin (figure 4.8).
109

Length Linéaire of the watercourse d'eau (en (in km) km)
45
40
35
30
25
20
15
10
5
0
Affl.1
Affl.5
Affl.2
Affl.4
Affl.3
Passable Obstacle barrier franchissable
Barrier Obstacle to be à équiper equipped
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Barrier Numéro number de l'obstacle
Figure 4.8. Relationship between the number of barriers and the watercourse length on the
Garonne and Dordogne tributaries (source: MIGADO).
With this approach, it soon becomes clear to the manager that 4 obstacles have to be equipped
on tributary n°1 in order to gain 40km of watercourse length whilst 20 would have to be equipped on
tributary n°5 to achieve the same result (figure 4.9). Similarly, it becomes clear that, on tributary n°1, the
equipment of the 4th obstacle alone gains 30km in length, provided that the 3 downstream barriers are
equipped.
Recovered Linéaire cours watercourse d'eau reconquis length (in (en km) km)
45
40
35
30
25
20
15
10
5
0
Affl.1
Affl.5
Affl.2
Affl.4
Affl.3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Number Nombre of d'obstacles barriers to be à équiper equipped
Figure 4.9. Relationship between the number of barriers to be equipped and the potential length
to be recovered on the Garonne and Dordogne tributaries (source: MIGADO).
110

Example of work carried out in the Principality of Asturias (Spain)
An inventory of barriers was established taking into account their passability for eels (figure 4.10),
enabling the identification of basin sections that could be colonised by eels and led to appropriate
restoration measures, based on the approaches developed previously, being taken.
Eel barriers
Barrier passability
Passable
Passable but difficult
Impassable
No data
Figure 4.10. Inventory of the barriers to eel migration in the Asturias river basins in 2006
(Source: Government of the Principality of Asturias).
Summary by river basin of information concerning this descriptor
It is noteworthy that, wherever this descriptor is relevant, it has been recorded, but it is not always
listed exhaustively.
111

Table 4.3. Information summary on barriers to upstream migration.
112

4.2.1.2. Potential habitat areas
Context and objective
The ubiquitous eel colonises lagoons and salt marshes, brackish estuaries or watercourses,
lakes, and freshwater ponds 9 . However, their accessibility is not guaranteed because of anthropogenic
constraints (flow management, wetland drainage, abstractions …. ) and natural constraints (long term
climate change or unexpected weather) which contribute to reducing these potential habitats.
As early as 2001, the WGEEL (Working Group on Eel: a European working group of ICES -
International Council for the Exploration of the Sea), responsible for assessing the state and the
evolution of the European eel population, emphasised the worrying consequences of habitat loss
and, in 2002, highlighted that half of the wetlands in Europe had disappeared or had been significantly
damaged over the past 30 years. This conclusion was reiterated in the 2006 report which stated that eel
stock restoration depends in part on the restoration of currently inaccessible habitats.
The objective of this descriptor is to determine the potential water surface areas available for eels
whatever their life stage, without taking into account possible barriers to migration (developed in the
previous part). This assessment is based on potential surface areas of rivers and also on flows, as
water quantity is an interesting parameter in more ways than one: spring and summer river flows are a
good indicator of the conditions in which the majority of basin colonisation occurs. In hydrosystems
where this attraction flow practically disappears by the middle of June, the colonisation process can be
significantly impaired. Moreover, summer flows and water levels also reflect the conditions in which the
growth phase of the species occurs. Dried-up stretches or zero flow generally lead to major
deteriorations in the quality of the species’ habitat.
Knowledge of this descriptor highlights the importance of zones that are susceptible to drought,
which must be linked to the indicator concerning "mortality due to water abstraction".
Study scale of the descriptor
The zoning is identical to that described for the “barriers to anadromous migration” indicator so
please see above.
However, in order to carry out a historical analysis of this descriptor at altitudes below 1,000m,
surface areas situated above impassable dams are included.
Five geographical entities can then be defined:
• Estuaries (E): from the transversal limit of the sea to the limit of the dynamic tide;
• Rivers (R): from the dynamic tide limit to the first impassable barrier or else to an altitude of 1,000m;
• Connected lakes (L) (altitude below 1,000m);
9 See chapter 2.
113

• Water impoundment upstream of an impassable barrier (LA) and up to an altitude of 1,000m;
• Rivers upstream of an impassable barrier (RA) and up to an altitude of 1,000m.
Data acquisition
The following information is collected for each basin:
• the length of each watercourse (km);
• the estimated width (km);
• the estimated area (km²);
• the flow when the water level is low (m 3 /s) for France (the flow that ensures the normal coexistence
of all uses and a sound aquatic environment);
• the surface area which is frequently subject to drought (km²); in the case of French river basins, the
“Assecs” (drought-sensitive river beds) Observation Network (ROCA) of ONEMA can be consulted
at the MISES (Inter-Departmental Water Body)
• monthly flows (m 3 /s) to be compared in France with flows in times of crisis.
The average width of watercourses can be extrapolated from the Strahler stream order (Souchon
et al., 2000). In the Loire basin, for example, measurements from a sample of watercourses have shown
that for watercourses ranking from 1 to 5, widths ranged from 1 to 20m.
This relationship is useful to establish quickly the rough size of watercourses in a given area
using a 1:50,000 map.
The collected information is then entered according to the following fields and formats (table 4.4).
114

Table 4.4. Name of data fields and formats concerning the “potential habitat area” descriptor.
Name of the field
Format / Type of
data
Country Text Specify the name of the country
Description
River basin Text Specify the code of the river basin
Habitat typology
Text
Enter the letter relevant to the following typology: E = estuary; R =
river, ZH = wetlands, LA = impoundment, RA = river upstream of
the 1st impassable dam
Name of the zone Text Specify the name of the estuary, the river, etc.
Total length Numerical Specify the total length of the estuary or the river in km
Width at the river mouth
(estuary)
Numerical For estuaries, specify the width at the river mouth in km
Width at the dynamic
tide limit
Numerical For estuaries, specify the width at the dynamic tide limit in km
Strahler stream order
(river)
Numerical For rivers, specify the Strahler stream order
Average width (river)
estimated according to
the Strahler stream
Numerical For rivers, specify the average estimated width in km
order
Total area
Numerical
Specify the total calculated or estimated area of the estuary, the
river, the wetland etc. in km 2
Low water level flow Yes/No For rivers, specify if low water level flows are known
Monthly flows Yes/No For rivers, specify if monthly flows are known
Flow in crisis situation
(case of France)
Yes/No For rivers, specify if flows in crisis situation are known
Areas susceptible to
drought
Numerical
For rivers and wetlands, specify, if possible, the drought-sensitive
areas in km²
Flood plain area Numerical For wetlands, specify, if possible, the flood plain area in km ²
Data exploitation
The collected information is presented according to habitat typology. Further information is
provided on rivers, which are divided into groups according to the Strahler classification:
Strahler
stream
order
Length in km
Width in km
Area in km²
Measured
Estimated
1
2
A map representation can be envisaged to provide a quick and global summary of the situation in
the relevant river basin.
Comparing this descriptor with the “barriers to anadromous migration” descriptor makes
it possible to infer, for example, the percentage of surface areas which are available, not easily
available and unavailable and to derive an indicator for eel habitat availability.
115

Furthermore, the quantitative dimension of “habitat availability" correlated with elements of water
quality and abstraction can be used to guide management strategies.
Summary by river basin of information concerning these descriptors
Table 4.5 highlights the lack of quantitative information concerning some areas: estuaries and
wetlands. This failing should be partially resolved, particularly by increasing data-gathering efforts in
these areas.
116

Table 4.5. Information summary on potential habitat areas.
117

4.2.2. Indicator of habitat functionality and quality
4.2.2.1. Habitat quality
Context and objective
The important phenotypic and behavioural plasticity of eels means that they can occupy and use
various types of aquatic habitat 10 . Their capacity to resist extreme environmental conditions also
enables them to colonise deteriorated environments. However, resisting and surviving extreme
conditions does not mean that the animal is able to effect the whole of its biological cycle
correctly and, in particular, its reproduction.
For several years, the increase in agricultural production, for example, has led to an increase in
the use of xenobiotics which build up very easily in the fat tissue of eels. During their life in fresh water,
eels build up fat reserves which are indispensable for their transoceanic crossing towards the Sargasso
Sea (fasting period) and for their reproduction. Unfortunately, most organic contaminants concentrate in
this fat tissue which is then catabolised to produce the energy necessary for migration and reproduction
(Robinet et al., 2002). As indicated by the WGEEL (2003), the increased use of pesticides in agriculture
must have an impact on eels although no data on these effects are available in the literature.
Currently, some work is beginning to target the true impact of pollutants on eels. For example,
Palstra et al. (2005) tested the impact of organic pollutants (of PCB type) on the survival and
reproduction of adults (artificial stimulation of gonadal maturation and reproduction). They concluded
that negative effects occurred at levels below those authorised for consumption. However, tests on
dosage-related effects have yet to be carried out in order to establish reference levels.
The EELREP programme (Thillart et al., 2005) also concluded that dioxin-type pollutants,
including PCBs, caused some deterioration in embryonic development.
Although work has been undertaken on the impact of some pollutants in the laboratory, only
limited information is available about the impact of these pollutants in the natural environment,
especially concerning the minimal concentrations that affect eels.
Furthermore, the synergic effects of pollutants on the species are not yet well understood (buildup
of nutriments, heavy metals, organochlorides (PCBs), hydrocarbons, pesticides, phytosanitary
products …..). In fact, the general degradation of water quality may have an impact on individuals.
Bruslé (1990, 1994) showed the increased sensitivity of eels exposed to heavy metals and to
pathogenic organisms. Similarly, for Girard (1998), organic and mineral pollutions are often involved in
the development of pathological processes.
10 See chapter 2.
118

It is therefore essential to monitor habitat quality in river basins in order to improve our
understanding of individuals’ growth conditions and thereby assess the physiological state of future
spawners.
Hence, the objective is to monitor the general quality of the environment until such time as eel
pollutants can be specifically targeted and their impact checked and quantified.
Study scale of the descriptor
It covers the whole river basin and distinguishes the different phases of the biological cycle
affected by the quality of the environment.
Data acquisition
Links were established with existing monitoring networks (in France using data from the Water
Agency in particular) set up within the context of the E.U. Water Framework Directive (WFD).
The general aim of the WFD is for all inland and coastal waters (watercourses, lakes, coastal
waters, groundwater) to reach “good status” by 2015, and its specific objectives are:
• to manage water resources sustainably;
• to prevent any deterioration of aquatic ecosystems;
• to supply sufficient and good quality drinking water;
• to reduce groundwater pollution and the discharge of hazardous substances;
• to stop the discharge of priority hazardous substances.
Member States had to set up water status monitoring networks by the end of 2006. Together with
a typology of surface waters and the calibration of water status evaluation methods, this system should
make it possible to compare aquatic environment quality between the management territories of
Member States. A map of global water quality was therefore prepared in all the river basins and is the
minimum requirement when assessing water quality in relation to eel requirements.
This system of water status monitoring and evaluation is based on the concept of chemical and
ecological status, drawing on all existing interactions between the elements that comprise the
ecosystem.
By 2015 at the latest, all inland and coastal waters must have reached “good status”, defined as
the combination of a good ecological and chemical status.
Good ecological status is defined by:
• its biological quality (phytoplankton, macrophytes and phytobentos, invertebrate benthic fauna,
ichtyofauna, the composition, abundance and diversity of which must be completely or nearly
equivalent to the conditions found when there is no anthropogenic disruption),
• its hydromorphological quality (hydrology, river continuity, morphological conditions) and its physicochemical
quality, both of which must be completely or nearly equivalent to disruption-free conditions.
119

The chemical quality of water is composed of three components:
• general parameters that may be called macropollutants (organic matter, nutriments, etc…),
• non-synthetic micropollutants (essentially metals),
• synthetic micropollutants (pesticides, chlorinated solvents, aromatic hydrocarbons, etc.).
As a complement it is recommended to take into account the results from countries or river basins
where more detailed studies have been undertaken. This is particularly the case in France where a new
Water Quality Evaluation System (SEQ-Eau) has been developed since 1999 (for further information on
SEQ-Eau, please refer to publications from the Water Agencies or visit the site of the National
Administration for Water Data and References (Sandre) at: http://sandre.eaufrance.fr). Water can be
assessed by its physico-chemical quality, depending on its various uses (drinking water, etc.) as well as
by biology. Concentrations are measured and compared with class limits, in particular those established
on the basis of World Health Organisation (WHO) recommendations, and then converted into quality
indices. These indices make it possible to judge water quality on the basis of a parameter, a
deterioration (using the smallest index derived from the set of parameters) or a set of deteriorations
(using the smallest index from all relevant deteriorations).
Parameters
Paramètres
Deterioration Altération
Water quality classes
and indices
Classes et indices de qualité de
l’eau
Classes d’aptitude of water de suitability l’eau aux
usages
for uses
ou
or
à la
biology
biologie
Figure 4.11. Principle for the evaluation of water qaulity through “SEQ-eau”.
Water quality is therefore described on the basis of each deterioration, with 5 quality classes
going from blue for the best to red for the worst and with an index varying continuously from 0 (the
worst) to 100 (the best) (table 4.6).
120

Table 4.6. Indices and quality classes of “SEQ-eau”.
Class Quality index Definition of the quality class
Blue 80 to 100 Very good quality water
Green 60 to 79 Good quality water
Yellow 40 to 59 Average quality water
Orange 20 to 39 Poor quality water
Red 0 to 19 Very poor quality water
Each level of water quality is determined by the least favourable parameter, whatever the
medium, i.e. the one which defines the worst class and the lowest quality index. In order to assess
annual or interannual surface water quality, a minimum number of samples and their optimal distribution
throughout the year is needed to qualify each deterioration. Annual quality is then determined by the
worst results, provided that they represent at least 10% of all the samples.
With the help of “SEQ-Eau”, water quality in France could therefore be assessed on the basis of
its ability to ensure biological equilibrium. A distinction may therefore be made, in the spirit of the Water
Framework Directive, between:
• the physico-chemical quality of water, taking into account the 8 biologically-relevant macropollutant
deteriorations;
• the water quality as regards all organic micropollutants, including the pesticides;
• the water quality as regards all mineral micropollutants (metals).
Data exploitation
Information concerning the water quality of a basin is presented in the form of a map whether
within the framework of the WFD or within the SEQ-Eau network. For each type of habitat (estuary,
river, etc.), it is also of interest to specify the number of control sites as well as the percentage of these
sites for each quality category (very good quality, good quality, etc.).
By way of example, figures 4.12 and 4.13 present a series of cartographic data recorded in
French river basins.
121

Very good quality (58)
Good quality (646)
Average quality (447)
Poor quality (179)
Very poor quality (174)
Figure 4.12. Example of the mapping of organic and oxydable substances in French
watercourses from 1997 to 1999 (Source: Water Agency/RNDE).
122

Class of water
suitability for biology
Very good
Good
Average
Mediocre
Poor
Figure 4.13. Map of surface water quality as regards organic and oxydable substances (source:
Adour Garonne basin Committee).
Between 1999 and 2001, measurements and analyses were carried out in 545 sites of the
monitoring network in the Adour-Garonne hydrographic basin. The measurement network shows the
global organic loading to be relatively moderate with some differences between hydrographic subbasins.
Globally, 50% of the monitored sites show water of good or very good quality. Nitrate
concentrations remain particularly worrying in the highly cultivated areas indicated by the red circles in
figure 4.14.
123

Qualité moyenne
Surface sur la période water 1999-2001 quality
Very Très bonne good
Good Bonne
Average Moyenne
Mediocre Médiocre
Poor Mauvaise
Figure 4.14. Map of surface water quality concerning nitrates in the Adour-Garonne basin from
1999 to 2001 (source: Adour-Garonne basin Committee).
Summary by river basin of information concerning these descriptors
Table 4.7 shows that a lot of information is missing and that some information could not be
recorded. A very significant effort is required to improve understanding of the ecological quality of
different eel habitats. This should be one of the priority research areas to be established in most river
basins.
124

Table 4.7. Information summary concerning habitat quality.
125

4.2.2.2. Impact of water abstraction on water level
Context and objective
Since the 1950s, industrial and urban development as well as the intensification of agricultural
practices have increased water resource usage. Hence in Europe (Commission of the European
Communities, 2000) the flow regimes of some watercourses have been significantly reduced, which has
resulted in the degradation of their ecological state.
For example, in 2002, 33.1 billion m 3 of water were abstracted in Metropolitan France, of which
81% came from surface waters (rivers, canals, lakes, impoundments ….). Although this distribution
varies with users and geographical sectors, according to the French Institute for the Environment
(IFEN), 55% of the water abstracted is used to produce energy, 12% goes to industry, 14% to
agriculture and 19% is used as drinking water.
The aim of this indicator of water abstraction intensity is to assess, localise and quantify any
water abstraction from the basin which causes indirect pressure on the various habitats of sedentary
eel populations Water abstraction can contribute to the reduction of habitat area [“assecs” (dried up
riverbeds)] or to the deterioration of environmental quality by undermining the self-cleaning ability of
watercourses during periods of low water flow.
Study scale
The study covers the whole basin in order to understand the water abstraction impact,
differentiating between abstractions from the active zone (individuals

In France, the Water Agencies usually hold this type of information, as is the case for the Adour
Garonne Water Agency (figure 4.15).
(A)
Agriculture
12%
(B)
Energie Energy
12%
Community
Collectivité
18%
Industrie Industry
14%
Agriculture
45%
Energy Energie
60%
Industrie Industry
10%
Community Collectivité
29%
Figure 4.15. Example of water abstraction by activity in 2003: (A) in France (34 billion m³); (B) in
the Adour-Garonne basin (2.7 billion m³) (source: The Adour Garonne Water
Agency).
Information obtained from the bibliography or from the field must be entered into a database with
the following characteristics (table 4.8).
Table 4.8. Field entry format for the “water abstraction intensity” indicator.
Field Format Expected response
Type of water abstraction Text Agriculture / Drinking water / Process / Energy
Water abstraction longitude Numerical Decimal degrees
Water abstraction latitude Numerical Decimal degrees
Zoning Text Active zone / Colonised zone
Habitat Text Estuary / River / Wetland / Lake
Annual water volume abstracted
Numerical
Specify in m 3 the total water volume abstracted
over the year
Month of maximum water abstraction
Text
Specify the month when water abstraction is
maximum
Beginning of the activity
Text
Specify the month when water abstraction
begins
End of the activity Text Specify the month when water abstraction ends
Data exploitation
Information is summarised in a geographic information system which indicates the annual
volumes abstracted and the month when abstraction reaches a peak. It is important to take the
127

volumes abstracted by season into account when comparing with other indicators (water quality,
accidental mortality, etc.).
Zones that are sensitive to drought and zones that become sensitive when their water flow is
reduced as they lose their self-cleaning capacity can be identified by comparing natural river flows with
water abstraction rates.
Maps of water abstraction and accidental mortality must be compared in order to evaluate the
potential links between drought-sensitive zones and the sensitivity of populations to unexpected
environmental changes and to the degradation of their health condition.
This can improve the management of juvenile resources used in stock enhancement programmes
where the sensitivity of these zones to drought has to be considered.
Applied examples: Quantitative monitoring of water abstraction in the Adour Garonne
hydrographic basin
In France, the Water Agencies centralise the information concerning water abstraction in the river
basins. This allows the production of annual reports on the trend in water abstraction (figure 4.16).
Water Prélèvfements abstraction d'eau in en millions Millions of de m³ m 3
1500
1200
900
600
300
0
Agriculture
Communit Collectivité
y
Industrie
Industry
Energie
1996 1997 1998 1999 2000 2001 2002 2003 2004
Figure 4.16. Evolution of water abstraction in the Adour-Garonne basin (source: The Adour-
Garonne Water Agency).
A map of the volumes used during the month with the lowest water flow (figure 4.17) highlights
watercourses which are the most sensitive to water abstraction during low water flow periods.
128

Watercourses in deficit according to
the SDAGE
Watercourses with water
deficit
Watercourses with
significant water deficit
Total volume used during the driest
month (all uses included) compared
to QMNA5 m³/m³ (the 5yr minimum
monthly flow
No abstraction or no
data
Figure 4.17. Map of water consumption during low water flows in the Adour-Garonne basin
(source: Adour-Garonne Water Agency).
These so-called sensitive zones should be exempt from stock enhancement and efforts on these
axes should focus on monitoring the mortality and health condition of the population.
Summary by river basin of information concerning these descriptors
Table 4.9 outlines the information collected on these descriptors.
It shows that, generally speaking, where water abstraction occurs, information is available which
gives some idea of the sensitivity of the productive area to such abstraction.
129

Table 4.9. Information summary by river basin on water abstraction intensity and its impact.
130

4.3. Indicators concerning the level of anthropogenic mortality
They must be defined for each biological stage: glass eel/elver, yellow eel and silver eel, and
indicate the volume of eels extracted from the production system or lost through the development of
various human activities (hydroelectricity production, water abstraction, spreading of toxic products or
fertilisers .... ).
4.3.1. Fishing mortality
4.3.1.1. Context and objective
As specified in chapter 2, eels are exploited at every stage of their biological cycle in various
environments and with various fishing gears and this varies in intensity depending on the river basin,
both in terms of biological stage and intensity. The exploitation rate can, in fact, vary from 0% where
fishing is prohibited to more than 90% in glass eel fishing below estuarine dams which completely block
the migratory flux. Exploitation is sometimes affected by constraints imposed by uses other than fishing
activity (poor water quality prohibiting harvesting and consumption, dams, etc.). Work undertaken within
the framework of the Indicang project also showed that the glass eel exploitation rate depended on
hydroclimatic components which affected in particular their estuarine transit time and their density in the
water column 11 .
Hence, it is natural that descriptors related to exploitation characteristics (catch level, fishing
effort, fishing gear characteristics) should be contributory elements in defining fishing pressure and also
in characterizing the environment within which eels develop.
For further details on fishery descriptors, see the methodological complements 12 and the work
undertaken in particular on the evaluation of estuarine recruitment intensity 13 .
4.3.1.2. Study scale for descriptors
The study must be carried out by river basin and take into account at least the spatial
segmentations defined by the WFD concerning surface waters: transitional waters, rivers and lakes, to
which we add wetlands. A transitional water mass is a distinct and significant part of the surface waters
located close to the mouthes of small or large rivers, which are partially saline due to their proximity to
coastal waters but are strongly affected by freshwater currents.
11
See chapter 2.
12
See chapter 6.
13
See chapter 7.
131

This is the minimal study scale. It can be further refined by including the biological stage and the
categories of fishers (professional or amateur for example in the case of France) exploiting the species.
4.3.1.3. Data acquisition
In order to maximise knowledge of fishing pressure in the river basin, it is recommended to collate
a certain amount of information by fishing season, concerning:
• the number of fishers by category (professionals, non professionals) and by zone;
• the number of fishers declaring their activity by type of fisher;
• the type of fishing gear used;
• the catch location according to the WFD segmentation defined previously;
• the weight caught by type of fisher and by biological stage;
• the existence and type of declaration (voluntary or obligatory; exhaustive or by sampling).
Two methods are currently used to obtain this type of information: an exhaustive method which
includes all fishers and is based on their declaration of activity (compulsory (in the majority of cases) or
voluntary); or a sampling method based on a representative sample of all fishers in activity, who will
provide the cited information, once again through compulsory or voluntary declaration.
In some countries, declaration systems record catches at national level. In this case, it is
imperative to agree upon the formats and structures of these bases in order to avoid dispersing
information in local bases which are unconnected with one another. This is particularly the case in
France with databases of fisheries statistics that are either marine (the “Harmonie” database for
example) or inland (the “SNPE” database for example). The information that is collected on a regular
basis contains a number of fisheries descriptors which go beyond the information required to
characterise eel exploitation. It is therefore recommended to collect at least the elements described in
table 4.10.
132

Table 4.10. Field entry format for the description of fishing mortality.
Name of the field Format / Type of data Description
Country Text Specify the name of the country
River basin Text Specify the code of the river basin
Zoning
Text
Enter the letter corresponding to the following typology: E =
transitional water; R = river, ZH = wetlands, L = lake
Name of the zone Text Specify the name of the estuary, the river, etc.
Name of the operator Text Specify the name of the person entering the data
Category of fisher Text Enter the relevant letter: P = Professional, A = Amateur, I = Illegal
Sub-category of fisher
Text
In the case of France only, specify M = Marine or F = River for
professionals; and E = Fishing gear or L = Line for amateurs
Official opening date Date Format dd/mm/yyyy
Official closing date Date Format dd/mm/yyyy
Type of data Text Enter the letter S = Statistical monitoring or E = Sampling
Nature of the data Text Enter the letter O = Obligatory or V = Voluntary
Total number of fishers
Numerical
Total number of fishers authorised to fish eels at the relevant life
stage
Number of fishers making
a declaration
Numerical Number of fishers having declared at least one catch
Fishing gear
Text
Enter the name of the fishing gear (cf. the international
classification
Targeted stage
Text
Enter the letter E = Glass eel / Elver; Y = Yellow eel; S = Silver
eel
Catches in kg Numerical Total weight of catch
It should be noted that this summary information is insufficient for abundance assessments for a
given biological stage (in particular the assessment of glass eel flux) 14 .
4.3.1.4. Data exploitation
All available information is summarised in a geographic information system giving global catch by
type of fisher and by fishing zone, according to the targeted stage. The resulting trend-type indicator
illustrates the evolution of catch by type of fisher and by zone over the years.
14 For this topic, please refer in particular to the guidelines in chapter 7 (estuarine recruitment evaluation) and to chapter 2 (further
information).
133

4.3.1.5. Summary example of fishery activity in the
Gironde-Garonne-Dordogne river basin
Since 1979, the “Cemagref” of Bordeaux has undertaken monitoring studies of the piscicultural
fauna, of fishing and of the production of the principal exploited species. Some of this work consists in
monitoring catches of glass eels / elvers and yellow eels by commercial fishers from a network of fishersamplers.
Table 4.11 shows the information provided by this sampling work which is necessary to
monitor this indicator.
Table 4.11. Distribution of commercial fisheries in the Gironde Garonne Dordogne basin
(source : CEMAGREF – G. Castelnaud, pers. comm.)
Country
River
basin
Year
Zoning
Name of
the zone
Fisher
category
Fisher
subcategory
Official
opening
date
Official
closing
date
Type of
data
Nature of
data
Total
number
of fishers
F Garonne 2004 E estuary Gironde P M 15 Nov 31 Mar E V 67 10
F Garonne 2004
E tidal
watercourse
Gironde P R 15 Nov 15 Apr E V 33 5
Number
of fishers
declaring
catches
Fishing
gear
Targeted
life stage
Catch in
kg
“pibalour
” type
push net 5
to 14 m 2 GE/E 13 300
Hand
scoop net
GE/E 102
F Garonne 2004
E tidal
watercourse
Gironde P R 15 Nov 15 Apr E V 72 18
2 push
nets
GE/E 1 044
F Garonne 2004
E estuary +
tidal
watercourse
Gironde P M+R 01 Jan 31 Dec E V 66 13
Basket
traps
Y 14 400
These tables can be established either from the simple summation of information collected from
an “exhaustive” source or else from a sampling process which implies defining a sample of
fishers representing the diversity of methods and fishing time in a given zone 15 .
4.3.1.6. Summary by river basin of information
concerning these descriptors
Knowledge concerning fisheries descriptors is highly heterogeneous for both biological stages
and river basins (table 4.12). Generally speaking, wherever a commercial fishery exists in the Indicang
river basin network, it seems that professional fisheries descriptors are monitored but amateur fisheries
monitoring is far less frequent, an issue that requires substantial effort in many river basins.
15 For further information, please refer to chapters 3 and 6 which cover sampling foundations and the collection and definition of
fisheries descriptors.
134

Table 4.12. Information summary on fisheries activity and its impact by river basin.
135

4.3.2. Downstream migration mortality
4.3.2.1. Context and objective
In 2002, the ICES/EIFAC Working Group on Eel (WGEEL) (ICES, 2002) highlighted the
importance of hydroelectric power station turbines as barriers to downstream migration. Individuals
passing through these turbines are likely to suffer significant, possibly deadly, physical and physiological
damage. The consequences are even more dramatic when there are several dams along the river
basin. Hence, the WGEEL concluded in its 2006 report that the physical and physiological impact on
eels passing through hydroelectric power stations can seriously compromise their ability to reach the
spawning grounds 16 .
The transit through a turbine means a significant increase in the flow regime, with an increase in
pressure followed by a sudden decompression on exit and also, for the migrant, a high risk of collision
with a fixed or moving part of the turbine (vane or blade). Given their size, eels are one of the
downstream migrants most exposed to mechanical collision leading to cuts, fractures, perforations,
lacerations and even severed individuals.
The seriousness of the impact caused by the passage through a turbine depends on the speed at
collision. The mortality rate varies from between 5 and 25% on large turbines (Hadderingh, 1982;
Hadderingh et al., 1992) to 100% on small turbines (Monten, 1985). Therefore, the objective is to record
all the barriers to downstream migration present in the basin, to collect the necessary information to
assess the barrier-induced mortality rate on downstream migrants and to complete the information
provided by fisheries descriptors concerning silver eel harvesting intensity.
4.3.2.2. Study scale
The upper limit of the relevant zone is set by the first totally impassable barrier (a barrier
described as totally impassable in the passability ranking grid defined for the “barrier to anadromous
migration” descriptor and for which no improvement is conceivable). All areas situated above that
construction are considered to be permanently lost for eels.
The purpose of the global assessment of a construction is to ascertain its passability for the
colonizing juvenile fraction of the population and to evaluate the mortality rate of spawners passing
through the turbines on their downstream migration. These migratory periods are asynchronous as
juveniles generally colonise the basin from May to July whereas future spawners typically migrate
downstream from October to December.
16 See chapter 2.
136

In an ideal situation, the field survey would coincide with downstream migration. However, for
cost-efficiency reasons and to optimize field surveys, it is recommended to survey a dam once a year,
during the colonisation period, i.e. from the 1st of May to 31st of July.
But in order to assess downstream migration, flows and lengths of individuals should also be
monitored over the October to December period.
4.3.2.3. Data acquisition
Experiments have been carried out in various countries (USA, Canada, Sweden, Scotland,
Germany, France), although mainly on salmonid juveniles and less frequently on eels, in order to
ascertain the mortality resulting from their transit through several types of turbine. Various predictive
mortality models have been proposed for Francis and Kaplan turbines. Pelton turbines, which result in a
100% mortality rate, are restricted to very high drops and are not located on watercourses with
migrants.
Larinier et al (1989) developed equations to explain salmonid and eel juvenile mortality, taking
into account the absolute and relative velocity at rotor entry, the rotor rotational speed, the length of the
fish and the inter-vane space measured half-way up the vane. On average, because of their size,
mortality is 3 to 5 times higher in eels than in juvenile salmonids migrating downstream.
In 2006, the GHAAPPE (Group for hydraulics applied to aquaculture development and
environmental protection) worked on the cumulative impact of dams in several river basins (Larinier et
al., 2006). Potential mortality (expressed in %) during eel transit through each turbine was defined in
relation to several factors: the type of turbine, the nominal flow, the head-drop, the rotor diameter, the
number of blades or vanes and the rotational speed). The mortality percentage is given for the various
eel sizes on the site, as damage increases with the size of the fish for a given turbine.
In order to calculate the potential global mortality attributable to a power station, the effects of all
the turbines are summed and weighted in relation to their turbined outflow. This potential mortality rate
is equivalent to the maximum rate when the entire watercourse flow passes through the turbines.
However, some individuals pass through spillways during water release periods and this must be
included in the result. A simplifying hypothesis is to consider that the respective percentages of
individuals passing through the turbines and the dam are close to those in the turbined water released
during downstream migration.
Given the work undertaken on eel mortality during downstream migration, and whilst awaiting
more precise protocols, we suggest that information relating to the characteristics of the power station
and the turbines be collected on a field sheet presented in appendix 8.1 17 .
Whatever the type of turbine, the following general information must be collected:
137

• number of turbines;
• type of turbine;
• presence of a downstream migration facility (yes/no);
• height of the drop of the downstream migration facility (m);
• presence of a protective device (yes/no);
• spacing between the gaps of the protection device;
• approach water velocity against the protective device (m/s);
• eel length upstream (minimum, average, maximum) or 1st 3rd quartiles and median of the length;
• nominal flow (m 3 /s) (definition and clarification in the field sheet completion guide - appendix 8.1) 18 ;
• available head water in m 3 /s;
• turbined outflow (m 3 /s)
• maximum flow (October - December period);
• the flow relating to quartile 3 of the October to December period (m 3 /s).
Depending on the type of turbine (Kaplan or Francis) the following descriptors are recorded:
• height of the drop (m);
• nominal flow (m 3 /s);
• rotor diameter (m);
• rotor rotational speed (at entry for Kaplan turbines);
• number of blades (Kaplan turbine) or vanes (Francis turbine);
• power (kW).
Table 4.13 shows the field characteristics in the database of hydroelectric constructions
potentially responsible for downstream migration mortality. These fields must be fully completed in order
to show the impact due to this activity. Completion instructions can be found in appendix 8.1 at the
following address http://www.ifremer.fr/indicang 19 .
17
See § Figure 4.3.
18
See § Figure 4.3.
19
Op.cit.
138

Table 4.13. Field entry format for the “downstream migration mortality” indicator.
Name of the field Format / Type of
Description
data
Number of turbines Numerical Specify the number of turbines
Type of turbines Text Specify the type of turbine (Francis, Kaplan, Pelton, Banki, ….
Rotor diameter Numerical Specify the rotor diameter in metres
Number of blades or vanes Numerical Specify the number of blades or vanes
Rotor rotational speed Numerical Specify the rotor rotational speed in revolutions / minute
Head drop Numerical Specify the head drop in metres
Power Numerical Specify the power output in kilowatts
Presence of a downstream
migration facility
If answer is yes, height of the
drop
Presence of a protective
device
If answer is yes, spacing
between the gaps
If answer is yes, approach
water velocity against the
protective device
Yes/No
Numerical
Yes/No
Numerical
Numerical
Tick if the answer is yes
If a downstream migration facility exists, specify its height in metres
Tick if the answer is yes
If a protective device exists, specify the spacing between the gaps in
millimetres
If a protective device exists, note the approach water velocity against it
in metres / seconds
Nominal flow Numerical Specify the nominal flow in cubic metres / second
Available head water Numerical Specify the available head water in cubic metres / second
Turbined flow Numerical Specify the turbined flow in cubic metres / second
Maximum flow (October –
December period);
Quartile 3 flow (October –
December period);
Observations
Size distribution of eel
population
Numerical
Numerical
Hypertext link
Yes/No
Specify the maximum flow over the October–December period in cubic
metres / second
Specify the flow corresponding to quartile 3 of the October–December
period in cubic metres / second
Note daily flow profiles over the October – December period when they
are available.
Specify if there are data on the size of eels likely to transit through the
turbines
4.3.2.4. Data exploitation
The data can be presented in two ways. For a global overview of the basin, it is recommended to
use a GIS to map all barriers to downstream migration by turbine type. As a complement, it is useful to
produce graphs, including the “potential habitat surface area” indicator mentioned previously, in order to
assess the joint effect of improving free circulation not only on new production areas but also for
downstream migration without undue mortality. Management decisions concerning a hydroelectric
construction should only be contemplated if upstream and downstream migration are considered in
parallel.
Furthermore, in coordination with the “yellow eel” and “silver eel” thematic boxes, which provide
an estimate of the abundance of these two stages, the results may be given as total mortality rate
and/or mortality rate by dam as well as in number of individuals and/or in weight.
139

4.3.2.5. Applied example
Larinier et al. (2006) assessed migratory axes on the basis of surface areas actually accessible to
eels, weighted by the downstream migration survival percentage. This method quantifies the benefits
from equipping the various stations for upstream migration whilst taking into account downstream
migration. It also helps to define equipment priorities for downstream migration and to determine the
upper equipment limit for upstream migration.
Figures 4.18 and 4.19 illustrate the approach using 2 curves: the upper curve (CBV) represents
the maximum habitat surface area in the river basin around the power station, the lower curve (CBVr)
represents the recoverable habitat surface area, weighted by downstream migration mortality (when
power stations have no particular facilities for downstream migration).
3000,00
Watercourse
2500,00
Surface area of the river basin (km2)
2000,00
1500,00
1000,00
500,00
0,00
1 2 3 4 5 6 7
Constructions
CBV
CBVr
Figure 4.18. Maximum river basin surface area, weighted by downstream migration mortality on
watercourse 1 (Larinier et al., 2006).
On the first watercourse (figure 4.18), downstream migration mortality is moderate given the low
level of equipment in the plants. It would seem, therefore, to be of interest to re-open the whole axis to
upstream migration as far as the 5 th power station as the gain in habitat is significant at that level and
downstream migration equipment is not a priority. For an optimal cost-recovery ratio, opening the first
power station may be sufficient initially.
140

800,00
Watercourse
700,00
600,00
Surface area of river basin (km2)
500,00
400,00
300,00
200,00
100,00
0,00
1 2 3 4 5 6
Constructions
CBV
CBVr
Figure 4.19. Maximum river basin surface area, weighted by downstream migration mortality on
watercourse 2 (Larinier et al., 2006).
On the second watercourse (figure 4.19), opening the first two power stations to upstream
migration would add little to the available habitat area. Downstream migration mortality in the remaining
power stations is such that, despite a significant potential gain in colonisable area, their equipment to
facilitate upstream migration is only justified if they are simultaneously equipped to facilitate risk-free
downstream migration.
4.3.2.6. Summary by river basin of information
concerning barriers to anadromous migration
Table 4.14 shows that a quantitative effort is still required whilst the qualitative side of these
descriptors has been addressed for the area covered by this study.
141

Table 4.14. Information summary by river basin concerning “downstream migration mortality” descriptors related to hydroelectric use.
142

4.3.3. Accidental mortality
4.3.3.1. Context and objective
Eels are sensitive to pollution because of their highly developed cutaneous respiratory capacity
and their benthic behaviour. The Sandoz incident raised awareness of eel vulnerability, when toxic
waste dumped into the Rhine caused the death of 400 tonnes of eels (Claus and Meunier, 1998).
During the 2003 heat wave in France, the “Conseil Supérieur de la Pêche” (Higher Fisheries
Council) reported locally significant eel mortality as early as July. Some examples, amongst others,
include the cases of the Aveyron, the Rhine (several thousand individuals along its whole length) and
some of its tributaries, the Oust (a thousand individuals), the Orne (several hundred), the Adour, the
Gaves (in particular of Oloron with several thousand dead individuals), the Nives, the Bidouze, the
Charente and some of its tributaries, the Loir, the Loire around Decize and the Poitou marshes.
This phenomenon was generally specific (only eels were affected) and continued over a long
period of time (until the beginning of September in most cases). Rivers of very different sizes were
affected across a large part of France. This mortality was due to several convergent factors such as
parasitism, water and sediment pollution, high temperatures and the development of bacteria (such as
Aeromonas). Eels may have been affected by toxic elements discharged from the sediments as they do
not readily flee and live in direct contact with the substrate. Some conditions, in particular high
temperatures, can lead to anoxia within the sediment which is then likely to release adsorbed elements
(some heavy metals or organic xenobiotics) in a toxic and absorbable form.
These exceptional or accidental episodes must therefore be taken into account as their local
impact on eel populations and their survival is far from negligible.
Mapping the frequency of these events with their causes and their consequences for eel
populations allows highly sensitive zones and periods for eels to be targeted. This information can be
added to that available from maps of zones that are sensitive to water abstraction.
4.3.3.2. Study scale
The method can be applied to all the territory currently colonised by eels. Previously-described
zoning must be respected: estuaries / watercourses / wetlands / lakes.
This type of survey must be done routinely every time abnormal pollution or mortality is noted.
However, vigilance must be increased when water flow is low and during heat waves, because low
water flows reduce the self-cleaning capacity of watercourses and together with high temperatures lead
to highly critical hydrological and physico-chemical conditions for the survival of species including the
eel.
143

4.3.3.3. Data acquisition
Zones of mortality, numbers of eels affected and mortality causes are recorded on a field sheet 20 .
Completion instructions for the field sheet, which are also included in appendix 8.2, ensure the
homogeneous recording of information in this type of sheet. Furthermore, in order to optimize these
surveys, the individuals collected should be sufficiently “fresh” 21 in order to be able to obtain biometric
data, note external anomalies, measure the dose of contaminants and check for the presence of
Anguillicola crassus. The INDICANG partners recommend that a water and sediment analysis be added
to the evaluation. The information collected is entered into a database, the characteristics of which are
described in table 4.15.
Table 4.15. Field entry format for the “accidental mortality” indicator.
Name of the field Format / Type of data Description
Country Text Specify the name of the country
River basin Text Specify the code of the river basin
Organisation centralizing the
data
Text
Specify the name of the organisation recording the information in the river
basin
Operating organisation Text Specify the name of the organisation in charge of field observations.
Name of the enumerator Text Specify the name of the person who recorded the information
Zoning Text Specify the typology letter
Name of the zone Text Specify the name of the relevant estuary, river, wetland or lake
Event longitude
Numerical
Decimal degree (cf. conversion formula and format in the completion
instructions appendix 8.2)
Event latitude
Numerical
Decimal degree (cf. conversion formula and format in the completion
instructions appendix 8.2)
Date recorded Date/Time Format dd/mm/yyyy
Life stage of affected eels Text Enter the relevant typology letter
Number of dead individuals Numerical Specify the number of eels which died following the event
Weight of dead individuals Numerical Specify the weight of eels, in kilogrammes, which died following the event
Affected watercourse length Numerical Specify the distance, in kilometres, affected by the impact of the event
Total surface area affected
Numerical
Specify the surface area in square kilometres affected by the impact of the
incident
Type of event Text Specify the type of event: pollution, dam failure, drought, eutrophication, etc.)
Type of pollution Text If applicable, specify the type of pollution (organic, chemical, ...)
Number of samples collected Numerical
Specify the number of individuals collected and sent to the laboratory for
further analyses.
Requested analyses Text Specify requested tests (parasitology, heavy metals, PCBs, viruses, etc.)
Monitoring cost Monetary Specify the cost of post-event analyses in Euros
4.3.3.4. Data exploitation
As the data are presented in the form of a geographic information system, accidental mortality
observation sites, estimated numbers of affected eels, relevant dates and specific causes can all be
20
Soulier L., Muchiot S., Susperregui N., Oroz-Urrizalki I., 2007. Guide de remplissage des fiches terrain – Mortalité accidentelle,
Ima/IKOLUR, appendix 8.2 of the Indicang report, http://ifremer.fr/indicang.
21
Girard P., Élie P., 2007. Manuel d’identification des principales lésions anatomo-morphologiques et des principaux parasites
externes des anguilles, Cemagref report number 110, appendix 4 of the Indicang report, http://www.ifremer.fr/indicang.
144

ecorded. Multi-year monitoring will reveal mortality frequencies in sensitive locations, which may
change perceptions of the “accidental” character of recurring events.
It is also interesting to compare the accidental mortality rate to the total eel population by stage,
with respect to the various development stages of the species: “glass eel/elver” “yellow eel" and “silver
eel”.
This mapping of accidental mortality must be compared with those of habitat quality/quantity,
water abstraction and health condition of the population. This can highlight a possible relationship
between mortality due to drought and mortality due to water abstraction, and also weight the water
quality in an acute situation context.
4.3.3.5. Applied example in the Loire river basin
In total, 55 cases of eel mortality were reported by ONEMA departmental brigades (figure 4.20)
during the summer of 2003 drought period. All the losses occurred during the low water flow
period, between the end of April and the end of September, especially in August (70% of observations)
when temperatures were very high. 9% of this mortality was caused by pollution sources, 20% was due
to drought in the watercourses and 71% occurred in water environments with no clear cause.
Occasional mortality
(A few individuals)
Significant mortality
(Tens of individuals)
Figure 4.20. Map of eel mortality in the Loire in 2003 (source: ONEMA).
145

In this basin, records show that mortality was spread across most of the species’ distribution area,
from the brackish environment of the Atlantic coastline to the small watercourses at the head of the
basin. Observations concerned principally the downstream half of the basin and the large colonisation
axes where eels were still present. More than 80% of the mortality was observed less than 400 km from
the sea.
In most cases, a few large individuals (60 cm and more), mainly females, were affected. Massive
mortality, i.e. more than around one hundred individuals, occurred in only 8% of recorded cases.
Given eels’ acclimatisation potential to high temperatures and stagnant water, this mortality was
most likely due to the combination of several factors that caused physiological weakness and aquatic
environmental degradation, rather than to the direct impact of temperature or water deficit.
It might also be noted that, in the central part of this basin, the Loir was one the watercourses the
most affected by eel mortality. And yet it is the watercourse the least affected by water deficit as it is
supplied by the Beauce aquifer. Hence, it would seem that mortality was more likely to be related to
water quality and toxic build-up in the sediments rather than to a strict water deficit problem.
4.3.3.6. Summary by river basin of information
concerning accidental mortality.
Table 4.16 shows that a significant data collection effort is required concerning this type of
mortality. Few basins in the Indicang network have easy access to this type of data.
This type of information is useful to highlight the sensitivity of a zone and the communities that
populate it, to the impact of other uses. This impact often becomes clear during periods of rainfall deficit,
associated with the intensive use of water resources.
146

Table 4.16. Information summary by river basin concerning the “accidental mortality intensity” descriptors.
147

4.3.4. Mortality due to water abstraction
4.3.4.1. Context and objective
Water abstraction in the lower reaches of the basin, especially in estuarine zones, can have a
substantial impact on glass eels and elvers, and in particular on silver eels migrating downstream
(ICES, 2002).
The aim of this water abstraction intensity indicator is to evaluate eel removal in the active zone
of the basin, where individuals smaller than 30cm can be found.
4.3.4.2. Study scale
The study covers the whole basin in order to address the impact of water abstraction,
differentiating between abstractions from the active zone (individuals

Table 4.17. Field entry format for the “water abstraction intensity” indicator.
Field Format Expected response
Type of water abstraction Text Agriculture / Drinking water / Process / Energy
Water abstraction longitude Numerical Decimal degrees
Water abstraction latitude Numerical Decimal degrees
Zoning Text Active zone / Colonised zone
Habitat Text Estuary / River / Wetland / Lake
Annual water volume abstracted
Month of maximum water abstraction
Beginning of the activity
Numerical
Text
Text
Specify in m 3 the total water volume abstracted
over the year
Specify the month when water abstraction is
maximum
Specify the month when water abstraction
begins
End of the activity Text Specify the month when water abstraction ends
Presence of filtration structures
Yes/No
Openings of filtration structures Numerical Specify the opening size in mm
Presence of a protective device
Yes/No
Protective device spacing Numerical Specify the spacing of gaps in mm
Abstraction velocity Numerical Specify the abstracted flow in m 3
Life stage of removed eels
Text
For abstraction in active zone glass eel/elver /
juvenile eel
Assessment of quantity removed Numerical Specify the weight in kg
4.3.4.4. Data exploitation
Information is collated in a geographic information system which gives the estimated volumes of
individuals smaller than 30cm removed as a result of water abstraction in the active zone, the
annual volume removed and the month when removals reach a peak. The seasonal aspect is
important for potential comparison with other indicators (water quality, accidental mortality, etc.).
4.3.4.5. Applied examples
Direct effect of water abstraction on eel populations: the case of the Blayais Station in the
Gironde estuary
The evaluation principle is shown in figure 4.21.
149

Sample fishing trips
with ESTURIAL
Fisheries sampling
Migratory models derived
from the team’s work
CPUE density
Estimation of elver flux at water
abstraction sites of the Blayais
nuclear power station
Volumes of water
abstracted by the power
station
Evaluation of numbers likely to
be destroyed
Figure 4.21. Diagram of the estimation principle for removals due to water abstraction in the
Blayais nuclear station (source: CEMAGREF).
The trend in elver density during night fluxes was determined from samples taken at the Blayais
nuclear station abstracted water intakes.
The analysis of commercial fishing data in the area gives the general trend of elver migration over
the entire fishing season. Using the information obtained from commercial fishing and from the samples,
a complete series of average daily densities can be estimated.
Knowledge of the daily abstracted water volumes, which depend on power station activity, is
essential for the final estimation.
Work undertaken by Cemagref in 1994–1995 (Debenay et al., 1995) showed that 5.2 tonnes
gross weight of elvers per year were removed by the power station, i.e. around 30 kg of elvers a day
over a 6-month migratory season.
In 2000, the objective of a second study (Roqueplo et al,. 2000) was to estimate the mortality of
elvers that were sucked in by the station and passed through the cooling system (figure 4.22). In order
to achieve this, elvers were marked with various dyes selected in the laboratory (Rhodamin, Bismarck
brown, Neutral red).
150

BATIMENT
REACTOR
REACTEUR
BUILDING
vapeur steam d'eau
turbine
conventional auxiliaires
conventionnels auxilaries
1 Injection of des dyed civelles elvers colorées
2 Fishing Zone de zone pêche
Cooling Eau de refroidissement water abstracted pompée from dans the l'estuaire estuary
Discharged Eau rejettée water
refrigerant réfrigérant
condenser condenseur
circuit RIS circuit S.R.I.
circuit RRI circuit R.R.I.
nuclear auxiliaires unit
auxilaries ilot
nucléaire
water eau
Condenser eau refroidissement cooling
du condenseur water
PUMPING STATION DE STATION POMPAGE
Filtering Tambour filtrant drum
mesh maille 3 size mm3mm
1
9.64 380 mcm
WATER PRISE D'EAU INTAKE
refrigerant réfrigérant
80m 3 /s
Camoin Philippe
SEC circuit circuit S.E.C.
SEC pompe pump S.E.C.
spillway
déversoir
WATER REJET DISCHARGE D'EAU
2
10.91 2500 cmm
Figure 4.22. Diagram showing the principle underlying the estimation of mortality: a known
quantity of dyed elvers are introduced into the Blayais nuclear power station
cooling water (source: Cemagref).
Following their transit through the power station cooling system, elver mortality rate was
estimated at a maximum of 9%, 6 hours after being sucked in. This mortality increased 1 day later, then
levelled off. Over a week, the cumulative mortality rate reached 15%.
This method, used in the Gironde estuary, can be repeated in other estuaries. Its cost was
estimated to be 10 k€ for one tagging study.
4.3.4.6. Summary by river basin of information
concerning these descriptors
Table 4.9 outlines the information collected concerning these descriptors.
151

4.4. Stock enhancement and transfer of individuals
4.4.1. Context and objective
All available information shows that elver recruitment has been extremely low for 25 years and
reached its lowest ever level in 2001 22 . Urgent action is essential to restore the species. Stock
enhancement is one of the measures considered in the management plans.
Many European Union Member States have already resorted to transfer and enhancement
operations but mainly in order to support fisheries rather than to improve stock levels (Jacobsen et al.,
2004; Allen et al., 2006; Shiao et al., 2006). Currently, the critical state of the species and the urgency of
the situation dictate that such measures be taken with the sole aim of restoring the species and that
they be combined with measures aiming to reduce all kinds of anthropogenic mortality.
The eel’s biological cycle cannot be fully reproduced in a controlled environment, even though in
the last few years some progress has been made in producing leptocephalus larvae of different eel
species (Tanaka et al., 2003; Nomura et al 2004; Okamura et al., 2004). In this situation, stock
enhancement can only be achieved by collecting some individuals from their natural environment and
transferring them to restocking zones, sometimes with a pre-fattening phase in extensive or intensive
aquaculture. To simplify, the term “enhancement” will refer to a situation with a pre-fattening phase in a
fish-farm whilst the term “population transfer” will refer to the capture, transportation and release of
individuals held temporarily.
This is one component of the European regulation which establishes European eel stock
restoration measures 23 . However, effective stock enhancement must be part of an eel management
plan and the number of eels smaller than 20cm required to increase the escapement rate of silver eels
(and not to support one or more fisheries) must be specified.
There is no question of recommending enhancement operations in river basins, and still less the
transfer of individuals between river basins. The decision to include supporting measures into
management plans belongs to river basin managers. However, as stock enhancement operations tend
to be poorly managed and understood, it seems necessary, from a scientific point of view, to specify the
requirements of these operations, to quantify them, and to suggest a post-evaluation protocol in order to
measure their effectiveness.
22
See chapters 1 and 2.
23
Regulation (CE) No 1100/2007 of the Council on 18 September 2007, Official Journal of the European Union, appendix 2 of the
Indicang report, http://www.ifremer.fr/indicang.
152

A precautionary approach is essential as regards the transfer of individuals. Indeed, great care
must be taken in the transfer of individuals and the related risk of dissemination of pathogenic agents,
which means that these operations must be closely supervised.
These transfers must also be localised and quantified, as they may explain size-structure
anomalies already observed in the eel population or the suspicious presence of eels in naturally
inaccessible territories. This helps to distinguish between natural and artificial colonisation.
Furthermore, it is essential to ensure these actions do not stop or slow down developments aimed
at ensuring free circulation in the river basin.
4.4.2. Study scale
The relevant scale includes not only the river basin but also the transfer locations in order to
ensure the traceability of transferred individuals.
It is recommended to take eels from sites situated in the mesohaline part of the estuary (15 to 18
‰), which is usually the area where 5A to 6A3 pigmentation stages are caught, stages defined by the
identification method established by Elie et al. (1983) which is detasiled in appendix 6 24 of the Indicang
report..
Release sites must be carefully selected taking into account water quality, type of habitat,
presence of barriers to migration and the existing natural population (cf. specifications appendix 8.3 25 ).
As the objective is to increase the potential number of spawners produced by the hydrographic network,
the focus is on freshwater and downstream migration survival and these measures must therefore target
sectors where anthropogenic pressures are limited.
For a rapid transfer, glass eels/elvers are collected preferably from the estuarine zone, ideally in
the mesohaline part of the estuary (degree of salinity between 15 and 18 ‰). However, the risk of
osmotic shock means that they must be released in an environment where the difference in salinity (D)
with the initial environment does not exceed 10 ‰.
Otherwise, a 48-hour preliminary acclimatisation period in a quarantine tank is essential. For
example, if glass eels/elvers are collected in 15 ‰ water, salinity in the reception tank must be at least
equal to 5 ‰.
A storage time equal to or greater than 48 hours is quite possible if natural or artificial conditions
(aeration, filtration) allow it.
24
Grellier P., Huet J., Desaunay Y., 1991. Stades pigmentaires de la civelle Anguilla anguilla (L.) dans les estuaires de la Loire et
de la Villaine, Ifremer, appendix 6 of the Indicang report, http://ifremer.fr/indicang.
25
Soulier L., Muchiot S., Susperregui N., Urrizalki Oroz I., Girard P., 2007. Guide de remplissage des fiches terrain
recommandations pour le >, Ima-EKOLUR, appendiox 8.3 of the Indicang report,
http://www.ifremer.fr/indicang.
153

4.4.3. Data acquisition
A field-based protocol is used to acquire the descriptors needed to establish the “restocking and
enhancement” indicator.
The field sheet (figure 4.23 and appendix 8.3) on biological, geographical, and sanitary aspects of
the transfer must be completed for each alevin transfer operation. The coherence of this operation with
the river basin management policy must be specified.
FIELD SHEET -
GENERAL DATA
RIVER BASIN
ORGANISATION CENTRALISING DATA:
NAME OF THE OPERATOR :
OPERATING ORGANISATION:
DATE COLLECTED:
DATE RELEASED:
TRANSFER OPERATION
Stage transferred
Pigmentation stage of the glass eel / elver (if possible, otherwise
specify if transparent or black)
Health checks before release?
If yes, which ones
Young eel
Quarantine period before release?
If yes, for how long ?
Weights transferred (Kg)
Collection location
Name of the river
Coordinate
Release location
Name of the river
Coordinate
X
Y
X
Y
Treatment before release?
If yes, which
ones?
Cost of the operation
Reasons for eel transfer
Explain the reasons behind eel transfer, justify the stages transferred, the collection and release
What place does it occupy within the management policy of the river
basin?
POST EVALUATION
Type of monitoring put in place
Monitoring frequency
Monitoring duration
Monitoring cost
Figure 4.23. Information sheet concerning enhancement operations in the field.
Figure 4.23 outlines the descriptors and notes required. Information recorded on the sheets must
be entered into a database. Fields to be completed and their format are described in table 4.18.
154

Table 4.18. Field entry format for the “restocking and enhancement” indicator
Name of the field
Format/Type of
data
Description
Organisation centralising the
Specify the name of the organisation recording the information about the
Text
data
river basin
Operational organisation Text Specify the name of the organisation in charge of the operation
Name of the operator Text Specify the name of the person in charge of the operation
Date collected Date/Time Format dd/mm/yyyy
Date released Date/Time Format dd/mm/yyyy
Country Text Specify the country where the operation is carried out
River basin Text Specify the code of the river basin
Collection longitude Numerical Decimal degrees
Collection latitude Numerical Decimal degrees
River basin where released Text Specify the code of the receiving river basin
Release longitude Numerical Decimal degrees
Release latitude Numerical Decimal degrees
Life stage of transferred eels Text Enter the code CT / CN / AN or the pigmentation stage
Weight transferred Numerical Specify the weight transferred in kilogrammes
Health check before release? Yes/No Tick if the answer is yes
Type of health check Text Specify the health checks carried out
Quarantine before release? Yes/No Tick if the answer is yes
Quarantine duration Text Specify the quarantine duration
Treatment prior to release Yes/No Tick if the answer is yes
Type of treatment Text Specify the type of treatment
Transfer operation cost Monetary Specify the total cost of the operation in Euros
Reasons for alevin transfer
Text
Explain the reasons behind alevin transfer, justify the quantities and the
stages being transferred, relate the operation to the basin policy
Type of monitoring set up
Text
Specify the type of monitoring set up: biometrics, parasitology, pollutant
dosage, ….
Monitoring frequency Numerical Specify the frequency of monitoring operations (in months)
Monitoring cost Monetary Specify the cost of one monitoring operation in Euros
Monitoring duration Numerical Specify monitoring duration in number of months
Further information, in particular on conversion formulae, can be found in the completion
instructions for this sheet in appendix 8.3.
4.4.4. Data exploitation
Information is held in a geographic information system which provides collection and release
locations and dates, health treatment, the stage and weight of transferred eels and the types of
monitoring put into place.
As river basins are highly heterogeneous, it is essential to specify quantities in relation to the total
length and surface area of the river basin.
Maps of “stock enhancement” operations must be compared with a map of the network of
monitoring stations for the "presence/absence" of eels and with the diagrams of size distribution in order
to explain anomalies found in this distribution and in the abundance distribution within a basin 26 .
26
See chapter 8.
155

4.4.5. Summary by river basin of information concerning
restocking and population enhancement
Table 4.19 shows that, within the Indicang network, recorded enhancement or transfer operations
have been undertaken only in the French river basins.
156

Table 4.19. Information summary by river basin concerning the “stock enhancement” descriptors.
157

158

Chapter 5
Eel biological quality indicators
Patrick Girard, Patrick Prouzet, Nicolas Susperregui, Tony Robinet,
Éric Feunteun, Pascal Laffaille, Christian Rigaud.
159

Information concerning sampling strategies can be found in the chapter on additional
methodology 1 . It is of course important to note once again that sampling is a major operation which
fundamentally determines the observations available and the conclusions that can be drawn from them.
The sample must be representative of a larger set which is the population established in the sampled
area at a given time. This sample is characterized by a set of variables or descriptors for a given
quantitative characteristic (such as length, weight, or age, or a qualitative one, such as pigmentation
stage, degree of silvering, or health condition etc)
These descriptors can be combined into summary variables which are then used as indicators of
the biological quality of the collected sample, for example: condition or similarity coefficients, population
age structure or mean age, proportion of individuals below a threshold size, or the number of parasited
individuals.
5.1. Indicators concerning individuals’ biometric characteristics
5.1.1. The elver biological stage
5.1.1.1. Context and objective
In a given river basin, migrating elvers can be characterized by biological descriptors which take
into account the factors affecting the variability of morphometric and pigmentary features, and even
genetic variability (although this is not addressed here). Based on the various factors identified in
chapter 3, samples can be obtained which can be used, in particular, for inter-basin comparisons. This
is because, for a given basin, intraseasonal variability in weights, lengths and in the proportion of
pigmentation stages in the sampled scientific or commercial catches has already been recorded.
5.1.1.2. Study scale
The sampling area is located in the estuary, but it is important to know where individuals are
collected in order to compare the samples from one river basin to the next, as biological characteristics
evolve with the time spent in the estuary and therefore with the distance between the collection point
and the river mouth. Thus, information on the origin of the elvers must be given in relation to four
sectors where environmental parameters differ and where elvers have their own characteristics: the
river mouth, the saline and non-saline tidal zones, and the fluvial zone.
1 See Chapter 3 for information on sampling strategies.
160

The samples must be collected over a long period of time given intra- and inter-annual variability
of the biological characteristics at a given site. Their geographical variability must also be specified.
Hence the need for a trend analysis of the length, weight and pigmentation stage for a site, which is
considered to be representative of the relevant geographical zone (southern or northern Bay of Biscay
for example). Where many sites are concerned, it is recommended to collect samples over a long
period of time using stations with equivalent hydrological characteristics (for example, the non-stratified
propagation zone of the dynamic tide) and during the main migratory period (for example, Dec-Jan in
the south of the Bay and Feb-March in the north of the Bay), in order to be able to make pertinent intersite
comparisons.
5.1.1.3. Data acquisition
The biometric characteristics are studied on live or very recently deceased (less than 6 hours)
individuals caught during commercial or experimental fishing trips made during the migratory season.
Elvers are stored in the refrigerator, for a few hours only, to avoid fluctuations in the studied
parameters before their laboratory examination. Their size (in mm) and weight (to the nearest 0.01g) are
recorded.
Sampling frequency must be specified. Glass eels migrate throughout the year with a main
migratory period which depends on the latitude of the particular river basin: they arrive in Southern
Europe much earlier than in Northern Europe. When there is a fishery in the river basin, this arrival
usually coincides with the opening of the fishing season. It is important to specify how frequently
individuals are sampled during the main migratory season or over the whole year: weekly, monthly or
seasonal. Catch data show that lunar phases are synchronous with glass eel arrivals in the estuary. For
river basins involved in sampling, it is important to know whether the collection of these samples
coincides with the lunar phases and, if so, with which phase: new moon, full moon, first or last quarter.
A minimum number of individuals must be sampled but no more than around a hundred. During
sampling, for a given pigmentation stage, significant individual variability in terms of weight and length
can be observed. A minimum of around fifty individuals should be collected for the sample to be
sufficiently representative of the population diversity. In the case of very heterogeneous individuals
(advanced pigmentation stages and stage 5B not in the majority) sampling could be extended to 100
individuals to take into account the high biological variability.
5.1.1.4. Data exploitation
During sampling, the factors to be taken into account concern both the individuals and the
environmental characteristics at the time of collection. Therefore, the relevant descriptors are presented
according to these 2 main themes in table 5.1.
161

Table 5.1. Descriptors of the biological characteristics and their environmental background.
Themes Descriptors Objectives Guidelines
Characterisation of individuals
Size
Weight
Pigmentation stage
Condition or similarity
coefficient
Indicator of new glass eel flux
arrival in the estuary
Pigmentation speed
In mm
In mg
According to the definition in
the reference document *
Water temperature
Daily measurements of water
Environmental parameters at
Tidal coefficient
temperature, flow and tidal
the time of collection
Migratory speed
coefficient.
Flow
* Azpiroz I, Cuende F.-X., Damasceno-Oliveira A., Diaz E., Lauronce V., Mendiola I., Urrizalqui Oroz I., 2007. Indicang Glossary –
French language with transaltions of terms in English, Spanish and Portugese, annex 5 of the Indicang site,
http://www.ifremer.fr/indicang.
We will not elaborate on the classical statistical methods used to compare mean weights and
lengths. They derive from the classical mono- or multi-factorial analysis of variance (Sokal and Rohlf,
2001) or deviance analysis when using the general linear model (Venables and Ripley, 2002), as well
as from classical covariance analysis when comparing allometric relationships between length and
weight (Mayrat, 1959; Draper and Smith, 1966; Cox and McCullagh, 1982). De Casamajor et al. 2000,
provide an illustration of the use of these methods to identify glass eel or elver fluxes from these
biometric characteristics.
Condition coefficient
This first index is the power in the allometric relationship between the individuals’ weight Y(t) and
length X(t), at a given stage and sampled at a given time (t), such that:
Jolicoeur and Heusner, 1971).
k
Y = bX (Teyssier, 1960 ;
The term “isometry”, i.e. proportionality in the relative growth of each part of an individual or
invariant density of an organism, is used when k is not significantly different from 3 (as is the case when
the weight varies with the cube of the length). The terms negative and positive allometry are used
respectively when k is below and above 3. The following approximation is often used to calculate this
Y
Y
coefficient: K = cste ∗ or more generally K = cste ∗ .
3
b
X
In order to test the differences in the slope of allometric relationships, about a hundred pairs of
values are used, as widely dispersed as possible. In this case, significant variations of the allometric
coefficient can be observed as noted by de Casamajor et al. (2000). This coefficient ranges from 2.4 to
3.11 and for some groups in the sample at an early stage of pigmentation (VA or VB) is significantly
different from 3. This coefficient serves to differentiate the various glass eel groups entering the estuary,
which at that time are characterised by an isometric relationship between weight and size, then by
negative allometry indicating weight loss.
a
X
162

Similarity coefficient
Gould (1971) suggested a similarity index s
significantly different for the various collected samples.
⎡bt
+ 1
⎤
s = ⎢ ⎥⎦
⎣ bt
1
( 3−k
)
, it is used when the k slopes are not
Gould (1966, 1971) refers to “static allometry” for a given life stage. This is why, if the positions of
the b curves differ, it is of interest to compare the weights for a given length at instants t and (t+1). This
is equivalent to comparing the positions of the centres of gravity in the scatters of points of the pairs
[Y(t),X(t)] and [Y(t+1),X(t+1)] with a correction related to the difference in slope between k (mean slope
of all allometric relationships) and 3. If s increases between (t) and (t+1), the relative weight of glass
eels/elvers at exit point (t+1) is larger than that at exit point (t) and vice versa.
Figure 5.1 illustrates the use of this index during the sampling programme carried out from
November 1997 to March 1998. The arrival of 4 different groups of glass eels into the estuary during
that period was detected through the very significant variations in the similarity index.
Similarity index (s)
Indice de similarité (s)
3
2,5 2.5
2
1,5 1.5
1
0,5 0.5
0
16-nov-97
21-nov-97
26-nov-97
01-déc-97
06-déc-97
11-déc-97
16-déc-97
21-déc-97
26-déc-97
31-déc-97
05-janv-98
10-janv-98
15-janv-98
20-janv-98
25-janv-98
30-janv-98
04-févr-98
09-févr-98
14-févr-98
19-févr-98
24-févr-98
01-mars-98
06-mars-98
(?)
Figure 5.1 Evolution in the similarity index during the 1997/1998 programme in the Adour
estuary. Arrows show the arrival of a new glass eel group into the estuary - (?) indicates poorly
identified flux (from De Casamajor et al., 2000).
163

5.1.2. Biological stages: the yellow and silver eel
5.1.2.1. Indicators characterizing size and age structures
Size and weight
These parameters are important for the success of various analyses:
• size-group monitoring analysis (and in particular identification of the active and colonisation zone
within a river basin, location of female individuals, etc.);
• analysis of the selectivity of the fishing gear used;
• calculation of the condition coefficient (body-condition of individuals);
• evaluation of the total mortality rate and of the anthropogenic mortality rate. In this context, it might
be noted that the significant presence of large-size individuals (>70-75 cm) in all types of deep
environment, or close to deep environments, should be taken to indicate that anthropogenic
mortality is not excessive.
In tidal areas or in rivers, the access to this information generally relies on the specific
monitoring of fisheries using yellow eel fishing gear. Observations should be spread across various
points of the relevant area, covering, when possible, several fishing days (as catches of large individuals
are proportionally higher during the early harvests) and two different seasons (spring and end of
summer/beginning of autumn), as autumn monitoring provides the silver index.
For electrofishing monitoring, the very heterogeneous distribution of various sizes in the
different types of habitats must be kept in mind. Thus, although the exploration of shallow waters by day
leads to efficient fishing, the likelihood of observing individuals larger than 30cm is very limited. This
heterogeneity, which also concerns observed abundance, makes it very difficult to assign a size
structure, even to a section of the watercourse between two constructions.
As regards weight, once the relationship between size and weight has been established for each
large compartment of the river basin during the initial operations (200 individuals minimum), further
interventions can be limited to collecting the size of individuals.
Age
It is, of course, of interest to know the relationship between age and size within a river basin. The
difference between the saline estuarine zone, the tidal fluvial zone and the upstream zones must be
kept in mind. Within the upstream zones, marked differences can also emerge between sub-basins
(water quality, productivity).
This relationship can be determined in two ways:
164

• sacrifice of individuals followed by examination of their otoliths after extraction, mounting, sanding
and colouring. This bony structure must be examined by three persons trained to read the age and
working independently. In all, these operations take about ¾ of an hour to an hour per otolith. A
minimum of 5 individuals by centimetre of linear size and by compartment must be examined. When
no information on the local growth level is available for a basin, it is recommended to start with the
tidal zone, using the samples collected for sex ratio observation.
• individual PIT tagging and monitoring over an area. This approach is limited to small streams or
canals which generally have significant recapture rates (5 to 10%) from one year to the next.
5.1.2.2. Sex ratio
The sex ratio of silver eels produced by a given zone more or less determines the size structure
observed in that zone. Males (size at silvering ranging from 27 to 45cm in our regions compared to
40cm to more than 1m for females) are known to be present in heavily populated zones, therefore often
close to the sea.
Customarily, male spawner production occurred for the major part, and in some cases
exclusively, in estuaries, small coastal watercourses, coastal dyked marshes and the downstream
reaches of river basins. The fall in total and fluvial recruitment, in parallel with the fall in density or in
abundance indices, including in the zones that are close to the sea, has often resulted in an increasingly
significant number of females. At present, it is important therefore to focus on tidal zones for periodic
sex-ratio observations by a specialised laboratory. These observations focus on the 30-45cm size range
and, within this group, on the proportion of males among the differentiated individuals. In the estuarine
zone, a size structure practically truncated at 45cm with a low % of males (less than 1/3 of the
differentiated individuals) is considered to be a clear signal of a dysfunctioning system (low recruitment
and excessive anthropogenic mortality).
In order to obtain this information, 75 to 100 individuals, 30 to 45cm long (5 to 6 individuals by cm
of linear size) have to be collected every 3 years, particularly when monitoring fisheries using basket
traps. In order to optimise sample use, an external and internal health examination, age determination
and the measurement of concentrations of various contaminants (PCB, heavy metals, pesticides)
should be scheduled on these same individuals. The storage of gonads by freezing should be avoided.
All, or a sample, of the gonad is extracted from a fresh individual in good condition and stored in Bouin
solution (picric acid, formaldehyde) so that the cells, which will be examined at a later stage, are not
damaged.
165

5.2. Indicators concerning the pigmentation status or the silvering
process.
5.2.1. The elver biological stage
5.2.1.1. Context and objective
It is important to note that the relationship between the pigmentation stage and the residence time
in the estuary can only be demonstrated if the temperature of the estuary is taken into account as it
plays a role in the speed of pigmentation 2 . However, an early stage of pigmentation, such as in stage 5A
or even 5B, can testify to the recent arrival of elvers in the river. The aim here is not to age elvers
migrating upstream accurately but to define an indicator for the manager (% of stage 5A elvers for
example) which indicates whether elver migratory speed has slowed down in the estuary because of a
natural blockage (hydrological phenomenon) or a physical one (dam).
5.2.1.2. Study scale
This is both spatial and temporal and similar to that defined for the analysis of the evolution of
elver sizes and weights. The collection zone is situated in the parts of the estuaries affected by the
propagation of the dynamic tide. The main period of elver migration is selected as the sampling period,
This spreads from November to April and from the South to the North of the central colonization zone
which corresponds to the INDICANG river basin network.
5.2.1.3. Data acquisition
Once the biometric measurements have been made, the pigmentation stage is determined
according to the established criteria 3 . Individuals caught a few hours earlier and stored at low
temperatures, in order to avoid the development of pigmentation, are examined through a binocular
magnifier. The stages identified range from the totally transparent glass eel, 5A, with no melanisation
whatsoever on the cephalic spot, to the most advanced stage, 6A4, when melanisation covers the whole
tegument, which becomes opaque. The young eel stage is a transition towards a change in life style.
This stage was not found in the samples taken from estuaries. The most frequent stage found in
estuaries is the stage 5B in a proportion generally greater than 80%. However, this percentage
fluctuates very rapidly depending on harvesting and environmental parameters hence the importance of
standardised sampling.
2 See Chapter 2, § .
3 Grellier P., Huet J., Desaunay Y., 1991. Pigmentation stages of glass eel Anguilla anguilla (L.) in the Loire and
Villaine estuaries, Ifemer, Annex 6 of the Indicang report, http://www.ifremer/indicang.
166

5.2.1.4. Data exploitation
The aim is to obtain a sample of about thirty individuals (a sample size generally acceptable to
biologists), which should, however, be increased if the variability of the criterion (here the pigmentation
stage) increases, or if we are looking for a rare event. Simple synopses, such as the ones presented in
table 5.2, then give an idea of the variation in this proportion within several samples collected at the
same place at different times or in different places at the same time. In particular, they show an increase
in the more advanced pigmentation stages in the estuary at the beginning of the main upstream journey
in this estuary, which takes place from February to April.
Table 5.2 Evolution in the proportion of the different pigmentation stages in the Loire sampling
station in 2005 (from Ifremer / Adera, 2005).
% 4 Jan 5 Jan 12 Jan 13 Jan 18 Jan 19 Jan 15 Feb 16 Feb
5A 25,4 24 7 21 13 15 10 7,5
5B 74,6 75 91 76 83 81 78,2 77,5
6A0 1 2 3 3 4 6,4 7,5
6A1 1 3,6 7,5
6A2 1,8
5.2.2. Silver eel biological stage
5.2.2.1. Context and objective
Silvering marks the end of the growth phase and energy, until then allocated to growth, is now rechanneled
into reproduction (Fontaine, 1994). Silvering is also characterised by a spectacular increase
in the weight of the ovaries in females, as shown by the increase in the gonadosomatic index (GSI) from
0.3 to 1.5 (Dufour, 1985). The follicular volume is multiplied by 125, whilst the apparent number of
follicles drops by up to 96% (follicular atresia). The consequence of this atresia, at the precise time of
silvering, is to reduce the share of energy dedicated to gonad maturation and allocate it, almost entirely,
to migration (Fontaine, 1994). However, eels are incapable of spawning while they are still inland
because their production of gonadotropic hormone is blocked (Dufour, 1985; Kah et al., 1989).
5.2.2.2. Study scale
The scale of the study is the river basin. The characteristics of the river basin where eels grow
from the time of their fluvial recruitment to their silver metamorphosis comprise the River Basin
167

Context. Given the key influence of the environment on the sex-ratio 4 , it is of course the decisive factor
affecting the population structure of silver eels produced by the basin and therefore it’s Reproductive
Potential. The Basin Context also includes the chemical effects (water and sediment pollution)
potentially able to jeopardise successful future reproduction.
5.2.2.3. Data acquisition
Silvering (the name comes from a change in pigmentation to adjust to marine life) is characterised
by 3 striking external signs (Durif et al., 2005 ; Acou et al., 2005 ; figure 5.2):
Figure 5.2. The 3 external signs of silvering. (a) Middle portion of a silver eel (Anguilla anguilla)
showing a lateral differentiated line studded with visible neuromasts and a very clear contrast
between the dark back and the silver white belly. (b) A silver eel’s head (A. anguilla) showing the
typical ocular hypertrophy. A. horizontal ocular diameter; B. vertical ocular diameter; Bl.br.ba.,
black-brown back; D.f. Dorsal fin; Nm. Neuromast; Si.wh.bel. Silver white belly; Un.cj.
unpigmented conjunctiva (from Acou et al., 2005).
• a full lateral line, studded with visible neuromasts (Acou et al., 2005);
4 See Chapter 2.
168

• ocular hypertrophy, raising the Pankhurst index (1982) above 8.0 (Ancona, 1927; Todd, 1981a;
Pankhurst and Lythgoe, 1982);
• a pigmentary contrast between the back, generally black (increase in dorsal melanin) and the belly,
generally silver white (increase in ventral purine).
Furthermore, pectoral fins are generally overdeveloped and a silvery gold colour.
Thus, within a population of eels, those who are ready to migrate downstream can be identified
and thereby the proportion of eels who are potential spawners 5 .
5.3. Indicators concerning the quality of the individuals: health condition
and chemical contamination
5.3.1. Health condition
5.3.1.1. Context and objective
The objective of this indicator is to assess, without sacrificing individuals, the health condition of
the eel population in the various river basins. The diagnosis of the health condition of the local
population is, of course, the preliminary stage of any rational management plan, for several reasons:
• particularly when transferring elvers or young eels: transfer of healthy individuals into healthy
ecological zones confirmed.
• spawner quality determines the probability that they will reproduce successfully, in the broad sense
of the term (including sexual maturation, migration towards spawning areas, reproduction and
survival of offspring). Their quality is reduced by anthropogenic effects, pollution sources or by the
threat of new pathogenic agents. A qualitative diagnosis of silver eels escaping from a basin must
be established in order to apply an Escapement Potential reduction coefficient, proportional to the
likely decrease in the reproductive success of the eels that escape.
• external anomalies are likely to result from the global quality of the environment colonised by the
species. The INDICANG partners recommend, therefore, the continuation of research aiming to
establish the relationships between environmental quality and the health condition of the
populations.
This sub-chapter is not intended to replace the health manual 6 produced by the Indicang project
(Girard and Elie, 2007), but it provides more precise guidance, in particular on the collection of
5 For a discussion of these observations, see Chapter 9.
6 Girard P., Elie P., 2007. Manuel d’identification des principales lesions anatomo-morphologiques et des principaux
parasites externes des anguilles, Cemagref, report 110, annex 4 of the Indicang report,
http://www.ifremer.fr/indicang.
169

supplementary data to define the environmental context and standardise data input, with the aim of
creating local databases, which are usable and comparable at larger scales.
5.3.1.2. Study scale
This method can be applied to all of the area currently colonized by eels. However, a careful
distinction must be made between the different types of zone, i.e. estuary/watercourse, wetlands/lakes
and their location in relation to the dynamic tide. Water salinity is an important factor to record when
describing the environment.
This method can be applied all year round but extra effort must be applied to the yellow and silver
stages during the summer seasons, as low water levels, often associated with an increase in water
temperature and a decrease in oxygen concentration, promote the development of lesions and
parasites.
5.3.1.3. Data acquisition
Given the critical status of the species, it is important to limit the sacrifice of individuals, and
essential to obtain the maximum information when samples are taken (table 5.3).
Table 5.3. Identification of descriptors obtained with or without sacrifice (Grisam, 2006).
Without animal sacrifice
• Size and weight
• Ocular diameters and length of pectoral fin
silvering stages
• Anguillicola crassus infestation rate by ultra-sound
• External health condition
With animal sacrifice
• Contamination by xenobiotics
• Internal health condition
• Age (otoliths)
• Sex (for individuals < 45 cm long)
• Degree of gonadic maturation
The sacrifice of individuals to monitor the sex-ratio can be used to check the internal health condition.
The observation and rating protocol suggested by Lefebvre et al. (2002) is used to look for a present or
past contamination by Anguillicola crassus.
Spawners’ quality can be assessed through the establishment of a health status by management
unit (hydrographic basin). The main components of this health status are as follows:
• parasitic diseases, bacterioses and mycoses, which generally indicate a decline in the eels’
health condition while some parasites (Anguillicola crassus, Pseudodactylogyrus sp.) and
viruses (EVEX, Herpes Virus) can potentially compromise the reproduction outcome directly;
• characterisation of the individual physical state (various physiological/histological/cytological
parameters);
170

Figure 5.3. Information on the environmental context of catches.
172

Figure 5.4. Description of pathological lesions: individual data collection form (number of
individuals

Information recorded on the forms must be entered into a database. The fields to be completed
and their format, as shown in tables 5.4 and 5.5, are described in the completion instructions for the
data collection forms > 7 .
Figure 5.5. Description of pathological lesions: general data collection form (number of
individuals >30).
Information recorded on the forms must be entered into a database. The fields to be completed
and their format, as shown in tables 5.4 and 5.5, are described in the completion instructions for the
data collection forms > 8 .
7 Soulier L., Muchiout S., Susperregui N., Urrizalki Oroz I., Girard P., 2007. Guide de remplissage des fiches terrain
et recommendations pour le >, Ima-EKOLUR, annex 8.3 of the Indicang
report, http://ifremer.fr/indicang.
174

Table 5.4. Field entry format for "Health status" indicator: environment form
Name of the field Format/Type of
Description
data
Country Text Specify the name of the country
River basin Text Specify the code of the river basin
Organization centralising
the data
Text Specify the name of the organisation recording the information
about the river basin
Operating organisation Text Specify the name of the organisation responsible for field
observations.
Name of the enumerator Text Specify the name of the person who recorded the information
Zoning Text Specify the typology letter
Name of the zone Text Specify the name of the relevant estuary, river, wetland or lake
Longitude X Numerical Decimal degrees
Latitude Y Numerical Decimal degrees
Date of the study Date/Time Format dd/mm/yyyy
Mode of harvesting Text In the case of sampling by fishing, specify the type of fishing gear
used
Targeted stage Text Enter the letter corresponding to the eel stage
Numbers harvested Numerical Specify the number of yellow and silver eels and the weight of
elvers caught
Water temperature Numerical Specify the water temperature in degrees Celsius
Dissolved oxygen Numerical Specify the concentration in dissolved oxygen in milligrammes / litre
Oxygen saturation Numerical Specify the oxygen saturation ratio in %
Conductivity Numerical Specify water conductivity in micro-Siemens/centimetre
Salinity Numerical Specify salinity in grammes/litre
pH Numerical Specify the pH value
Flow Text Specify if the flow is fast/average/slow
Colour Text Specify if the water is green/brown/other
Turbidity Text Specify if turbidity is high/low/non-existent
Algae Yes/No Tick the box if answer is yes
Mosses Yes/No Tick the box if answer is yes
Smell Yes/No Tick the box if answer is yes
Pollution Yes/No Tick the box if answer is yes
Presence of sick eels Yes/No Tick the box if answer is yes
Stages of the sick eels Text Enter the letter corresponding to the eel stage
Number of sick eels Numerical Specify the number of sick eels observed
Presence of dead eels Yes/No Tick the box if answer is yes
Stages of the dead eels Text Enter the letter corresponding to the eel stage
Number of dead eels Numerical Specify the number of dead eels observed
Known pathological
antecedents
If answer is yes, dates and
references
Observation
Yes/No
Text
Text
Tick the box if answer is yes
Specify the month and year and the reasons
8 Soulier L., Muchiout S., Susperregui N., Urrizalki Oroz I., Girard P., 2007. Guide de remplissage des fiches terrain
et recommendations pour le >, Ima-EKOLUR, annex 8.3 of the Indicang
report, http://ifremer.fr/indicang.
175

Table 5.5. Field entry format for the "Health status" indicator: pathology form
Name of the field Format/Type of data Description
Country Text Specify the name of the country
River basin Text Specify the code of the river basin
Organisation centralising the
Specify the name of the organisation recording the
Text
data
information about the river basin
Operating organisation
Text
Specify the name of the organisation responsible for field
observations.
Name of the enumerator
Text
Specify the name of the person who recorded the
information
Zoning Text Specify the typology letter
Name of the zone
Text
Specify the name of the relevant estuary, river, wetland or
lake
Longitude X Numerical Decimal degrees
Latitude Y Numerical Decimal degrees
Date of the measurement Date/Time Format dd/mm/yyyy
Eel number Numerical Allocate a number to the eel
Life stage of the eel
Enter the letter corresponding to the eel stage
Length Numerical Enter the size of the eel in millimetres
Weight Numerical Enter the weight of the eel in grammes
Sex Text Enter the letter corresponding to the sex of the eel
Lesions Text Enter the 2 letters corresponding to the type of lesion
Anatomical localisation of the
Enter the letter corresponding to the anatomical
Text
lesions
localisation of the lesion
Code for lesions / anatomical
Enter the 3 letter code: 2 letters for the type of lesion + 1
Text
localisation
letter for the localisation
Importance of the lesions
Numerical
Enter the class number corresponding to the importance of
the lesion
Parasitism Text Enter the 2 letters corresponding to the type of parasite
Anatomical localisation of the
Text
Enter the letter corresponding to the anatomical
parasites
Code for parasitism /
anatomical localisation
Parasite abundance
Text
Numerical
localisation of the parasite
Enter the 3 letter code: 2 letters for the type of parasite + 1
letter for the anatomical localisation
Enter the class number corresponding to the abundance of
the parasite
Photo number Text Enter the photo code = basin code + eel n°
5.3.1.4. Data exploitation
In a global health approach, the prevalence of external alterations can be calculated from the
analysis of collected data, which provides information on the global health condition of the population
in different points of the basin.
In a specific health approach, this method can provide the prevalence of a distinctive alteration
(e.g. necroses). The emergence of certain lesions may indicate the presence of some undesirable
substances in the environment and help with their detection (table 5.6).
176

Table 5.6. Main causes of anatomo-morphological alterations (chemical causes).
Anatomo-morphological alterations
Many pathological manifestations
Thinness
Tumours and other lumps
Malformations, deformities
Colour alteration
Absence of organs
Ocular lesions
Haemorrhages
Haemorrhagic ulcers
Necroses
Erosion
Potential causes
Acute bacterial or viral septicaemia
Internal parasitism, micropollutants, chronic infections,
nutritional deficiency
Parasites (myxobolus), chronic infections, viruses, pollution: oil,
PAH, DDT, PCB, amines, As, X rays
PAH, organochlorides (pesticides, herbicides), heavy
metals (Pb, Cd), vitamin deficiencies, gas oversaturation,
parasitism, tumours
Viral, bacterial or parasitic infection, stress, CO2 excess,
hypoxic state, irritations
Generalised bacteriosis, injuries, cannibalism, traumatism, gas
oversaturation, parasite
Generalised bacteriosis, virosis, parasitosis, gas oversaturation,
metabolic disorder (nephrocalcinosis), traumatism,
micropollutants (PAH)
Infectious diseases (bacteriosis, virosis), parasitism, mycoses,
irritations, injuries, vitamin A deficiency
Traumatism (predators), parasitism, infection, chemical
pollutions, ammonia, hydrocarbons (flat fish)
Crude oil, Cd, Cr, Hg, paper pulp waste, bacteriosis, virosis,
external parasite, burns (UV), traumatism, cannibalism,
deficiencies
Bacteriosis, external parasite, nutrition or vitamin deficiency,
unfavourable environmental factors and chemical
pollutions (crude HAP, Cd), burns (sun UVs)
With this specific approach, the issue of “contaminants” can be prioritised with respect to the eel
in the different basins and the zones and molecules to monitor can be targeted. Furthermore, routine
use of this method highlights the frequency and the period of emergence of these pathological
manifestations.
All this information is summarised in a geographical information system, which makes it possible
to target (i) the zones presenting a health risk where a special effort on environmental quality (water -
sediment) is needed, and (ii) the virgin zones to protect and to recommend for transfer operations.
This map of the populations' health status must be correlated with that of habitat quality in order
to highlight the relationship between the prevalence of anatomo-morphological alterations and the
environmental quality.
The medium-term objective is to construct a table similar to the one presented in table 5.7.
177

Table 5.7. Example of quality table to be set up (source: Girard, 1998).
Classes Prevalence Water quality
0 - 1 % NS Excellent
1 - 5 % Low Good
5 - 20 % Average Average
20 - 35 % High Poor
> 35 % Very high Very poor
These health aspects must be taken into account in stock enhancement operations. A control
sample of elvers must be examined before any transfer operation. If the prevalence of lesions exceeds
5 %, the health risk is too high for the transfer to be undertaken.
This prevalence chart can also help identify virgin zones suitable to receive healthy elvers.
Some diseases can be deadly (e.g. Aeromonas, Pseudomonas, Vibrio for example ). It is
therefore necessary to correlate the prevalence chart of alterations due to potentially lethal pathogenic
agents with that of accidental mortality.
5.3.1.5. Applied example
To date, there is no example available on the observation of external lesions as this approach is
new. However, similar work was undertaken on the Anguillicola crassus parasite in the Adour in 1998.
Maps (figures 5.6 and 5.7) of Anguillicola crassus infestation prevalence and intensity were produced,
and these show the degree of infestation in relation to the geographical location of the samples.
178

Prevalence of
Anguillicola crassus
Prevalence of Anguillicola crassus infestation in
harvested eels
Figure 5.6. Map of Anguillicola crassus infestation prevalence (source : Bellet et al., 1998 -
Migradour).
179

Anguillicola crassus
infestation intensity
2 parasites per eel
10 parasites per eel
20 parasites per eel
Anguillicola crassus infestation intensity in
harvested eels
Figure 5.7. Map of Anguillicola crassus infestation intensity (source : Bellet et al., 1998 -
Migradour).
180

5.3.2. A synopsis of the quality of individuals and current
knowledge by basin
Table 5.8 presents the observations that must be collected in order to have an idea of the quality
of spawners or, more generally, of the individuals produced in a given basin. Currently, analyses carried
out on chemical contaminants can only reveal the quantities that have built up in the various tissues of
the eel but do not really allow conclusions to be drawn about the consequences and effects of this toxic
accumulation in contaminated animals. The present state of knowledge, in particular of the cellular and
genotoxic effects of chemical substances, is not yet sufficient to establish the bases of future
management plans of the species. At present, the eel is used as a kind of “ecosystemic probe”, which
accumulates many chemical substances, in particular those that are liposoluble, in a given environment.
Table 5.9 shows that much work remains to be done in order to improve our understanding of
individuals’ health condition and their aptitude to produce quality spawners.
However, it is important to acquire this knowledge so that population transfers, which could well
intensify in future years, do not become the cause of parasite and disease dissemination. The
knowledge is also needed so that spawner quality can be taken into account through the Escape
Potential reduction coefficient 9 . The calibration of this coefficient requires a national health evaluation,
as recommended by the GRISAM Eel group (Scientific Interest Group for Amphihaline Fish - France).
9 See Chapter 9.
181

Table 5.8. Information required to asess eel quality at various stages of its biological cycle.
Post Analysis Relevant organ(s) Methods Estimated cost per
individual
Search for parasites Anguillicola crassus prevalence Gas bladder Opening and counting 38 €
38 €
Pseudodactylogyrus sp. prevalence Branchiae Dissection and counting
Sampling constraints
Fixation (alcohol)
Search for viral and bacterial
infections
Physio-histocytological
diagnoses
Presence of viruses (EVEX,
herpes virus…)
Blood Serology seroneutralisation 15 € Specimen taken from fresh
individuals
Pathogenic bacteria
Spleen, kidney, blood cells and any - In situ: diagnostic guidance (pathology codes) Training of staff On fresh individuals
organ or tissue presenting lesions - Laboratory :agar cultures
collecting specimens
Pathology codes (indicative of the
general health condition)
Skin, fins, branchiae Ante-mortem diagnosis Training of staff
collecting specimens
On fresh individuals
Condition coefficient Whole Individual Size and weight measurement On fresh individuals
Intra muscular lipid concentration Muscles Tissue percolation with petroleum-ether On fresh individuals
Hepatic necroses Liver Microscopic analysis of a thin liver section On fresh individuals
Gonado-somatic index Gonads Gonad weight / body weight On fresh individuals
Growth rate Otolith Otolith annuli analysis 15 € No
Nucleic acid content Blood Flow cytometry To be completed Sample from fresh
individuals
Circulating vitellogenin Blood ELISA dosage Blood taken on fresh
individuals
Female fecundity Gonad Ovocytes – diameter and number On fresh individuals
Dosage of organic and
metallic pollutants
Dosage of organic pesticides Muscles Gas chromatography ~100€ Freezing
TEQ dosage of dioxin-like
Muscles ** (+ gonads for some) DR-CALUX® method ~120 € Freezing
compounds (PCBs, dioxins, furans).
Dosage of PAHs Gall bladder Gall-bladder spectrofluorimetry Low (less than 5€ On fresh individuals
Dosage of metals Liver and muscles Mass spectrometry on delipidised tissues Low ~10€ On fresh individuals
182

Table 5.9. Summary of knowledge on the health condition of eel populations by catchment.
183

Chapter 6
Fisheries abundance and pressure
indicators
Gérard Castelnaud, Laurent Beaulaton
184

6.1. Presentation and methodological contents
By its very nature, fisheries management science (Stephenson and Lane, 1995; Richards and
Maguire, 1998) only applies to species exploited by fishing. The “fishery system” (i.e. a fishery in its
ecological, economic and social environment) is studied because the question frequently arises
concerning pressure on the resource and its capacity to renew itself. This question interacts with the
sustainability issue of commercial or recreational fishing activity and its future in terms of users and
employment, vessels, investments, markets and consumption. Fisheries management science must be
based on the precautionary (Garcia, 1994; FAO, 1993,1996a, 1999; Brethes, 1999) and the
sustainability principles (Legay, 1993; Caddy and Griffith, 1995; FAO, 1996a) and lead to responsible
fishing (Bourg, 1993; FAO, 1995, 1998).
The eel (Anguilla anguilla) is a species that is heavily exploited at every stage of its biological
cycle in its distribution area, by professional fishers, by amateurs with nets, traps, and lines and by
poachers. Within the framework of the Indicang project, it was recognised that this fishery (which exists
on most of the selected river catchments) plays an essential social, economic and cultural role. However
this methodological handbook focuses principally on the evaluation of fishing pressure on eel
populations and on using fishing to monitor the relative abundance of the different eel ecophases.
Hence, two groups of fisheries descriptors and indicators relating to fisheries biology and
fisheries socio-economics were selected, although the second group is only presented briefly. In fact,
no fisheries sociologists or economists were involved in Indicang. Such experts are generally few in
number in fisheries and quasi-non-existent in inland fisheries monitoring.
Furthermore, it must be noted here that some of the fisheries biological descriptors become
(within the framework of inland fisheries) socio-economic descriptors with a simple change in
terminology: nominal effort is usually the number of fishers; the total catch of a species is its production;
multiplying production by an average price gives the species’ value or turnover.
Chapters 7, 8 and 9 on the various ecophases (recruitment, colonisation and sedentarisation, and
escapement) and chapter 4 on the environment cross-refer to this chapter (6) for the theoretical
definitions and bases of population dynamics, for the collection and estimation methods, for the analysis
and interpretation of relative abundance descriptors and indicators, and for fishing mortality indicators
(fishing pressure).
We deal here with fisheries biological descriptors and indicators, attempting to show their
foundations and usefulness, to explain and clarify the concepts and their limits, and to introduce
definitions and their practical consequences, relying for this on the reference works on the theory of
185

population dynamics (Beverton and Holt, 1957; Gulland, 1969; Ricker, 1980; Laurec and Le Guen,
1981).
Methods to study the fishery system and to collect data, the conditions for successful outcomes,
together with data processing and descriptor estimation methods are discussed below. The indicators
which emerge from this discussion are given along with their use, analysis and interpretation. The
evaluation of the trend in eel abundance by life stage is used to provide concrete examples of these
methods.
A general outline on statistical fisheries monitoring methods and on necessary data can be found
in the FAO reference document (1999) on the fisheries data collection programme, which in part,
provided the basis for chapter 5 of the Handbook of Fish Biology and Fisheries by Evans and Grainger
(2002). The older Caddy and Bazigos FAO handbook (1988) develops sampling-based fisheries control
and monitoring methods. The highly topical forthcoming thesis of Beaulaton (2007) discusses the
application of these methods to inland waters, and presents complex data processing and analyses of
fisheries descriptors. Illustrations are mostly based on the middle reaches of the Loire for silver eel
fishing and on the Gironde for the other two life stages. The “research” statistical monitoring system of
catches in the Gironde and its results were reviewed using the more efficient mathematical tools from
the publications of Castelnaud et al. (2001), Beaulaton and Castelnaud (2005), and Beaulaton et al
(2006) and the thesis cited above.
Please refer to Appendix 9 1 for a succinct presentation of some fisheries socio-economic
descriptors and indicators. These are particularly useful to managers when assessing the importance
of eel fishing in relation to overall fishing activity in a catchment, using, for instance, fishers’ turnover
figures. The main source is the recent work on coastal fisheries “PECOSUDE” (Léauté and Cail-Milly,
2003), which concerned the Indicang area countries (except for the United Kingdom) and involved some
of the French Indicang scientists in pilot catchments.
6.2. Fisheries statistical monitoring
According to FAO reference guidelines (Holden and Rait, 1974; FAO, 1999), fisheries
assessments should combine biological, economic, socio-cultural and compliance indicators to guide
management decisions. The production of indicators requires the collection of data from fisheries and
their environment using appropriate methodologies. “Fisheries (or catch) statistics” encompass basic
data on catch and fishing effort, together with the descriptors and indicators which derive therefrom and
their analysis in the context (and operation) of the fishery.
They cannot be dissociated from fisheries (statistical) monitoring systems which define data
collection. Unlike conventional statistics which focus on the activity, fisheries statistics, in addition to
186

their socio-economic dimension, have a biological dimension that focuses on harvested species and it
is this that makes them of interest to the fisheries biologist.
Depending on data accuracy and reliability, fisheries statistics with a biological and socioeconomic
focus should provide:
* the relative abundance level of species;
* information on some of the biological processes;
* an estimate of production - tonnage and value.
These biological and socio-economic aspects are complementary and hierarchical because if
basic catch and effort data are obtained of a quality sufficient for biological usage (which is at the métier
level) then socio-economic indicators (which are best situated at the species level) can necessarily be
calculated.
Fisheries statistics
Socio-economic
objective
X
Biological
objective
Figure 6.1 -
The objectives of fisheries statistics and their inter-relationship
At the beginning of the 21 st century, it is no longer necessary to demonstrate the importance of
fisheries statistical monitoring: it has long been emphasized by researchers and fisheries managers, by
international organizations such as FAO and ICES and by national organizations such as the
COGEPOMI in France. It is accepted by fishers and their representatives. Yet the lack of efficiency in
data collection systems often seems to be at odds with the importance given to fisheries data for the
management of stocks and of the fisheries themselves (FAO 2002). This is a particular problem in the
case of inland fisheries (FAO 1996b; FAO 2003). In the case of eels in France, this became apparent as
early as the beginning of the 1980s within the framework of the National Think Tank on Eels
(Castelnaud and Gascuel 1984). Catch statistics depend fundamentally on data quality, i.e. their
accuracy and exhaustivity (FAO 1996b; FAO 2002). A critical analysis of available official data (Muchiut,
2001; Castelnaud and Cauvin, 2002; Castelnaud et al, 2006) shows biases, omissions, and
1 Castelnaud G., 2008. Les descripteurs et indicateurs de socio-economie des peches, Cemagref, annex 9 of the Indicang report,
187

discrepancies which affect published results (when they are published) due to non-reporting (or flawed
reporting), under-reporting, or fanciful reporting and declarations.
Fishing data belong to fishers who are central to the data collection process. The rather
administrative conception of catch statistics, the available resources, the logistics and the few
observable results are both confusing and dissuasive for fishers. “Catch statistics” are often perceived
by fishers to be a constraint with no credible purpose and an indirect means of checking their turnover,
leading to rejection or under-reporting. The trick is to obtain the trust and cooperation (FAO 1999) of
fishers, with their particular sociological characteristics, in order to limit so far as possible the excuses
to do nothing or do it badly, which they have found and continue to find within the statistical monitoring
systems themselves.
6.3. Fisheries biology theory
The theory has developed from the general observation that the fishery could be considered as a
measurement tool which intercepts, extracts (without restitution), a part of the flux, thereby giving a
signal that provides an abundance index under some conditions as well as information on the species’
relative abundance (as opposed to absolute abundance which quantifies the total flux). The validity
conditions for this abundance index, the CPUE (catch-per-unit-effort), are discussed below. They apply
to scientific survey fishing as well as to commercial and recreational fishing. Indeed, the latter two:
• may generate significant catches of fish and crustaceans with efficient fishing techniques (often
copied in scientific fishing) but they are limited to, and defined, by target species and life stages, by
“profitable” fishing periods and by restrictive regulations;
• are generally sustained (commercial fishing) in space and time, but their coverage tends to be less
than experimental fishing (periods), although sometimes more (space).
6.3.1. Catch-per-unit-effort (CPUE), abundance and fishing mortality
coefficient
Consider a fishing zone A of uniform density where all fish have the same vulnerability*; N is the
abundance in the zone and N/A is the density. Consider a fishing gear extracting a constant number N
of animals; the catches ∆C, resulting from a set of fishing activities during a very short time interval, are
proportional to ∆f the fishing effort 2 and to fish density. If q’ denotes the constant of proportionality, this
N
leads, following GULLAND (1969), to the basic equation: Δ C = q' Δf
(1)
A
http://www.ifremer.fr/indicang
2 See § >
188

From this equation, it follows that the CPUE (catch per unit of (fishing) effort) is proportional to
ΔC
N
density and abundance: CPUE = = q' = qN with q=q’/A.
Δf
A
The instantaneous fishing mortality coefficient F(t) is defined by the relationship:
F( t)
=
1 dC(
t)
N(
t)
dt
That is, over the brief time interval ∆t :
Δ C = FNΔt
(2)
From (1) and (2) we obtain:
N
Δ C = q' Δf
= qΔfN
= FNΔt
hence qΔ f = FΔt
(3)
A
During a time period t, (3) becomes:
t
∑
0
FΔt
=
t
∑
0
qΔf
that is
Ft = qf , with f = ∑Δf
t
0
, effort
applied during the period t.
F is the fishing mortality coefficient applied during a unit of time (Laurec and Le Guen 1981)
which leads to: F=qf (4)
Generalising, CPUE becomes: CPUE = C/f = qN and C = FN.
6.3.2. Catchability and its components
Catchability or q is the probability that a fish will be caught by a unit of fishing effort 3 . It is the
result of several components (figure 6.2) that Laurec and Le Guen (1981) organise into two groups:
availability and efficiency. Availability itself is subdivided into accessibility and vulnerability.
Vulnerability can be separated into factors that depend on the fish and factors that depend on the
interaction between fishing gear and fish. Efficiency encompasses all factors that depend on the fisher
in the fishing action, rather than the fishing gear. The efficacy of the fishing gear itself is integrated into
vulnerability.
3 See § >
189

Fisher
Efficiency
Capability
Fishing gear
(efficacy)
Environment
(influence)
Availablility
Fish
(behaviour)
Accessibility
Figure 6.2 - Proposed diagram of the concept of catchability and its components (after Laurec
and Le Guen, 1981).
6.3.2.1. Availability
For a fishing unit to catch a fish, it must be accessible and vulnerable.
Accessibility means that the fish is physically present in the fishing zone and this can therefore
vary with migratory patterns (spawning or trophic in particular).
Vulnerability depends on the behaviour of animals and on their interactions with the fishing gear
but it is not related to fisher behaviour (see efficiency).
Fish vulnerability depends on various factors:
• those linked to animal behaviour, life rhythms, sex (reproduction, adaptation to a new environment,
searching for food) and those, such as bathymetry or hydrodynamics, that have an impact on its
behaviour. Various examples are given in previous chapters for glass eels / elvers or silver eels
(impact of hydrodynamic conditions on migratory behaviour for instance);
190

• those related to fishing gear characteristics and its capacity to explore the zones where fish is
located, with the notions of avoidance, escapement and selectivity. The most characteristic example
is the use of glass eel / elver scoop nets fitted with long handles to allow fishing at greater depths
during clear water periods. Another example is the mesh size in a basket trap or the hook size of
paternosters which target a well-defined size category 4 .
6.3.2.2. Efficiency
Efficiency covers anything that depends on the fisher when using fishing equipment i.e. tactics,
experience, skill, intuition. The overall performance of fishing gear is therefore a combination of this
extrinsic efficiency and its intrinsic efficiency, which is related to vulnerability.
6.3.3. Fishing effort and fishing power
According to Laurec & Le Guen (1981), the fishing effort applied to a stock of aquatic animals is
a measure of all fishing gear deployed by fishers on this stock, during a given time interval. This
definition means that the number of fishing boats and their characteristics, the fishing gear, the level of
activity and the human capacities, etc. must be taken into account. This is summed over a selected
period of time. This primary measurement is the nominal fishing effort.
Whilst effort is a cumulative measurement in space and time, fishing intensity is local and
instantaneous; it is total if it concerns the whole fishing area. Beverton and Holt (1957), Ricker (1980)
specify that fishing intensity is defined by unit of surface area. The passage from local intensity to
effort is obtained by integrating over space and time.
6.3.3.1. Nominal effort and effective effort
A fishing métier in its broad definition is a fishing gear or a fishing practice (targeting one or
several species). In its more precise usage, which suits fisheries for migratory species, it associates a
fishing technique and a target species or ecophase. Examples: (species) allis shad (Alosa alosa) – net;
(ecophase) glass eel/elver – scoop net.
The most basic fishing effort (primary), usually called “nominal” in France, is the number of
vessels or, in the case of inland fisheries for migratory species, the number of fishers (meaning the
number of active fishers undertaking, whether by boat or on foot, a particular fishing métier during the
authorized fishing season). The nominal effort unit is then: one fisher (commercial) who undertakes (in
4 Some examples are given in the chapter on fishing gear selectivity for yellow eel fishing and particularly on their
mesh size. The age or size at which the fish becomes vulnerable to the fishing gear is the age at first capture and of
recruitment.
191

a sustained manner) a fishing métier, by boat or on foot, using the standard gear, during the entire
fishing season which is authorised for the given species or ecophase.
Fishing effort is generally said to be effective when it is a more refined measurement
(secondary) than the nominal effort and is closer to the fishing action, such as the number of fishing
days (meaning the number of fishing days of active fishers, undertaking a particular fishing métier,
fishing with a boat or on foot, during the authorized fishing season). The effective effort unit is then:
one fishing day of a fisher who undertakes a fishing métier, by boat or on foot, using the standard gear,
during the entire fishing season which is authorised for the given species or ecophase. Where possible,
it is the effective effort that should be used to estimate CPUE.
6.3.3.2. Relationships between catchability and fishing
effort
Catchability is a global concept which makes it possible to link the notions of fishing effort and
fishing mortality using equation (4): F = qf.
The effective fishing effort is an intermediate concept between nominal effort and the coefficient
of fishing mortality, and should be as close as possible to the latter (Laurec and Le Guen 1981).
If f is defined in terms of nominal effort, in order to maintain equation (4), q must vary with time.
Effective effort is used in the hope that catchability can be held constant so that CPUE can be
considered to be proportional to abundance.
When using nominal effort, any variation in efficiency generates a variation in catchability. Hence,
Laurec and Le Guen (1981) write: F = q * f = q * T * f where fn is nominal effort, fe is effective
effort and T is efficiency.
n
n
Assuming that efficiency is “extracted” from qn and included in fe, then qe is the catchability and
fe the effective effort, giving: F = q e
* f
e
Hence, the effective effort fe makes it possible to remove the efficiency component from
catchability as it is included in the effective effort (qn when efficiency is removed becomes qe,
catchability limited solely to its availability component).
e
n
6.3.3.3. Units of effort and study scale
The nominal effort in number of vessels or fishers in activity during a season is of interest at an
inter-seasonal scale (the scale of abundance monitoring using the fishery tool for comparisons). As
soon as it becomes more precise, at the level of fishing methods and/or more commonly at the level of
the reference period, it becomes effective effort, for example the number of fishing days using a scoop
net or a basket trap by fishing month, etc.
192

If a study is undertaken on an intra-seasonal scale, the primary level of fishing effort
measurement cannot be the previous inter-seasonal nominal effort because this latter is, due to its
reference period, beyond the temporal scale and inoperative. To be operational, its reference period
must allow comparisons and calculations and therefore represent an appropriate subdivision of the
season: generally a fishing day, but possibly a month, a fortnight, a week, the tidal fortnight, etc. If the
month is selected for example, the nominal effort is then the number of fishers (commercial) who
undertake (in a sustained way) a fishing métier, by boat or on foot, using the standard gear for one
month. This effort can also be used inter-annually. As in the inter-seasonal case, the fishing day will be
the effective effort unit.
It should be noted that if total effective effort is available, then the less refined measurements that
are usually necessary to calculate it must also be available (e.g. if it is possible to calculate the total
number of fishing days, then the number of fishers must be known).
6.3.3.4. Fishing power
According to Beverton and Holt (1957), Ricker (1980) and Laurec and Le Guen (1981), fishing
power p is the catching power of a vessel, measured per unit of fishing time compared to a standard
vessel (taken as a reference) using standard fishing gear and fishing in the same area.
For Laurec and Le Guen (1981), the power of a vessel is given by the ratio of its catch C to that of
the standard vessel Co fishing in the same context; conventionally, the standard vessel has a fishing
power (p) equal to 1. The notion of fishing power makes it possible to:
weight, in a fleet and for a given fishing métier, and within a season, the effort of each fishing unit
compared to the “standard” one.
take into account, between seasons, the global trend (usually the increase) in efficacy for a given
métier in the whole fishing system (fleet) .
Remark : fishing power (p) weights the effort (f) but not catchability (q) in the CPUE formula
C = q * N
f
Example: In the Gironde estuary when the “pibalour” was first authorized between 1975 and 1979,
dories were fitted with 5 m 2 “pibalours” . Most fishers progressively increased their net area and in the
1980s there were three types of boats: those which had stayed at 5 m², those which had doubled to 8-
10 m², and those which had increased to the maximum authorised area of 14 m². From the 1990s,
virtually only the last two groups could be found. During the season, in the 1980s the catches of the
three types of boats could have been compared using the 5 m 2 pibalour as the reference (standard
boat). Hence, within and between seasons, the fishing power differences could have been estimated
and the total effort could have been expressed in standard units. This was not done at the time because
193

this methodological approach had not been integrated due to scientists’ inexperience in fisheries biology
and their focus on setting up and consolidating the Cemagref network for the collection of representative
and reliable data, which was considered the main priority.
According to Gulland (1969), the fishing power of a gear, i.e. its catch per unit of fishing time, is
the proportion p of animals caught by the fishing gear over the swept area a. If A is the total fishing area
and N the existing stock, catch is equal to: C p a a
= *
A
* N , as C = F * N , the product p *
A
gives a
direct measurement of the fishing mortality coefficient.
Gascuel (1995) gives his own interpretation: fishing power is the ratio between effective effort and
nominal effort and catchability is the product of availability (d) (which depends on fish) and fishing power
(p) (which depends on the fisher and the fishing technique):
fe
p =
f
and d p
n
q = * .
In this case, the fishing power corresponds to the efficiency as presented by Laurec and Le Guen
(1981). This subdivision by Gascuel (1995) is simpler than the previous authors’ but restrictive as it
excludes the interaction of the fish with the fishing technique which is included in the vulnerability
component of availability. It leads to a major difficulty, which is how to integrate an estimation of fishing
power which is a catch rate (quantifying efficacy differences) into effective effort where effort units are
standard by definition.
6.3.4. Practical use of the concepts for abundance monitoring with CPUE
The use of CPUE (catch per unit of (fishing) effort) provided by the “fisheries” tool to monitor the
abundance of an exploited species or ecophase of this species, is based on the proportional
relationship between catch and stock. This assumes that the relationship F= qf can be used, because
either catchability is constant or its variability is a function of easily measurable external parameters 5 .
Most of the time, catchability is related to several factors, some of which are completely
uncontrollable. This is the case for accessibility (fish is present or absent from the fishing zone) but the
fisher chooses his fishing zone according to the presence of fish and whether it is possible practically to
catch it (with his fishing gear). It is also the case for fish behaviour which is related to its physiology and
affected by bathymetric and hydroclimatic conditions; and the performance of the fishing gear (efficacy)
faced with these same conditions, which can vary during the season but are broadly similar between
seasons.
These sources of unavoidable variations in catchability are a kind of constant in the fisheries
system. On the other hand, other factors affecting fishing gear efficacy and efficiency are of human
origin and can, in principle, be controlled or at least limited, and avoided or at least evaluated. This is
attempted using fishing effort, which is related to catchability and onto which is transferred the element
5 See Chapter 7.
194

of variability due to fishing gear efficacy and efficiency. Nominal fishing effort is a kind of "black box"
which has to be opened and examined in order to define the most precise and accurate effective
effort possible, given the level of detail and precision of the available database.
Incidentally, Laurec and Le Guen (1981) point out that effective fishing effort is often a rather
abstract notion which requires refined data and that it is difficult to find a definition of effective effort and
a unit that resolve all problems. This is why in practice, attempts are made to “correct nominal effort”, to
refine the effective effort unit and “to limit catchability variations as far as possible”.
6.3.5. Summary
At an inter-annual level, quite long series have to be collected, a minimum of 5 to 10 years,
making sure that CPUE is evaluated consistently: reliability being more important than accuracy.
* A certain number of simplifying hypotheses are used that are hopefully not too far from reality:
the behaviour and accessibility of animals broadly the same from one year to the next; independence of
fishers’ catches in the same zone or from one zone to the next during the season; minor variations in
the number of fishers per zone from one year to the next with no impact on individual catches.
* Insofar as possible, catchability variations that condition the proportionality between CPUE and
absolute abundance are limited by using appropriate stratifications (differentiating fishing métiers, and
dividing the sector under study into homogeneous fishing zones); by taking into account external factors
affecting fish or fisher behaviour; by analyzing the structure of fishing effort and using the most refined
effort units possible; by estimating CPUE using GLM-type modeling of impact 6 .
* All changes are identified in fishing power, in efficiency (fishing tactics), in fishers' strategies and
techniques (often starting with the underlying causes: resource dispersion, sales price and market for
example).
* Abnormal hydroclimatic and other (economic, regulatory) events in a given season are identified
in order to improve the understanding and interpretation of inter-annual variations in the CPUE curve;
such events lead to errors in CPUE estimations but, in principle, they should not hide the general trend
of the curve over a relatively long period of time.
* At any time, especially when the CPUE curve is stable and to the extent that collected data
allow it (improvement in their precision), effort can be corrected by the fishing power if it has changed
between two periods or/and a second and more refined effective effort unit for future CPUE
comparisons can be used. In this latter case, past years are re-calculated when possible or the CPUE
continues to be calculated in parallel using the primary effort unit in order to maintain coherence and the
possibility of diagnoses using the complete available series.
6 See § >
195

6.4. Method used to monitor fisheries and obtain basic data.
6.4.1. Description of the approach
The aim of, and the reasons behind, fisheries statistics and hence fisheries monitoring have been
explained. The mathematical theory and the concepts of fisheries biology showed more particularly why
CPUE was calculated from the fisheries tool, what was the point of the concepts and how they are used
to undertake this calculation.
Now, we will show how to obtain fisheries descriptors (of fisheries biology), i.e. how to obtain
catch statistics through fisheries monitoring.
Reminder: catch statistics, in a given fishing system, are established over time using people, methods
and tools which together comprise the (statistical) fisheries or catch monitoring system. The various
stages of implementation and operation of this system are explained in detail below.
The results of fisheries statistics, and in particular the reliability of fisheries descriptors, depend
first and foremost on basic fisheries data collected from the fishers and on the knowledge of the
fishery’s characteristics and functioning (Neis et al., 1999; Hilborn, 1985; Brethes, 1990). It is therefore
necessary to know what to collect (which data), from whom (which fishers), where (fisheries sector)
and in what context (fishing system with its history). Once the data have been collected, they must be
entered, stored, checked, corrected, completed and validated before being sorted, extracted, combined
and analysed with mathematical tools to give the descriptors and then indicators that will be analysed
and interpreted.
6.4.2. Definition of the study project and preliminary analysis of the
statistical fisheries monitoring context
Generally speaking, the ultimate objective, i.e. statistical monitoring with a biological and/or socioeconomic
focus, should be defined a priori as it partly conditions the level of accuracy in the required
data and therefore the methods and tools used, which can be simplified in the second case.
Having said this, the priority objective of Indicang was to collect and analyse fisheries statistics for
biological purposes, although it is important to bear in mind that a socio-economic statistical monitoring
system is only a variation (or a continuation) of a biological statistical monitoring system.
The theoretical objectives and the feasibility of a project depend on the initial situation, the type of
fishery, and on existing statistical monitoring systems (or their absence).
The scientific approach as regards biological and socio-economic fisheries statistics, i.e. the
theoretical bases and the methods of collection, processing and analysis of the fisheries descriptors, is
common to all species and concerns both mono- and multi-species fisheries. Depending on the
196

species or the ecophases monitored, differences may be found in the units used (effort units), in the
values of specific descriptors and in the analyses and interpretations. In the case of multi-species
fisheries, statistical monitoring usually includes all exploited species or at least the principal ones, for
obvious reasons relating to fisheries management, to logistics, to the efficiency of field surveys and
hence to human and material costs. From the scientific and methodological viewpoints, it is helpful and
necessary to be able to place the species or the ecophase into the context of fishing for the other
species, to understand the links in fishing practices (calendar, switching of fishing effort, etc.) in order to
optimise the descriptors and explain their values and their trends. All this must be taken into
consideration when perfecting and implementing a (system of) statistical catch monitoring.
Any legal (professional or amateur), or illegal fishery must be assessed a minima using basic
fisheries descriptors in order to establish the number of fishers (nominal fishing effort) and production
(fishing mortality). On the other hand, when monitoring the relative abundance of a species, it is
recommended to verify that the fishery is capable of providing the necessary information, in particular
given the fishing technique used and the spatial (fishing zones) and temporal (fishing season) coverage,
depending on whether the stock is sedentary or mobile. The further away one moves from the
necessary conditions of representativeness and reliability, the more difficult and risky it becomes to trust
the results and therefore the abundance index.
However, unless particularly restricted by the hydrodynamic conditions of the fishery sector or by
regulations (and this must be verified), a fisher will always be in the best place to catch the target
species, at the best time and using an appropriate fishing technique. Consequently, the structural
conditions for the observed phenomenon to be representative are often fulfilled, especially if the fisher is
assiduous (which is a quality of the professional).
If there are several fisher categories, a choice can be made based on the apparent aptitude of the
different types of fishing that they undertake and on the restrictive and decisive factor: fishers'
cooperation. If there is only one category of fishers, there is no choice but to use it.
Whether beginning from a “blank page” with no fisheries statistical monitoring, or from an existing
system which appears to be insufficient or to offer room for improvement, we must establish a baseline
for the fishery. When fisheries statistical monitoring exists, its functioning and results must be analysed
to identify weaknesses and gaps in order to consolidate and improve it.
6.4.3. The characteristics and functioning of a fishery
Often, knowledge of the baseline of the fishery, its structure and operation, are progressively
developed from the results of the statistical monitoring system as well as from samples and surveys
conducted either as part of the monitoring or independently.
197

The more accurate the knowledge of the “fishing system”, the easier it is to make precise
strategic choices (fisher categories, stratifications, fisheries statistical monitoring based on fisher
populations or samples…) and to implement the monitoring logistics. This also simplifies the
identification and collation of peripheral administrative information, concerning fisher groups and other
stakeholders (economic…). It also makes it possible and easier to undertake targeted surveys.
Information from official bodies usually concerns:
* the fishing zone and its administrative and regulatory limits;
* the fishing regulations by category of fisher, by fishing zone, by species and ecophase, by
fishing gear and by period (making it possible to establish a fishing calendar);
* the numbers of fishers by category and by status;
* the vessels and their characteristics (type, length, tonnage, engine power, etc.);
* sometimes the catch volumes with variable reliability.
This information is rarely comprehensive, precise and clear as the data-recording approach of
these bodies is based on sectoral, statutory and regulatory requirements which do not cover the fishing
activity as a whole. Consolidation of fishing zone information, especially in the lower reaches of rivers,
by licence type, status and professional and amateur fisher categories has yet to be done.
Field surveys, by direct contact, by post or by telephone are essential or at least very useful in
many cases. They must be supported by a simple and concise structured questionnaire which clarifies
the fishing activity of the fisher and the factors that affect it: status, licences, multi-activity, help with the
work, fishing gear, fishing métier, period, fishing area and intensity, marketing method.
This information makes it possible to establish:
* the different fishing métiers, fishing zones and seasons and thereby the fishing calendar(s)
by zone and by fisher group;
* the number of fishers by métier and by zone, and the related number of vessels;
* possible stratifications by homogeneous fishing zone, and time period;
* the extent of multiple activities as well as the selling strategies which determine the effort
used;
* the existence of differences between fishing effort and fishing power by métier within the
season and between seasons (if historical data are available).
It must be noted that information on fishing practice will be available though the results from the
statistical monitoring of catches only if the latter is exhaustive (and reliable). Often this information is
198

lacking and must be obtained otherwise, as shown here, in order to correct compulsory declaration
systems. In voluntary declaration systems, this operation is essential if extrapolations are to be made.
6.5. Basic data, collection tools and methods
As well as requests to official bodies and the general surveys mentioned in the previous
paragraph, two other methods of collecting data on fishing practice are possible:
* instantaneous data collection, often on an occasional basis (i) concerning fishing practice and
fishing effort in the fishing zone, by visually counting the boats or by GPS beacons and (ii) concerning
the vessel, by on-board recording of catch and fishing effort parameters;
* deferred data collection on the fishing activity, catch and fishing effort parameters, by deferred
contact with the fisher.
The basic central datum of fisheries statistics is the catch or harvest, in weight or in numbers if
the relevant species, or ecophase, makes it possible for individuals to be counted and if the fisher or the
enumerator is required to do so. A catch depends on a fishing action which is determined by the fishing
effort and involves:
* a fisher on foot or a fisher or crew on a boat;
* some fishing equipment, the characteristics of which can be specified (length, surface,
number…);
* the filtered volume and the swept area;
* a fishing date for which fishing time may be available (trip, effective fishing time);
* a site or fishing zone, recorded and codified.
These data must therefore be organized logically starting from the fisher or the boat and cover the
whole fishing season of the relevant species or ecophase. This means that a recording system must be
established. This refers to the deferred data collection mode which is the most frequently used to obtain
data continuously over the fishing season, the instantaneous mode being generally used (if at all) to
complement or verify data collected in the former mode. The fisher can be approached in two ways,
which determine two data collection options:
* fishing logbook, given to the fisher, to be completed and recovered periodically in various
ways, such as by post or through collection points at fishers’ representatives, etc.;
* occasional direct meetings with fishers, on a voluntary basis, at home, in a café, on the dock,
etc.
This leads to two types of catch monitoring statistical systems:
* a compulsory declaration system (usually administrative) using logbooks or returns which, in
theory, concerns the entire population of fishers by category, in the fishing area;
199

* a voluntary declaration system (usually scientific – in France, CEMAGREF for example) by
direct and personalized recording of a sample of fishers who are prepared to co-operate. The system
should be as representative as possible.
The compulsory declaration system aims to be exhaustive, but this is rarely the case. The
voluntary declaration system targets a sample comprised a priori of most of the fishers ready to
cooperate (around 25 to 30% of the population for the Cemagref system in Gironde). It can be
exhaustive when, in a fishing sector, the total population of fishers is low.
These two types of monitoring system can be combined into a mixed system based on
compulsory declarations by the entire fisher population, which are validated by the voluntary
declarations from a sample of fishers.
These fisheries statistical monitoring systems rely on:
* raising awareness and gaining the trust of the fisher community, through information on setting
up the statistical monitoring system, its objectives, the constraints and the advantages for fishers,
officials and scientists.
* a recording medium. The fishing catch return sheet must meet a certain number of standards
and conditions to be operational. In order for the enumerator to complete the fishing return, it must be
designed in such a way that data given orally by the fisher or read from any of his records can be noted
in a practical and rational way with, for example, lines where days are pre-filled and columns with fishing
métiers by month. Figure 6.4 gives an example of such a catch return sheet.
200

Fisher XXXXXXXXXX Origin
:
Season 2004- Fishing
Zones : Z2 (MAI) et Z3 (CIV,
EPERS Quality : 1 Survey date
.. / .. / 2005
Fishing
type:
‘pibalour’, net
DE JA FE MA APRI MA JUN AU SEP OCT
CIVP (kg) CIV CIV CIV ALAF (nb) ALA ALA MAIF (kg) MAI MAI
P
1 0,35 2,18 0,5+0,3 1 10 15 0,9 1
2 0,48 1,22+1,48 0,58 55 2 10 2
3 0,42 0,4 0,3 3 7 17,2 3,2 3
4 0,47 1,32 4 4 1,6 4
5 0,5 1,26 0,38 4 5 0,4 5
6 0,25+0,25 0,4 5 6 5,8 1,3 fin de saison 6
7 0,32 1 7 3,5 3,6 7
8 0,26 0,6 0,48 8 3 8
9 0,25 0,97 0,25+0,23 8 9 8,2 9
10 0,18 1+1,05 11 10 7 10
11 0,87+0,64 0,44 6 11 10,3 11
12 0,36 0,82 0,66 3 12 17,4 17,5 12
13 0,12+0,20 0,74 3 13 2 3,6 13
14 0,22+0,10 1,16 0,7 14 14
15 0,3 1,06 0,32+0,6 15 15
16 2,1 1,34 0,30+0,26 0,4+0,3 24 16 16
17 1,18+0,3 0,18+0,2 0,2+0,4 20 17 17
18 0,6 0,4 8 18 18
19 1,48 11 19 19
20 0,75 1,32 1,42 7 23 20 20
21 2,0+0,72 0,64 0,44 14 10 21 1,3 21
22 0,44 1,64 0,24+0,4 0,11+0,13 22 3,6 22
23 0,58+0,7 2 0,6 0,23+0,27 1 23 3,5 23
24 1,52 1,56 0,6 0,3 24 24
25 1,3 0,86 0,4 0,44 25 1,8 25
26 2+0,12 3 0 26 4,7 26
27 1,02 FROID 3 27 12,3 27
28 0,38 35 5 28 28
29 0,92+1,34 29 5 29 19,6 29
30 0,64+0,76 0,3+0,5 25 4 30 1,5 30
31 2,24 0,96 0,92 3 31 31
TOT 23,22 31,85 8,58 7,64 113 217 TOT 29 107,4 57,9 2,5 TOT
Saisie
.. / .. /
Figure 6.3 - Multi-species record sheet from the Cemagref statistical monitoring system in
Gironde with a fictitious example of codified field data entry (Z= fishing zone; MAI=
meager; CIPV = elver – “pibalour”; 0.25 + 0.25 = 0.25kg of elver 1st tide + 0.25kg of
elver 2nd tide) (source : Cemagref).
The fisher's return sheet must be simple, clear and precise. In order not to be discouraging, it
should not ask for too many details, unless the fisher, having been made aware of their
importance, is prepared to provide them. The fisher must understand easily the nature of the
requested data and find quickly where to record it on the sheet with the minimum of writing and
deliberation. The return must relate to one day, a fortnight or a month and include, if possible, the
maximum fixed pre-filled information such as for example the days (1 to 31) in line with two trips a
day (1, 2) in the case of a monthly return and the fishing métiers which succeed one another
during the year in the columns. The logbook can be mono or multi-species (this is often the case
to optimise monitoring and the fisher’s collaboration) and in the latter case, it can focus on the
main fishing métiers as it is impossible to have everything: there is always a trade-off between
quantity and quality. Figure 6.4 gives an example of a page from a multi-species logbook.
201

Fishing
zone
Elver
scoop
net
Elver
'drossgae
' small
push net
Lamprey net
Example
Second period of the month of:..............March. ...................................Year:..............2008..................
Tide
Lamprey 'bourgnes'
basket traps
Shad net
Eel
basket
traps
Prawn
net
Other
(specify)
Kg Kg Kg Kg Kg Kg Kg Kg Kg Kg Kg
18
19
20
21
Tide 1
Tide 2
Tide 1
1D1 Tide 2
Tide 1
Tide 1
1D1 Tide 2
1
2
Tide 2
7
0.5 0.5
D1 Tide 1 8
22
1 Tide 2 0 0
23
24
25
26
Tide 1
Tide 2
No fishing - river in spate
Tide 1
Tide 2
Tide 1
Tide 2
Tide 1
Tide 2
1D1 Tide 1 1.5 7
27
D1 Tide 2 10
D1 Tide 1 5
28
Tide 2
29
30
31
Tide 1
Tide 2
Tide 1
Tide 2
Tide 1
Tide 2
Details concerning the fishing gear used during the month:
Lamprey 'bourgones'
Eel basket traps
Prawn basket traps
Average number
80
Lamprey net
Shad nets
Other net
Mesh (in mm)
36
60
Length (in mm)
140
150
Various observations: (e.g. Spate, water release, pollution, temperature, mortaility and/or fish diseases):
Flood for 2 days - basket trap disease - sick shads - large dead eels - catch of 2
sturgeons - too much wind - too cold
Figure 6.4 - Page from the multi-species logbook for Gironde commercial river fishers with a
fictitious example of codified data entry by a fisher (I, D1= fishing zones) (source :
Association agreee departementale des pecheurs professionnels en eau douce de
la Gironde).
The expected result from completion of this record card is to identify the catch in numbers or in
weight with the date of fishing (day or trip), the fishing zone, the fishing métier and the fishing
gear characteristics, for example the number (basket traps) or the dimension (net).
202

* Logistical support:
Annual census of fishers with periodic updating of knowledge of the fishery, highlighting changes
in fishing techniques (new métiers), fishing zones, nominal fishing effort (number of fishers),
fishing power (length and area of nets, new material, number of nets and gears, type of vessel,
engine power, etc.);
Fishers are regularly given a logbook and periodically return it or data are periodically recorded
there and then;
In the case of a voluntary monitoring system, regular search for new cooperative fishers in order
to improve the representativeness of the sample by fishing zone and to counteract interruptions
due to retirement or untimely refusal to cooperate further;
Entry of collected data into a database, storage, checking with fishers in the case of anomalies
(quite frequent; various omissions, erroneous weight or date or zone or fishing gear, etc.)
crosschecking of data and information on the fishery in order to detect mistakes and false
declarations, corrections.
* Data processing, sorting and stratification, estimation of descriptors, checking of the results at
each processing stage and rechecking already processed data for final validation.
* Analysis of descriptors and apparent trends, diagnosis.
* Presentation of the results, as a file, report, dissemination of information to the public and
back to the fishers, possible discussion of the results.
* Human resources and skills as appropriate (ideally a field enumerator and personalised
contacts).
These statistical monitoring systems, although different in their design and their results, are based
on common methods and tools and have common strengths and weaknesses. It is hence necessary to:
* find the total number of fishers by métier (and by category), to define standard nominal and
effective effort units;
* go into the field and collect basic data regularly and over a long period of time;
* take into account fluctuations in the number of fishers between seasons (this being a
particular issue with the voluntary system);
* take into account the under-sampling of the fishery (cooperative fishers in the voluntary
system; return rate

Depending on the context of the fishery and its objectives, the methods and tools can be
adjusted and adapted from this presentation. In particular, depending on whether the fishing métier
concerns a sedentary stock (ex: yellow eels) or a migratory stock (ex: migratory glass eel flux, silver
eels migrating downstream), the fishery can be monitored using different approaches in the selection of
fishers (category, population or more or less reduced sample size), the spatial scale (fishing zone of
varying size, stratification), the temporal scale (one or several periods during the fishing season) and
when taking into account the fishing gear and the way it is used (basket traps, fyke-nets, hand scoop
nets, push nets, barrier-nets, etc.)
The fisher is at the centre of the system and the success and sustainability of the statistical
catch monitoring system depend on him. The relationship must be based on trust and data
confidentiality must be preserved at all levels, otherwise turnover (income) could be indirectly revealed.
For example, it is a mistake (difficult to correct) to allow an official from a fisher association to collect the
logbooks as the latter will be considered by the fishers to be first and foremost another fisher and not
someone to whom they wish to show their logbooks. No personal data must appear in the results.
6.6. Entering, storing and validating basic data
Once collected from the fishers, the data have to be entered, stored and organized into a
database. The structure of, and the number of tables and fields in, this database depends on the level of
detail of the data and of the fishing trips, as well as the combinations and calculations that are planned
in order to obtain the descriptors.
It is important first to stress the need for the data to be as centralised as possible so to ensure
controlled data entry and secure storage. This is not always possible with local databases which often
cannot be connected to one another as they are incompatible.
As an example, consider the Cemagref database (called GIRPECH), which was developed under
Access and comprises two groups of inter-connected tables (figure 6.5):one group of tables (fisher,
residential history, administrative management, status, assiduity) where annual entries are made of the
geographical (place of residence and changes) and administrative (status, licence) situations and also
of the activity of cooperative fishers (assiduity by season, species, métiers and fishing zones);
* another group of tables (catch, species, métier) which stores all the basic catch and fishing
effort data (from cooperative fishers) by species, fishing métier, level of data accuracy and quality.
The “fisher” table plays a central role which shows in the input masks.
The “pechmet” table serves as an intermediary between the "fisher” (pêcheur) and “métier” tables
and the "catch" and "assid" tables which use them and allows requests to be made including both
“catch” and “assid” tables at the same time.
204

Data can be entered into the base in three groups using input masks: a “geographic and
administrative situation” group, a “fishing calendar and assiduity” group and a “catch and effort” group.
Figure 6.5 -
Relational diagram of the Girpech database.
Basic data can be verified at several stages:
* at the time of entry, the first potential anomalies (absence of data, unmeaningful or unreliable
data, etc.) are identified on the field sheets and in the various listings held by the administrative or
associative bodies;
* after entry, catch or individual effort totals and related information, fishing zones, dates and
details are obtained by specific requests so that basic data anomalies and entry errors can be identified;
* finally, initial processing and the calculations of mean catch, mean effort, total catch and total
effort may reveal incoherent data relating to some fishers.
Once identified, an attempt is made to correct these errors. It is clear that basic data veracity
can best be appreciated through voluntary declarations at the time of collection from the fisher and that,
in the case of compulsory declarations, data can only be reviewed and corrected by returning to the
fisher (which can, of course, also be done in the case of voluntary declarations).
Data validation is the end result of successive verifications and corrections together with any
complementary historical data that may have been recovered.
205

6.7. Estimation of fisheries biological descriptors
6.7.1. Métiers and fishing effort units
The basic data lead to descriptors and require the definition of effort units by fishing métier.
The reference used here as an example is the statistical monitoring based on the Cemagref voluntary
declarations for eels on the Gironde.
Reminder: a fishing métier is the combination of a fishing technique and a target species or
ecophase. Examples on the Gironde:
* eel at glass-eel/elver stage - hand scoop net = CIVT;
* eel at glass-eel/elver stage – “drossage” (two circular push nets) = CIVD;
* eel at glass-eel/elver stage – “pibalour” (one or two pushed leader nets, 5 to 14 m 2 area) =
CIVP;
* eel at sub-adult stage – basket traps = ANGN.
Reminder: The nominal effort unit is a (commercial) fisher who undertakes (in a sustained
manner) a fishing métier, by boat or on foot, using the standard gear, during (the entire) fishing season
which is authorised for the given species or ecophase. Examples on the Gironde:
* the nominal fishing effort unit for the glass-eel/elver – “pibalour” métier is an assiduous
commercial fisher using his boat to push a “pibalour” of 5 to 14m 2 from the 15 th of November to the 31 st
of March;
* the nominal fishing effort unit for the eel-basket trap métier is an assiduous commercial fisher
using 60 to 150 basket traps for at least 3 months between the 1st of March and the 31st of October.
Reminder: The effective effort unit is one fishing day of a (commercial) fisher who undertakes
(in a sustained manner) a fishing métier, by boat or on foot, using the standard gear, during (the entire)
fishing season which is authorised for the given species or ecophase. When possible, CPUE are
estimated using the effective effort. Examples on the Gironde:
* the effective effort unit used to calculate CPUE is the “pibalour” fishing day (1 boat rigged with a
“pibalour” for one fishing day). The mean effective effort of cooperative fishers is therefore expressed in
terms of “pibalour” x fishing days (1 boat rigged with a pibalour x the number of fishing days);
* the effective effort unit used to calculate CPUE is the basket trap by fishing month (1 basket trap
used for 1 month). Hence, the mean effective effort of cooperative fishers is expressed in terms of
basket traps x month (mean number of basket traps used by month x the average number of fishing
months).
206

6.7.2. Fisheries biological descriptors
The fisheries biological descriptors required are:
* total catch (C): total weight in kg for each life stage of the eel;
* total fishing effort (f): season, number of fishers, number of fishing days or trips, number of
fishing gears, total surface area exploited; total filtrated volume (units: 1 fisher-season, 1 fishing day or
trip, 1 fishing gear or equipment, m 2 of push net or “pibalour” surface area, m 3 of filtrated water);
* catch-per-unit-effort (CPUE): weight or number / season (= total catch), weight or number /
fisher group, weight or number / fisher, weight or number / fishing day or trip, weight or number / gear,
weight or number / m 2 of push net or “pibalour” surface area, weight or number / m 3 of filtrated water.
6.7.3. Descriptor estimation methods
In general, the level of detail in the collected and stored data determines the possibilities for data
stratification, aggregation, calculation and modelling. The optimal option is discussed below so as to
provide a reference point, which can be adapted if necessary.
6.7.3.1. Compulsory declaration system
If the declarations of fishers in the monitored population can be considered to be exhaustive, total
catch and total effort by fishing métier are obtained by simply adding the recorded catch per fisher, in
terms of the chosen and feasible stratifications / aggregations (by season, month or any other period of
time, by fishing zone).
If the declarations are not exhaustive, some correction has to be attempted. This should be
possible using the declaration return rate (received logbooks / total number of fishers), which of course
is related to the number of fishers who have not declared their catch. Hence the interest in
understanding the fishery as mentioned above.
The catch and effort of fishers who have not made a declaration can be estimated by multiplying
respectively the mean catch and the mean effort of fishers who have done so by the number of fishers
who have not. This calculation can be improved if there is stratification into homogeneous fishing zones.
It is then possible to use, in each zone, the mean catch and mean effort of the fishers who declare their
catch and the number of fishers who do not.
The CPUE for the fishing métier is calculated from the ratio of total catch to total effort also
according to the chosen and feasible stratifications/aggregations (by season, month or any other period
of time, by fishing zone). It is also possible to calculate the mean of the CPUEs per fisher but beware,
the result is different. However, this approach allows the variability of mean CPUE estimations to be
calculated and statistical comparisons between fishing zones and between time periods for example are
then possible.
207

6.7.3.2. Voluntary declaration system
In order to estimate total catch and effort, we start from the mean catch and effort of the
cooperative fisher sample, by fishing zone if this stratification exists, and we extrapolate to the whole
population. In order to do this, nominal effort obtained from the total number of fishers by fishing métier
and the nominal effort unit is used. The classic stratified sampling theory of Cochran (1977) can be used
for these estimations.
The CPUE by fishing métier can be calculated according to “classic” methods in several ways:
* either from the estimated total catch and effort as in the compulsory declaration system;
* or by the ratio of total catch to total effort for the sample;
* or by calculating the mean of the CPUEs per cooperative fisher in the sample.
The General Linear Model can explain the observations through a certain number of effects and
permits the processing of data from unbalanced sampling plans, as is often the case in voluntary
declaration systems: it corrects inter-fisher variations.
A few examples are given below. For further information on their use for the statistical monitoring
of fisheries for migratory species, please refer to Castelnaud et al (2001), Beaulaton and Castelnaud
(2005), and Beaulaton et al. (2006) where these examples can be found, as well as to Beaulaton
(2007).
* Case of the allis shad fished with a net in the Gironde:
10
22
7
+
0 ∑ 1z [ Zone=
z]
∑ 2a
[ year=
a]
∑ 3q
[ fortnight=
q]
+
z=
1
a=
1
q=
1
ln( CPUE 1) = β + β I + β I + β I ε (1)
⎧1
if Y = y
I[
Y = y]
= ⎨
⎩0
if Y ≠ y
2
ε iid ~ N(0,
σ ) (iid = independent and identically distributed) (2)
208

* Case of the elver fished with a scoop net in the Gironde:
Ln(CPUE+1)= season + fisher + month + tide + season x month
* Case of the sea lamprey fished with a net in the Gironde:
Ln(CPUE+1)= season + month + fisher + zone
6.8. Fisheries biological indicators, analysis and interpretation
The three descriptors, C (total catch), f (fishing effort), CPUE (catch per unit of effort), can be
used to develop a fishing mortality indicator, and by extension a fishing pressure indicator, as well as an
abundance trend indicator, for a fishing sector.
The fishing mortality indicator is simply the total catch, which is the mortality due to fishing
within the framework of anthropogenic impacts. It is frequently called “fishing mortality” (which can be
confusing as the fishing mortality coefficient is also called fishing mortality to simplify). It is obtained from
the descriptor(s) - seasonal “total catch”:
* it is equal to the value of the relevant descriptor if only one stage for a fishing métier is
concerned;
* it is equal to the sum of the values of the descriptors if several fishing métiers, several stages
and therefore the species are concerned.
The total catch is used to build and interpret abundance indices, the CPUE. As “fishing mortality”
it is indispensable for models of population dynamics. This fishing mortality is the result of fishing
pressure on the stock.
A fishing pressure indicator can be defined, composed, on the one hand, of the “fishing effort”
descriptor which is indicative of pressure characteristics and measures the amount of pressure exerted
on the stock and, on the other hand, of the “fishing mortality” descriptor which measures the effect of
pressure on the stock and the amount destroyed.
When the total biomass (absolute abundance) has been assessed, the exploitation rate and
therefore the impact of the fishery on the stock can be calculated using the total catch or fishing
mortality.
The relative abundance trend indicator of a given stage for a given fishing métier derives from
the comparison of chronological series of the three descriptors C, f, CPUE over at least 5 to 10 years.
This is because, first, comparing the trend of the CPUE series with those of total catch and effort means
209

that the coherence between the three descriptors, and the validity of the CPUEs, can be verified,
particularly when these CPUEs originate from a sample of fishers. And, second, in some cases (the
Vilaine estuary, Briand et al., 2003) when fishing effort and the resulting exploitation rate are very high,
CPUE is no longer a good indicator of abundance and the total catch must be used (Gascuel et al.,
1995).
Often, the abundance trend is derived solely from the CPUE series but if (as is often the case)
the CPUE trend appears to be stable, there is a serious risk of being in error (Mullon et al., 2005). The
CPUE alone cannot always be used as an abundance indicator because it only indicates “how” the
trend seems to evolve (a stable CPUE trend may mask a decline in abundance).
If the trends in total catch and in fishing effort and its structure (fishing power, fishers' tactics) are
known then the factors explaining these trends can be understood, the risk of error can be assessed
and limited, and the diagnosis can be improved (Gulland, 1969; Laurec and Le Guen, 1981; Chapman,
1990; Kleiber and Perin, 1991; Myers et al., 1997; Neis et al., 1999).
For example, if stable CPUE are the result of total catch and total effort decreasing
proportionately, and if the true effort is poorly measured and represented by the effort unit used, there
may be a failure to recognise that catch levels are being maintained by a change in individual effort. The
stock may reduce in size but be organised differently (with movements of individuals) and the remaining
fishers may be distributed differently over the fishing zone, moving more, and searching for fish more
intensively and more judiciously with their fishing gear.
The three descriptors must be collated over several consecutive years (chronological series) if
comparisons are to be made together with the beginnings of an evaluation of abundance trends for one
or more stages and for the species.
The various possible combinations of trends in CPUE, total catch and total effort are shown in
figure 6.6 (Girardin et al., 2006). This makes it clear that caution is essential in cases 6, 7, and 8.
N° 1 2 3 4 5 6 7 8 9 10 11 12 13
CPUE
Total catch (C)
Total effort (f)
Figure 6.6: Possible combinations of catch per unit of effort
Cases 1 to 5 represent situations where abundance is at least stable and almost certainly
increasing. Cases 1 and 5 imply, for the CPUE trend to be correct, that total catch either increases more
quickly or decreases more slowly than total effort.
Cases 9 to 13 represent the opposite case and can only really indicate a decrease in abundance.
In cases 9 and 13, total catch increases more slowly or decreases more quickly than total effort.
210

In conclusion, it should be recalled that the entire process leading to fisheries biological
descriptors in the proposed method is designed so as to respect the fundamental postulates (Beverton
and Holt 1957; Brethes 1990; Kleiber and Perrin 1991), and to limit, so far as possible, the sources of
absolute and systematic error (Caddy and Mahon 1996; Caddy 1998) and variations in catchability
(Chapman 1990; Chadwick and O’Boyle 1990; Kleiber and Perin 1991; Gillis et al., 1993; Gillis and
Peterman 1998; Harley et al., 2001) in an attempt to follow the principles and conditions established in
previous chapters.
6.9. Examples of indicators of the trend in relative abundance and of
diagnoses.
6.9.1. Indicators of relative abundance trends at the glass-eel/elver stage.
This example concerns indicators of the trend in relative abundance of glass eels / elvers fished
in the Gironde fishery sector over the 1978-2005 period, with a “pibalour” in the estuary and the hand
scoop net (T: “tamis à main”) and circular push nets (D: “drossage”) upstream in the fluvial zone.
Figures 59 and 60 show the three descriptors C, f, CPUE for the glass eel/elver – hand scoop net, and
glass eel/elver – “drossage” (from 1996) on the one hand and glass eel/elver – “pibalour” on the other
hand.
180
160
140
120
100
80
60
40
20
0
(1) t
(2) Day x 100 (3) kg / gear / day
(1) Catch T
(1) Catch D
(2) Effective effort T
(2) Effective effort D
(3) CPUE
(3) CPUE D
1977-1978
1978-1979
1979-1980
1980-1981
1981-1982
1982-1983
1983-1984
1984-1985
1985-1986
1986-1987
1987-1988
1988-1989
1989-1990
1990-1991
1991-1992
1992-1993
1993-1994
1994-1995
1995-1996
1996-1997
1997-1998
1998-1999
1999-2000
2000-2001
2001-2002
2002-2003
2003-2004
2004-2005
30
25
20
15
10
5
0
Figure 6.7 - Glass eel / elver – scoop net and glass eel / elver – “drossage”: total catch, total
effective effort and CPUE of commercial fishers between 1978 and 2005 (source :
Giradin et al., 2006).
211

The break in CPUE for the “pibalour” métier between 1981 and 1982 is not as clear as for the
"scoop net" métier but it shows a similar if more progressive downward trend.
Whilst catches from the “scoop net” métier also fell between 1981 and 1982, the catches from the
“pibalour” métier periodically reached identical levels between 1978 and 2002. But unlike the “scoop
net” métier where effort decreases imperceptibly, the effort used in the "pibalour" métier nearly doubled
from 1989. This explains the decrease in the "pibalour" CPUE mentioned previously.
The movements in CPUE and catches of the “drossage” métier do not change the results of the
analysis based on the “scoop net” métier.
Over the past few years, the CPUE of the three métiers have remained low, as have catches,
despite a sustained effort with the “pibalour”.
Globally, following a two-stage decline in glass eel / elver abundance between 1980 and 1985
shown successively by both main métiers, the situation is currently stable but at a low level due to the
results of the last five years.
In this example, the series of descriptors show that there is a marked drop in CPUE at the
beginning of the monitoring period for the two oldest métiers, a result which is not contradicted, and is
even supported, by trends in the other two descriptors C and f.
There is a strong linear correlation between “pibalour” and “scoop net” CPUE tested by the
classic method (Beaulaton & Castelnaud, 2006) and confirmed by GLM which, incidentally, shows its
advantages (Beaulaton & Castelnaud, 2006).
212

6
5
4
3
2
1
0
(1) t
(3) kg / ‘pibalour’ / day (2) day x 1000
7
6
5
4
3
2
1
0
1977-
1978-
1979-
1980-
1981-
1982-
1983-
1984-
1985-
1986-
1987-
1988-
1989-
1990-
1991-
1992-
1993-
1994-
1995-
1996-
1997-
1998-
1999-
2000-
2001-
2002-
2003-
2004-
(1) (2) Effective effort (3)
Figure 6.8 -
Glass eel / elver – “pibalour”: total catch, total effective effort and CPUE between
1978 and 2005 (source : Giradin et al., 2006).
Remark: As explained previously, the obvious increase in fishing power for the “pibalour” métier in the
1980s was not assessed but in this case, the increase in nominal and effective efforts between the
1980s and the 1990s is such that the trend in this descriptor is clear and contributes to an undeniable
interpretation of trend in abundance. In a different setting, with for example a stable nominal effort, this
exercise in the evaluation of the change in fishing power would be very useful as in the case of the
yellow eel, presented previously.
Indicators of relative abundance trends at the yellow eel stage.
This example concerns the indicator of the trend in relative abundance of yellow eels fished with
basket traps in the Gironde fishing sector over the 1978 - 2005 period. Figure 6.9 shows the three
descriptors C, f, CPUE for the yellow eel - basket trap métier.
213

400
350
300
250
200
150
100
50
(1) t
(2) ANGN – eel basket trap
(3) kg / basket trap / month
2
1,8
1,6
1,4
1,2
1
0,8
0,6
0,4
0,2
0
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
0
(1) Catch (2) Nominal effort (3) CPUE
Figure 6.9 - Eel–basket traps: total catch, total nominal effort and CPUE of commercial fishers
in the Catchment between 1978 and 2005 (source : Giradin et al., 2006).
Nominal effort, having fallen in two stages between 1982 and 1983 and between 1990 and 1991,
has increased slightly and levelled off since 2001. Total catch, having dropped over the same periods,
has decreased faster than effort since 2000. A slight increase in 2004 was followed by a further fall in
2005, as was also the case with the CPUE.
The fishers, in reduced numbers for the last ten years, benefit from eel movements from a more
significant non-exploited surface area whilst covering more ground over the fishing territory as the
resource has become more dispersed. There are justifiable fears that the current CPUE trend,
calculated from sample data, may well not be representative of true abundance and may mask a further
significant decline in the eel resource, which is suggested by the trend in the other two indicators, even
if the effort unit used (the basket trap by fishing month) cannot include some of the true effort
characteristics indicated above.
An interesting parallel can be drawn with the interpretation of the trend in the three yellow eel
descriptors C, f, CPUE on the Adour by Cuende and Marty (2005) in the Indicang progress report. On
the basis of the graph from Lissardy et al. (2005), the latter conclude, for a similar period to that of the
Gironde (1986-2003), that “production shows a downward trend, which can be explained by a fall in the
number of fishers whilst fishing yield remains globally stable”.
The yellow eel example in the Gironde is instructive in two ways. It shows even better than the
glass eel / elver example that had no data been available before the abundance index fell, i.e. 1989, it
214

would have been possible to conclude that there was a slightly upward trend in abundance in the
following decade, followed by a stable trend more recently. A simple extrapolation could have led to the
mistaken conclusion that this was merely a prolongation of the past, even if this were unknown. Hence
the interest in having chronological series of descriptors which are as long as possible (Pauly D.,1995).
This also justifies the interest in a reconsideration of the effort units and measurements used as
well as of course the changes in fishing power and tactics as explained in the previous chapters.
6.9.2. Indicators of relative abundance trends at the silver eel stage.
This example concerns the indicator of the trend in the relative abundance of silver eels migrating
downstream and fished with a “guideau” in the “middle reach of the Loire” sector over the 1987 - 2006
period.
The catch and fishing effort data used are those reported by Boisneau and Boisneau, in chapter
10, which come from a sample of 4 commercial fishers from the 15 using the "guideau" in the fishing
sector. The CPUE is calculated here. The sample is considered to be representative of the total
commercial fisher population using the “guideau”.
Figure 6.10 shows the three descriptors C, f, CPUE for the silver eel-“guideau” métier.
50000
45000
40000
35000
30000
25000
20000
15000
10000
5000
0
Production (individuals)
Effort (fishing days)
CPUE (individuals / fishing day)
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
475
450
425
400
375
350
325
300
275
250
225
200
175
150
125
100
75
50
25
0
Figure 6.10 - Silver eel - "guideau": total catch, total effort and CPUE from the commercial fisher
sample between 1987 and 2006 (source : G.Castelnaud).
215

Between the beginning and the end of the study period, effort decreased by half and catch by
more than two-thirds. CPUE, which remained rather stable between 1987 and 2003, has decreased by
half since. Recently, catch has decreased faster than effort, leading to a decrease in CPUE and to a
current, worrying diagnosis of a downward trend in silver eel abundance in this sector of the middle
Loire.
216

Part III
Evaluation by inland life stage
217

Chapter 7
Estuarine recruitment indicators
Patrick Prouzet, Jean-Charles Bouvet, Noëlle Bru, Elise Duquesne,
José-Carlos Antunes, Alfredo Demasceno-Oliveira, Ahmed Boussouar,
Marie-Noëlle De-Casamajor, Florence Sanchez, Murièle Lissardy
218

7.1. Context and objective
This chapter deals with a particular biological stage: the “civelle” (glass eel and elver), i.e. a
migratory individual caught or observed on the coastline (surfcasting on the Landes coastline for
example) or in the marine and fluvial parts of tidal rivers and streams.
Individuals observed and counted at this life stage are usually caught by either professional or
amateur fishermen using distinct gears as defined by regulations in force, or else by scientists during
monitoring programmes, the objective of which is to evaluate the characteristics and behaviour of
individuals or to count them.
This biological stage is exploited mainly by southern Atlantic Arc countries, particularly France
and Spain. This exploitation generates significant socio-economic activity 1 . Glass eel abundance has
fallen significantly since the beginning of the 1980s but exploitation remains widespread in many
estuaries of the Atlantic zone due to high Asian demand and very high market prices in recent years
(Nielsen and Prouzet, 2007).
Estuarine recruitment can be assessed by estimating the extent of glass eel arrival into estuaries
affected by the dynamic tide. Fluvial recruitment 2 is what is left of estuarine recruitment after fishing, the
sedentarisation of some glass eels in zones under marine influence and the reduction in numbers due
to natural or (non-fishing) anthropogenic causes 3 .
The objective of this chapter is to define indicators of estuarine recruitment concerning its level,
its variability, the hydrodynamic factors affecting glass eels' entry into the estuary and their presence
near the surface, their level of extraction by various uses (including fishing) as well as the intrinsic
quality of migratory fish and the length of their stay in the estuary.
Another part of this handbook 4 deals with catch recording and descriptors required in fisheries
statistics. Commercial or amateur fishing activities are a precious source of information (sometimes all
that is available) but the information they provide needs to be used in such a way that it is
representative and inferences can be drawn concerning the entire glass eel population migrating into
the estuary.
Two different methods are suggested in order to define:
1 Léauté J.-P., Coord., 2002. Contract DG/FISH PECOSUD, http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/site-atelieradour-rapport-5-5.pdf
2 See Chapter 8.
3 See Chapter 4. §
4 See Chapter 6.
219

• a relative indicator showing trends through time (during the season) and in space (in several
estuaries or at several points of an estuary for the same period). This indicator should represent the
relative evolution of abundance in the target population at a given place and time. In order to
achieve this, the (scientific or commercial) catch per unit effort is generally used, as theoretically it is
possible to eliminate effort variation by using an indicator which is a ratio of catch to effort i.e. Catch-
Per-Unit-Effort: CPUE = qN according to this equation. This requires a sound analysis of what
fishing effort really is, an analysis that is too often neglected in fisheries studies, whether comparing
this index in a given estuary (evolution of fishing techniques) or comparing several estuaries
(calibration of fishing effort). This is far from being trivial when fishing effort is simply defined, using
only the number of fishers. For example, the fall in the number of eel fishers on the Adour has not
been uniform over the last ten years. Some fishers who specialised on this life stage have ceased
their activity at the same time as other, less efficient, fishers, seeking a complementary activity,
have begun. As long as this initial analysis has been carefully conducted, the variation in CPUE will
be representative of the trend in abundance provided catchability q remains constant. Hence, the
hypothesis that catchability remains constant must be verified 5 .
• an absolute indicator describing a number or a mass of individuals present in a given place and at
a given time. This can be assessed from the densities present in a given identifiable place, or
volume, from which the catch originates. This biomass of glass eels can be estimated from their
density during the incoming tide, either through scientific research or by the detailed monitoring of
commercial vessels' catch and effort (for a given sample), and extrapolated to the circulating
volume 6 . This extrapolation may be achieved by using an equation relating the exploitation rate to a
set of variables so that variations in the catchability of glass eels on a given day and in relation to
the gear used can be taken into account.
7.2. Scale of the study
7.2.1. General framework
This covers the part of the estuary affected by the dynamic tide, which is generally subject to a
variety of administrative arrangements (fishing or navigation regulations) 7 .
The study period is the glass eel migration period which depends on the latitude. In the central
zone of the distribution area (centered on the Bay of Biscay), it generally occurs between November and
April and somewhat later further northwards.
5 See Chapter 6.
6 § >
7 See Chapter 6.
220

#
#
e
7.2.2. Further details on the administrative and environmental
context
7.2.2.1. Geographical and administrative framework
The Adour and Gaves basins will be used as examples and their geographical location is shown
in figure 7.1. This figure comprises two maps which provide a sufficiently accurate overview of the
information available on French estuaries.
Once the general context is established, it is necessary to identify and characterise the various
environmental parameters which affect migratory glass eel behaviour. Various descriptors can be used
to characterise the zone being studied and to follow the fluctuations in environmental records.
Having selected the zone where observations are to be carried out, existing information and
organisations holding data must first be identified and the following must be checked: the accuracy,
quality and frequency of measurements, the geographical location of the sample and its location within
the water column, the unit of measurement, and the type of device and its accuracy.
The Adour Catchment
N
Location
‘Commune’ boundaries
OCEAN
ATLANTIQUE
St Jean #
de Marsacq
Rivière saas
et gourby
#
Saubusse
#
Josse #
#
#
Pey
Orist
Luy
Transversal Limite limit
of transversale the sea
de la mer
#
St Martin-de
Hinx
Ste-Marie-de-Gosse
#
# St-Etienne
d'Orthe
#Port-de-
Lanne
Gaves ré u nis
‘Cantons’
#
Anglet
Nive
#
Tarnos
Boucau
Bayonne
#
Adour
Saint-martin
-de-seignanx
St-Barthélémy
#
Lahonce Urcuit
# #
#
Urt
#
St-Laurent
de Gosse
Joyeu s
Guiche #
Bidouze
#
Sames
Limite Public du maritime Domaine
domain Public Maritime boundary
Limite Salt de water salure limit des eaux
Adour basin
‘Commune’ boundary
‘Canton’ boundary
‘Department’ boundary
A
10 0 10 20 Kilomètres
Administrative Division administrative division de of l'estuaire the estuary
Adour
Bassin
river
versant
basin
de l'Adour
Maritime Zone maritime zone (22km) Km) under sous the control of
Maritime contrôle des Affairs Affaires maritimes
Mean Trait de high côtewater mark
Zone mixte (21.9 Km)
Mixed zone (21.9km) under DDAF control
sous contrôle de la DDAF
(Agriculture & Forest Dept)
Fluvial Zone fluviale, zone, under sous DDAF control
contrôle de la DDAF
Boundary Limite entre between chaque each zone zone
B
Hydrology Hydrologie
‘Communes’ of the Pyrenees-
Atlantiques Communes des Pyrénées-Atlantiques
‘Communes’
Communes des
of the
Landes
Landes
Figure 7.1. General features of the Adour basin and its estuary: A) geographic and hydrographic
context; B) administrative and hydrodynamic context of the Adour estuary (source : Atlas of the Adour
Watershed, Adour watershed observatory: IGN BD, 5/2002 map and S.Gharbi, 2002).
221

7.2.2.2. Environmental context
Environmental descriptors are organised using four main criteria (table 7.1). It is important to note
immediately that precision concerning the hydrodynamic context (flow, tide coefficient, thermal and
saline homogeneity of water) is particularly important as this context 8 plays a predominant role in glass
eel behaviour and transit times, and hence on their vulnerability to fishing gear in the estuary. Examples
simulating this migration within the part of the estuary affected by the dynamic tide are given in the
following paragraph.
Table 7.1. List and nature of the estuary environmental descriptors.
Criteria Descriptors Objectives Guidelines
PHYSICAL STRUCTURE Bathymetry (metres) Calculation of the depth and the wet cross-section
at a given point of the estuary in order to estimate
the circulating volume or to implement a
hydrodynamic model
HYDRODYNAMISM
PHYSICO-CHEMISTRY
CLIMATE
Granulometry
(descriptive)
Flow velocity (in
m/s)
Flow (in m 3 /s)
Tide coefficient or
water level (in cm)
Estimate glass eels’ burrowing capacity and their
aptitude to move towards the banks at times of
flood or during ebb tides.
When coupled with the wet cross-section, the
circulating water volume or the volume filtered by
fishing gear can be calculated.
Hydrodynamic features: definition of the tidal limit
or a hydrodynamic barrier
Additional information on flow used in
hydrodynamic models to estimate the tidal
propagation speed and distance.
Lunar cycle Characterise nocturnal luminosity and also
favourable migration periods for glass eels.
Water temperature Characterise the thermal barrier to glass eel
(in °Celsius) migration or estimate the pigmentation speed.
Turbidity (in NTU or
mg/l)
Salinity (in ppm or in
g/l)
Pluviometry (in mm
per day)
Air temperatures (in
°Celsius)
Cloud cover (in
oktas)
Characterise the suspended load which affects the
amount of light within the water column for a given
nocturnal or diurnal luminous intensity.
Highlight potential haline stratification of the water
column
Gives some idea of water clarity when turbidity
measurements are not available or can be
estimated using a model
When data on water temperature are lacking, can
be used to estimate this temperature using a model
based on smoothed data with a time lag
Characterise a reduction in external luminosity:
degree of cloud cover to be taken into account
particularly during quadratures
Mean widths and depths every 10 metres
along a transect.
Substrate description: friable (mud, sand,
gravel) or hard (schist, blocks, …)
Some idea of the filtrating speed of a fishing
gear can be obtained by using a flow meter
but a hydrodynamic model can only be
validated with a fixed measuring device
(Doppler rosette for example).
Record daily information provided regularly on
most watercourses or else use the specific
module in l/s/km 2 and pluviometry
Record information about daily events for
each tide. The water level is a reference point
(level 0 of the maps) but can also be
measured at a given point of the estuary
compared to a terrestrial reference point.
Qualitative descriptor with 4 classes: full
moon, new moon and the two quadratures.
At a depth of around 2 metres by permanent
sensor otherwise use an approximate
estimate by air temperature and flow.
Choose depths between 1 and 2 metres
(depth layer that can be explored by surface
scoop nets)
Occasional measurement to check that there
is no deep salt area (which could trap glass
eels) within the surveyed zone.
Generally provided regularly by
meteorological stations
Generally provided regularly by
meteorological stations
Generally provided regularly by
meteorological stations
7.2.2.3. Analysis of the estuarine propagation speed of a
glass eel flux
In order to fully understand this parameter, a behavioural model designed for the Adour
(Boussouar et al, 2005) was adapted to the Isle (Duquesne, 2007) and to the Loire (Prouzet et al, 2008).
222

The conceptual diagram shown in figure 7.2 summarises the observations collected. They are
discussed here briefly.
The longitudinal transit of a glass eel flux in the estuary depends significantly on the
hydrodynamic conditions and the propagation of the dynamic tide into the estuary. This transit is passive
in the part of the estuary affected by the tidal front. The actions of the tide coefficient and the river flow
on glass eel behaviour cannot be considered separately. Current speed is an important factor in glass
eel migration, blocking it when maximum downstream speed is higher than 0.3 m/s.
The vertical behaviour of glass eels may be studied from densities observed near the surface and
at depth. The ways in which the fluxes transit vary greatly according to environmental conditions and in
particular to the light. The vertical position of glass eels is determined by two main factors: turbidity and
lunar phase. In murky water glass eels use all of the water column, regardless of the phase of the lunar
cycle. In clear water, they transit at depth, especially during the full moon and the first and last quarters.
Cloud cover is a factor which modulates nocturnal luminosity. Vertical movements within the water
column are quite active and are related to light avoidance.
Figure 7.2. Conceptualisation of glass eel behaviour in the estuarine area in relation to
external factorsin the Adour estuary (source ; Boussouar et al., 2005).
8 See Chapter 2.
223

This conceptual diagram of glass eel migration is only valid for the part of the estuary affected by
the propagation of the dynamic tide and not hydrologically stratified (no marked halocline and
thermocline). The hydrology may vary greatly from one estuary to the next, depending, in particular, on
the river regime and the morphology of the downstream reaches. For example, the diagram may be
simplified if the estuary is murky (more than 100 NTU). In this case, this factor, which is important for the
glass eel vertical distribution in the Adour estuary, is not as relevant.
A behavioural model (Prouzet et al, 2003; Boussouar et al, 2005) was developed to model the
impact of hydroclimatic components on glass eel behaviour and their speed of transit in the estuary.
Appendix 10 9 of the Indicang report gives the numerical version of the conceptual model outlined in
figure 7.2. It combines a 1D hydrodynamic model, which must be adapted to the bathymetry of the
relevant estuary, to the records of relevant flows, and the calculated or observed water levels, with a
model of light in the water column which determines the vertical distribution of the glass eel flux. A
number of glass eels is then introduced at the mouth as either a dense or a scattered volume, and as
either a symmetrical or an asymmetrical distribution during each incoming tide.
This model can be applied in many different ways and helps to test various behavioural scenarios
and compare them with observations collected from either fisheries data or scientific work undertaken
on the Adour, the Isle or the Loire; the 3 river basins in the Indicang project for which this behavioural
model was adapted (Boussouar and Prouzet, 2007).
On the Adour, the question was the following: do glass eels migrate upstream during the day with
the incoming tide, given that sampling in open water did not produce any glass eels during the day?
Logbook analysis showed the successive peaks in total daily landings in 2 distinct fisheries about
15 km apart (figure 7.3). The first one is located in the estuarine area between 10 and 20 km from the
estuary. The boats use two push nets (called “pibalour”) to catch the glass eels. The second one is
located on the Adour river, upstream of the confluence with the Gaves Réunis, about 30 km from the
mouth. This fishery mainly uses hand-held scoop nets from anchored boats. Figure 7.3 gives a precise
example of the distribution of peak landings in these 2 zones and the time interval between these peaks.
9 After Boussouar A., Arino O., Prouzet P., 2005. Formulation du modele comportemental elabore pour la civelle, Annex 10 of the
Indicang report, http://www.ifremer.fr/indicang
224

250
200
08-déc-99
10-déc-99
Pibalour
Push net
Tamis
Scoop net
12-nov-99
28-déc-99
Catches in Kg captures en kg
150
100
50
09-nov-99
29-déc-99
0
01/11/1999
05/11/1999
09/11/1999
13/11/1999
17/11/1999
21/11/1999
25/11/1999
29/11/1999
03/12/1999
07/12/1999
11/12/1999
15/12/1999
19/12/1999
23/12/1999
27/12/1999
31/12/1999
04/01/2000
08/01/2000
12/01/2000
16/01/2000
20/01/2000
24/01/2000
28/01/2000
01/02/2000
05/02/2000
09/02/2000
13/02/2000
17/02/2000
21/02/2000
25/02/2000
29/02/2000
04/03/2000
08/03/2000
Figure 7.3. Fluctuations in total daily catches using the “pibalour” (push net) in the maritime
zone and the “tamis à main” (hand scoop net) in the mixed zone during the 1999-
2000 fishing season (from Prouzet et al, 2003).
This figure shows in particular a first landing peak on the 9 th of November in the maritime zone
followed by a second on the 12 th of November in the mixed zone on the Adour just past the location
called “Horgaves”. Detailed data concerning the evolution of catches and hydroclimatic conditions are
recorded in table 7.2.
Table 7.2. Evolution of glass eel catches using the “pibalour” (push net) and the “tamis à main”
(hand scoop net) and relevant hydroclimatic characteristics during the period
between the 9th and the 12th of November 1999 on the Adour (after Prouzet et al.,
2003).
Date
“pibalour” catches
(kg)
Scoop net catches (kg) ratio Flow (m 3 .s -1 ) moon Water t° (°C)
09/11/99 92.37 68.86 85 139.83 NM-FQ 11.62
10/11/99 73.29 92.04 80 145.43 NM-FQ 11.28
11/11/99 50.74 138.25 74 137.55 NM-FQ 10.92
12/11/99 35.76 157.54 66 128.95 NM-FQ 10.62
If it is assumed that the peak observed in the fluvial zone belongs to the same flux caught in the
maritime zone for the dates mentioned in table 43 (between 9/11 and 12/11), the glass eel journey to
reach the fluvial zone (a distance of about 20 km) takes 2 to 3 days. It should be noted that this was a
period when the moon was not very bright which assisted catches by the 2 gears used as glass eels
were closer to the water surface.
Given these conditions, we attempted to reproduce the observed propagation speed using the
behavioural model with 2 simulation hypotheses: daytime burrowing (figure 7.4) or migration at depth
during the day with the incoming tide (figure 7.5).
225

These simulations showed that the glass eel flux must migrate by day during the incoming tide if
the successive landing peaks in the two fisheries are to be explained. Various examples were taken and
confirmed this simulation (Prouzet et al, 2003). Glass eels, given the luminosity, move on or near the
bed and hence cannot be fished by push nets operating at the surface.
Simulation du déplacement d'un flux de civelles avec enfouissement des civelles le jour
(A)
(B)
(C)
Arrivée du flux de civelle le 09/11/1999 à 5h00
entre 8 et 13 km (zone maritime)
(D)
Présence du flux de civelle le 10/11/1999 à 3h00
entre 10 et 15 km (zone maritime)
Figure 7.4. Results from the behavioural simulation model with glass eel diurnal burrowing: (A) Arrival of
the glass eel flux on the 9/11/1999 at 5.00am between 8 and 13km (marine zone); (B)
Presence of the glass eel flux on the 10/11/1999 at 3.00am between 10 and 15km (marine
zone); (C) Position of the glass eel flux in the water column between 17 and 22km (marine
zone) on the 11/11/1999 at 4.00am; (D) Glass eel migration resumes during the incoming tide
on the 13/11/1999 at 2.00am. The flux has not yet reached Horgaves 30km away (source :
Prouzet et al, 2003).
Simulation du déplacement d'un flux de civelles avec migration près du fond durant le jour
(A)
(B)
(C)
Arrivée du flux de civelle le 09/11/1999 à 5h00
entre 13 et 20 km (zone maritime)
(D)
Déplacement du flux le 10/11/1999 :: localisation
entre 20 et 25 km de la mer à 3h00.
f 11/11/1999 00 L fl d i ll tt i t l fl i l l 12/11/1999
Figure 7.5. Results from the behavioural simulation model without glass eel diurnal burrowing: (A)
Arrival of the glass eel flux on the 9/11/1999 at 5.00am between 13 and 20km (marine zone); (B)
Movement of the flux on the 10/11/1999: position between 20 and 25km away from the sea at 3.00am;
(C) Presence of the flux on the 11/11/1999 beyond Horgaves (30km away from the sea); (D) The glass
eel flux reaches the fluvial zone on the 12/11/1999 (source : Prouzet et al, 2003).
226

Another question asked by managers concerns the accumulation of glass eel groups carried by
successive incoming tides in the propagation zone of the dynamic tide as a result of the hydrodynamic
processes (difference in the passive migratory speeds from one incoming tide to the next or merely a
reduction in such speeds as eels move upstream).
Evolution des cohortes de civelles sur l’Isle du 05/01/06 au 08/01/06
01/06
Coefficient
: 74
Flow
Débit : 53 m 3 /s
08/01/2006 2h
6 5
4
1+2+3
4
18-21 Km
18
21 Km
2
12-16 16 Km
4
3
2
07/01/2006 0h
1
8-12 Km
6-10 Km
3
0-2 2 Km
2
1
06/01/2006 0h
0-2 2 Km
2
Distance = 2-6 Km
1
05/01/2006 22h
Profondeur (m)
1
Figure 7.6. Simulation of the migration of 6 successive glass eel arrivals on the Isle with the
hydrodynamic conditions observed between the 5th and 8th of January 2006 (from Duquesne,
2007).
227

The behavioural model can be used to address this question as shown by the example in figure
7.6 concerning the migration of glass eels carried by successive incoming tides at the mouth of the Isle
(Duquesne, 2007).
Assuming that a group of glass eels arrives with each tide and assuming average hydrodynamic
conditions (close to the mean tide coefficient and flow), the fusion between the different groups starts to
occur around 20km from the confluence with the Dordogne, i.e. around Saint-Denis de Pile, after 5
successive incoming tides.
It is also important to know the glass eel flux propagation speed if the aim is to allocate the daily
catches of a fishery located on the axis of the estuary to a glass eel group, the biomass of which is
estimated from a sampling station that is outside the exploited zone.
Figure 7.7 illustrates this case for the Loire estuary. The sampling station, located near Le Pellerin
for sampling protocol reasons and to respect the hypotheses underpinning the statistical model used to
estimate glass eel flux, is situated further upstream than the main fishing zone, which is found between
Paimboeuf and Cordemais.
Zone de pêche
‘Reference’
dite de
fishing zone:
référence : lot
lot LM
LM
Lot 5
Zone
d’échantillonnage
used for sampling
PK
0
• 15
40
Figure 7.7. Position of the fisheries on the lower reach of the Loire estuary and of the station
sampling the flux of glass eels (from Prouzet et al, 2008).
The analysis of catch series from the 2 consecutive fisheries (figure 7.7) located on the fishing
zones LM and L5 (where the flux sampling station is located) shows staggered, and sometimes multiple,
landing peaks, which makes it difficult to associate one landing peak in "LM" with another one in the
fishing zone further upstream “L5” without knowing how this glass eel flux moves in the estuary.
Hence the use of the behavioural model to assess the lag in the hydrodynamic conditions
observed between catches made at “LM” (at kilometre 20) and the sampling zone situated upstream (at
kilometre 35).
228

Estimations show that the lag between the flux passing through the fishery “LM” and the fishery
"L5" is about one day under average conditions of tide coefficient and flow as shown in figure 7.8.
Using this simulation of the flux progression, “LM” landings on day (d-1) and “L5” landings on day
d can be related to the estimated biomass on day d during the incoming tide.
A
B
C
D
Figure 7.8. Simulation of the migration of a glass eel flux in the Loire estuary on the 14 th of
February 2006 (flow= 452m 3 /s and tide coefficient = 83). (A) 3.00 am arrival at
Cordemais; (B) 6.00 am the centre Of the flux is at the 30 th km; (C) 14.00h
burrowing starts at the turn of the tide at the 35 th km; (D) 19.00h arrival of a new
flux in Cordemais and the first one passes through the reference station (from
Prouzet et al., 2008).
7.3. Data acquisition: definition of descriptors characterising the
abundance of a glass eel flux
7.3.1. Descriptors characterising the absolute abundance of a daily
or seasonal glass eel flux
The aim is to provide a methodology to characterise the descriptors used to estimate the absolute
abundance of a glass eel flux (or group) in a given space and time interval. These can be obtained
though scientific sampling in the estuary or by trapping and counting migrating individuals (for example,
using counting structures located on estuarine dams, as at Arzal on the Vilaine). They can also be
229

obtained using commercial fishing gear (subject to their instrumentation) when luminosity conditions are
homogeneous in the water column.
It will be recalled that a glass eel flux is a glass eel group migrating in an estuary during a given
time interval: incoming tide, a day or a season.
The abundance of the flux can be defined instantaneously as a density (weight or number per unit
volume), a biomass or as a number which integrates this density over the whole circulating volume, or
as a simple count in specially designed structures located in the zone of tidal sway.
7.3.1.1. Estimation of the abundance of a glass eel flux by
the sampling of densities in the water column
The selected study zone (also called sampling station) must be located in the estuary as close as
possible to the mouth in order to provide the fullest possible characterisation of an entering flux.
Likewise, if the aim is to quantify glass eels migrating upstream on a particular watercourse, the study
zone must be located as close as possible to the confluence of this river with the connected
watercourse.
For a successful quantitative evaluation of densities of glass eels entering a zone, it is
recommended to sample this fixed zone over a period of time.
The space (volume) in which glass eels are sampled at this station must also be clearly defined.
The default space corresponds to the watercourse wet cross-section of the sampling zone and hence to
the volume of water transiting through this section during the relevant time interval. Spatial
heterogeneity related to physical or physico-chemical environmental factors can lead to glass eels
concentrating in a particular zone of the water column. These variations must be identified so that an
effective sampling protocol can be devised.
Thus, as far as possible, the following recommendations should be followed. The sampling zone
must:
• be under the influence of the dynamic tide to avoid the merging of the various glass eel groups
entering the estuary with each incoming tide;
• be linear and, if possible, canalised so as to avoid flux aggregation behaviour related purely to
transversal current heterogeneity;
• present neither haline nor thermal stratifications to avoid vertical trapping of glass eels below the
shear zone, the volume of which cannot be modelled simply by a uni-dimensional hydrodynamic
model (horizontally and vertically homogenous average speed) or by sinusoidal approximation.
Sampling protocol
This protocol must, so as to allow effective sampling of the flux in the water column, produce
information:
230

• over the whole water column or, failing that, for two surface layers and the middle part of the water
column 10 because, luminosity rapidly decreases beyond depths of three metres;
• across the whole width of the river to test the potential transversal heterogeneity of the flux (edge or
channel effect). The number of lateral sampling points will vary according to the width of the
watercourse. Three default sampling points are, nonetheless, recommended (on the right bank, in
the middle and on the left bank).
The time scale chosen for sampling may vary (one hour, one tide, one day).
Once calibrated, the sampling protocol (sampling points, types of fishing gear used, equipment
used) must be followed without changes.
A descriptor of abundance per unit of volume: density
This descriptor concerns one of the components which, together with fishing effort 11 , has a direct
impact on catch abundance. In the case of glass eels, fishing effort is calculated from the volume filtered
by fishing gear, which is the most accurate definition of effective effort.
Calculation of the volume filtered (m³) during a certain period of time.
The aim is to quantify the volume of water entering the scoop net during a given time interval.
Volume filtré = vit _ filtration × aire _ filtrée × durée
This is calculated as follows:
3
2
m = m / s × m × s
where
vit _ filtration is the filtration speed, which may be equal simply to the speed of the
current entering the zone or to the velocity of this current combined with the speed of the boat;
aire _ filtrée is the filtration surface area, which may be equal to the area of the opening of the fishing
gear; durée is the filtration time, which corresponds to the fishing period.
Calculation of glass eel density D (in g/100m 3 )
Density (in g/100m 3 ) is calculated as follows:
D
Poids
=
Volume filtré
=
100 × P
v × S × Δt
where P is the weight of glass eels caught per scoop net tow; v is the average filtration speed (in
m/s) during the fishing time interval Δ t (in s) ; S is the area of the fishing gear (in m 2 ).
10 See Chapter 2, § >.
11 See Chapter 6.
231

By monitoring the density within the water column over a period of time at a specified sampling
station (figure 7.9), the evolution of glass eel relative abundance over the duration of the rising tide can
be described. The biomass that migrates during the incoming tide can be estimated by combining
recorded densities at different times and the evolution of the volume which transits through the station.
Figure 7.9. Depiction of the boat route at a sampling station located in the middle reach of the
Adour estuary. The 3 transects are visible and coincide with the repeated passage
of the boat along the right and left banks and in the middle of the estuary (source :
Prouzet et al., 2003).
Elements required to calculate the circulating volume and its characteristics
The circulating volume is defined as the volume of water where glass eels may be found and are
carried upstream in different layers of the water column during a given period of time. This volume is the
variable A in the catchability (q) calculation.
232

Table 7.3. List of elements to be collected in relation to the circulating volume.
Calculation elements Objectives Guidelines
Study period: beginning, end,
duration
• Define the temporal scale of the study (one
tide, one day, one season)
• In the case of one tide, to be defined
precisely with respect to respective intensity
of flow and tide coefficient
Section • Average profile of the water section in the
catch zone
• Characterise the potential glass eel
dispersion area (wet cross-section )
Water current velocity curve
over the relevant period of time
or mean water velocity during
incoming tide
Temperature
Turbidity
Calculate instantaneous velocity during sampling
or an average velocity during the chosen time
period (usually the duration of the incoming tide)
Obtain water temperature at the time of glass eel
sampling or else define the water temperature
affecting the pigmentation rate by more frequent
measurements
Turbidity at the time of sampling at 2 to 3 metres
from the surface
Assess according to tide calendars
and hydrological records. Implement a
1D hydrodynamic model if possible
Use bathymetry and the average depth
at a fixed point
To be collected with care at various
moments during the incoming tide.
The calculation of the biomass
depends on its measurement and
accuracy
At least one average value to check if
there is a thermal barrier to glass eel
migration at the time of sampling
At the scale of a single tide, record an
average value and, when turbidity
varies during the tide, record at the
time of each sampling if possible
The velocities measured directly affect the circulating volume and density calculations. Great care
has to be taken when collecting and calculating this variable, making as many accurate measurements
as possible.
Two examples of data acquisition
The Adour estuary: a comprehensive protocol
Figure 7.10 shows the selected study station in the Adour river basin. It is situated about fifteen
kilometres from the mouth, in a zone under maritime regulations with a commercial push net fishery.
The bathymetry is regular and there is no haline stratification in the zone. The average width is 200m
with an average depth of 6.5m.
233

Mouth
Mixed zone
Marina
Motorway
Bridge
Chemical
products
Marine zone
Station 4: downstream of the Ile de Berenx
Figure 7.10. Map of the Adour estuary and identification of the sampling zone (source : Prouzet
et al, 2003).
Preliminary studies showed the heterogeneous dispersion of glass eels in the entire water
column. The objective was to collect information on glass eel density in the whole circulating volume at
a given moment at the study point.
Two sections were defined: a surface section and a deeper section and three horizontal transects
(right bank, middle and left bank).
In order to simultaneously sample the surface and the deeper section along a transect, a boat
equipped with two 0.65m diameter push nets was used. The surface scoop net was equipped with a
flow meter at the opening in order to measure filtration velocity and the deeper second one with a
Temperature, Depth and Salinity probe (“TPS”) (figure 7.11).
234

GPS
Flow meter
‘TPS’ probe
50Htz sounder
Figure 7.11. Diagram showing the position of the fishing gear on the boat during sampling
(source : Prouzet et al., 2003) (‘TPS’ probe measures temp, depth, salinity).
Sampling occurs at night during the incoming tide period identified as the migration period of
glass eels in open water. Five-minute trawls were undertaken alternately along the banks and in the
middle of the watercourse as shown in figure 7.12. Each run of the boat with the push nets was made in
a downstream direction against the current. Some twenty tows can be made during a tide which
amounts to about 8 runs across the whole width of the river.
Downstream
Aval
LB
RG
M
RB
RD
Surface
At depth
Fond
S1 S2 S3
S4 S5 S6
Left Rive bank
gauche Middle
Milieu Right Rive droite bank
Tidal Courant current
(incoming de marée (flot) tide)
Transect
Tamis poussés
Push nets
(points d’échantillon-
(sampling points)
-nage)
S1,.. Compartments
Compartiments
Figure 7.12. Diagram of the sampling protocol (adapted from de Bru and Lejeune, 2004).
235

For each sampling run, the following parameters were recorded:
• position in real time (geographical coordinates) supplied by the boat’s GPS by direct reading or via
the use of a route tracker;
• the weight of glass eels caught (in g) measured in the laboratory, using a precision scale after
drying;
• the filtration velocity calculated using a flow meter placed at the surface net aperture. The number of
revolutions is calculated for each measurement and expressed in the chosen unit using the formula:
flow _ fin − flow _ debut
vit _ surf ( m / s)
= (2.625*
+ 4) /100
durée _ flow
where, ‘flow_beginning’ (‘flow_début’) and ‘flow_end’ (‘flow_fin’) are the number of revolutions
made by the propeller of the flow meter at the beginning and at the end respectively of each
measurement; ‘duration_flow’ (‘durée_flow’) is the measurement period;
• three velocity measurements are made during each run for a period of thirty seconds each;
• the boat speed (in knots) called "speed over ground” is supplied by the boat’s GPS;
• the speeds in m/s can be obtained using the following formula:
vit _ fond(
m / s)
= vit _ fond(
Nd) *0.5144 ; (vit_fond = speed over ground SoG)
• this speed can be calculated using the data from the route tracker. Each transect is identified by its
duration and the distance covered using the geographical coordinates. It is given by:
vit _ courant(
m / s)
= vit _ surf − vit _ fond (vit_courant= current velocity; vit_surf= surface
velocity;vit_fond=velocity at depth) .
It corresponds to the velocity of the incoming tide (positive velocities when the tide is rising);
• the depth (in m) of the bed is collected in real time by the boat’s sounder. The depth at which the
deeper push net is operated is measured using the attached "TPS" (temperature-depth-salinity)
probe;
• the water temperature (°C) is recorded by the “TPS" (temperature –depth-salinity) probe placed on
the deep push net during sampling. The surface water temperature is recorded by a permanent
probe situated on the bank which stores time-referenced data;
• Turbidity (NTU) is recorded occasionally by a Doppler turbidimeter during incoming tide sampling
trips;
• the flow (m 3 /s), the tide coefficient, the lunar phase, other hydrological and climatic conditions are
recovered according to the guidelines defined in table 7.1.
236

(a)
(b)
(c)
(d)
(e)
Figure 7.13. Equipment used for scientific sampling and physical data collection: (a) GPS, route
tracker and sounder used during sampling ; (b) fishing gear (push net) used to
catch glass eels (diameter=60cm) ; (c) Doppler-effect turbidimeter (Anderaa) ; (d)
Mechanical (and electronic) flow meter used to record instantaneous velocity; (e)
“TPS” probe used to record water temperature and depth (photos : Ifremer).
237

Example of data and analysis following a sampling trip on the Adour:
Table 7.4 gives the values of the parameters collected during the trip undertaken in the Adour
estuary on the 8/12/2004:
Table 7.4. Environmental parameters collected during the scientific sampling trip of the 8/12/2004
(data From Ifremer).
Day
Tide
coefficient
River flow
(m 3 /s)
Lunar
phase
Mean
turbidity
Water
temperature
(°C)
Cloud cover
Hydrological
conditions
08/12/2004 58 104,5 NM 6 8 No barrier to
migration
Table 7.5. Data collected or estimated for the 15 runs (data from Ifremer).
Date Run Position pos GPS
time
Speed
over
ground
m/s
Surface
speed
m/s
Surface
weight (g)
Weight at
depth (g)
Time (s)
Surface
area of the
gear (m 2 )
08/12/2004 1 M 1 22:40:05 1.13 1.298 1.31 1.12 300 0.33
08/12/2004 2 RB 2 22:52:05 1.13 1.407 2.677 4.269 300 0.33
08/12/2004 3 LB 3 23:02:35 1.03 1.411 0.965 1.818 300 0.33
08/12/2004 4 M 1 23:11:50 1.03 1.496 0.679 5.586 300 0.33
08/12/2004 5 RB 2 23:20:45 0.98 1.496 1.231 2.715 300 0.33
08/12/2004 6 LB 3 23:29:55 0.98 1.583 1.141 1.431 300 0.33
08/12/2004 7 M 1 23:38:40 0.98 1.632 2.046 3.8 300 0.33
08/12/2004 8 RB 2 23:48:25 0.87 1.561 2.631 3.664 300 0.33
08/12/2004 9 LB 3 23:57:50 0.87 1.435 0 1.509 300 0.33
08/12/2004 10 M 1 00:07:25 0.93 1.695 2.899 11.54 300 0.33
08/12/2004 11 RB 2 00:16:35 0.87 1.587 1.47 5.084 300 0.33
08/12/2004 12 LB 3 00:27:10 0.98 1.544 1.529 3.25 300 0.33
08/12/2004 13 M 1 00:36:55 0.93 1.609 1.157 11.967 300 0.33
08/12/2004 14 RB 2 00:47:30 0.87 1.661 0.785 4.842 300 0.33
08/12/2004 15 LB 3 01:00:00 0.93 1.418 0.558 3.316 300 0.33
Calculations of filtered volume and glass eel density for each tow (example of tow 1):
Filtered volume=surf_sp*time*area_gear=1.298*300*0.33=128.5 m³
Surf.density=100*weight_surf/filtered volume=100*1.31/128.5=1.02 g/100m³
Density at depth=100*weight_depth/filtered volume=100*1.12/128.5=0.87 g/100m³
A simpler and more general acquisition protocol: the Minho estuary
The study zone is situated around 7.5 km from the mouth in the part of the Minho that is common
to both Spain and Portugal. Average width is 620m for an average depth of 3.7m. When the tide rises,
the variation in the water level is 0.6 m.
Three zones over the width of the river were selected (right bank, middle and left bank). On each
of these, 3 boats equipped with the same type of fishing gear, the “tela”, simultaneously sampled the
entire water column, covering a width of about 10m (figure 7.14).
238

FLOW
Figure 7.14. Photo and diagram showing the fishing gear used: the “tela” in the Minho estuary in
Portugal (photo courtesy: Antunes J.-C./CIMAR).
Samples were collected at night during the incoming tide. The fishing gear was kept in the water
column during the whole period of the incoming tide. Glass eels were regularly collected from the finemeshed
set net (1–2 mm), sorted on a mesh grid (4–5 mm) in order to discard by-catches, and then
stored in viviers.
Figure 7.15. Glass eel catch and by-catch inside the ‘tela’ (photo courtesy: Antunes J.C./CIMAR).
At the end of the study trip undertaken during the new moon, the following calculation elements
(table 7.6) had been obtained.
239

Table 7.6. Summary of calculation elements (data from CIMAR).
Position Catch (g) Average depth (m) Average speed (m/s) Duration (mn)
Left bank 280 2.25 0.23 135
Right bank 150 2.25 0.21 135
Middle 220 4.75 0.2 135
The graph of current during the rising tide (figure 7.16) shows velocities to be relatively stable
during the sampling period.
Current vitesse velocity du courant (in m/s) (en m/s)
0,3
0,25
0,2
0,15
0,1
0,05
0
0,00 20,00 40,00 60,00 80,00 100,00 120,00
Time temps (in min) (en mn) from par the rapport beginning au of début the rising de la tide marée
montante
Figure 7.16. Graph of the evolution of the current during the incoming tide in the Minho estuary
(data from CIMAR).
An average velocity was estimated in order to assess the volume of water filtered by each fishing
gear during the tide (table 7.7). Together with the catch recorded for each gear at the end of the tide,
this velocity enables densities to be calculated.
Table 7.7. Estimation of densities of glass eels collected in the net
Position
Instantaneous volume filtered by the gear
(m 3 /s)
Filtered
volume (m 3 )
(m 3 )
Density
(g/100m 3 )
Left bank 11.5 93,150 0.30
Right bank 10.5 85,050 0.18
Middle 10 81,000 0.27
Instantaneous volume filtered by the gear (m 3 /s) = 50× average velocity
Filtered volume (m 3 ) = instantaneous volume filtered by the gear x time x 60
Density (g/100m 3 ) = catches / filtered volume x 100
During the time the tide rises (135 minutes), the volume circulating through the section is
estimated at 17,010,000 m 3 .
240

7.3.1.2. Estimation by trapping and counting of glass eel
fluxes migrating in the estuary
Beforehand, it is necessary to clarify whether the counting concerns some, or all, of the migratory
flux, and what are the hydrological conditions affecting the proportion of the flux which is concerned by
the trapping. If trapping is not exhaustive (which is often the case), this counting should be combined
with mark-recapture and double-trapping techniques.
If the aim is to measure estuarine recruitment rather than fluvial recruitment (i.e. individuals
migrating beyond the limit of the dynamic tide) individuals must be counted on barriers located very low
down the migratory axis.
The counting technique is simple, but if it is to provide a representative indicator (absolute or
relative) of estuarine recruitment (or of fluvial recruitment when used further upstream 12 ) it requires that
certain basic hypotheses be fulfilled which is often far from the case.
How to assess whether basic hypotheses are valid
The first hypothesis is that all migrating individuals pass through the counting installation. If this is
not the case, the proportion of migrants that may avoid this structure must be estimated.
A first approach is to assess the degree of passability of the barrier. If it is impassable (case of
the Vilaine estuary dam, figure 7.17) regardless of the hydroclimatic conditions at least during the main
period of upstream migration, counting will indicate the abundance of individuals passing through the
barrier but not the abundance of glass eels arriving at the barrier especially if the second basic
hypothesis does not hold.
Downstream
Figure 7.17. Photo of the the Arzal – Camoël dam and 3D representation of the pass (source :
IAV, June 2006) 13
12 See Chapter 8.
Upstream
241

This second hypothesis is that migrating individuals do not accumulate before the barrier. This
raises the issue of how attractive and how effective the pass is. An initial assessment can be made by
comparing the distribution of glass eel/elver pigmentation stages below the barrier and inside the pass.
These data can be complemented by analysing the arrival of glass eel fluxes into the open estuary
either by monitoring catch per unit effort or by assessing daily fluxes.
How to assess estimation biases
The most commonly used solution is to mark a certain number of individuals and observe their
dispersion into an unmarked population. Multiple mark-recaptures may be used to find out what
happens to marked individuals which are sampled within the counting structure if the aim is to compare
the numbers counted in the pass with the number of individuals which overcome the barrier. This
method may also be used to find out what happens to individuals sampled below the barrier if the aim is
to assess the importance of accumulation in front of the barrier.
The principle behind this method is as follows:
• a random sample of glass eels is drawn from the population caught in the pass or downstream of
the barrier (if the numbers caught are too high to use the entire population);
• the animals caught are marked (M) then released into the environment where the total population is
to be assessed (either upstream or downstream);
• a second sample (C) is then drawn from the release zone and the number of marked animals (R)
and unmarked animals (C-R) are counted;
M * C
• the total population is estimated using a simple formula such as N =
R
, which with modification
of the Petersen equation becomes (Ricker, 1975)
V
*2
* N
( )
( C − R)
N =
( C + 1)( R + 2)
* ( M + 1)( C+
1)
N = ( R+
1)
, with variance
There are more sophisticated techniques to assess biases linked to an open population. These
are based on multiple mark-recaptures and on structuring the sampled population into size strata, if
biological stages other than glass eels are being considered. They use the general linear model applied
to demographic studies by evaluating the multiplication factor (Venables and Ripley, 2002).
13 Collective, 2006. Suivi des passes à poissons du barrage D’Arzal – Camoël (Vilaine, Morbihan) – Suivi de la passe à anguilles –
Données de la migration 2005, Institution d’amenagement de la Viliane, June 2006 - http://www.ifremer.fr/indicang/boite-bassinsversants/pdf/suivi-passe-anguilles-arzal-2005.pdf).
242

7.3.2. Descriptors relating to the relative abundance of a daily or
seasonal glass eel flux
These may be obtained from monitoring carried out during scientific or commercial fishing
operations. They are usually obtained by assessing the number or weight of catches made during a
given period of time using a defined fishing effort 14 .
This paragraph only covers the elements required to improve understanding of the environmental
and fisheries context which characterises the catch of individuals by a commercial fishing unit.
The methodological elements discussed will be linked to those presented previously on the
sampling of glass eel densities.
Catch estimated through the analysis of fishing logbooks at a given site varies as a function of
several variables, which are summarised in the following formula (following Laurec and Le Guen (1981)
or Gulland (1969)) : Capture = f × q × N (equation 1)
where f = fishing effort; q = catchability and N = population abundance.
7.3.2.1. Notion of catchability: (q)
Studies of glass eel behaviour 15 show the very significant effect of turbidity on their presence near
the surface in a zone where they are vulnerable to hand scoop nets or push nets which usually only
target the superficial layer of the water column.
Hence, catchability (q) can be defined as the product of 3 probabilities:
a
q = ac × × s
A
(equation 2)
where:
ac: is the accessibility i.e. the probability that individuals will be present on the fishing grounds
(main period between November and April in the central zone of the distribution area) or present
near the surface;
a : the probability that the area (or volume) where the fish can be found will be swept (or
A
filtered) by the fishing gear with a the area swept (or volume filtered) by a unit of effort and A the
total area (or volume) of the zone explored by the fleet ;
s : the vulnerability of the fish to the gear or the efficacy of the gear.
14 For further information on descriptors of this type, please refer to chapters 3 and 6.
15 See Chapter 2.
243

A fishing power can then be defined for each gear used, which, following Gulland (1969), is the
catch of the fishing gear per unit of fishing time for a given density of aquatic animals.
In the case of glass eel fisheries, this can be related to the gear’s filtration area, to the boat’s
power which can increase the filtration speed and thereby the volume explored, or else to various
equipment that allows the entire water column to be explored.
7.3.2.2. Notions of fishing effort (f)
According to Poinsard and Le Guen (1975) and the 1970 ICES Charlottenlund conference:
“the fishing effort applied to a stock of aquatic animals is a measure of the complete set of fishing
methods used by fishers on this stock, during a given time interval”.
It can be measured according to: the number of fishers, of fishing gears, of fishing hours during
one trip or more generally during one fishing season.
This is the case for nominal fishing effort, which is applied by fishers to one or more targeted
fish populations.
The nominal fishing effort, i.e. the various means used by the fisher to catch the fish, is different
from the effective fishing effort, which is the effort actually applied to catching the fish (generally the
fishing trip time is equal to the time taken to travel to the fishing ground plus the search time plus the
fishing time).
Finally, fishing intensity is used to reduce effort (a cumulative measurement across the zone
occupied by the exploited stock and over a period of time) to an instantaneous and more local
measurement. This can be extended, with or without weighting, to the whole fishing zone. It is then
called global fishing intensity and by simple summation approaches the nominal fishing effort.
It is therefore necessary, especially if the aim is to compare the abundance of several river
basins, to define clearly the environmental context of the fishery and the characteristics of the fishing
gear used. The same is true when studying the evolution of catch over time in a given river basin in
order to ensure that the fishing gear used does not vary, and that the fisheries have exploited zones
where elver behaviour has not varied with time (table 7.8).
244

Table 7.8. Descriptors relating to fishing effort and individuals’ catchability.
Criteria Descriptors Objectives Guidelines
Turbidity
Measure the probability that the
flux is close to the surface
Characteristics of
individuals’
catchability
Fishing effort
characteristics
Quality of collected
information
Water temperature
Lunar cycle
Obstacles to free circulation
Hydrodynamism
Size of the estuary
Fishing gear characteristics
(fishing gear – boat)
Existence or absence of a
declaration system or of a
sample-based monitoring
system
Identify the periods with a
barrier to migration
Identify the periods of optimal
catch
Check that there is no obstacle
to migratory individuals
Analyse propagation speed of
the dynamic tide
Consider the size of the fishing
gear used compared to the
width of the exploited estuary
• Measure of nominal effort:
number of fishers per type
of fishing gear
• Measure of effective effort,
including the notion of
filtered volume relative to
circulating volume
Measure the reliability of the
information and the ability to
derive reliable information on
abundance evolution
Compare water clarity to a threshold value
of around 30 to 50 NTU. Below, consider
that elvers may avoid the surface.
Probable migratory barrier below 5°C
Favourable period during the new moon
provided there is no hydrodynamic barrier
Presence of a dam impossible or difficult to
pass or alternatively significant water
diversion towards abstraction or storage
facilities.
Check whether the flux of glass eels has
become concentrated because the
propagation movement of the tide has
slowed.
Estimate the transversal part of the estuary
where there is no or little fishing activity.
• Number of fishers and fishing gears
• Characteristics of fishing gear used
• Characteristics of the boats used and
their pushing or trawling power
Analyse accurately the fisheries data
collection system, its biases and the
accuracy of data that it provides (daily,
monthly, yearly, identification of effective
fishing days ….)
Relationship between effective effort and fishing gear characteristics: application to the
elver fishery
Many fishing gears are used to catch elvers. Their size and complexity vary greatly. They can be
operated by hand from the bank or from an anchored boat, or alternatively pushed by the boat or
anchored onto the bed.
Some are fitted with a handle, others are not. Those which are not can either be worked from the
gunwale of the boat or pulled by cables at different depths.
Given the description of glass eel or elver behaviour and its analysis as a function of a number of
hydroclimatic variables, the efficacy of a certain type of fishing gear will depend on the water volume it
can filter and its capacity to function at various depths and hence to adapt to changes in glass eel or
elver behaviour within the water column.
245

Hand-operated scoop nets
These are generally used from the bank or alternatively from a stationary boat. In France, 2 sizes
are authorised: 0.50m wide for amateurs (filtration area 0.2m²) and 1.20m wide for professionals
(filtration area 2.26m²) with 2 to 3m-long handles. They are either circular or oval.
In Spain, in the Basque Country or in the Asturias for example, these same practices are found
with circular scoop nets fitted with a handle and operated from the bank or from a boat, with diameters
ranging from 0.6 to 1.4m.
In Portugal, on the river Minho, this type of scoop net is used together with a fish aggregating
device called a “Tela” 16 . This aggregates some of the fish migrating during the flood tide, which are then
removed from the net using a scoop net fitted with a handle several metres long (figure 7.16).
Apart from this latter case where the gear is coupled with a device that aggregates fish in one part
of the water column, the filtrating capacity of this fishing gear is limited and it is only really effective in
narrow watercourses or when the fish concentrate near the banks seeking calmer water during floods. It
is still frequently used in fluvial parts of estuaries but has generally been replaced by push-nets in the
maritime areas given the increasing scarcity of glass eels in recent years.
Pushed or towed nets.
These are of variable dimension. They use the boat’s power to be pushed if they are attached to
the gunwale or towed if they are operated by cables. The fishing power of the fishing boat with the
fishing gear will depend on the size of the net opening and the speed at which the gear is pushed or
towed, which depends on the boat’s power (restricted to 150 kW in estuaries). Given these two
parameters of area and push or tow speed, an estimate can be made of the filtrating capacity of the
fishing gear per unit of time. The efficacy of the fishing gear may be further improved if the net can be
operated at different depths. This is the case for nets which are fitted with handles or towed by cables.
They are generally circular with a regulated diameter (1.20m) although some derogations have been
granted in some French estuaries to allow oblong shapes and greater filtration areas (between 3.6 and
14m² in total for one or two fishing gears used per boat).
Examples of various kinds of fishing gear are shown in figure 7.18.
16 Coimbra J., Antunes J.C., Damasceno-Oliviera A., Dias S., 2005. Indicang – Relatório de Etapa – Bacia Hidrográfica do Minho,
CIIMAR, Porto, Portugal, http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/rapport-etape-bv-minho.pdf
246

a
b
c
d
Figure 7.18. Photos of various kinds of fishing gear used in the river basins of the INDICANG
network for elver fishing: (a) Adour push net; (b) Loire push net; (c) Oria push net
(Gipuzkoa); (d) Oria hand net (photos : a & b Ifremer ; c & d AZTI).
7.4. Data acquisition: definition of indicators.
7.4.1. Trend indicator using CPUE
These indicators are more fully discussed in chapter 6 which deals with fisheries descriptors.
A seasonal Catch Per Unit Effort series (ratio of total catch to number of trips) on the Adour from
1928 to 2003 is used as an example (figure 7.19).
247

30
25
CPUE (kg/fishing trip)
CPUE (Kg/sortie)
20
15
10
Max CPUE
Mean CPUE
Min CPUE
5
0
1927/1928
1929/1930
1966/1967
1968/1969
1970/1971
1972/1973
1974/1975
1979/1980
1984/1985
1986/1987
1988/1989
1990/1991
1992/1993
1994/1995
1996/1997
1998/1999
2000/2001
2002/2003
CPUE moyenne CPUE Mini CPUE Maxi
Figure 7.19. Evolution of mean CPUE in kg/fishing trip on the Adour using a hand scoop net
(IFREMER data).
The trend is defined from catch data using the same gear (the hand scoop net). It excludes
catches made with push nets from 1995 because the available fisheries statistics distinguish catch
made by different gears. The trend is clear and shows 2 broad periods: one before and one after 1980
characterised by very different elver densities, which have been very much lower since the 1990s
compared to what they were in the 1980s and before.
7.4.2. Indicator of daily abundance using density
7.4.2.1. Method developed in the Adour estuary and
adapted to the Loire, the Isle, the Oria and the Minho
estuaries.
Firstly, the water section must be divided (horizontally and vertically), based on the principle
(which usually holds) that there is a heterogeneous horizontal and vertical distribution of glass eels
behind the tidal front.
In order to calculate the circulating volume, the water section is divided as follows:
248

Depth
Watercourse width
(m)
0 50 100 150 200
0
-1
-2
1 2
3
-3
-4 4 5
6
-5
-6
-7
-8
Figure 7.20. Diagram of the different divisions of the relevant water section at the sampling
station on the Adour (after prouzet et al., 2003).
Figure 7.20 shows that the method introduces some degree of uncertainty concerning the
extrapolation coefficient between the recorded density in the 6 compartments of the sampling station
(figure 7.10) and the volume circulating within this compartment. Four horizontal and three vertical
divisions are chosen. Twelve different divisions are therefore included in the calculation of elver
abundance in each sector of the water column. The impact of this division on biomass estimation can be
tested.
Table 7.9 shows the calculated surface areas of the six sectors of the water column for these
twelve divisions.
Table 7.9. Calculated surface areas of the 6 sectors in the 12 divisions selected on the Adour
(source : Prouzet et al., 2003).
Sector
Divisions
1 2 3 4 5 6 7 8 9 10 11 12
1 70 60 50 40 105 90 75 60 140 120 100 80
2 70 90 110 130 105 135 165 195 140 180 220 260
3 70 60 50 40 105 90 75 60 140 120 100 80
4 420 360 300 240 350 300 250 200 280 240 200 160
5 420 540 660 780 350 400 550 650 280 360 440 520
6 420 360 300 240 350 300 250 200 280 240 200 160
Two hypotheses are very important if the calculations are to be valid:
• elver distribution is homogeneous within a sector;
• instantaneous speed is the same anywhere within a sector.
A sinusoidal model is used to estimate the velocity of the current so as to obtain continuous
values at each transect (figure 7.12) during the whole time the tide is rising.
249

π
ν ( t ) = c sin( ( t − b))
+ ε
a
ν ( vit _ courant ) : current velocity calculated for the transect i’ (i= ‘RD (right bank)’ , ‘M
-
i i i
i
i
-
i
(middle)’ , ‘RG (left bank)’) (m/s)
- a : duration of the incoming tide (s)
- c : maximum current velocity calculated over the period ‘a’ (m/s)
- b : time at which the tide begins
-ε : error term
• A first correction of densities can be made by modelling current velocities by transect and by
deriving a corrected surface velocity (filtration speed of the gear).
• The corrected density is found by applying a correction factor given by the corrected surface
velocity divided by the initial surface velocity. The following formula clarifies this calculation:
D
corrigée
vit _ surf
= D ×
vit _ surf
corrigée
initiale
vit _ fond + vit _ courant
= D ×
vit _ fond + vit _ courant
corrigée
initiale
The derivation of corrected current values leads therefore to an initial correction of observed
densities. In order to optimise estimates of the evolution of glass eel densities in the various sectors of
the water column during the tide, a relationship is established between current velocity and the density
of glass eels at a given moment. This relationship is defined as follows:
where:
- d : density (g/100 m³)
- ν : velocity (m/s)
-α and β : linear equation coefficients
ln( d
- i : index identifying a transect (i=1=RD (right bank) ; i=2=M (middle) ; i=3=RG (left bank)
- j : index of position inside the water column (j=1=surface; j=2=at depth)
-ε : error related to the model
ij
) = α ln( ν ) + β + ε
By using this relationship, the velocity and the glass eel density can be calculated at a given
moment for each sector of the water column.
The calculation of the daily biomass relies on the integration over time of densities and velocities
for the various sectors of the water column. The following general formula can be used:
1
B ⎤ × S
⎥⎦
N 6
flot _ fin
⎡
journ. = ∑∑ d
s
t
s
t
N
flot début
S
⎢⎣ ∫ ( ). ν ( )
_
1 = 1
B
A
sn
ij
i
ij
ij
250

where
- B : daily biomass of the glass eel flux
- d : density (g/m³)
- ν : velocity (m/s)
- S : area of the relevant sector (m²)
- s : index related to the sectors; s = [1:6]
- n : division index ; n = [1 ;N]
The duration of the incoming tide is calculated using the velocity curve of the modelled current.
Therefore, densities are calculated over this estimated period. For a given division, the daily biomass is
equal to the sum of the biomasses calculated for the six sectors (A). A biomass for the average night
tide is then calculated including all the chosen divisions (B).
These calculations use a mathematical programme “EstimJour” developed under the software
©SPLUS 6.1 17 . This programme was developed by the Aquitaine Fisheries Resource laboratory of
Ifremer in partnership with UPPA-Laboratory of Applied Mathematics and the UPMF-Labsad.
Example of calculation of glass eel biomass migrating in the Loire estuary during the
nocturnal incoming tide
The daily estimation calculated here is based on scientific samples collected in the Loire estuary
on the 28 th of March 2006. The samples were collected at night during the incoming tide between 1.40
am and 5.00 am in the "commune" of Pellerin (Kilometre Point=35 km) (figure 7.7). Twenty one fiveminute
transects were made during this period according to the protocol described above.
Table 7.10. Hydrological and weather data for the 28 March in the Loire (Ifremer data).
Day
Tide
coefficient
River flow
(m 3 /s)
Lunar
phase
Mean
turbidity
Water
temperature
(°C)
Cloud
cover
Hydrological
conditions
28/03/2006 96 1 850 NM 49 11.9 Low No barrier to
migration
The velocities and densities calculated for each transect are recorded in table 7.11.
17 Bru N., 2004. Programme de calculs des estimations journalières de biomass de civelles à partir de campagnes scientifiques
sur une saison de pêche, appendix 11 of the Indicang report, http://www.ifremer.fr/indicang.
251

Table 7.11. Density and velocity data by transect for sampling undertaken on 28 March 2006
(Ifremer data). (LB = left bank; M = middle; RB = right bank).
Run
number
Transversal
position
Time
(mn)
Speed
over
ground
(m/s)
Surface
speed (m/s)
Density at
depth
(g/100m³)
Surf density
(g/100m³)
1 RB 0 1.25 -0.29 0.5 1.2
2 M 10 1.24 -0.12 0.1 1.5
3 LB 19 1.32 0.11 1.0 4.1
4 RB 29 1.30 0.07 0.8 4.9
5 M 38 1.29 0.44 0.3 2.2
6 LB 47 1.23 0.53 0.7 5.5
7 RB 56 1.24 0.38 1.7 4.3
8 M 66 1.23 0.68 0.8 3.7
9 LB 75 1.25 0.59 2.4 5.9
10 RB 85 1.33 0.45 0.7 3.2
11 M 95 1.22 0.72 0.8 6.2
12 LB 105 1.30 0.35 1.5 4.5
13 RB 115 1.28 0.43 1.0 4.2
14 M 124 1.22 0.59 0.8 3.1
15 LB 135 1.29 0.27 1.2 3.0
16 RB 145 1.30 0.40 0.8 3.4
17 M 154 1.27 0.27 0.6 3.7
18 LB 163 1.33 0.01 1.4 3.3
19 RB 175 1.29 0.01 1.2 2.7
20 M 184 1.32 0.11 0.3 1.4
21 LB 192 1.30 -0.01 0.7 3.1
Twelve different divisions were chosen to calculate the surface areas of the six sectors (table 7.12).
Table 7.12. Surface areas (in m²) of the sectors for the selected divisions (Ifremer data).
100 80 70 60 200 160 140 120 300 240 210 180
100 140 160 180 200 280 320 360 300 420 480 540
100 80 70 60 200 160 140 120 300 240 210 180
800 640 560 480 700 560 490 420 600 480 420 360
800 1120 1280 1440 700 980 1120 1260 600 840 960 1080
800 640 560 480 700 560 490 420 600 480 420 360
The complete data set is then entered into the calculation programme 18 . The simplified output is
described below as an example (figures 7.21 and 7.22):
18 Bru N., 2004. Programme de calculs des estimations journalières de biomass de civelles à partir de campagnes scientifiques
sur une saison de pêche, appendix 11 of the Indicang report, http://www.ifremer.fr/indicang.
252

0.6
0.6
Velocity of the current measured in situ (in m/s)
0.4
0.2
0.0
Velocity of the current measured in situ (in m/s)
0.4
0.2
0.0
-0.2
-0.2
Right bank
Middle
Left bank
0 50 100 150
Time
0 50 100 150
Time
Figure 7.21. Velocities measured (in m/s) against time (in min) across the watercourse (data from
Ifremer).
.
o
ox
O
X
Measured velocities
Estimated velocities
0.6
0.6
x
x
o
0.6
o
Velocity of the current: Right Bank (in m/s)
0.4
0.2
0.0
x
o
o
x
x
o
x
o
o
x
x
o
Velocity of the current: Middle (in m/s)
0.4
0.2
0.0
x
o
x
x
o
o
x
Velocity of the current: Left Bank (in m/s)
0.4
0.2
0.0
o
x
o
x
x
x
o
x
o
x
o
o
x
o
x
-0.2
-0.2
-0.2
o
0 50 100 150
Time (in min)
0 50 100 150
Time (in min)
0 50 100 150
Time (in min)
Figure 7.22. Velocities (in m/s), measured and modelled, against time (in min) across the
watercourse (data from Ifremer).
253

Having validated the velocity data obtained by modelling, biomasses are estimated near the
surface and at depth (figure 7.23).
Total Biomasse Biomass totale (kg): :
273
38.26
0.14
Biomass at the
Biomasse
surface of
en
the
surface
water
(kg)
(kg):
:
150
18.31
0.12
Biomasse au at depth fond (kg):
123
28.13
0.23
Figure 7.23. Programme output screen showing estimates of the biomass migrating during the nocturnal
incoming tide, the standard error and the coefficient of variation for the tide of 28 th March
2006.
The accuracy of the estimates is strongly related to the measurement of the velocities. Thus, it is
important not to neglect observations of this variable made with the flow meter or, in the case of
important inaccuracies (twigs in the water hindering the propeller rotation for example), to use the
velocity calculated by a 1D or 2D hydrodynamic model.
Figure 7.24 shows the correlation between the observed data and data simulated by a curvilinear
1D model using precise bathymetry which, however, does not take into account transverse
bathymetrical heterogeneity (2D transverse model required).
Vite Velocity s s e
1
0.75
0.5
0.25
0
-0.25
M
M continuation M
RB Ms uite
RB continuation RD
LB RDsuite
LB continuation RG
Simulated RGsuite
Velocity VitS im
-0.5
-0.75
-1
10 20 30 40
Te mps Simulation de s imulation time (hr) (h)
Velocity
Vitesse
0.75
0.5
0.25
M
Ms uite
RD
RDsuite
RG
RGsuite
VitS im
Vitesse Velocity
0.75
0.5
0.25
0
M
Ms uite
RD
RDsuite
RG
RGsuite
VitS im
0
-0.25
-0.5
-0.25
18 20 22 24
TeSimulation mps de s imulation time (hr)(h)
-0.75
42 44 46 48 50
TeSimulation mps de s imulation time (hr)(h)
Figure 7.24. Comparing velocities observed in the transverse section with the velocity simulated by the
unidimensional hydrodynamic model : (M) middle; (RB) right bank; (LB) left bank. (from Boussouar and
Prouzet, 2007).
254

For this reason, it is recommended, if possible, to position the sampling station in a linear
canalised section with regular bed morphology.
Method applied on the Minho
The mean estimate of the density is 0.25g/100 m 3 (minimum: 0.18 g/100 m 3 and maximum:
0.30 g/100 m 3 ). Extrapolating to the circulating volume (using the river section at the level of the
sampling station) during a rising tide gives an initial estimate of some 43 kg. This approximation may be
greatly improved by better definition of the hydrodynamic process in the estuary and by more
widespread sampling during the rising tide.
7.4.3. Indicator of seasonal abundance using densities and
commercial catches: the Adour example
The method used to estimate the seasonal flux of glass eels entering the estuarine area,
developed on the Adour by the Aquitaine Fisheries Resources Laboratory of Ifremer, in partnership with
UPPA-LMA and the LabSad (Bru and Lejeune, 2004 ; Bru, Lejeune and Prouzet, 2004), was applied on
the Adour to the series of catches and environmental variables collected since 1998.
This quantification of the seasonal glass eel biomass is based on a generalised linear regression
statistical model. This model takes into account on the one hand, the catches of commercial fishers and
on the other hand, data concerning environmental variables shown by previous studies to have an
impact on glass eel migratory behaviour (Prouzet et al., 2003) 19 .
The biomass and its precision may be obtained in two ways: by using a method based on the
simulation of a set of trajectories derived from a bootstrap-type approach 20 and by using a purely
analytical method. It is the latter method that is summarised below 21 .
The first stage consists of a series of surveys to estimate the biomass of a flux migrating during
a nocturnal incoming tide. Then the catches taken from this migratory flux are identified, either directly
(catches made close to the sampled zone) or by using simulation tools to estimate the time taken by
glass eels to migrate from the sampling zone to the fishing zone (or vice versa).
Once the relationship between catches and the target glass eel flux is established, it becomes
possible to define a fishing exploitation rate, which is simply the volume of glass eels extracted from the
19 Chapter 3 gives further information on the nature and the effects of these variables.
20 Bouvet et al., 2006. Indicang – Quantitfication de la biomasse saisonnière de civelles (Anguilla anguilla) dans l’estuaire de
l’Adour et estimation du taux d’exploitation saisonniere de la pêche professionnelle au tamis poussé, Ifremer,
http://www.ifremer.fr/indicang/sites-thematiques/pdf/flux-saison.pdf.
21 A comprehensive explanation of the method can be found in the following document: Bru et al., 2006. . Indicang –
Quantitfication par une methode analytique de la biomasse saisonnière de civelles (Anguilla anguilla) dans l’estuaire de l’Adour et
estimation du taux d’exploitation saisonniere de la pêche professionnelle de 1998 à 2005, Ifremer,
http://www.ifremer.fr/indicang/sites-thematiques/pdf/estim-civelle-1998-2005.pdf
255

flux by fishing, with the exploitation rate for a given day j being:
taken from the flux on day j and
ΔC
ΔB
j
k
with
Δ Bk
: biomass estimated on day k (where k=or ≠ j).
Δ C
j
: volume of catches
A database can thus be developed including the estimated biomass of the glass eel flux, which
corresponds to the catches made, the exploitation rate and the environmental variables that affect glass
eel behaviour and hence their catchability.
These variables can be coded or raw, something which can only be decided after an exploratory
data analysis.
A minimum of 30 scientific surveys is estimated to be necessary to build the prediction model
relating the level and the variation of the exploitation rate (prediction variables) to the commercial catch
and to selected environmental variables (explanatory variables) (table 7.13). The complete validated
database is given in appendix 12 of the Indicang report 22 .
Table 7.13. Outline of the database necessary to build the prediction model (Ifremer data).
Season
Date
Daily
biomass
(kg)
Exploitation
rate (%)
Comm.
catch
(kg)
Index
(catchability)
Tide
coefficient Turbidity(NTU)
Flow
(m 3 /s)
Moon
Coded
Coded
tide
turbidity
coefficient
Coded
flow
S9899 12/01/99 138 4.35 6 1.79 40 32.2 403.5 NM 1 2 3
S9899 14/01/99 412 1.99 8.2 2.10 57 12.7 377.8 NM 1 1 2
S9899 22/01/99 157 7.13 11.2 2.42 81 19.1 270.4 FQ 2 2 2
S9899 28/01/99 42 10.71 4.5 1.50 66 29.8 745.9 FM 2 2 3
S9899 11/02/99 79 5.82 4.6 1.53 38 21.1 428.8 NM 1 2 3
,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,,
S0405 14/12/04 198 8.89 17.6 2.87 93 9.0 94.55 NM 3 1 1
Following the descriptive analysis of the data in table 7.13 and a study of the various possible
combinations of explanatory variables (table 7.14), the selected model is characterised by the following
variables and shows the significant impact of turbidity on the exploitation rate:
texp ~ turbcod + capture 0.678 + mareecod + debcod:lune + capture:lune (equation 3)
where:
« texp » : exploitation rate
« capture » : commercial catch in kilos
« mareecod » : coded tide coefficient
« turbcod » : coded water turbidity
« capture :lune » : interaction parameter between the ‘catch’ and the ‘moon’ variables
« debcod :lune » : interaction parameter between the ‘coded flow’ and the ‘moon’ variables
22 Ifremer, 2005. Base de civelle Adour : tableau de données des campagnes d’échantillonnages réalisées en zone maritime sur
l’Adour au cours des saisons de pêche comprise entre 1998 et 2005, avec estimation des taux d’exploitation, annex 12 0of the
Indicang report, http://ww.ifremer.fr/indicang.
256

The “capture” exponent is calculated from the linear relationship between the natural logarithms
of the “texp” and “capture” variables.
The selected coding is the following:
- The tide is coded into three classes: 1 if x

Figure 7.25 shows the relationship between observed and predicted exploitation rates and their
dispersion around the 45° line, which is when there is a perfect correlation between the two.
The coefficient of determination R 2 , which measures the proportion of explained variance, is high,
R 2 =SSq(regression)/SSq(total) =0.90, meaning that 90% of the total variance is explained by this model
(r = 0.95). The size of the coefficient of determination is not surprising as catch appears on both sides of
the equation. However, what is interesting is the significant proportion of the variance explained by the
“turbidity” factor. This shows that, in a clear estuary (low average turbidity), the water column opacity
has a significant influence on the presence of glass eels near the water surface and hence on their
vulnerability to fishing gears that only explore depths between the surface and a depth of 2 metres.
The second stage consists of using calculations based on daily information concerning catch
and relevant environmental variables (table 7.15). The application of the model then gives estimates of
the daily exploitation rates of the glass eel fishery operating at night during incoming tides.
Table 7.15. Database used to predict exploitation rates during a nocturnal tide with the predictive
model (equation 3) (from Ifremer data).
Date
Catch
Catchability
index
Coef Turbidity Flow Moon Coded
flow
Coded
turbidity
Coded tide
01/11/2004 0 64 6.7 158.48 FM 1 1 2
02/11/2004 0.2 -1.6094 52 6.5 140.14 LQ: 1 1 1
03/11/2004 0.2 -1.6094 41 6.4 128.128 LQ: 1 1 1
04/11/2004 9.3 2.23 31 6.4 111.917 LQ: 1 1 1
05/11/2004 0 28 6.5 105.206 LQ: 1 1 1
06/11/2004 7.1 1.9601 29 6.6 100.193 LQ: 1 1 1
,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, ,,,
30/03/2005 0 77 9.3 245.822 LQ: 1 1 2
31/03/2005 63 8.8 235.22 LQ: 1 1 2
Given daily catches and the corresponding exploitation rates, it is possible to estimate the
migrating biomass during a nocturnal tide and its fluctuations. After smoothing and correction, a series
of values may be derived for the abundance of the nocturnally-migrating glass eel flux over the relevant
fishing season (figure 7.26).
The seasonal biomass can be estimated by summing daily biomasses estimated during nocturnal
tides using either raw or smoothed data.
258

Estimated smoothed biomass (kg)
4 00
3 00
2 00
1 00
0
No
v
De Ja Fe
1999 - 2000
Figure 7.26. Smoothed series of glass eel biomasses migrating during nocturnal tides during the
1999-2000 fishing season on the Adour (source : Bru et al., 2004).
7.4.3.1. Estimation based on the raw series of nocturnal
biomasses
This first method consists of summing biomasses estimated exclusively on days when push net
catches are recorded.
n
∑
B sais
= Bˆ
i
ˆ
i =1,…,n : days of commercial landings over the season
i=
1
7.4.3.2. Estimation based on the smoothed series of
nocturnal biomasses
Here the estimation is based on the most extensive series of daily biomasses by summing the
smoothed values of daily biomasses 23 . This gives the seasonal biomass estimated using all the days in
the season (from the first to the last day of effective fishing):
n
∑
B ˆ = Bˆ
( lissée)
(lissée=smoothed) n : number of days in the season
sais
i=
1
i
23 Please refer to the Indicang report for further information on the smoothing method and the size of the smoothing window used,
which depends on glass eel migration speed and the time they take to cross the fishing zone. Bru et al., 2006. . Indicang –
Quantitfication par une methode analytique de la biomasse saisonnière de civelles (Anguilla anguilla) dans l’estuaire de l’Adour et
estimation du taux d’exploitation saisonniere de la pêche professionnelle de 1998 à 2005, Ifremer,
http://www.ifremer.fr/indicang/sites-thematiques/pdf/estim-civelle-1998-2005.pdf
259

A 95% confidence interval may be calculated for both of the seasonal biomass estimation
methods (Bru et al., 2004).
The general equation is given by:
V
⎛ n ⎞ n
( ) ∑ ( ) ∑ ⎜
⎛
) )
⎟
⎞
n
)
Bˆ
= V ⎜ ∑= Bˆ
⎟ = V Bˆ
+ 2 cov B , B = V ( Bˆ
) + 2 × 0.18×
V ( B) × ( n −1)
= < ⎠
sais
⎝i
1
i
⎠
i
1
i
i
j
⎝
i
j
∑
i = 1
i
where “n” is the number of days in the season when a daily biomass
Hence the confidence interval is: IC ˆ (95%) = Bˆ
− 2σ ( Bˆ ); Bˆ
+ 2σ
( Bˆ
)
B sais
Bˆ i
could be estimated.
Figure 7.27 shows the series of calculations required to estimate the seasonal biomass using the
statistical model that has been developed, and its application to a database of daily catches which also
includes relevant environmental characteristics.
260

Nocturnal Les captures professional professionnelles catch de in the nuit main
sur la
pêcherie fishery principale (LM) and (LM) its relevant et la biomasse nocturnal
de nuit
biomass qui lui correspond at the sampling à la station station d’échantillonnage
which is
based permettent the calculation de calculer of un the taux exploitation d’exploitation. rate.
Celui The ci est latter mis is en related relation to avec catches les captures and
et
environmental les variables environnementales variables which affect qui influent fishing
sur
l’intensité intensity
du prélèvement
Statistical
modélisation
statistique
Modelling
Définition Define an d’une equation équation linking liant the le
taux exploitation d’exploitation rate with aux nocturnal captures
catch
de
and
nuit
hydraulic
et aux conditions
conditions (tide
and flow)
hydrauliques (marée et débit)
3000
100
2000
50
1000
0
0 67.5
0
135
débit Flow en in m3/s m³/s
captures Daily catch journalières in kg
en kg
3000
2000
0.2
1000
0
0 67.5
0
135
Flow débit en in m3/s m³/s
Mean exploitation rate in %
Texpl moyen en %
150
100
75
50
0
0 67.5
0
135
Tide coefficient coefficient
de marée
Daily captures catch journalières in kg
en kg
150
100
0.2
50
0
0 67.5
0
135
Tide coefficient coefficient
de marée
Mean Texpl moyen exploitation en % rate in %
Base de données saisonnières
Seasonal sur les captures database journalières concerning au the
daily lot catch LM et at les LM conditions and hydraulic
conditions
hydrauliques
extrapolation
Estimation Estimation of nocturnal de la biomasse migratory
migrant biomass saisonnièrement by adding up nocturnal de nuit
par cumul de
biomasses
biomasses de nuit
Figure 7.27. Series of calculations and data necessary to estimate the nocturnal migratory biomass in an estuary (here the Loire) during
the fishing season (after Prouzet et al., 2008).
261

The seasonal biomass, which represents the lion’s share of the estuarine recruitment (the fishing
season being synchronous with the main migratory period), is then estimated by doubling the quantity of
glass eels estimated at night (based on the plausible hypothesis that the intensity of glass eel entry is
the same during the day as during the night).
7.4.3.3. Some estimation examples.
Further information on the calculations can be found on the Indicang site 24 .
A 5 kg catch threshold constraint was set on the Adour and the Loire. Below this threshold, the
nocturnal migratory biomass was considered to be equal to the catch which of course minimises the
biomass estimated over a season and is equivalent to an exploitation rate equal to 1.
Glass eel recruitment was then estimated to be twice the nocturnal biomass estimation and
corrected, on days when algorithmic constraints applied (catch below the minimum 5 kg threshold or
hydrodynamic conditions that were incompatible with the defined behavioural model 25 ), by applying an
average or median exploitation rate estimated from unconstrained days.
The Adour (from Bouvet et al, 2006).
The methods to estimate the abundance of glass eels moving up an estuary on each tide and
during the fishing season were perfected on this estuary. The tools allow the volume of glass eels to be
estimated taking into account a certain number of constraints, for example those related to the daily
catch declarations or to a high flow regime preventing or slowing tidal progression into the estuary.
This work makes it possible to estimate a series of estuarine recruitments, which as already
noted, are approximately double the glass eel biomass entering the estuary on each nocturnal tide,
mainly during the fishing season. The precision of these estimations will be greater, the greater is the
number of days processed (in compliance with the algorithmic constraints defined above) compared to
the number of days taken into account. Two corrected estimations are provided: BIOCOR1 (using, for
excluded days, the average exploitation rates estimated for processed days) and BIOCOR2 (using
median exploitation rates).
Table 7.16 shows that the nocturnal biomasses of glass eels migrating into the Adour estuary,
after the confluence with the Nive, are from 3 tonnes (2002-2003 fishing season) to 66 tonnes (1999-
2000 fishing season). The 3-tonne estimate is questionable as the number of days calculated using the
seasonal estimation model is low (8 processed days out of 131 possible days i.e. 6%). On the other
hand, the 66-tonne estimate for the 1999-2000 fishing season is probably more robust given the
24 http://www.ifremer.fr/indicang/.
262

percentage of processed days (72.7%). The exploitation by marine fishers can be estimated therefore to
be between 2% and 6.6% for a recruitment ranging from 6.5 tonnes (2002-2003 season) to some 130
tonnes (1999-2000 season). River catches must be added and, in recent years, they have been slightly
higher than marine catches in the Adour basin. During the 7 years when it was monitored, the
commercial marine and river exploitation rate can be estimated to have ranged between 5% and 15% of
the migratory glass eel flux (except for 2004-2005 when the commercial river fishers' catch was much
higher than their marine counterparts at around 4 tonnes 26 which means a 12% exploitation rate for this
season, i.e. within the estimated 5% to 15% range).
Table 7.16. Corrected estimates of estuarine recruitment and the global exploitation rate of
marine fishers in the Adour basin for fishing seasons from 1998-1999 to 2004-2005
(from Ifremer data).
Fishing
seasons
Estimated
estuarine
recruitment
(in tonnes)
Allocated
catches
in kg
BIOCOR1
(in tonnes)
BIOCOR2
(in tonnes)
Estimated
exploitatio
n rate
Global
exploitation
rate
BIOCOR1
Global
exploitation
rate
BIOCOR2
1998 – 1999 40.0 1655 43.8 48.7 4.1% 3.8% 3.4%
1999 – 2000 127.7 4579 129.5 133.7 3.6% 3.5% 3.4%
2000 – 2001 29.8 1446 33.2 37.8 4.9% 4.3% 3.8%
2001 - 2002 40.6 770 39.4 40.8 1.9% 1.9% 1.9%
2002 – 2003 3.5 388 6.5 6.6 11.1% 6% 5.9%
2003 – 2004 14.8 1093 16.5 18.9 7.4% 6.6% 5.8%
2004 - 2005 43.1 1398 43.9 44.3 3.2% 3.2% 3.1%
The Loire (from Prouzet et al., 2008).
The method perfected on the Adour was directly adapted to this estuary.
The underlying hypotheses are as follows:
• It is estimated that 2 fluxes a day arrive at the mouth and that the calculated nocturnal biomass only
represents about half of the numbers entering the estuary. Catches from the downstream fishing
zone 23E7LO are added to those made in fishing zone 23E7LM in order to estimate the biomass. It
is assumed that there is no mortality between the 2 adjacent zones. Figure 7.28 shows the fishing
zones. The reference station used to estimate the entering flux is situated upstream of the 23E7L5
zone (figure 7.7).
• A first observation may be made: the under-estimate of estuarine recruitment will worsen as the
number of days processed by the calculation programme declines (imposition of the algorithmic
25 See Chapter 2.
26 Collective, 2005. Caractérisation des captures de la pêcherie aux engins sur le domaine fluvial de l’Adour et des côtiers landais,
Migradour, Pau, France, http://w3.ifremer.fr/indicang/boite-bassins-versants/pdf/bilan-relais-adour-snpe-2004.pdf.
263

constraints relating to the daily minimum catch volume and the threshold flow allowing tidal
progression).
The global exploitation rate is defined as the ratio of catches declared in the estuary to the
estuarine recruitment (or global flux over 2 tides). It was higher during the 2003-2004 season. This was
due to the number of days processed during this fishing season, which was much lower than in other
seasons (58% against 75% and 65% respectively during the 2004-2005 and 2005-2006 seasons – table
7.17).
Fishing saison de season pêche %LO %LM %L5 %L4 %L3 Catch captures allocated allouées
2003 - 2004 40.2 45.4 6.9 3.6 3.9 14 515 kg
2004 - 2005 17 52 6 6 13 7 146 kg
2005 - 2006 29 50 8 6 7 9 512 kg
Figure 7.28. Distribution of glass eel landings in the Loire estuary (source: Harmonie DPMA-
IFREMER database) for 3 recent fishing seasons.
264

Table 7.17. Estimation of biomasses and exploitation rates in the Loire for fishing seasons from
2003 2004 to 2005-2006 (source : Prouzet et al., 2008).
Zones
Captures Allocated
de nuit
Biomasse Estimated
Global Flux global flux
Tau Mean
nocturnal allouées catch
Estimée nocturnal de biomass
nuit (extrapolation to aux the
2
d’exploitation
rate
2 marées) tides)
moyen
Lot LM
6,692 tonnes
24,69 tonnes
16,4%
2003 - 2004
[23,9 – 26,0]
[14,2% et 18,4%]
Lot LM
2004 - 2005
3,716 tonnes
26,03 tonnes
[25,6 – 27,0]
12,5%
[11,5% et 13,5%]
Lot LM
2005 -2006
4,780 tonnes
25,7 tonnes
[24,7 – 26,9]
14,9%
[13,1% et 16,6%]
Estuaire Estuary
2003 - 2004
14,515 tonnes
= 2*24,69+5,834
55,21 tonnes
Global Texp
exploitation
global
rate 26,3% 26.3%
Estuaire Estuary
2004 -2005
7,146 tonnes
=2*26,03+1,181
53,24 tonnes
Texp global
Global exploitation
13,4%
rate 13.4%
Estuaire Estuary
2005 -2006
9,512 tonnes
=2*25,7+2,729
54,12 tonnes
Global Texpexploitation
global
rate 17.5%
17,5%
In order to rectify incomplete recruitment estimations, the mean and median exploitation rates
estimated for the 27E3LM fishery were used for the days when the algorithmic constraints were
imposed. Hence, 2 corrected estimations are given in table 7.18: one called BIOCOR1 (which uses the
mean exploitation rate) and the other one called BIOCOR2 (which uses the median exploitation rate).
Table 7.18. Corrected estimates of the Loire estuarine recruitment and global exploitation rate
for the 2003-2004, 2004-2005 and 2005-2006 fishing seasons.
Fishing season
Estimated
estuarine
recruitment (in
tonnes)
BIOCOR1
BIOCOR1 (in
tonnes)
BIOCOR2
BIOCOR2 (in
tonnes)
Estimated
global
exploitation
rate
Global
exploitation
rate
BIOCOR1
Global
exploitation rate
BIOCOR2
2003-2004 55.2 76.6 103.6 26.3% 18.9% 14%
2004-2005 53.2 52.6 52.9 13.4% 13.6% 13.5%
2005-2006 54.1 61.1 67.9 17.5% 15.6% 14%
Table 7.18 shows that estuarine recruitment was much higher during the 2003-2004 season than
in the following two seasons after correction of the values. The exploitation rate appears to be between
around 14% to 19% for these 3 seasons if the corrected estimates calculated from mean or median
exploitation rates are used, with estuarine recruitments during the fishing season between some 53 and
103 tonnes.
265

The Isle (from Duquesne, 2007 and Susperregui et al., 2007).
The fishing activity on this tributary of the Dordogne, the mouth of which is situated in the
propagation zone of the dynamic tide, is not continuous, requiring therefore an adaptation of the
estimation method. Information on fishing on the Dordogne was used to complete the otherwise
fragmented series.
Figure 7.29 outlines the series of calculations undertaken. The biomass, by nocturnal tide, is
estimated using the same general sampling protocol as perfected on the Adour and the Loire, but the
way in which biomass, catches and hydrodynamic conditions are related differs. First, a statistical model
is established between the CPUE observed on the Dordogne downstream of the confluence with the
Isle (CPUE.D) and the nocturnal biomass (BNE) estimated during scientific surveys undertaken on the
Isle on the same day, taking into account the flow (which affects glass eel entry into the tributary and
fisher behaviour). This relationship is then used to estimate a series of nocturnal biomasses smoothed
over a 3-day window (the time taken on average for glass eels to travel the distance between the Isle
and Dordogne confluence and the fishery further upstream). The diurnal biomass is then added to the
nocturnal biomass (the former being estimated as the mean of the previous and following days’
nocturnal biomasses). Hence a series of diurnal biomasses is obtained, making it possible to determine
the seasonal biomass or the biomass over a defined period of time (between 2 floods for example). It is
then simple to derive the exploitation rate of the push-net fishery on the Isle.
266

‘CPUE Dordogne’
signal
Dordogne
fishery
Estimation of nocturnal
biomass entering the Isle
(ENB)
ENB = f (CPUE.D, flow)
Modelling + smoothing
Contraints
Days without CPUE.D and flows >120m³/s
Saturdays & Sundays withdrawn
Sampled
Zone
Mean of (d+1) and (d-1)
D o r d o g n e
Isle
‘Isle catch’
signal
Biomass
estimation
Isle fishery
Exploitation
rate
Figure 7.29. Diagram showing the principle underlying the calculation of the glass eel biomass
entering the Isle during the fishing season (from Duquesne, 2007 and Susperregui
et al, 2007).
Table 7.19 shows that the Isle recruitment seems to fluctuate significantly and varies from 3.2
tonnes during the 2000-2001 fishing season to 17.5 tonnes during that of 1996-1997. On this Dordogne
tributary, fishing is frequent when the hydroclimatic conditions are favourable. This explains the
significant variations in exploitation rates between years ranging from 33.2% in 1998-1999 to less than
1% in 2000-2001.
267

Table 7.19. Summary of estimated seasonal biomasses of glass eels moving up the Isle and of
the exploitation rates achieved.
Season
Number of days Global seasonal Estimated seasonal Corresponding
include
exploitation
biomass
Isle catch(kg)
1996/199 10 26,5 17475,4 4646,6
1997/199 5 7,0 6928,2 396,1
Period 1 12,9 1049,1 136,0 5
Period 7 4,5 843,4 038,7
Period 8 9,9 443,8 44,2
Period 1 3,4 1938,5 65,8
Period 9 4,2 2653,3 114,4
1998/199 7 33,2 9944,8 3215,7 0
Period 3 43,4 4556,3 1977,4
Period 3 22,9 5388,5 1238,2
1999/200 10 19,5 9619,8 1647,7
Period 2 23,2 1618,1 375,6
Period 8 15,9 8001,7 1272,0 5
2000/200 5 0,7 3173,5 34,3
Period 4 1,2 2577,6 33,1
Period 1 0,2 595,9 1,2
2001/200 9 13,0 13745,5 2064,2 0
Period 4 20,8 6887,9 1436,3
Period 1 9,4 3000,6 283,2
Period 2 8,9 3856,8 344,6 0
2002/200 5 3,1 5411,7 171,4 6
2003/200 11 6,5 8538,9 595,0 1
Period 4 5,1 2963,9 153,31
Period 7 7,9 5574,9 441,6 7
2004/200 12 15,1 12369,8 1866,9 4
Period 5 18,2 6035,8 1103,1
Period 6 12,0 6334,4 763,7
2005/200 6 5,9 4930,0 356,8 4
Period 2 3,3 1207,4 939,9
Period 4 8,5 3722,6 316,9
2006/200 5 1,3 10318,9 166,2 7
Period 4 1,6 9028,1 152,8 9
Period 1 1,0 1290,7 513,4
7.5. Synopsis of indicators measuring relative and absolute abundance
and fishing intensity
7.5.1. Relative or absolute abundance indicators
The methods suggested and described in this chapter are diverse and of varying complexity
(table 7.20). Some, such as counting or catch per unit effort monitoring, are relatively simple to
implement but demand that certain hypotheses be fulfilled, which are sometimes very difficult to fulfil
(constant catchability, constant attraction of counting structures, reliability of trapping installations …).
268

Others are based on sampling a known water volume during the upstream journey of elvers. They are
more demanding but give a better description of the variability of individuals’ catchability and a better
calibration of fishing gear efficiency. Finally, it must be noted that extrapolating abundance to all of the
main upstream migration season is only possible if trapping and fishing efforts are frequent. If there
were no commercial fishery, it is clear that estimating such indicators of the absolute abundance of a
seasonal flux of glass eels would be particularly costly.
Table 7.20. Summary of abundance indicators and their interpretation.
Criteria Indicators Descriptors needed for
calculation
Relative CPUE from • Total catches
logbooks
• Nominal or effective
effort characteristics
Absolute
CPUE from
scientific surveys
Daily or seasonal
abundance
Total counting or
counting by unit of
volume
• Scientific catches.
• Characteristics of
the filtered volume.
• Density in the
estuary
• Circulating volume
during incoming tide.
• Equation linking
catchability to
exploitation rate
Number counted in a
trapping or water
abstraction installation
and identification of the
passing or shunted
volume
Expression
Catch/f = q.N
Density of
individuals
Density x
circulating volume
or
B= Catch/ ( E.
Δ t)
Number or density
extrapolated to the
explored volume
Interpretation
Abundance trend indicator over one season, provided
that catchability remains constant and the effort unit is
comparable from one period to another.
Indicator used to measure the importance of the flux
by extrapolation to a circulating volume, provided that
it is verified that the sampling takes into account the
transverse and vertical heterogeneous distribution of
migratory glass eels
• Daily abundance estimated by combining
sampled densities and water volume passing
through the sampling station during the incoming
tide.
• Seasonal abundance can be obtained by linking
daily exploitation rate ( E.
Δt)
modulated by
environmental conditions which affect the
catchability of the moving glass eel flux and
consequently their catch or else by linking CPUE
and the biomass of glass eels migrating during
one chosen unit of time (for example the tide).
• Simple in principle as long as the installation is
placed low enough on the watercourse and if the
dam can only be passed through the fish pass,
the efficacy of which has to be tested.
• Delicate if the flux can take another pathway
(water abstraction of an electrical power station
on a by-pass branch for example) or if the dam
can more or less be passed when the water
reaches a certain level or when the edges are
flooded. In these conditions, double trapping with
marking is essential but its interpretation is
difficult
7.5.2. Indicators measuring fishing intensity
Table 7.21 summarises how to define indicators which measure the impact of extraction for a
particular use on the migratory flux. This mainly concerns fishing but it could also relate to a pumping
station for cooling electrical power plants for example. A rate of extraction, i.e. an instantaneous
exploitation rate (E), is defined over a given time interval ( Δ t)
.
269

Table 7.21. Synopsis of indicators measuring the extraction intensity due to an activity such as
fishing or pumping.
Indicator Descriptors needed for its calculation Interpretation
Rate of extraction • Abundance of exploited stock and catch or other
( E.
Δ t)
extraction from it must be known
• For a given day d, the exploitation rate will
• The natural mortality coefficient (M) is not used as the
ΔC
exploitation rate is assessed over a short period of
j
be approximated using: .
time during which M is considered to be insignificant
ΔB
(case of one tide for example)
j
• Global or partial extraction rate (daily or
seasonal %)
Instantaneous
fishing mortality
coefficient (F)
or instantaneous
exploitation rate (E)
• This instantaneous value requires for a given period of
time the volume of the catch and the abundance of the
exploited population for the same period. It is
combined with 2 other coefficients M: the natural
mortality coefficient and Z: the total mortality coefficient
such that Z=F+M
Effective effort • Volume extracted by fishing or water abstraction /
volume where the exploited population can be found
• Effective effort can be estimated using
A
a
or more
roughly the total number of trips if their mean
extraction capacity has been calibrated
() t
() t
dC
F () t = 1 × and E =
N dt
F
Z
Directly derived from the catch equation (equation
1) it compares the volume explored by a fishing
gear (a) with the volume occupied by the
population (A)
Of interest when extrapolating the exploitation
rate from one river to another
This extraction rate can be measured directly or approximated by measuring the effective fishing
effort calibrated according to the filtered volume or more generally according to the number of trips.
Figure 7.30 gives, for the example of the Isle, the relationship between the extraction rate (or
exploitation rate over a defined period) and the total number of fishing trips undertaken on this
Dordogne tributary.
The equation relating the seasonal exploitation rate to the number of fishing trips undertaken by
the “elver fleet” operating on the Isle during a fishing season is given by:
T exp l ≈ 0,0117( σ = 0,0016) Sorties − 4,961( σ = 2,704) which explains 85.13% of the variance.
This enables the manager to define a maximum fishing effort in order not to exceed a chosen
exploitation rate over the fishing season. For example, what is the maximum total effort in order not to
exceed a 20% exploitation rate? Answer: 2,133 fishing trips over 81 fishing days on average, i.e. around
26 fishers per tide over an average season.
270

Exploitation rate in %
Fishing trips
Figure 7.30. Relationship between the global exploitation rate and the number of fishing trips
undertaken during the fishing season on the Isle.
271

Chapter 8
Colonisation and sedentarisation
indicators
Pascal Laffaille, Christian Rigaud
272

8.1. Context and objective
Whether in order to establish initial diagnoses or to assess the impacts of management
measures, monitoring the yellow eel stage can provide important information. In particular, such
monitoring is likely to provide information a posteriori on the colonisation phase process and to give
indications a priori concerning the silvering and downstream migration phases 1 .
The relative evolution of the species’ presence can also be monitored within the relevant
system (at spatial and temporal scales) and this trend can be compared to those of other areas. The
data collected can also contribute to identifying major dysfunctions within each hydrographic unit,
catchment or sub-catchment.
The design of the monitoring arrangements and the analysis of the collected data must integrate
three important factors:
• the behavioural diversity of yellow eels;
• the diversity of the colonised or colonisable aquatic compartments (estuaries, rivers, watercourses,
waterbodies, wetlands, etc.), each having a particular impact on the local dynamics of the species
and being more or less restrictive in terms of sampling and monitoring;
• the knowledge of gear and of sampling strategies and of their respective limitations and
advantages.
The ultimate indicator within the framework of a management process aiming to restore the
species inside a river catchment must focus on the escapement of quality spawners. The goal must be
to ensure that this escapement is positive and sustainable, which can only result from co-coordinated
actions and efforts aimed at:
• optimising the colonisation of the catchment, in relation to the number of glass eels entering that
catchment;
• improving habitat accessibility and quality;
• increasing significantly the survival rate during the entire growth phase;
• and controlling the level of impacts on the downstream migratory phase of produced spawners.
Two distinct domains must be analysed and monitored in order to optimise the approach:
• the state of the species or of a particular size group within the relevant catchment;
1 See Chapter 2.
273

• the level of impact of human activities on the species’ local dynamics 2 .
In each of these two large domains, the actions undertaken, the observation methods used and
the data collected in an area or in a river catchment must sit within the following approach:
• begin with an initial diagnosis (observation and comparison with a reference situation);
• contribute to the concerted choice of one (or several) objective(s) to be attained and of adaptive
management actions;
• and lastly, make it possible to observe the consequences of management decisions concerning this
(or these) objective(s), i.e. how the situation evolves.
8.2. Estuarine and fluvial recruitments
In order to separate the total recruitment (i.e. the estuarine recruitment 3 ) from the fluvial
recruitment (which colonises the river catchment; and is dealt with in this chapter), the tidal zones must
be identified 4 .
The glass eel flux entering the estuarine zone comprises the total recruitment of the catchment
(figure 8.1). This recruitment is highly dependent on the general state of the European eel stock and on
the oceanic conditions for reproduction and larval migration. The geographical location of a river
catchment in the inland distribution zone and its main characteristics (flow and water quality in
particular) also have a significant impact on the order of magnitude of this total recruitment 5 . Information
on the evolution of this total recruitment and on the level of catches made from it can be obtained by
monitoring glass eel/elver fisheries 6 .
2
See Chapter 6 which concerns the impact of local fisheries.
3
See Chapter 7.
4
See General Introduction, § >.
5
See Chapter 2.
6
See Chapters 6 and 7.
274

Anthropogenic mortality
(abstraction, fishing…)
Natural
mortality
Sedentarisation
N total
Tidal zones
N fluvial
A particular context (zone size, uses, water quality etc.).
For example distance between the tidal limit and the sea:
• 7km for the Sévre niortaise and the Vilaine (dam)
• 70km for the Loire (narrow estuary)
• 140km for the Gironde-Garonne-Dordogne
• European population and reproduction
status
• Geographical location of the RB
• Size of the RB – Attraction flows
• Water quality
For each river catchment
Figure 8.1. Total and fluvial recruitment downstream of a river catchment. Factors affecting
these two values.
Natural mortality, the sedentarisation of some individuals in tidal zones and in coastal wetlands,
fishing and other anthropogenic mortalities (abstractions, etc.) then determine the fraction of this total
recruitment which leaves the zone affected by the dynamic tide and hence comprises the fluvial
recruitment. This fluvial recruitment can be estimated, and/or its evolution assessed, in various ways
(see following sections). In any case, the data on glass eel/elver fisheries never reflects the level of
fluvial recruitment as sedentarisation downstream of the fishery and natural mortality, at least, must be
considered.
The dynamic tide limit can be natural or artificial (for example, the presence of a construction
blocking the effect of the tide). This remark applies both to the main axis of a river catchment and to the
tributaries which merge with the main river inside the tidal zone. Whatever the situation encountered,
the zone affected by the dynamic tide is populated by individuals that have not had to overcome a
barrier. In this zone, therefore, the abundance and quality characteristics of individuals relate only to
total recruitment and human pressures exerted on the estuarine stock.
Further upstream, eel abundance along the axis and the dispersion of the stock of individuals
entering this axis depend on factors such as the distance to the dynamic tide, the number and
passability of barriers and the gradient of the watercourse. This initial stock, in turn, depends on the
position of the axis within the river catchment (distance to the tidal zone, altitude, etc.) and on its degree
of attractiveness (water quality, flow, etc.).
275

The very different context within which these two areas function (tidal and upstream zones) and
their very different levels of heterogeneity (water levels, distance to the sea, habitat type, etc.) mean that
these two large compartments within a river catchment require very different approaches 7 . In the
upstream zone, it is also useful to distinguish the axes and the deep environments, where very few
methods and data are currently available, from other environments which are much more accessible to
eel sampling.
8.3. Data acquisition
8.3.1. The need to use size classes
The term Yellow eel within a river catchment comprises individuals ranging from 6cm to more
than 1 metre long. The heterogeneous distribution and behaviour of these differently sized individuals
within habitats, the difference in selectivity of sampling and observation tools, the emergence of the
silvering process (in preparation for reproductory migration) and the sexual dimorphism are all factors
that militate in favour of analysing the data by size class.
We suggest identifying 6 groups of 15cm long whilst acknowledging the fact that, as in any
segmentation, the limits are never perfect 8 . However, this prevents major errors in data analysis and
interpretation by avoiding an irrelevant and unhelpful global analysis.
Generally, we assume that individuals less than 30cm are yellow eels in their growth and/or
colonisation phases. This size range is often observed in pass facilities downstream of river catchments.
In the great majority of cases, eels less than 15cm long have been in the river catchment for one or two
years. And in the great majority of cases, sexual differentiation occurs between 15 and 30cm (fewer
than 6 inland summers).
Over 30cm, silvering individuals can be caught which means that some individuals from the
populations observed may be leaving for the sea. The size range between 30 and 45cm mainly
concerns males likely to silver and migrate downstream. In order to identify the sex of yellow eels of this
size group, a macroscopic and microscopic examination of the gonads is necessary (see the Colombo
and Grandi protocol, 1996) which requires the sacrifice of a representative sample. This sacrifice must
be optimised in order to reduce its impact. This information on the quality of individuals may be
important when analysing the data from a site 9 . Above 45cm, only females, in their growth or silvering
phase, can be observed; their size seems to be strongly related to water levels in particular. However,
some favourable habitat may be devoid of large sizes if the mortality rate (natural or anthropogenic) is
too high, preventing individuals from surviving to such sizes.
7
See General Introduction.
8
See Chapter 2.
9
See Chapter 9.
276

8.3.2. Sampling tools
Three broad categories of tools can be used to collect data and information on yellow eels during
their colonisation or sedentarisation phase:
• passive gear (fisheries or specific approaches);
• passes on dams;
• electrofishing operations.
Each tool has strengths but also significant weaknesses which have to be taken into account,
particularly when interpreting the data.
8.3.2.1. Passive gear
Passive gear includes basket traps, deep lines, fyke-nets, “capetchades” (fixed nets which
channel eels into basket traps), etc. Within the regions and the very diverse environments colonised by
the species, there is a great variety in the types and shapes of gear and in the materials used for their
construction (Brandt, 1971; Luneau et al., 2003).
Passive gear takes advantage of some of the features of yellow eel behaviour during its active
period from Spring to Autumn. Hence, trapping is based on their search for shelter, contact or food.
Their acute olfactive sensitivity can be utilised but can also hamper their catch, at least temporarily.
Incidentally, the “acclimatisation” of new gear, or of gear coming from another zone, is often necessary
and recommended in order to achieve a significant level of catch in a fishing zone (Mohr, 1971).
These are often the only tools that can be used in deep zones (major axes downstream of
catchments, waterbodies, etc. – Adam, 1997), in zones that are difficult to access (in particular, dyked
marshes – Baisez 2001) or in compartments where conductivity is high (salt marshes, etc. – Laffaille et
al., 2000a). The catch level observed for a given level of fishing effort should reflect the species'
abundance in this zone, as should the qualitative characteristics of catches. However, this relationship
between abundance levels and catch quality is complicated by the individuals’ activity rhythms and
behaviour patterns as well as by the gear characteristics because these factors introduce the notions of
accessibility, vulnerability, selectivity and efficacy, which have been discussed elsewhere 10 (Baisez,
2001).
Impact of the characteristics of the gear used
In all size structures observed by trapping (figure 8.2), Berg (1990) and Naismith and Knights
(1990) highlight the existence around the dominant-size class (or mode):
• of a fraction which is totally dependent on the selectivity of the gear used (i.e. the smaller sizes);
10
See Chapter 6.
277

• and of a fraction whose characteristics are related to those of the existing stock (i.e. the larger
sizes).
Gear selectivity effect
Fraction depending on the structure of the existing stock
% of catches
Size
Figure 8.2. Example of size structure observed by passive gear showing the selective effect of
the gear (on the left) and the part which depends on the existing stock (adapted
from Berg, 1990).
As regards the selective effect of the gear, the mesh used is the important element to take into
consideration. For each mesh, the selectivity curve allows three characteristic values to be taken into
account (figure 8.3):
• L 0 , size below which all individuals can escape ;
• L 50 , size corresponding to every second individual escaping from the trap;
• L 100 , size above which no individual can escape from the trap.
278

% retained
Size
Figure 8.3. Theoretical graph showing passive gear selectivity with L 0 , L 50 and L 100 .
These three characteristic values have a quasi-linear relationship with the mesh size, recognising
the need to distinguish between net meshes and rigid meshes (plastic or metal). In order to illustrate
this, the data available on L 100 , obtained by controlling escapement during specific tests, were analysed
(figure 8.4). The 60-70mm gap which appears between the two curves relates to the fact that net
meshes can change shape (to some extent), unlike metal or plastic meshes.
In the case of basket traps, the mesh size does not vary; however, Ximenes (1986) highlights the
fact that the large majority of net gears have different mesh size from the “paradière” (guiding net) to the
bottom of the eel pot, with each section of the gear having a different selectivity.
And she notes, particularly in some Mediterranean lagoons, the cumulative impact of this mesh
size heterogeneity on the size structures observed. In such a context, it would be a mistake to take into
account only the smallest mesh size (beginning with the guiding net, some individuals larger than the
L 100 of the smallest mesh size of the gear may escape). Logically therefore, analyses should only focus
on sizes larger than the L 100 of the largest mesh in the gear. When possible, it is therefore
recommended to choose gear with a unique mesh size in order to make standardised
observations.
279

Net mesh
Rigid mesh
L 100 (in mm)
Mesh size (in mm)
Figure 8.4. Relationship between the L 100 and the dimension of rigid meshes (Ximenes, 1986;
Baisez, 2001) or net meshes (Adam, 1997; Naismith and Knights, 1990).
Another gear selectivity issue was highlighted by the monitoring work of Mohr (1971). He states
that eel basket traps must be as long as possible in order to increase the distance between the
successive funnels (“empêches”, “anchons”, etc.). This is because if the eel is still in contact with the
first funnel when it reaches the second one, it has the ability to reverse and come out of the trap.
Therefore, this author is not referring to selectivity on the left-hand side of the size structures but on the
right. Hence, any individual longer than the distance between the first two funnels is able to escape the
trap. Unfortunately, and unlike mesh selectivity, the impact of this gear parameter on the escapement
level has never been specifically tested to our knowledge.
Finally, Chisnall and West (1996) raise the problem of comparing CPUEs obtained with different
gear. When comparing catches using basket traps and fyke-nets, they suggest weighting the CPUEs
observed with fyke-nets or “capetchades” using the length of their guiding net which, unlike basket
traps, often intercept the eel at a significant distance from the trapping device. This method of analysing
and processing the results is, however, not very satisfactory. It should be recalled that only data
originating from very similar gear should be compared; or alternatively the evolution over time of the
characteristics of catches originating from each type of gear can be analysed without seeking to
compare them.
Impact of individuals' behaviour
The significant temporal variability of passive gear catches has been mentioned on many
occasions. Periods when there is little moonlight are known to be the most favourable for yellow eel
activity outside of winter (Morh, 1971; Corsi and Ardizzone, 1985) at temperatures above 12-13°C
(Baras et al., 1998).
280

Data collected monthly for 2 years, in a closed zone, on a constant stock and in the same lunar
conditions, showed clearly the very close relationship (figure 8.5) between temperature variations and
the level of basket trap catch (Baisez, 2001). It seems that this relationship depends on the level of
activity and the magnitude of the movements of individuals more than on their actual number. Clearly,
the greater the intensity and magnitude of movements, the higher the probability of an encounter with a
fixed gear. Ximenes (1986) made the same observation in the case of lagoons.
Number of catches per site
Temperature ⁰C
Month
Figure 8.5. Basket trap fishing. Evolution of mean catches () by site with associated standard
deviations and of mean water temperature () during 18 surveys in 1998 and 1999 (adapted from
Baisez, 2001).
In both cases, with fixed fishing stations, catches dwindled after a few days except when an
environmental event (water movement, storm, etc.) caused the redistribution of individuals within
habitats.
Baisez (2001) also highlights two other interesting points. First, it is possible to observe a fall in
catch when temperature conditions are a priori optimal (the June 98 example in figure 8.5). In this case,
other factors such as the very low concentration of dissolved oxygen in the environment during the
summer were to blame. Second, comparing electrofishing and basket trap data collected at the same
time and on the same sites (figure 8.6) highlights a passive gear selectivity which is not related to the
mesh size of the gear (the sizes shown in the figure exceed the L 100 of the gear). In all likelihood, this
selectivity is related to the size of the areas covered by different-sized eels. Hence, small individuals
with a small range of action appear to be under-represented in catches made during the first few days.
A contrario, large individuals with a greater range of action, relative to the scale of the habitat, are much
more likely to come into contact with the gear and are, therefore, over-represented in the initial landings.
281

Basket traps
Electrofishing
% of catches
Total length class (in mm)
Figure 8.6. Size structures observed on the same site and at the same periods when monitoring
by basket trap (5mm mesh, L 100 = 200mm) and by electrofishing (adapted from
Baisez, 2001).
In fact, Chisnall and West (1996) queried the low levels of small eel catches despite small mesh
sizes being used to trap them and had already raised the behavioural issue. Observation over a long
period of time (a week or more) can sometimes lessen this problem. This size structure distortion
related to individuals’ behaviour can be corrected with more certainty by using a mark-recapture
procedure over several weeks. Knights et al (1996) report on this type of approach in small-sized
environments with recapture rates ranging between 5% and 18%. On the other hand, they record very
low recapture rates (below 1%) in the estuarine and fluvial parts; and this rate cannot reliably be used to
correct the signal obtained.
Impact of the operator’s strategy and know-how
Often overlooked, incorrectly, in the list of factors significantly affecting the quality of the results,
operators contribute another level of variability when applying their strategy and know-how to the use of
a given gear. Doesn’t everyone drive differently even though they all have the same driving licence?
Hence, the positioning of the gear, the time elapsing between two readings, the frequency of gear
movements, the use or non-use of bait and the type of bait used are all factors that will affect the fishing
yield and the size structure of the catch.
Overview of passive gear
In the end, if the impact due to the strategy of the fisher using the gear is added to the distortion,
caused by gear diversity, in the observed abundance of different-sized individuals compared to the
existing stock, it becomes very difficult to compare signals collected with different gear in different
areas. A trend, in terms of abundance index and of observed size structures, can only be established
by observing the relative evolution of the signal collected in a given territory using a given gear and a
282

given strategy. This is the case, in particular, in annual fisheries monitoring (analysis by metier or gear,
and by compartment 11 ).
8.3.2.2. Pass facilities
Unlike Northern European countries (Rigaud et al., 1988), there has been very little development
in the installation of eel-specific fishpass facilities in France, Spain and Portugal.
Selectivity of the pass facilities
Even though eels can use standard fishpasses, particularly the Denil-type fishpass (baffle-type)
with energy dissipation structures (Baras et al., 1994) or fishlifts (Legault, 1992), their poor ability to
jump or swim against strong currents places them at a disadvantage when using these facilities,
especially in the case of small size classes (figure 8.7).
Brush-type passess (n=6,276)
Lift (n=202)
% of catches
Total length class (in mm)
Figure 8.7. Size classes of eels passing over a bristle ramp and in the fishlift on the site of
Tuilières (Dordogne) over the same period of time (adapted from Legault, 1992).
Specific passes remain few compared to the number of recorded barriers but they can
significantly help in overcoming many types of dam if they are properly maintained over time. The
principle is to offer migrants wetted ramps covered with a substrate which facilitates their climbing.
Covering substrates vary greatly (Rigaud et al., 1988), they can be natural (straw or twigs twined into
braids, bundles of wood, blocks, etc.) or artificial (brush mats, wire mesh, concrete blocks, etc.).
Legault (1992) initiated some work on the selectivity factors of these ramps which confirmed the
impact of the nature of the substrate, previously highlighted by Moriarty (1986b), and of the slope of the
ramp (15 to 30° are recommended). This type of analysis was taken further by diversifying the substrate
11
See Chapter 6.
283

under study and the longitudinal as well as the lateral slopes of the ramp (Voetgle and Larinier, 2000)
and confirmed the major impact of these various factors on the selectivity of the facility. Migrant size
structure, which varies significantly with the distance to the tidal limit, must also be taken into account
when defining an optimum facility at a given site (Legault et al., 2004). Added factors are the location of
the facility in relation to the barrier and the characteristics of the attraction flow which determine the
degree of attractiveness of the pass to the flux of migrants.
All these elements can affect the efficacy of the facility (numbers that have passed and waiting to
do so) and its selectivity concerning the size of passing eels. This means that the data collected can
only be considered as indices of the relative evolution of the migratory flux at a given site. On the
other hand, the representativeness of these passes in relation to the total flux actually passing through
the barrier (other pathways are possible) and to the potential migratory flux arriving at the foot of the
dam can only be assessed case by case through procedures such as mark-recapture, downstream
monitoring or using several successive trap-passes (Baras et al.,1994; Legault et al., 2004; Briand et al.,
2005; Laffaille et al., 2007).
Facility monitoring
Monitoring can usefully be implemented on sites equipped with passes, whether these are
specific or not. It is therefore important to standardise the protocols and methods used and integrate the
highly transient nature (on an annual scale) of the most significant eel runs (these occur over a 2-month
period, see Laffaille et al., 2000b for example). This monitoring must quantify the runs (minimum
number of individuals colonising the headwater) and show their evolution over time (year, month, week),
specify the most important factors affecting the intensity (temperature, attraction flow, etc.) and
characterise migrants (size, weight, health condition, body condition, etc.).
In the case of multi-specific passes, occasional operations can be organised to monitor and
estimate transitory fluxes by trapping and mark-recapture (Baras et al., 1994). Video recordings of runs
are a very precious tool (Castignolles, 1995) although they have not yet been perfected especially for
smaller individuals. Examples show that the distinctive behaviour of eels (progression along the edges
and at depth) can lead to difficulties in their detection by video systems and these need therefore to be
adapted accordingly (Carry and Delpeyroux, 2003).
In the case of specific ramps, counting the runs can be done in different ways:
• by using resistivity counters which are progressively perfected and operational on a few specific
sites (Pallo and Travade, 2001). Subject to a calibration phase, they can estimate run intensity
with a level of reliability that has yet to be defined.
• by trapping and weighing (drained weight), total counting rarely being possible especially during
very intensive migratory periods (Legault, 1994).
284

In both cases, regular biological observation is essential in order to collect data on the size,
weight and health condition of the migrants using the ramp. This must be done weekly a minima and
adapted to the run intensity, migratory peaks should of course be monitored more frequently than
periods of low migratory activity 12 .
The method used to collect the sample from the total catch over a given period of time is an
important issue, especially in sites located in the downstream sections of the river catchment where
small sizes are present 13 . Random sampling which often uses gear with inadequate mesh size should
be avoided. Instead, sorting on plastic grids, under running water, is to be preferred, allowing 2 or 3
batches to be quickly separated. Hence, with respect to the relationship between the L 100 of a rigid mesh
and the dimension of this mesh, we note that:
• a 5mm mesh will quite efficiently allow the selection of most individuals smaller than 20cm;
• a 10mm mesh can separate individuals of 20 to 30cm in length and the larger ones.
These successive sorting devices can be made with PVC pipes which fit into one another. This
preliminary sorting process means, in particular, that the small sizes are not underestimated in the runs.
Once sorted, the group of individuals is weighed (or counted) and at least 50 in each batch are
examined individually (size, weight, health condition).
These data, for a given site and device, and standardising the observation process, make it
possible to monitor the inter-annual evolution of the characteristics of the run (number, size, health
condition) which relate to the development of the level of inland colonisation by the species in the
relevant area (Legault et al., 2004). On the other hand, the comparison between sites is much harder
due to previously mentioned selectivity and efficiency problems with the devices. Bearing this in mind,
comparisons are still possible but always by relating the size of the run observed to the surface area of
the river catchment located upstream of the barrier, i.e. in number of individuals per km 2 of the drainage
catchment.
8.3.2.3. Electrofishing
As with any observation technique, electrofishing has some advantages, in particular the speed
with which information can be collected, but also some limitations (significant cost, selectivity,
inaccuracy, risks for operators, etc.) that vary with the methods used and the environments explored.
General principle and types of current used
Catching fish with electricity is based on a certain number of reactions induced by polarised
current, the most important of which is electrotaxis (Lamarque, 1976). When placed in a direct current
12 Please refer to Chapter 2 for current knowledge on the factors which seem to cause a significant increase in run intensity. If the
run size is too great, a sample can be examined.
13 See Chapter 3.
285

field, the fish points, and then swims, towards the anode (immersed positive pole, which may be
handled by an operator) when the potential gradient (V.cm -1 ) exceeds a certain value (the anodic taxis
threshold). This phenomenon occurs at a certain distance from the electrode in a zone where the
current density is quite low. The closer to the anode, the greater is the potential gradient and when it
exceeds the narcosis or tetanus threshold, the fish stiffens and stops swimming. Therefore, an
effective fishing method uses the smallest possible zone of electrotetanus around the anode so that the
fish is forced to swim towards this electrode and can be detected and caught by the operator.
Incidentally, Lamarque (1976) emphasises that this efficacy factor is very important for eels. In most
other species, once stunned the individual usually comes closer to the anode due to the momentum
from its swimming in the electrotaxis zone, whereas, when stunned, the eel curls up and comes to a
total standstill.
There is no doubt that a constant direct current is the most effective with a very small zone of
electrotetanus. On the other hand, alternative current creates stacked zones of electrotetanus and
electrotaxis which lead to very low fishing efficacy if operators are not efficiently trained to detect and
catch the eels that are immobilised. Between these two extremes, different types of current can be used
such as, in decreasing order of efficiency, a three-phase full-wave rectified DC (waveform ripple ratio
4%) or half-wave (17%), a 400 Hz rectangular (pulse) wave current and 10% ratio and lastly a 100 Hz
rectangular (pulse) wave current and 10% ratio (Lamarque, 1976). The choice is often the result of a
compromise between the characteristics of the sampled environment (in particular water conductivity
and accessibility of the catch zone) and the practicability of the device.
Eel catchability with electrofishing
Generally speaking, the level of catchability of fish with an electric current is a much debated
issue (see for example Zalewski and Cowx, 1990; Bohlin and Cowx, 1990).
In order to be caught and counted, an individual must be:
• accessible, i.e. present in the study area;
• attracted and/or immobilised by the electric field (influence of the search strategy, the nature of
the current used, water depth, water conductivity, water temperature, the season, the size, the
species, etc.);
• detected by the operator (influence of turbidity, shelter and vegetation density, water depth,
current velocity, operator’s concentration, etc.);
• caught in the hand net and unable to escape subsequently (influence of the operator's
experience, the shape of the hand net, the mesh used, etc.).
All these elements determine the overall efficacy of the fishing operation, so it is clear that each
fishing operation (station, fishing team, date, particular environmental conditions, etc.) is unique.
286

In terms of eel accessibility, Lambert (1997) showed the seasonal variation in eel occurrence
(percentage of stations where eels can be found) by analysing the general data from the Hydrobiological
and Piscicultural Network of the “Conseil Supérieur de la Pêche” (Higher Fisheries Council). This trend
(figure 8.8), surprisingly for an active gear, strongly resembles that of passive gear which seemed
closely related to the level of activity of the eel (influence of water temperature and oxygenation).
Eel occurence (in %)
Dec
Jan Feb Mar Apr May June July Aug Sept Oct Nov
Figure 8.8. Seasonal evolution of eel occurrence during operations undertaken within the
framework of the Onema Hydrobiological and Piscicultural Network (adapted from
Lambert, 1997).
The fishing team will also affect fishing efficacy due to its level of experience and/or cohesion,
and also to its degree of focus on finding eels (rather than other species such as salmonids). Eels do
not react well to electric stimulation and are often burrowed (and hence depend on substrate and
shelters) during the day. Greater persistence is therefore necessary than for many other fish (salmonids
for example) to entice them into the field of the electrode. The time taken to explore a station will
therefore strongly affect the efficacy of the operation. A minimum fishing time of 30 seconds is required
for each sampled micro-habitat if any eel present is to be attracted, immobilised or detected (Laffaille et
al., 2005b). Similarly, the exploration of shallow and/or vegetation-dense zones within the fishing station
is essential to maximise efficacy, especially in the case of small individuals as these are their preferred
micro-habitats (Laffaille et al., 2003).
Moreover, for a given operation, all species and sizes within a species do not present the same
electrofishing catchability characteristics. Several authors (Bohlin and Sundström, 1977; Philippart,
1979; Mahon, 1980) recommended some time ago that the data be analysed by biological unit
(species, sizes within a species). In the case of eels, although the results vary by author, it seems that
catchability is lower for small sizes.
287

Lambert et al (1994), in the Breton marshes, observed that catch probabilities did not vary a great
deal between sizes but were generally lower for individuals less than 10cm (figure 8.9).
Catch probability
Mid-point of size class (mm)
Figure 8.9. Eel catch probabilities from 28 surveys (3 runs) in the Breton marshes (adapted from
Lambert et al., 1994).
During operations undertaken on Welsh and English rivers, Aprahamian (1986) observed a
marked and significant increase in catch probability with eel age and hence size (table 8.1). Taking into
account the mean growth observed by Aprahamian (2000) in these rivers (about 2cm a year), the
significant increase in catch probability seems to occur here around 25-30cm. Naismith and Knights
(1990) agree with this observation and conclude that electrofishing is not really effective on individuals
smaller than 30cm in operations which do not target and are not standardised for eels.
Table 8.1. Electrofishing and eel catch probability by age (Aprahamian, 1986).
Age group
Catch probability
0 to 4 years old 0.36 (±0.13)
5 to 9 years old 0.39 (±0.10)
10 to 14 years old 0.54 (±0.10)
Over 14 years old 0.59 (±0.19)
Furthermore, even when working by species or by size group, most studies show that catches
decrease with successive runs, leading to an underestimate of the stock by about 15 to 25% (Mahon,
1980; Bohlin and Cowx, 1990). This fall is most probably due to the stress caused by successive electric
shocks (Cross and Stott, 1975) and/or to a difference in sensitivity to electricity within the original
population, in particular because the most sensitive and catchable individuals are collected during the
first run (Bohlin and Cowx, 1990). As regards eels, Naismith and Knights (1990) showed that a
significant number of operations resulted in higher catches on the second run than on the first one.
However, this observation cannot be generalised because in the Breton marshes, in shallow and silted
up canals where vegetation density is high, Lambert et al (1994) observed that, on average, the first run
288

accounted for 57 % (± 17) of the catches made over 3 runs, with the first two runs amounting to
84.3 % (± 10.5) of this total. Similarly, Callaghan and McCarthy (1992) observing the same 3-run
protocol on an Irish river found the first run to amount on average to 58% of total eel catches on each
station (variation ranging between 44 and 80% according to the station).
This discussion confirms, therefore, what was mentioned earlier, namely the unique nature of
each operation (environment, date, team, strategy) and the resulting varying efficacy of each fishing
operation.
Different methods for using the tool
The term “electrofishing” encompasses a wide diversity of uses. Steinmetz (1990) noted, from an
FAO survey of 32 national representatives, that all countries use backpack electrofishing, ¾ of them
also use it during vessel-based surveys whilst only 15% use electrified trawls for sampling. This latter
method will not be dealt with here. It is highly specialised and can only be used in deep environments
with flat and even beds (no branches, blocks, etc.).
Whether from a boat or not, electrofishing can be implemented in different ways to survey
different stations and this can have an impact on the nature of the signal collected and the way in which
it is analysed. The following fishing strategies are often found:
• successive runs and net-based collection in an isolated station;
• survey (1 run) by station or by representative river sector “ambiance”(aquatic plant habitat,
riparian zone, etc.) or by habitat (rapids, shallow fast flowing waters, etc.);
• strip-based fishing along the bank;
• point-based fishing with a standardised effort.
Naturally the mesh used in the equipment used to isolate stations (nets) and to collect individuals
(hand nets) has a significant impact on the size structure observed. Please refer to the relationship
between mesh size and L 100 described in the paragraph on passive gear for the limits of the gear used
in each operation.
8.3.2.4. Impact of the characteristics of sampling tools
The stock inside a river catchment cannot currently be estimated
The summary of knowledge on the yellow stage of the eel 14 and of the characteristics of sampling
tools highlights the wide variety of inter-related factors that affect:
• the characteristics (density and size structure in particular) of the fraction of the eel population
present in a given site or habitat (characteristics of the relevant axis, of the relevant section of
that axis, of the sampled habitats within that section, etc.).
14
See Chapter 2
289

• the characteristics of this population fraction observed via a sampling tool or strategy (notions of
selectivity, efficacy, etc.) which gives a modified picture.
These many sources of variability should not be seen as obstacles to monitoring or estimating
the number of individuals present. They must simply be kept in mind and integrated into the design of
monitoring arrangements or strategies and the analyses of the results concerning this stage in a river
catchment.
Hence, the fact that it is impossible, currently, to estimate the existing stock at the scale of
a river catchment must be noted:
• as there are almost no data on yellow eel abundance in deep environments;
• as it is very difficult to allocate a given signal (abundance, size structure, etc.) to a shallow
compartment;
• as the appropriate weighting factor to apply to the average signal collected in each shallow
environment is uncertain (inclusion of respective lengths?, of respective areas?, etc.).
The fact is that there are significant spatial and technical constraints in the comparison of
observed abundances, which mean that it is very difficult to compare signals collected on yellow eels
using different gear and from stations with different habitats. The analysis becomes more reliable when
restricted to comparable habitat sampled in the same way (same tools, same use). This is of course the
case in stations where the monitoring team returns regularly at the same period with the same protocol.
This is also the case when identical habitats are monitored using the same methods (shallow zones at
the foot of the barrier, bank exploration, etc.).
In addition to this methodological issue, the investment required, in terms of human and financial
resources, to access this type of information in all the catchments of the distribution area cannot
reasonably be maintained over a long period of time. But this is the time scale required to monitor the
recovery of a species such as eels in a river catchment.
Reasons to monitor the relative evolution of observed signals
A trend, in terms of abundance index and/or observed size structures, can only be
established by observing the relative evolution of the signal collected in a given territory with a
given gear and strategy. This is the case in annual fisheries monitoring (analysis by metier or gear,
and by compartment 15 ) and in permanent monitoring networks involving electrofishing or fishpass
facilities.
The analysis station by station (or compartment by compartment) of the evolution of the
signal collected over a given period using the same standardised protocol (date, tool, strategy) seems
of interest to visualise the relative evolution of the species in a catchment, a sub-catchment or a given
15
See Chapter 6.
290

compartment (see following sections). A second stage might consist of a spatial analysis of this
approach (identification of the zones in the catchment which evolve in the same way). In any case, this
method of analysis is not limited by the significant variability observed between the stations and by the
absence of abundance data in deep axes.
8.4. Methods used to estimate the stock present in a sampling station
These different sampling methods and their methodologies can be used either within the
framework of short electrofishing operations (a few hours) or in monitoring operations using passive
gear over much longer periods (several days and even several weeks) 16 .
8.4.1. Mark-recapture procedures
This evaluation technique was initiated by Petersen (1896 in Kendall, 1999). It is based on
catching and marking individuals and releasing them back into the environment between operations.
The proportion of marked individuals in the total number of individuals caught by fishing at a later date
can be used to estimate the most likely number in the initial population.
This technique is based on a certain number of assumptions, of which only the main ones
concerning eels are mentioned here:
• the stock remains constant over the observation period (no emigration, immigration, mortality or
birth);
• no marks are lost over the observation period and marks are clearly identifiable;
• no change occurs in the survival rate or in behaviour due to marking;
• the distribution of marked and unmarked individuals is homogeneous within the population
being estimated;
• catchability is identical for marked and unmarked individuals. The hypothesis that marked and
unmarked individuals have an equal probability of being caught has never really been verified.
Yet, it seems likely that this probability must be lower for individuals that have been caught,
handled, marked and released into the environment especially because of the stress involved.
Either electrofishing or fishing with passive gear in closed sectors can be used for these markrecapture
operations. In the latter case, Moriarty (1986a) found that recapture rates varied between 5.5
and 18.5% when restricted and densely populated environments were monitored. This level of recapture
rate allows a priori reliable estimations of the existing stock.
16 Please refer to Chapter 6 which, from a different angle, also deals with capture-recapture models and, more
generally, with the use of yellow eel data from fisheries. These data may be used to derive descriptors of absolute or
relative abundance over time and space, or simple indications of the presence of the species in a given zone, at a
given time.
291

However, the territorial character of yellow eels must be taken into account (Laffaille et al., 2005a)
as it can seriously compromise the hypothesis of the homogeneous distribution of marked individuals
within the area, which is required for this method to be used (Rigaud, unpublished data). Two cases are
then possible:
• either the operational range of each gear is known and releasing marked individuals on the
harvest site is not a major issue. The estimated stock must then be related to the sum of all
the areas corresponding to the various gear used;
• or this information is not available and marked individuals must be released as randomly and
widely as possible in order to optimise their homogeneous distribution among their unmarked
congeners. In this case the stock can be related to the whole area under study.
Naismith and Knights (1990) also raise the issue of the very low recapture rates (0.1-0.2%)
observed in wide and open spaces (example of estuaries or fluvial zones) as these rates do not lead to
reliable estimates.
The methods used to calculate the most probable stock vary according to recapture techniques.
8.4.1.1. Case of 2 runs
In a population where N is the unknown number of individuals, M individuals are caught, marked
and released into the environment. A second sampling survey leads to R individuals being caught of
which m are marked.
According to Petersen (1896 in Kendall, 1999), the estimated number of individuals N is then
N est = M x m / R
The confidence interval equals: t x
M ²( m + 1)( m − R)
( R + 1)²( R + 2)
with t =1.96 for α=0.05 and t =2.58 for α=0.01
For values of m below or close to 10, Bailey (1951) recommends an adjustment using
N est = M x (m+1) / (R+1) which reduces the estimation bias.
8.4.1.2. Case of several runs or capture phases
For each sample (or run) i, there is a potential marked stock M i (the stock initially marked minus
recaptures of previous runs) and m i marked individuals are observed among R i catches.
The different runs, once completed, give the Schnabel estimator modified by Chapman (1952)
that is: N est = Σ (M i x m i ) / Σ m i .
For low values of Σ m i , the less-biased formula N est = Σ (M i x m i ) / ((Σ m i ) + 1) is preferred.
292

This method can be useful when recaptures are low; information from several periods or runs can
be collated thereby reducing estimation biases.
8.4.1.3. Other cases
Other data processing methods, of increasing complexity, have been developed on the basis of
this principle, in an attempt to adapt to various working contexts (long observation periods with nonconstant
stock, integration of new marked individuals during the survey, etc.). Rochard’s review of these
methods (1992) is of great interest.
8.4.2. Stock depletion
The technique is used in a given isolated station, based on successive runs during each of which
animals are caught but not released. This technique of progressive depletion is the basis of many
methods used to estimate the most likely size of the existing stock (Cowx, 1983; Gerdeaux, 1987). Of
course, this estimation is all the more robust if successive catches decrease regularly.
The catch probability is an important element to take into account in this method. It is equal to the
ratio of the number of individuals caught over the number available in the relevant station (i.e. the
unknown number being investigated). It can be calculated for each run or for the entire operation but
only a posteriori.
As with mark-recapture methods, a certain number of assumptions must hold to be able to carry
out these estimations:
• the stock to be estimated must remain constant over the study period (no entry into or exit from
the system);
• the catch probability must be the same for each fish;
• this catch probability must be constant from one sampling to the next;
• individual catch probabilities must be totally independent.
If the first assumption can be made to hold, in particular through the physical isolation of fishing
stations, the others are much more problematical.
Various methods of data analysis are available to estimate the most likely existing stock but often
the assumptions hold partially if at all.
8.4.2.1. Regression-based estimation methods (Leslie,
1939 and De Lury, 1947 in Seber and Le Cren, 1967)
This method is based on the existence of a linear relationship between the sum of the catches
made before the i th run (x axis) and the catches of the i th run (Ci on the y axis). The intersection of this
line with the x axis indicates the most likely stock present in the station. Laurent and Lamarque (1975 in
293

Rochard, 1992) undertook a detailed analysis of the application of this method to piscicultural
populations. In the case of two successive harvests, of course only one line passes through the two
points (0, C 1 ) and (C 1 +C 2 , C 2 ) (figure 8.10).
However, two principles must hold in order to be able to carry out the estimation: C 1 >C 2 and
C1
²( C1
− C2
)²
> 16. Otherwise, the only thing that can be said is that the stock in situ is higher (and
C ²( C + C )
2
1
2
often very much higher) than the sum of the catches from the two runs.
When the calculation is possible, we obtain N est = C 1 ² / (C 1 – C 2 ) and the confidence interval is
then t ×
C C
1
2
( C − C )² 2
1
C
1
+ C
2
with t =1.96 for α=0.05 and t =2.58 for α=0.01
If both conditions which make this calculation possible are not met at the end of the two runs,
supplementary runs must be made in order to obtain a reliable estimate.
Catch per run
N (estimated)
Cumulative catch
Figure 8.10. “Graphical” estimation of stock numbers based on cumulative catches (x axis) and
the catches per run (y axis).
In this case, the calculation phase is slightly more complicated as a simple linear regression is
necessary to determine the line which will give the most likely N. In order to obtain an estimation with a
confidence interval of reasonable size, the regression must be robust with a minimum r 2 of 0.99 for 3
harvests, 0.94 for 4 harvests, 0.77 for 5 harvests and 0.64 for 6 harvests. Given these very significant
statistical constraints, Laurent and Lamarque (1978 in Rochard, 1992) conclude that this method is
difficult to use for more than two runs and for second-sample catches which are not significantly lower
than those in the first sample.
294

8.4.2.2. Maximum likelihood estimation method
The general principle is based on the analysis of the event E which corresponds to the
succession of catches observed over k runs at a station with E = { C 1 , C 2 , C 3 ,…. , C k } during an
operation. The probability of observing this event p(E) is equal to the product of the probabilities of
observing the C 1 , C 2 , C 3 ,….., C k which comprise it. These latter probabilities depend on N (initial stock in
place) and on p, the probability of catching the species, which is assumed to remain constant over the
complete set of runs. The event E observed during the operation is considered to be the event with the
highest probability of occurring. The various calculations seek therefore to define the values of the
parameters (N and p) which correspond to the maximum value of p(E), hence the name given to these
methods. Different calculation methods exist.
The Moran-Zippin method
Zippin (1956; 1958) used Moran’s method (1951) but suggested some changes in the calculation
strategy used to estimate N and p. Assuming that the catch probability p remains constant over time, the
method defines the probability of observing C i (the catch in number on the i th run) as being equal to
drawing C i individuals with a catch probability p from (N-n i ) individuals (n i being the sum of catches
before the i th run).
This leads to P (Ci) = p Ci * (1-p) N-ni-Ci where p and N are unknown.
Higgins (1985) wrote a programme in BASIC to calculate the (N, p) pairing maximising this
probability and the associated standard error.
Carle and Strub method
Carle and Strub (1978) retain the assumption of a constant catch probability but make it possible
for users to guide the search for p using prior knowledge on the level of catchability of a species, a size
group, etc. Hence, these authors integrate a 2-parameter beta function (α, β) into the Moran (1951) and
Zippin (1956; 1958) equations, allocating a particular weight to each possible value of p. The wide range
of possible (α, β) pairs corresponds to a wide variety of weights allocated to each value of p (uniform,
normal distribution, range of preferred values compared to others, etc.) Starting from this general
theoretical design, they mainly developed the case where α=β=1, which corresponds to an identical
weight being allocated to all catchability levels. Gerdeaux (1987) used this hypothesis when producing a
computerised calculation programme.
Compared to the Moran (1951) and Zippin (1956; 1958) method, the Carle and Strub (1978)
method leads to lower biases and variances for catch probabilities >0.3. It can also provide an estimate
in the case of successive catches where other methods are ineffective (larger numbers in the second
sample for instance). Even though it relies on the ever-fragile hypothesis of catch probability remaining
295

constant, the Carle and Strub method (1978) actually appears more robust than preceding ones. It can
also be easily integrated as a macro-function into a spreadsheet.
Schnute method
Currently, this method is the only one which does not systematically make the assumption that
catch probability remains constant over time. Schnute (1983) suggests testing and comparing three
hypotheses and using the one which best fits the observed data so that an estimate can be made of the
existing stock, including a confidence interval. The three main variants of this method are:
• the standard model where the catch probability remains constant over the runs;
• the model where one probability is used for the first run and then a lower and constant one for
other runs;
• the model where the catch probability decreases monotonically with each run.
Lambert et al (1994) tested this method on 29 trips of 3 runs each targeting eels. In this analysis,
the authors created another variant with a constant probability over 2 runs, and a lower probability on
the third one. Their analysis showed that:
• nearly always, the monotonic decrease was rejected;
• the constant probability held in 82% of cases;
• the variant with a high probability on the first run held in 4 stations;
• the one with a high and constant probability for the first two runs held in only one station.
Despite its interest, dissemination of this type of tool for calculation and analysis has yet to occur.
That kind of method has been also used by Truong and Prouzet (2001) to estimate the upstream run of
adult salmon in the Adour river from catches landed by the professional fisheries in the estuary.
8.4.3. Single runs
Single runs mainly concern point-based surveys and abundance indices. In many contexts, a
successive run depletion procedure cannot be implemented. This is particularly the case in deep and/or
large environments (estuaries, rivers, large waterbodies, connected waterbodies, etc.).
One monitoring strategy option is to sample quickly many random points or sites, over a given
period of time, rather than to collect more precise information on a limited number of stations. Blondel et
al (1970) and Copp (1990) stress that this type of collection strategy is statistically more robust.
These various survey methods correspond to a single run procedure on one site, one “ambiance”
(representative river sector), one habitat or one point and are of two types:
• exhaustive survey of the entire station or of the whole or part of the “ambiances” or habitat
which constitute it (riparian zones, blocks, aquatic plant habitats, riffles, etc.);
• point-based sampling, with a standardised fishing effort, especially in terms of time and surface
area, applied to each point.
296

8.4.3.1. Exhaustive and single run on one station or one
“ambiance” (representative river sector)
Even if in some monitoring cases, the average percentage collected on the first run, during
multiple run procedures, appeared to be significant a posteriori (Callaghan and Mc Carthy, 1992 ;
Lambert et al., 1994 ; Laffaille et al., 2005b), a non-negligible variability was also noted between the
sampled sites. Moreover, Naismith and Knights (1990) showed that a significant number of second runs
resulted in higher catches than the first one in the area they studied. Therefore no general statement
can be made on the efficacy of a single run; and in any event, this procedure cannot highlight a potential
problem of efficacy related to a particular factor (current intensity, water temperature, fishing team, etc.).
When an exhaustive survey of the station is the initial choice, the time gained using a single run
compared to a two or multiple run strategy is not very significant, especially because of the large
amount of time spent (which is not proportional to the number of runs) in setting up the initial fishing and
gauging operation and in tidying at the end. A double run is therefore often justified when undertaking
an exhaustive survey of a station. The issue of the representative river sectors (“ambiances”) that make
up this station is more delicate because the first run invariably disturbs the fish and leads to their
reorganisation within the habitats, so that the second run in a non-isolated “ambiance” no longer has a
genuine meaning, making the results very difficult to interpret particularly in terms of spatial or temporal
comparisons.
8.4.3.2. Point-based fishing
These point-based fishing methods aim either to gain time in a station or to collect a large number
of observations. In both cases, the fishing effort that is applied must be standardised as must be the
surface area of the "point” (Laffaille et al., 2005b).
Point-based surveying of a station
The objective is to obtain an abundance index proportional to the true abundance in a station with
a procedure which is synonymous with time saving. In shallow (less than 60cm) and relatively narrow
(less than 7m) environments, Laffaille et al. (2005b) assessed the relevance of point-based surveying
(1m² of surface sampled per point, with a minimum fishing time of 30 seconds, 1 point for about 5m² of
river) in 35 stations. In the case of the eel, a very good relationship emerged between the abundance
index derived from the sampled points and the density estimated by two successive and exhaustive
runs in these stations. Furthermore, the size structures observed by the two methods were not
significantly different. Unfortunately, this type of definitive calibration is unavailable in other types of
environment but the team is currently working on this problem in association with local technical
structures.
297

Point-based abundance indices
Point-based abundance indices aim to collect data at the scale of one point (or micro-habitat). In
this case, the collected data must relate to catches made and to the description of the sampled point
and of its more or less immediate surrounding environment (Brosse et al., 2001). This procedure is of
great interest for the analysis of species/size/life-stage relationships in relation to the available habitats
(Copp, 1990; Persat and Copp, 1990) in various environments. It is also the only truly efficient
procedure that can be implemented to undertake intra- and inter-sites comparisons in certain large and
deep environments (Nelva et al., 1979).
It does not however provide any means to assess the efficacy and/or the selectivity of the fishing
action at a given point. The characteristics of the electrofishing tools and the degree of eel reaction as
well as its benthic behaviour, mentioned in previous paragraphs, mean that for this species, points
where water levels are above 1.50m are of little importance for its presence or the level of this presence.
On the other hand, point-based surveying of riparian zones makes it possible to collect relevant signals
or indices of the presence of eels, which can be monitored over time in a given zone or can be
compared to indices collected in an identical way on other banks. It seems, therefore, that this pointbased
abundance index technique is not a strategy that will enable the estimation of the abundance of
the species and the various size groups in the large watercourses or in the large and deep
environments of the river catchment. On the other hand, this procedure can be used in the riparian
zones of these deep environments to collect abundance indices for the different size classes, which are
present and whose analysis is of interest (Feunteun et al., 2000).
8.4.4. Overview of the stock assessment methods
In summary, point-based fishing remains the most efficient method for the rapid collection of
information on an existing eel stock at a given station. As with any technique, there are limits to its
efficacy and it is somewhat selective. These characteristics must be integrated into the analysis of the
collected data and/or into the design of the monitoring network.
Hence, in deep environments (water level over 1.50m - 2m), it is clear that only point-based
surveys in riparian zones can be undertaken and that no overall abundance can be estimated in this
type of environment even by mark-recapture.
In shallow environments, if certain procedures are respected, it is possible to estimate the most
likely stocks in the stations surveyed by depletion or mark-recapture and to obtain an element of
comparison between the various sites of a river catchment and of different systems (using the density of
individuals per surface unit).
If, at a given station, the survey strategy (tool and method) and season are standardised,
monitoring the trend of observed or estimated abundance seems relevant and useful. But between sites,
298

it is only conceivable to compare stock estimation results from stations of the same type, using an
identical survey strategy during a similar period. Comparing quantitative inter-site information collected
at different periods of time with different techniques and in very dissimilar environments seems very
dubious. Only occurrence analysis of the species and of the various size groups seems to be relevant at
this observation scale (Lasne and Laffaille, 2008).
8.5. Analysis of results
8.5.1. Approaches in terms of abundance or biomass
8.5.1.1. With respect to a pristine target
The objective (Article 2 – Paragraph 4) of the European regulation establishing measures for the
recovery of the stocks of European eel 17 is “to reduce anthropogenic mortality in order to permit with
high probability the escapement to the sea of at least 40% of the biomass of adult eel relative to the best
estimate of the potential escapement from the river catchment in the absence of human activities
affecting the fishing area or the stock”. Annex 13 gives a worked and theoretical example of this
approach 18 . It is merely an illustration as this target - which refers to the pristine production of silver
eels derived from a strong glass eel total recruitment several years beforehand - has to be a long term
objective which, to be more definite, means returning to and maintaining the 1960s and 1970s
recruitment levels (ICES, 2006).
Although difficult to use in the short and medium terms, it is important to keep this objective in
mind as a reminder that despite the strong recruitments and high biomass levels that existed in the
1960s, the failure to allow a sufficient escapement of good quality silver eels relative to the recruitment
levels at the time certainly contributed to the rapid decline observed later on.
If the recovery plan(s) is (are) to be successful with a return to these recruitment levels in several
decades time, the inland anthropogenic impact must be controlled to comply with the rule that an
adequate escapement of high quality spawners is essential to sustain the dynamics of the species.
8.5.1.2. With respect to historical data
In order to set a target, a second option is possible and consists of referring to historical
observations concerning in particular the yellow stage. Given that current abundance is low and that it
is difficult to determine the pristine production of a catchment and to use it as a reference point, the
objective could be, at least in the early years, a progressive return to the yellow eel biomass and
17 Article (CE) No 1100/2007 of the Council of 18 September 2007, Official Journal of European Union, annex 2 of the Indicang
report, http://ifremer.fr/indicang
18
Rigaud C., Laffaille P. Cible par rapport a un stock pristine, annex 13 of the Indicang Report, http://www.ifremer.fr/indicang.
299

abundance levels observed during the 1960-1970 period or at least 1970-1985. Current management
plans are supposed to include these historical reference levels. These historical data might involve
observed abundance levels or indicators of the species occurrence within the river catchment, or even
indicators of their quality, in particular in terms of sex ratio which is strongly related to the density of
individuals. When there is absolutely no information concerning a catchment, it is possible to extrapolate
from data derived from a nearby and relatively comparable catchment.
It must be noted that this historical situation must of course be clearly distinguished from the
pristine situation mentioned above, as these historical references derive from contexts where
anthropogenic pressures may sometimes be significant. Having said this, and at least to begin with,
these historical references can be used to set meaningful objectives for stakeholders (with respect to
the situations that some stakeholders have experienced). Given the current state of the species, a
return to these more recent abundance or biomass levels associated with an optimal downstream
migration context (where mortality is controlled during the return run to the sea) would surely lead to a
very significant improvement in terms of spawner escapement.
Along the same lines in the case of yellow eels, we must emphasise the interest in, and the
importance of :
• data obtained from electrofishing or counting at barriers implemented during this period;
• long fisheries series which, although not providing abundance, give reference CPUE levels that
can be used to evaluate the trend in the corresponding abundance levels. For further
information on fisheries monitoring data, please refer to the specific report on fisheries
descriptors 19 .
In the case of yellow eels, the way to find historical references in a river catchment differs
according to the compartment under consideration.
In tidal zones that are accessible without having to overcome any barrier, the abundance level is
closely correlated to the total glass eel recruitment at the entry of the estuary and to the anthropogenic
pressures exerted on this fraction of the river catchment stock. In exploited tidal zones, the 1960-70s
CPUEs associated with a type of gear and use may furnish the best historical local target to include
within the framework of the recovery plan.
In the Atlantic coast tidal zones, including these different elements should, in the great
majority of cases, achieve a target of increasing CPUE ten-fold compared to current levels,
whilst, at the same time, leading to the observation of a significant presence of males in these
downstream zones (a minimum sex ratio of 1:1, i.e. around 1/3 of males among the differentiated
individuals of the 30-45cm class).
19
Please see Chapter 6 for further information concerning fisheries indicators.
300

In the rest of the river catchment, we saw that the distribution of individuals depends on many
factors (distance to the tidal limit, attractiveness of each axis, number and passability of barriers,
gradient, fluvial recruitment at the entry of the axis, etc.). In this context, it seems very difficult to identify
a historical reference point of general validity. If there are data over the 1970-1985 period for some
stations, they can be used to set the desired target on these same sites using the same observation
strategy and during the same season. But the number of stations is unlikely to be very significant.
So the primary objective for the next 15 to 20 years must be a continuous and significant
improvement in the occurrence levels of individuals less than 30cm in the river catchment - a size which
reveals the colonisation intensity of a river catchment 20 . The historical upstream limits of the
presence of these young individuals can be established from oral history or written accounts. Finding
the position of old fisheries, usually through administrative records, can also help. For a large-sized
catchment such as the Loire, currently-available accounts place the historical upstream limit of
individuals less than 30cm at about 500 to 600km from the limit of the dynamic tide.
In stations where this size range is already present, and in keeping with the proposition for the
tidal zone, a ten-fold improvement in current densities would be a coherent initial target, keeping, of
course, to the same observation protocols.
These objectives can only be achieved after several decades of structured and co-coordinated
actions, aiming first to profit to the greatest extent possible from current low recruitments and then
gradually to build on the increasingly important recruitments that may emerge.
Specific pass facilities could also provide interesting historical data related to migratory
intensity, but generally they have only very recently been installed in Atlantic catchments (ten years
maximum). As for the implementation of standardised and reliable eel monitoring on these specific
facilities or on multi-specific passes, it is even more recent and insufficiently developed.
Data obtained from passes can never, therefore, be used to provide historical reference points
but instead should be interpreted as indicators of the early signals observed at the very beginning of the
implementation of the local management plan, and it is of even greater interest to observe and analyse
their relative development over the coming years and decades.
In the end, these historical elements concerning the upstream limit of the species’ presence and
abundance levels make it possible to set a primary target level starting from a situation which is known
to be very poor. Their use can concretely raise awareness and provide meaningful objectives to local
stakeholders concerning the efforts required to restore the situation. Realistically it will require several
decades to achieve them, which leaves time to specify optimum targets with respect to the pristine
situation and to improve evaluation techniques.
20
See Chapter 2.
301

8.5.1.3. With respect to the relative evolution of observed
abundance levels
The knowledge acquired on size distribution within habitats and on fishing efficiency concerning
different size ranges and observation contexts (water levels in particular) shows that it is impossible to
compare the collected signals (catch levels, size ranges) in very different stations, using very
heterogeneous sampling strategies. On the other hand, the analysis station by station (or group of
stations by group of stations across homogeneous areas) of the trend in an abundance signal
collected over a given period of time with an unchanged procedure (date, tool, strategy) is of interest
for the visualisation of the relative evolution of the species' status.
This analysis of the complete set of size classes or by size group using the same observation
strategies can be adapted to the high variability of fishing efficiency according to the methods and the
contexts highlighted in the previous report. Only stations where fishing or observation methods have not
varied from one period to the next can be used for such an analysis. The Loire catchment analysis can
be cited as an example. This analysis considered the evolution, by station or by group of stations, of the
signals from each 15cm size group 21 . The maximum catch over a series of monitoring periods was
given a value of 100 and catches from other years were related to this reference point (% compared to
the Maximum Observed or %MO).
If each new year of monitoring gives a new maximum score for a given size group, the trend is
obviously increasing for this group in this station. This can be seen in the example of the Brière marshes
(north of the Loire estuary). This type of analysis also shows that the situation improved significantly
after a pass facility was installed on the downstream construction (figure 8.11). However, the
environment is not yet saturated.
% in relation to the maximum observed
Figure 8.11. Overall trend in relative abundance indices in the Brière sub-catchment between 1995 and
2003 following the installation of a specific pass on the downstream construction [RHP (hydrobiological
and piscicultural network) Onema data; Laffaille and Lasne analysis, Rennes 1 University; source:
Laffaille, Lasne, Steinbach, Vigneron, 2004, Cogepomi Loire].
21 See § >.
Years
302

On the other hand, if the maximum reference corresponds to the first year or the early years of
monitoring, then a declining trend becomes obvious. For example, in the Maine sub-catchment (Loir-
Sarthe-Mayenne), the general trend is decreasing and currently the level is 30% of the maximum
densities observed since 1995 (figure 8.12) and things were already at a very low ebb in Europe in
1995!
% in relation to the maximum observed
Years
Figure 8.12. Overall trend in relative abundance indices between 1995 and 2003 in the Maine sub
catchment (Loir-Sarthe-Mayenne) [RHP (hydrobiological and piscicultural network) Onema data;
Laffaille and Lasne analysis, Rennes 1 University; source: Laffaille, Lasne,
Steinbach, Vigneron, 2004, Cogepomi Loire].
Stable situations can also be identified with some fluctuation in catch levels at a station taking into
account inter-annual variations. This is the case for example in the middle reach of the Loire (figure
8.13) where the abundance index levels off at about 50% of the maximum values observed during the
1995-2003 period.
303

% in relation to the maximum observed
Years
Figure 8.13. Overall trend in relative abundance indices between 1995 and 2003 in the middle
reach of the Loire sub-catchment [RHP (hydrobiological and piscicultural network)
Onema data; Laffaille and Lasne analysis, Rennes 1 University; source: Laffaille,
Lasne, Steinbach, Vigneron, 2004, Cogepomi Loire].
In a second phase, the analysis can focus on the spatial distribution of the trends observed at
station level, by identifying compartments or sub-catchments where the situation is stable, improving or
deteriorating (figure 8.14).
Trends 1995 – 2003 :
no, or erroneous, data (40%)
decreasing (28%)
stable (17%)
increasing (15%)
Figure 8.14. Distribution of trends (all size classes) in the Loire river catchment between 1995
and 2003 [RHP (hydrobiological and piscicultural network) Onema data; Laffaille
and Lasne analysis, Rennes 1 University; source: Laffaille, Lasne, Steinbach,
Vigneron, 2004, Cogepomi Loire].
304

In the case of the Loire catchment, only 30% of the stations monitored between 1995 and 2003
show densities increasing or levelling off. These are situated mainly in Vendean coastal rivers, Sèvre
Niortaise, Sèvre Nantaise, and the downstream and middle reaches of the Loire. Only the downstream
reach of the Loire sub-catchment shows a general increase in densities. All the other sectors indicate
that densities have decreased (30% of stations) or that eels have practically disappeared (40% of
stations): Maine (Loir, Sarthe and Mayenne), Vienne, Creuse, Cher, Indre, upstream reach of the Loire
and Allier.
When there is no electrofishing network, an eel-specific electrofishing network could be set up
in order to monitor the trend in relative abundance of the various size groups.
Except where very small river catchments are concerned, the objective cannot be to get a
representative signal of the situation concerning the entire stock in a river catchment, but rather to adopt
simple basic rules such as:
• distribute stations in the various sub-catchments or zones of the area concerned with the
presence of the different classes of distance to the dynamic tide limit;
• select stations whose characteristics optimise fishing efficiency (maximum width 6 to 7m,
maximum depth 1m but preferably < 60cm) avoiding areas at the base of barriers (potential
accumulation phenomena);
• use a standardised method which is then reproduced on every annual visit to the stations, and
is sustainable in each station, but without solely seeking to obtain densities; abundance
indices such as those described in this chapter 22 , being sufficient (Laffaille et al., 2005a);
• work, if possible, in September, before the fall in temperature and the first heavy rains, in order
to collect information on future silver eels and work with acceptable levels of efficiency 23 .
However, it is not always possible to choose this period. Hence, in dyked coastal marshes, the
summer concentration of the system (increase in water conductivity, covering of lenses, etc.)
makes any intervention problematical beyond the end of June.
A 2-3 yearly review seems sufficient to evaluate the trend which in any case only becomes
apparent over a long period of time. This frequency also makes it possible to establish three groups of
stations over an area, each worked every third year, thereby increasing the number of observation sites
for the same annual monitoring effort.
Data can then be analysed in the same way as previously discussed and the results of
abundance indices can be compared between stations which, this time, will have comparable profiles in
terms of habitat and work protocols.
22 See § >
23 See Chapter 2.
305

A monitoring network by passive gear can also be envisaged, in particular in deep
environments. However, the following must be taken into account 24 (Baisez, 2001):
• gear selectivity related partly to the physical characteristics of the gear, i.e. the mesh size which
sets the minimum catch size (L 100 is around 270mm for a 10cm mesh) and the spacing
between the different funnels which set the maximum catch size, and partly to the behaviour of
the various sizes (greater mobility of large individuals leading to greater catchability);
• the very marked time trend in catch levels (lunar cycle, water movements, storms, temperature
trends, etc.) meaning that the sampling periods used must always be the same;
• the strongly heterogeneous distribution of individuals and the noticeable influence of gear
positioning strategy.
In order to obtain comparable abundance indices from one site to another and from one year to
the next, it is therefore very difficult to use short catch periods (1 to 2 weeks) and a limited number of
sites. By monitoring the results over a season of a fishery operating in one compartment of the
catchment, and taking care to collect size structure data periodically to complement the data on catch
levels, a reliable size index and structure can be obtained for a given compartment of a catchment.
8.5.1.4. Overview of abundance or biomass targets
In all cases, efforts must be pursued over time scales of the order of several decades if these
targets are to be achieved and this dimension must be integrated into the thinking, the strategies and
the action plans to be implemented.
The reference to the pristine situation (years when elvers are abundant and anthropogenic
mortality absent) is quite complex in its application. Research, by compartment and by life stage, on
abundance levels in the 1960-70s, or at least 1970-1985, seems more pragmatic and is sufficient
to set widely-disseminated objectives in a participatory way. Of course care must be taken only to
use reference values derived from observations made in an environment free of sizeable barriers and
not equipped with efficient pass facilities. These physical factors would significantly affect the signals
observed further upstream.
This initial phase of target definition is important but quite often requires special procedures which
can be lengthy. This phase must not hinder the rapid definition of actions aimed at recovering the
species. Simple methods can be used to show the evolution of abundance signals over time and
considering the very low current abundances, the first target to be set could be a significant, positive
and constant evolution of these signals over at least the next ten years. Research on historical
levels by compartment can be carried out in parallel and contribute to the definition of relevant final
targets.
24 See § >.
306

Finally, it must be noted that the use of abundance and biomass data can also help in formulating
diagnoses in particular of barrier passability levels along an axis. Observing the trend in abundance
indicators of particular size classes collected under standardised conditions along an axis can, in this
case, be very useful. It is not so much the reference to a target value but more the spatial and
temporal evolution of the signal which is used for the diagnosis, in this case.
8.5.2. Observation of the occurrence of the species or of size
groups
Targets expressed in terms of abundance or biomass, such as mentioned above, are in
accordance with the spirit of the draft European Regulation aiming to identify quantified and precise
objectives to be achieved.
These quantified targets require, however, the calibration of a certain number of methods and
indicators as well as the collection of a significant quantity of data at the scale of each catchment, within
a short period of time in order to quickly adopt management plans. But, in all cases, the current lack of
knowledge on the distribution of numbers within a catchment's compartments, particularly in deep
areas, significantly compromises any calculation at the scale of the whole inland growth phase and
consequently at the scale of the river catchment.
Therefore, it is highly likely that, in the short term and in order to be efficient and ready to
intervene quickly, managers will have to draw up action plans combining:
• the acquisition of the data necessary to describe the state of the resource and the pressures,
using methods which can identify the largest local black spots to be treated as a priority;
• the implementation of a certain number of actions taking into account in a balanced way, the
various uses which cause direct mortality of the species or which affect, to varying degrees,
water and habitat quality.
Starting from the initial diagnosis of the state of the resource and/or the pressures exerted on it,
monitoring the relative evolution of the situation can be useful to identify the level of effort required to
improve the local status of the species and/or the initial impacts of this approach. Hence, observing the
occurrence of the species or of some size groups can be useful at least in a first stage to visualise the
current state of the resource and the way in which the situation is tending to evolve.
The absence of the species from zones that used to be colonised or the absence very close to
the estuary, of small individuals which are very heavily involved in the colonisation of the river
catchment, are early, sufficiently explicit and unequivocal indicators meaning that quantifying
abundance and biomass might not be necessary, at least to begin with. Seeing the species reappear,
preferably through natural migration, in the upstream zones of a river catchment from where it had
disappeared, or seeing the active colonisation zone, characterised by the presence of individuals
307

smaller than 15cm and between 15 and 30cm, spreading upstream again are the signs that
management is going in the right direction.
This type of simple observation can be very useful at least in the early years. However, it must
also be associated with a defined target, which is set at the beginning of the process and widely
disseminated.
8.5.2.1. Principle
Various studies highlight the fact that the trend in the abundance of juveniles, starting from the
dynamic tide limit, is for the most part related to fluvial recruitment intensity (Smogor et al., 1995;
Ibbotson et al., 2002). The greater is the fluvial recruitment, the further upstream of the zone affected by
the tide before juvenile abundance declines significantly. Locating the zone where these juveniles have
significantly decreased in numbers, or even disappeared, may be an indirect way to measure the trend
in the level of colonisation of the catchment. Furthermore, colonisation also seems to be densitydependent,
i.e. young eels, recently arrived in the catchment, colonise upstream zones when
downstream zones are already occupied (Feunteun et al., 2003). Therefore, by combining the existing
stock in downstream zones and the fluvial recruitment, the upstream distribution limit of these small
individuals can be used as an indicator of space occupation (figure 8.15).
High
Fluvial recruitment
+
existing stock
Average
Low
Estuary Downstream Upstream
Inland
Figure 8.15. Theoretical distribution of small eels (

Hence, the limits of the species’ presence within a river catchment result largely from colonisation
phenomena, which may be more or less recent depending on the age of the individuals. The very likely
hypothesis of a robust relationship, at least in young individuals, between the spatial dimension of the
colonisation, the fluvial recruitment intensity and the existing stock makes it possible to relate the limits
of this presence to past and present colonisation levels.
Smogor et al (1995) concerning the American eel and Ibbotson et al (2002) concerning the
European eel suggest that inland habitat colonisation follows a simple dispersion model. These authors
show that eel abundance decreases along the longitudinal gradient according to a negative exponential
model. However, given the difficulty of obtaining standardised, accurate and precise eel data for the
whole river catchment, these methods cannot really be used in the conservation context, in particular for
eel populations in lentic environments or in large systems (Naismith and Knights, 1990; Jellyman and
Graynoth, 2005). Recent work in conservation biology suggests that data on the species’ presence or
absence (Vojta, 2005; Lasne et al., 2007), associated with logistic models (Oberdorff et al., 2001; Pont
et al., 2005) are appropriate for analysing the trends in the species’ distribution and for intra- and intercatchment
comparisons. This approach was used successfully for migratory species by Eikaas and
MacIntosh (2006) but it has rarely been applied to eels (Broad et al., 2001; Lasne and Laffaille, in
press). Reasoning in terms of occurrence is a way to use results obtained with a broad diversity of
observation methods and in a wide variety of habitats. It does not require accurate abundance estimates
and can rely on operations which are much less demanding in terms of equipment, sector isolation and
multiple runs. However, a minimum standardisation is desirable.
But although the observation of the species or of a size group at a given site can surely testify to
its presence, the reverse is not true, if it is not observed during a survey, its absence from the site
cannot be guaranteed; it may be more difficult to catch because of its scarcity for example. This type of
analysis therefore should be based on a large number of observations, which is statistically more robust
in any case (Laffaille et al., 2005a). Any hypotheses and conclusions reached will be supported more
firmly by the repeated non-detection of the species or of a size group in a catchment zone.
This analysis, based on the species’ presence or absence, is particularly relevant upstream of the
limit(s) of the dynamic tide (LDT) in a river catchment. The distance to this (or these) limit(s) is a simple
way to examine the distribution trend along an axis or in a catchment. Within the framework of this
approach, it is important to identify clearly, on each axis of the catchment, the natural or artificial
(barrier) limit to the influence of the dynamic tide. This limit is important as it makes it possible, first, to
identify the tidal zone within the catchment and, second, to establish reference points from which the
species' presence level may change under the influence of several factors (distance, gradient,
constructions, etc.) as we move upstream. Whatever the object being considered, each sampled point is
identified by its distance to the limit of the dynamic tide and by the presence (value 1) or absence (value
0) of the species or of a given size group of individuals. A logistic analysis is then undertaken using an
309

appropriate computational environment (Excelstat, Statlab, R, etc.). The logistic curve which best fits the
data set can be identified as well as a noteworthy theoretical point corresponding to a probability of 0.5.
The distance to the limit of the dynamic tide at which this probability occurs is compared with either an
absolute reference value or with a historical one (figure 8.16). Multi-year monitoring makes it possible to
evaluate the trend of this colonisation front and to estimate whether the fluvial recruitment associated
with the existing stock is increasing or decreasing. Inter-catchment comparisons can then be used to
estimate the trends in each of the European river catchments.
% Loss % Gain
Occurence probability
Distance to the tidal limit (km)
Figure 8.16. Principle underlying occurrence analysis based on a logistic model (adapted from
Lasne and Laffaille, 2008).
8.5.2.2. Analysis of existing data within a catchment
of:
The presence of yellow eels upstream of the dynamic tide limit in a catchment is mainly the result
• the current and past (20 years or more) intensity of the fluvial recruitment;
• the degree of migratory passability of axes.
Definition of large zones within catchments and areas
An important initial diagnosis must be undertaken using all “recent” (last decade) available data
from a catchment or an area (a group of small catchments or sub-catchments) and by including all
sizes. Hence, all the results from electrofishing, regardless of its original objective (multi-specific
network, WFD network, eel fishing, etc.), from pass monitoring, from commercial or recreational
fisheries’ data, etc. may be used.
The collection of historical data on the species’ presence must also contribute to identifying
concrete objectives in terms of the species returning to zones from which it had disappeared.
310

Four main zones can then be identified in the current eel situation (figures 8.14, 8.15 and 8.17):
the active zone with the established presence of eels less than 300mm.
the colonised zone with the established presence of eels of all sizes.
the potentially recolonisable zone where eels are currently absent but with a significant hosting
capacity and where it is possible to restore free circulation at reasonable cost given the expected
density of individuals.
the zone classed as non colonisable (or inaccessible) situated upstream of pragmatically set
upstream limits (altitude greater than 1,000m, constructions or successive constructions which cannot
be equipped for upstream or downstream migration at reasonable cost given the expected density of
individuals, etc.).
Active zone (10,000 km²)
Colonisable zone (96,000km²)
Inaccessible zone (19,000km²)
Figure 8.17. Active, colonisable and inaccessible zones in the Loire catchment.
Provided the number of observation points is significant relative to the size of the river catchment
(a minimum ratio of one station per 500km 2 of river catchment might be adopted), this spatial analysis
should also make it possible to identify significant differences between the sub-catchments of the
same hydrosystem which may be related in particular to accessibility issues. For example, on the Loire,
the negative impact of a high concentration of barriers to upstream migration in the Maine subcatchment
compared to the rest of the river catchment can be observed very easily (figure 8.18). Low
recruitment can be explained by a very high density of barriers (Lasne and Laffaille, 2008).
311

Occurrence probability :
Size classes:
Figure 8.18. Analysis of mean occurrences between 1995 and 2003 (RHP Onema data) of the
different size groups in the Loire catchment (Lasne and Laffaille, Rennes 1
University, 2006).
Analysis by size class
Following this general analysis, covering all sizes, it seems relevant and useful to use the same
approach by size group. Given the biological knowledge of the species 25 , the presence of individuals
less than 30cm must be examined with particular care. This size range corresponds to individuals
having quite recently entered the catchment (less than 5-6 summers in the inland zone) with a
significant proportion of colonisation behaviour and practically no silvering or downstream migration
behaviour.
The zone where this size class can be found in a river catchment is called the active zone. It is
the part of the hydrosystem where fractions of the existing population continue to be renewed by the
natural arrival of young individuals and where, therefore, a migrant potential exists (figures 8.15, 8.17
and 8.18). This active zone can be limited upstream by major barriers or by a natural limit where the
abundance of migrants is insufficient to induce a journey further upstream. As an example, we can
illustrate the upstream movement of the active zone that could occur in the case of a progressive
improvement in fluvial recruitment intensity (figure 8.15).
Using the previous example, this time illustrated by presence probabilities according to a logistic
model, the negative impact of a high concentration of barriers to upstream migration in the Maine sub-
25
See Chapter 2.
312

catchment compared to the rest of the river catchment can be seen clearly for eels less than 300mm
(figure 8.19) but decreases when the size of eels increases (Lasne and Laffaille, 2008).
Distance to the tidal limit (km)
Figure 8.19. Presence probabilities of eels smaller than 300mm in the Loire Allier and Maine sub
catchments estimated from RHP (hydrobiological and piscicultural network) Onema data
(adapted from Lasne and Lafaille, 2008).
The output of such an approach can take the form of either a distribution map of the different
zones or of a logistic curve covering all the size groups or individuals less than 30cm. When using a
logistic curve for individuals less than 30cm, the active zone then represents the colonisation front.
Importance of temporal trend observation
Temporal trend analysis is also an important element to take into account. Individuals less than
30cm in the Gironde river catchment can be used as an illustration over the 1990-1996 and 1997-2003
periods (figure 8.20).
Presence probability
Presence probability
Occurence probability
Distance to the limit of the
dynamic tide (km)
Distance to the limit of the
dynamic tide (km)
Figure 8.20. Presence probability of eels smaller than 30cm in the Garonne and Dordogne sub
catchments estimated from RHP (hydrobiological and piscicultural network) Onema data
(Cemagref and Onema, 2006).
313

Between the two periods, a change in the signal recorded on the Garonne axis can be seen, most
probably due to the installation of a specific pass system at Golfech in 2002. It must also be noted that
the distance to the limit of the dynamic tide corresponding to a probability of 0.5 (D 0.5 ) varies, according
to the periods and the axes, between 110 and 150km which seems low given historical observations on
these sizes.
When monitoring the desired recovery of the species in this catchment, the focus must be on
monitoring the movement upstream of the D 0.5 of individuals less than 30cm, which represent the
colonisation front. The ongoing analysis of historical data should make it possible to set a target but a
minimum gain of 100km can probably be defined already.
Importance of inter-catchment comparisons
By comparing the colonisation front, inter-catchment comparisons can easily be made and biases
reduced so far as possible. For example, on the Loire, specific sampling evenly spaced between 0 and
130km from the zone of the dynamic tide shows that eels < 300mm have a presence probability of 0.5
(D 0.5 ) at 90km from the limit of the dynamic tide. This monitoring leads to the conclusion that currently
there a 50% chance of observing individuals less than 30cm at 90km from the tidal zone limit on the
Loire axis. On the Aulne (West Brittany), the probability of observing these young colonising individuals
drops from the 40 th kilometre. This difference shows the negative impact on upstream eel migration of
the many dams along the Aulne, unlike the Loire which is free of them for over half of its length from the
sea. Similarly, the probability of young eel occurrence is equal to 1 in the Loire downstream sectors
whilst it is lower on the Aulne meaning that the downstream part of this river catchment is undersaturated
and the fluvial recruitment is low (figure 8.21).
It must also be noted that, on the Loire, which is free of barriers to migration, the fact that
individuals less than 30cm have practically disappeared around 120-130km from the limit of the dynamic
tide is totally abnormal. As previously shown in the knowledge summary 26 , high concentrations of
individuals less than 10-20cm were still observed on this axis in Orléans some 300km from the limit of
the dynamic tide only 25-30 years ago. This means that the fluvial recruitment is completely inadequate
leading to low upstream migration of young individuals in the river catchment.
26
See Chapter 2.
314

Presence probability
Distance to the tidal limit (km)
Distance to the tidal limit (km)
Figure 8.21. Probability of occurrence for eels less than 300mm on the Loire and the Aulne
(Laffaille and Lasne, Rennes 1 University, unpublished data).
8.5.2.3. Outlook in terms of management
When analysing the colonisation front, two essential parameters must be measured, representing
the two following targets:
• D 0.5 which corresponds to the distance from the tidal limit where the probability of observing eels
less than 30cm is equal to 0.5. It is a colonisation and accessibility index;
• the probability of downstream occurrence (the value of the ordinate at the origin of the
abscissa), which measures the “saturation” in the downstream zone and is an index of the
fluvial recruitment and of the existing downstream stock.
Monitoring the colonisation front and the two associated parameters, approximately every third
year, is a simple way to evaluate the impact of management in terms of fluvial recruitment and existing
stock intensity. The desired targets remain to be defined in each catchment.
Given the historical data currently available and the initial results of a predictive model
evaluating the colonisation front in the Loire in the absence of barriers (Laffaille and Lasne,
unpublished data), it seems that:
• a D 0.5 at 150km from the limit of the dynamic tide for eels less than 15cm and at 300km for eels
less than 30cm;
• and a probability of occurrence downstream equal to 1;
could become the targets. Attaining them will require significant efforts in the
management of the colonising flux and of the barriers to migration.
All this information about an axis must be considered both as diagnostic elements (abnormal
situation when D 0.5 is too close to the limit of the dynamic tide and a downstream D ≠ 1, given the
historical data on the presence of this size group) and as initial reference points for the evaluation of the
impact of management measures in the years and decades to come.
315

8.6. Interest of, and need for, specific approaches for small sizes
In certain zones of a river catchment, and even in the whole river catchment, there can be a total
absence of standard and non-specific electrofishing data. This is often the case on the large river axes
which are difficult to access or along small tributaries which are part of an overall management area
such as Brittany. In such a context, it is essential and important to initiate a monitoring network,
particularly of small sizes for all the reasons mentioned above.
Moreover, the selective nature of the information available particularly on individuals smaller than
30cm and even more on those smaller than 15cm (on average less than 2 inland summers) must be
noted; yet these groups are likely to provide a rapid indication of the impact that decision-making has
had on the colonisation phase in the downstream part of the catchment.
Hence, in the Gironde-Garonne-Dordogne river catchment, the analysis of the 1,403 non-specific
electrofishing operations [RHP (hydrobiological and piscicultural network) and others] carried out over
26 years in the river catchment (1977-2003) by Onema shows a significant selectivity of individuals
smaller than 15cm (figure 8.22) despite the fact that over half of the operations were carried out in the
downstream reach of the catchment.
Percentage of catches (%)
Size classes (mm)
Figure 8.22. Size structure of all eel catches of the CSP (Fisheries Higher Council) in the Gironde
Garonne-Dordogne catchment from 1977 to 2003.
As indicated earlier in this chapter 27 , this observation, derived from non-specific fishing in various
contexts, particularly different water levels, may arise from several factors:
• reduced sensitivity of these small sizes to electric current;
• greater difficulty in detection;
• poorly-adapted equipment (mesh size of hand nets and nets);
• non-systematic exploration of all habitats including the more shallow ones;
• too fast an exploration;
316

• etc.
Whatever their objective, specific operations must be carried out based on the principle of
working in shallow environments (less than 60cm deep). This is because, if eels less than 30cm are
present in a sector, they are present in these shallow environments which are sought-after habitats for
this particular size group (Laffaille et al., 2003), especially in riparian zones and in zones affected by the
current and offering some shelter (figure 8.23).
Deep pool
Percentage
Pool – riparian
zones
Run-Riffle
Size classes (mm)
Figure 8.23. Prevalence of individuals based on size and on the habitats explored (Cemagref,
2006).
Moreover, these shallow zones allow efficient fishing. This was found to be the case on
successive runs over such habitat at the base of constructions (figure 8.24).
Percentage
Size classes (mm)
Figure 8.24. Mean and quartile of fishing efficiency observed by size class in 2006 during
specific operations in shallow environments at the foot of dams (Cemagref, 2006).
Given the particular distribution of the smallest eels within habitats, the need to work with smallmeshed
hand nets (1-2mm), and the importance of the exploration strategy (slow progression on all
shallow habitat), we have seen that practically all standard electrofishing operations provide significantly
27 See § >.
317

iased information on the presence level of small eel size classes. The same is true in Atlantic
catchments where the mesh size of the passive gear used cannot detect them. Only specific pass
devices would be likely to provide information but they are still scarce within catchments and in any case
can rarely be found on successive constructions along the same axis.
Furthermore, there are many sampling methods (see for example Brosse et al., 2001); the most
common one is electrofishing (Lambert et al.,1994; Feunteun et al., 1998), associated with stock
estimation measurements by successive runs without release (Cowx, 1983). However, these are
biased, inaccurate (Zalewski, 1985; Cowx, 1983), labour-intensive and time-consuming (Lobon-Cervia
and Utrilla 1993; Prévost and Baglinière, 1995). Various specific monitoring strategies for these young
individuals (less than 30cm) can be envisaged over an area or along an axis.
For example, a specific methodology was developed for salmonid zones (Laffaille et al., 2005b)
which provides standardised data with a high probability of catching the smaller stages whilst reducing
cost and effort. The abundance point sampling technique was modified (EPA - Nelva et al., 1979) and
adapted to catch eels in small, shallow rivers. In addition to the fact that it can be carried out quickly and
at low cost, this sampling methodology seems to be more efficient because the multiplication of
numerous small samples (such as point samples) provides a much more precise and robust stock
estimate than a small number of large samples (Blondel et al., 1970; Copp, 1990). This sampling
technique was validated by comparison with a standard methodology using the Carle and Strub
estimator (1978). It proved to be highly predictive for densities < 150 individuals/100m -2 , regardless of
the river catchment considered, for habitats such as riffles, runs and pools (figure 8.25).
Predicted densities
Observed densities
Figure 8.25. Eel densities (in number per 100m 2 ) in the Fremur river catchment predicted using
the linear model developed in the Vilaine river catchment compared to densities
estimated by the Carle and Strub method (1998) (Laffaille et al., unpublished data).
Moreover, this methodology was adapted to different environments. Hence, on the Loire,
abundance point sampling (EPA) has been carried out by University of Rennes 1 in the shallow riparian
318

zones of 50 cut-off meanders regularly distributed between 0 and 130km from the limit of the dynamic
tide (figures 8.21 and 8.26).
Eel relative densities
Size classes
Distance to the tidal limit (kms)
Figure 8.26. Relative densities of different-sized eels in relation to the distance from the limit of
the dynamic tide in the Loire catchment (Lasne and Laffaille, Rennes 1 University).
Similar methods were developed along the barrier-free Garonne and Dordogne axes on the first
100km from the limits of the dynamic tide. An observation network (Migado, Cemagref, CSP) was set up
at the foot of the first constructions (zone where smaller eels concentrate) along around 30 small
tributaries which merge with the two major axes concerned over a distance spreading from the limit of
the dynamic tide to 100km upstream. The size group targeted comprised individuals less than 15cm
that, as mentioned previously, are poorly detected by the standard observation network. Double-run
surveys were undertaken using the “Martin-pêcheur” (the “Kingfisher”: a portable electrofishing unit) in
shallow environments located within a 50m radius downstream of each construction. Each site
represented 2 hours work for 4 people and the total amount of fishing corresponded to some 30 persondays,
an amount that must be doubled if time for exploration, administration and analysis is included.
The signals collected in June and in September-October at the same stations are not significantly
different, giving some freedom as to the season in which to intervene. Similar results were obtained on
the Loire. Maximum densities were observed in tidal stations with 4 individuals/m². The levels observed
then decreased rapidly at the entry of the various sampled tributaries the greater the distance from the
limit of the dynamic tide and became practically nil at 30-40km from this limit (figure 8.27) which again
means a totally insignificant current fluvial recruitment.
319

Presence probability
Distance to the limit of the dynamic tide (km)
Figure 8.27. Evolution of the probability of presence of eels less than 15cm in relation to the
distance to the limits of the dynamic tide on the Garonne and Dordogne axes
(Cemagref, 2004-2005).
In any event, it is clear that it is possible and of interest to set up a specific network for small sizes
in zones where data are insufficient or in order to refine the information available on individuals less
than 30cm and especially less than 15cm, a size which is difficult to detect by standard electrofishing.
Currently, this specific approach, centered on a size group which is difficult to detect with
standard observation strategies, has:
• demonstrated the efficiency of fishing procedures and substantiated the choice of the habitat
explored (shallow), of the tools used and of the type of survey carried out.
• identified optimal intervention periods. There seems to be a close relationship between
migratory intensity and the existence of attraction flows in the river. As soon as this attraction
flow decreases (June in the Gironde in 2005-2006), presence boundaries change very little
and autumn observations do not differ from those made in June.
• highlighted the very small difference that existed between seasons even though these showed
slightly different glass eel CPUE in the estuaries. This outcome confirms, if needs be, the
weak relationship between these glass eel CPUE and the level of fluvial recruitment in the
catchment, but also probably the very low saturation of the tidal zone.
• collected what can be considered to be initial reference levels against which the impact of
management measures in terms of fluvial recruitment can be assessed.
Returning every second or even third year in each selected station is in fact sufficient to assess
the impact of management in terms of the relative trend of observed densities. This type of information
in the downstream zone must, of course, be related to the monitoring of these young individuals’
upstream progression.
320

8.7. Synopsis and links with other indicators
The synopsis of available yellow eel information 28 highlights the complexity of this long growth
phase which occurs within very diverse compartments and habitats inside a river catchment, and
encompasses a significant variety of behaviours.
It has also emphasised the fact that both the quantitative characteristics of the existing
population fractions in an area and their qualitative characteristics (sex ratio, size and age structures,
growth stage, size-weight relationship, contamination levels, etc.) must be taken into account.
It is easy to see how these different factors, which must also include the temporal dimension of
the observation, will affect the yellow eel data collected at a given site (figure 8.28). As regards the
relationship between the sampling tool and the species (accessibility, vulnerability, efficiency), please
refer to the summary of fisheries descriptors 29 .
These many sources of variability in the signal observed should certainly not be considered
to be a barrier to all kinds of stock monitoring or estimation. They must simply be kept in mind and
integrated into the design of monitoring arrangements or strategies of this life stage in a river catchment.
This aspect is developed in this chapter. Hence, the fact that it is currently impossible to estimate
the existing stock in a river catchment must be noted; it is due on the one hand to the quasi-total
absence of abundance data from deep environments and on the other hand to the uncertainty of the
relevant weighting factor for the average signal collected in each compartment (taking into account
respective watercourse lengths, respective areas, etc.).
However, the analysis, station by station or compartment by compartment, of the evolution of the
signal collected over a given period using the same standardised procedure (date, tool, strategy) is of
great interest to visualise the relative evolution of the species in a catchment, a sub-catchment or a
given compartment. This approach could then be followed by a spatial analysis (identification of
catchment zones which evolve in a similar way). Whatever the approach adopted, it is important to
have access to a synoptic and reliable description of the environment where the species can be
found in a catchment, particularly in terms of colonisation, growth and downstream migration contexts.
To conclude, monitoring the yellow stage in a river catchment must allow:
• on the one hand, an initial diagnosis in terms of the status of the species in its growth
phase in the various catchment compartments and of the anthropogenic pressures exerted in
some of these compartments.
28
See Chapter 2.
29
See Chapter 6.
321

• on the other hand, the monitoring of how the situation is developing towards the chosen
target or objective. The pristine target appears to be the most relevant in terms of biology but
it is currently difficult to estimate in river catchments and management zones (unknown
pristine recruitment, natural mortality poorly known) and to use in the short term.
It seems therefore logical to aim at an intermediate target, which goes some way towards this
pristine objective. Local historical reference points (abundance levels, range of the species and of
young individuals) may contribute to the identification of this intermediate target. In the case of
abundance indices based on CPUE, the significant evolution of the context (type of gear used, number
of fishers, zone exploited by each fisher, etc.) must be kept in mind as it makes any comparison
between 30-year-old data and current data quite difficult. In the absence of abundance reference points,
a pertinent intermediate target may be a 10-fold improvement in all the compartments which are still
currently populated. This non-optimal intermediate target would take four or five decades to achieve
according to the most optimistic estimates. In any case, meeting this intermediate objective would
indicate a radical change in the species’ management and a reversal of the trend observed for at least
25 years as well as the first tangible signs of stock recovery.
Knowledge of how the yellow stage of eel works (river catchment colonisation and
sedentarisation) leads to four key conclusions:
• it is currently impossible to estimate reliably the existing yellow stock in a river catchment,
particularly in the deep zones.
• no recovery is possible in the river catchment without improvement in the downstream zones,
particularly those affected by the tide. These downstream zones must be monitored
particularly closely;
• starting from the particularly poor current situation, catchment recovery upstream of the limits of
the dynamic tide will begin with the recovery of fluvial recruitment. This phenomenon will have
to be monitored closely with appropriate methods;
• finally, as well as restoring significant presence and abundance levels inside the river
catchment, urgent attention must be paid to the individuals that are ready to migrate
downstream in the coming years by offering them a downstream migration context which is as
favourable as possible.
Table 8.2 summarises the different actions presented in this document with their area of
application, their objective, their major constraints and an estimate of their cost. They are then prioritised
in order to identify the minimal set of actions to implement within a river catchment and the optimal set.
322

Observation site
(habitats diversity, position within river
catchment or along an axis etc)
Gear used
(intrinsic selectivity and efficacy)
Level of presence of the species on a site or along an axis in a
nested relationship with:
• the abundance at the entry of the axis
• the location of the axis within the catchment (distance to the tidal
limit, number of barriers, etc.)
• the site the site characteristics (distance from the site to the entry
of the axis, number and passability of barriers, gradients, etc.)
• the quality of available habitats within (habitats, etc.).
Observed signal
(abundance, size, sex-ratio etc)
Yellow eel
(each size with its own features in terms of
activity rhythm, of behaviour, of distribution
t t )
Operator
(gear use strategy)
Season/Date
(temperature, trophic resource, flow etc)
Figure 8.28. Factors affecting the “yellow eel” signal observed at a given site and on a given
date.
323

Table 8.2. Diagnoses and potential indicators concerning the yellow eel stage
Relevant
zone
Theme Possible actions Constraints Frequency
Cost
elements
Priority level
Historical
references
Historical CPUE
Data on size structures
and sex-ratio
-Years preceding 1985 - Description
of the period’s context (number of
fishers, type of gear, fishing strategy,
etc.)
Initial diagnosis
Mainly detection of sizes >75 cm Initial diagnosis
Optional as the most
relevant intermediate
target = current CPUE
x 10
Optional but of
interest
Abundance indices/CPUE
monitoring
By zone and by type of gear and
metier
Yearly
monitoring
3 days a year
per monitored
fisher
Compulsory if fishery
See fisheries
descriptors
Tidal
zones
Current state and
trend monitoring
Objective:
- 10-fold increase in
CPUE compared to
current levels
– Re-emergence of
large sizes
- Minimum 1/3 males
in 30-45cm
differentiated
individuals
- Size and weight structure
- External health condition
- Silvering indices
- Sex-ratio
- Age
- Internal health condition
- Chemical contamination
- Grafted onto CPUE monitoring
with, if possible, 2 runs (springbeginning
of summer/September-
October). Silvering indices in
Autumn. 300 individuals minimum by
compartment (salt, fluvial tidal)
- 75 to 100 eels of 30-45cm (5 to 6
per cm)
- No freezing of gonads
Every third year
by a specialised
laboratory
15 person-days
every 2 years by
compartment
3/4 hour per
individual
3/4 hour per
individual
1/2 hour per
individual
Compulsory
Compulsory
Compulsory
Optional
See box on
environment
Evaluation of
anthropogenic
pressures
Analysis of the size
structure
Use of size structure and sex-ratio
data
Every second
year by a
specialised team
Contribution to the
calibration network
for the analytical
method (size / %
SPR)
Zones
upstream
of the
d i
Historical
references
Historical abundances
- Identification of data from stations
pre-1985 with reference to the
estimation method (passive gear in
deep zones, electrofishing) and to
the season of observation
Initial diagnosis
2 months /
10,000 km²
Compulsory
324

dynamic
tide limit
Historical occurrence
- Identification of the upstream limits
reached by the species on the one
hand and by individuals less than
30cm on the other hand
- Work on oral history or written
accounts as well as on fisheries
administration archives
Initial diagnosis
2 months /
10,000 km²
Compulsory
- In deep zones, CPUE monitoring, if
existing fishery the same procedure
is applied as in the tidal zone
Yearly CPUE
monitoring, twoyearly
for sizeweight
and
silvering indices
Tidal zone
Compulsory if fishery
State of the
existing stock and
monitoring
Evolution of abundance
indices (intermediate
target = current
abundance of different
size groups x 10 in
stations where they are
currently present)
- On stations where historical data
available, monitoring using the same
method
On existing multi-specific network
(minimum of 1 station per 500 km²),
analysis of the relative evolution of
observed abundance with
unchanged strategy (date,
technique). Spatial analysis of
observed trends
2-3 yearly
monitoring
According to
frequency of
network (2-3
yearly
recommended)
3 person-days
per site
% of cost of
existing network,
to be estimated
Compulsory if
historical stations exist
Compulsory analysis
if existing network
- If there is no network, one site per
500km² is set up including stations
where efficient electrofishing is
possible and that are evenly
distributed between the various subcatchments
and between the various
classes of distances from the
dynamic tide.
2-3 yearly
monitoring
2 person-days
per site per year
Compulsory if there is
no existing network
State of fluvial
recruitment and
monitoring
Monitoring of passes
close to the limits of the
dynamic tide
- Daily monitoring from April to July-
Size structure once a week (100
individuals minimum)- Monitoring
the number of passes expressed in
km² upstream of the site and the
type of passes
Yearly
monitoring
40 person-days
per site per year
Compulsory
325

Specific fishing : < 15 et
15-30 cm
Objective:
Upstream progression of
the 0.5 probability to observe
the size group
and
Occurrence probability
equal to 1 downstream
and
Significant increase (10 fold
minimum) in abundance
indices in the more
downstream sites (related to
fluvial recruitment) and of
abundances along the
observed axis (good
passability for migrants)
- Equipment ad hoc (mesh 1.5-2mm)
- Observations from mid-June
- 15 sites minimum by axis
- Monitoring of shallow zones (less
than 60cm of water) at intervals from
the limit of the dynamic tide to
150km upstream (banks of cut-off
meanders, foot of first barriers, small
tributaries)
- Work based either on
representative river sector sampling
or on stock estimation
- Maintenance of the same strategy
when returning
- Analysis of abundance indices
obtained and/or of the trend of the
presence probability of the size
group
- Separate analysis of

Chapter 9
Indicators of potential spawner
escapement
Eric Feunteun, Tony Robinet, Javier Lobon-Cervia, Pauline Boury,
Philippe Boisneau, Anthony Acou
327

9.1. Introduction
9.1.1. The silver eel issue
“ Review of the available information on the status of the stock and fisheries of the European eel
supports the view that the population as a whole has declined in most of the distribution area, that the
stock is outside safe biological limits and that current fisheries are not sustainable. Recruitment is at a
historical minimum and most recent observations do not indicate recovery. The level observed since
1990 is below 20% of the level observed not more than three generations ago” (ICES 2006).
In line with the proposal of the GRISAM eel group (2004, 2006), which reiterated some of the
points – expressed in terms of the precautionary approach (precautionary target and biological limit)–
put forward by the WGEEL (ICES, 2005, 2006) of the ICES/EIFAC working group (Garcia and De
Leiva Moreno 2005), this document presents the initial stage of the work of the INDICANG European
programme on the silver eel thematic box. This work seeks to construct the indicators required for the
definition of the management targets concerning the seawards escapement of the silver European eel
Anguilla anguilla and for their subsequent evaluation.
It does not seem possible that, in the foreseeable future, the number of spawners per recruit
can be assessed in the absence of any anthropogenic disruption. However, it is possible to
concentrate efforts on estimating the levels of spawner production at the scale of the river catchment,
given the various anthropogenic impacts affecting the catchment, and to identify the anthropogenic
mortality factors that affect this life stage.
The method currently used in the INDICANG programme in order to understand the impact of
anthropogenic activities on the spawner escapement rate is to compare this rate in a sufficiently large
number of catchments characterised by a variety of anthropogenic disruptions. These catchments are
reference catchments where estimates of target objectives may be standardised and then applied to
other catchments.
As silver eel production is the primary management target (ICES, 2005), it is essential to be
able to estimate quickly the proportion of individuals that escape from a river catchment by comparing
it to the theoretical spawner potential of the same catchment. The objective of the "silver eel" group is
therefore to establish indicators of the silver eel reproductive potential and escapement by reference
catchment, integrating the geophysical context (size and nature of the river catchment), the scientific
context (ongoing studies) and the administrative context (political and financial support).
It is agreed that, in river catchments where the available information does not allow for the
production of quantitative indicators (e.g. biomass production levels), the objective is simply to suggest
a temporal monitoring methodology for the relative variation in production levels and/or escapement,
from one year to the next for example, without referring quantitatively to biomass. Within the
328

framework of the INDICANG programme, the targets for these indicators will be adapted to the
available information.
The first stage 1 outlined the existing information and available methodologies enabling both
quantitative and qualitative monitoring of these estimates (Figure 9.1) and the combination of these
monitoring systems with the analysis of direct mortality due to the sum of human activities, according
to the specific context of each catchment.
WHAT LEVEL OF
REPRODUCTIVE
POTENTIAL?
3 estimation methods:
• Monitoring through
fishing and trapping
• Direct estimation (markrecapture)
• Indirect estimation (local
population structure)
WHAT SPAWNER
QUALITY?
1
WHAT DIRECT MORTALITY
OF HUMAN ORIGIN?
• Sex ratio, age at maturity
• Female fecundity
• Energy reserves (condition
index)
• Contamination rate
(pollutants, parasites)
2
SILVER EEL
ESCAPEMENT
FROM THE
RIVER CATCHMENT
3
• Hydroelectric turbines
• Fishing (all life stages)
• Reduced access to and
destruction of habitats
• Acute pollution and
problems related to
water quality
What is the true extent of
escapement compared to
the production level of the
river catchment?
(anthropogenic mortality
at this stage)
Figure 9.1 - In an ideal situation, silver eel escapement from a river catchment is correlated
with its potential spawner production (1) and their quality (2) but also with direct
anthropogenic mortality which reduces this escapement (3).
1
See Chapter 2
329

9.1.2. Indicators to indicate what?
In catchments where this is possible, three types of indicator are targeted:
“Reproductive potential” indicator. The methods used to assess the reproductive potential of a
river catchment, i.e. its silver eel productive capacity, must be adapted to the possibilities offered
by each river catchment (monitoring and/or fisheries or the geophysical context alone).
“Quality” indicator. The quality of silver eels determines their ability to reproduce.
They need the necessary energetic resources to migrate between Europe and their
spawning grounds and low contamination levels to produce viable gametes and larvae.
“Mortality of human origin” indicator concerning the silver stage. Estimate silver
eel mortality which is directly induced by the sum of human activities (hydro-electrical
turbines, fishing, acute pollution, even loss of habitat, barriers to migration…).
The construction of these indicators encompasses the particular characteristics of the
monitoring methods adopted in each reference catchment as well as the impact of local hydroclimatic
variations (temperature, water levels, atmospheric pressure, etc.) which play a preponderant role in
the annual escapement rate. It is currently difficult to understand the impact of such variations as there
is no standardised method to assess these indicators at the regional scale covered by the INDICANG
project, let alone at the scale of the European population. It remains, therefore, to define which
relevant descriptors are available in reference catchments, and then to implement regional synoptic
indicators integrating the descriptors derived from the different types of monitoring and information
levels.
9.2. Study scale: the river catchment
The characteristics of the river catchment where eels grow from the time of their fluvial
recruitment until their silver metamorphosis comprise the Catchment Context. Given the determinant
influence of the environment on the sex-ratio 2 , it is of course the decisive factor affecting the
population structure of silver eels produced by the catchment and therefore its reproductive potential.
The Catchment Context also includes the chemical effects (water and sediment pollution) that may
potentially jeopardise successful reproduction in the future.
The first stage of our approach is to describe simply and accurately the “Catchment Context”
and to understand the extent to which it affects the reproductive potential, calculated from the existing
eel population structure. This relationship should make it possible to identify, as clearly as possible,
the catchment factors that are responsible for a reduction in the (quantitative) Reproductive
Potential, as well as in the (quantitative) Escapement and the quality of the future spawners. The
approach establishes targets which aim to improve the quantity and the quality of silver eel
escapement (Feunteun, 2002; GRISAM, 2004 and 2006).
2
See Chapter 2.
330

Furthermore, the approach may be reversed if there are no catchment data on the migratory or
existing population structure. It is then simply the catchment context that gives some indication of
spawner reproductive and escapement potentials and their quality.
9.3. Building the indicators
The panel of indicators must provide a common methodology to estimate the Reproductive and
Escapement potentials and the Quality of silver eels in coastal and inland hydrosystems.
This methodology must be:
sensitive to changes in silver eel production levels and to the impact of management measures;
adaptable to each river catchment;
transposable to different scales, from local catchments and sub-catchments to regions (e.g. the
Bay of Biscay region) by providing a typology of the relationships between the catchment and the
silver eel indicators;
economically sustainable.
Furthermore, the indicators must provide information on the reproductive potential and the
effective reproductive biomass:
from living individuals:
relative and/or absolute spawner abundance;
maturity index (Pankhurst);
sex-ratio (based on the size of silver eels);
condition coefficient (based on individual size/weight).
from sacrificed individuals:
parasite load;
contaminant load;
age at migration.
9.3.1. Reproductive potential
9.3.1.1. Context and objective
A biologist seeking to predict the annual volume of silver eels migrating downstream in a
catchment is faced with the unavoidable climate stochasticity making any accurate forecast impossible
for a given year (Lambert, 2005). For this reason, it is preferable to base a river catchment production
level on the catchment Reproductive Potential instead of trying to estimate the true volume migrating
downstream (escapement) each year.
The reproductive potential reflects the growth conditions for yellow eels which depend on the
catchment context. The latter can be estimated from the data concerning the existing eel population.
This reproductive potential represents the number of silver eels migrating downstream from the
331

catchment that would be observed each year if downstream migration did not depend on
unpredictable climatic factors and if there were no anthropogenic mortality.
9.3.1.2. Data acquisition
To begin with, an inventory of abundance evaluation methods was developed within the
INDICANG network of reference catchments (and also outside of this network). A rapid overview
identified 2 categories of indicators which correspond to the level of information available on the silver
stage (poor or significant):
category 1. monitoring downstream migration (3 methods, significant level of information).
These methods are based on the more or less complete characterisation of the number of
individuals migrating downstream in a given location of the catchment and at a given time;
category 2. monitoring the existing stock (2 methods, little information).
When monitoring of the silver eel downstream migratory phase is not possible, information can be
derived from monitoring the existing stock.
9.3.1.3. Data exploitation
Category 1: monitoring downstream migration
Exhaustive monitoring
Monitoring downstream migration using traps or other devices placed on dams, barriers or weirs
or simply on rivers (watermill grates for example) enables practically all downstream migratory eels to
be intercepted, according to examples given in the literature (e.g. Vøllestad and Jonsson, 1988;
Feunteun and al., 2000). Experimental fisheries can be established to catch migratory silver eels at a
time and a place in the catchment. Such monitoring makes it possible to quantify the total silver eel
production of the river catchment located above a particular trap, during its operational period. If the
trap is located downstream in the river catchment and is not very size-selective, the estimated yield is
then relatively close to that of the whole river catchment. In some cases, it can be helpful to use
complementary methods to estimate escapement from the reproductive potential.
This is the case for a fish trap located on the Frémur (North Brittany, France) where annual
monitoring provides the structure (age, size and sex ratio) of the silver eel population leaving the
catchment and its abundance (number and biomass). The study began on 1st January 1997 and
continues thanks to a monitoring protocol set up by the Fish Pass company which is managed by local
authorities or by the Ile and Vilaine fishing federation, depending on the year.
At the beginning of their reproductive migration, silver eels pass through the downstream
migration trap located at the Pont es Omnès dam, about 6 km from the estuary. This trap was
designed in September 1996 by the Fish Pass company and became operational in 1997.
Downstream migratory eels longer than 200 mm can be caught (which is significantly smaller than the
smallest silver eels). The downstream migration trap consists of a grid (inter-space = 8 mm) which
traps eels migrating upstream and makes it possible to catch eels passing the cofferdams. The eels
332

are then directed along a channel taking them to a catch box (figure 9.2), which is fixed to a post and
can be winched up. Water falls into the trap by gravity from a water inlet in the spillway and the trap
operates non-stop.
Figure 9.2 - The "Pont es Omnes” dam (4m high) equipped with an upstream fish pass/trap
(on the left bank) and a downstream trap (on the right bank) (photo : E. Feunteun).
Between 1997 and December 2004, 1,246 trapping sessions were carried out, each lasting
between 1 and 10 days. {>The average time interval between two trappings was 2.5 ± 1.4 days.
During the migratory peaks, which occur between November and April, the downstream trap was lifted
every day. All the eels caught were anaesthetised with Eugenol and their size measured to the
nearest millimetre. They were released back into the water below the dam. The efficacy of the trap
was tested. It was found to catch 100 % of downstream migratory eels, except during the 1999
exceptional decennial floods during which the % of eels escaping was very high for 2 days at the time
of the downstream migratory peak.
Table 9.1 - Number of silver eels caught in the Frémur river catchment fish trap (Acou, 2006).
Year 1997 1998 1999 2000 2001 2002 2003 2004 Total
Number 550 676 1,101 705 392 366 517 299 4,606
This monitoring indicates that the number of silver eels caught in the river catchment varies from
299 to 1,101 individuals depending on the year. The table shows that the number of downstream
migratory eels increased until 1999 and has decreased since. Given this apparent population decline,
more extensive studies have revealed that these numbers depend principally on hydroclimatic
parameters (Acou, 2006).
333

Silver eel sex determination
Sex determination of a sample of 130 silver eels (min/max size: 319 - 794mm) caught in the
downstream trap between 2002 and 2004 was undertaken (figure 9.3). Sex was determined by the
macroscopic observation of paired gonads, according to the criteria described by Syrski (1976) for
males and by Colombo et al. (1984) for females. The female gonads are ribbon-shaped and develop
as the eel grows. The male gonad looks like a flattened filament containing lobular structures. In order
to "reveal" the gonads by “coagulation” of their albumen, the tissue surface was soaked with surgical
spirit.
100
80
Numbers (in %)
60
40
20
Males
Females
0
300 360 420 480 540 600 660 720 780 840
Size classes (mm)
Figure 9.3 - Evolution of the sex-ratio by size class of silver eels caught in the Pont es Omnes
downstream trap (♂ = 63 ; ♀ = 67).
The silver eel sex ratio changed during the study period falling from about 80% to less than 65%
of males between 1997 and 2004 (Acou, 2006; Laffaille et al., 2006).
334

Proportion of silver males (%)
Years
Figure 9.4 - Proportion of male silver eels observed in the sedentary fraction (electrofishing
and fyke net, white columns) and migrating downstream in the downstream trap
of Pont es Omnes (black columns).
Age estimation
The age of downstream migratory eels was estimated by using otolithometry to examine the
saggitae otoliths. This is based on counting and interpreting the hyaline rings around the nucleus
(Mounaix, 1992). The age is expressed as the number of years spent in inland waters and relates to
the number of annuli found on the otolith (Mounaix, 1992). These age estimates were validated by
comparing information from mass staining with calcein and individual PIT-tagging (Guillouët et al.,
2005). Ageing of eels migrating downstream was undertaken on 88 silver eels (size interval: 320 mm –
794 mm).
The youngest individuals migrating downstream were 3 years old while the oldest were at least
9. Males were 3 to 6 years old (on average 4.3 ± 0.9 years old, but around 75% were between 4 and
5) and females were between 4 and 9 years old (on average 5.5 ± 1.1 years old, but more than 80%
were between 4 and 6) (Laffaille et al., 2006).
Developing age-length keys
Age-length keys have been developed for each ecophase (Recruitment, Stock and Downstream
migration) using all the individuals of known age. The objective is to convert total annual catch
numbers in the Frémur into an age distribution. The age-length key is based on the hypothesis of a
normal size distribution at different ages (Adam, 1997).
335

Three stages are required to develop this key:
at each age a, the average size (µ a ) and the corresponding standard deviation (σ a ) are calculated ;
the values of normal distribution (Na (µa ; σa) by age for the centre of each length class are
calculated;
the age-key length is developed: the percentage of age a eels present in each size class l using a
30mm class width is calculated.
The percentage is equal to P l,a = 100 ×
N a(
μa;
σa)
Σ N a(
μa
; σa)
The age-length key of sedentary and downstream migratory eels was developed using all the
individuals analysed by otolithometry, regardless of the year they were collected. This was feasible
because, for these two ecophases, the years during which the samples were collected did not affect
the average size of eels analysed at different ages (Ancova, P>0,05). This approach had the dual
advantage of significantly increasing the range of sizes studied and of integrating interannual growth
variability.
1600
Numbers (figure)
1400
1200
1000
800
600
400
200
Age 9
Age 8
Age 7
Age 6
Age 5
Age 4
Age 3
0
300 360 420 480 540 600 660 720 780 840
Size classes (mm)
Figure 9.5 - Size structure stratified by age of migrating silver eels caught in the Pont es
Omnes downstream migration trap.
336

35
30
Frequency (%)
25
20
15
10
5
0
3 4 5 6 7 8 9
Age class (year)
Figure 9.6 - Age class frequency distribution of migrating silver eels caught in the Pont es
Omnes downstream migration trap.
The silver eel size distribution is bimodal with a first mode varying between the 360mm and
420mm size classes and a second mode between 510mm and 600mm. The first mode relates mainly
to males and the second one exclusively to females (figure 9.5). The sex ratio of silver eel populations
can be estimated by analysing the size-frequency distributions.
Flux estimation based on the monitoring of a fishery (tag-recapture)
Tag-recapture experiments can be carried out in commercial or experimental fisheries. The size
of the silver eel population can be estimated using numerous estimators, the simplest being the
Peterson estimator (or derivative) which calculates, at recapture, the dispersion ratio of the silver eel
population marked at first catch into the total silver eel population. Models relating CPUE to
abundance estimates can then be perfected.
For example, a tag-recapture experiment was carried out on the Loire during 5 successive
“guideau” commercial fishing seasons (Boury et al., unpublished data):
in all, this concerned 14 fishing sites, of which 13 are located on the downstream reach of the
Loire, between La Ménitré (upstream of Angers) and Ancenis, and one on the Mayenne. Concessions
for these sites, which are all located on the public fluvial domain, are granted by the State. They
assure the livelihoods of 16 commercial fishers. Annual silver eel catches have fluctuated around 50
tonnes per annum since the 1980s (Babin, 1992; SMPE). Catches from previous years and turnover
(income) figures remain unknown to this day.
The chosen tagging sites are located upstream of the study zone. Over the four seasons,
between three and four fisheries supplied the eels required for the various tagging studies. They were
located in La Ménitré (155km from the estuary), Denée (126km), La Possonnière (120km) on the Loire
and in Chambellay (186km) on the Mayenne.
337

Each tagging study was carried out after catching a sample of 150 to 500 eels per site and per
date. Tagging systems can be chosen in order to distinguish both marking sites and dates (dye and/or
acrylic ink injection into the dorsal fin for the sites and tattooing of the ventral aspect for the dates).
Eels are anaethetised using clove essential oil commonly known as Eugenol. Following the marking
session, and after passing through a recovery tank, the eels are released downstream from the
“guideau” net that caught them. This marking technique was perfected and tested under experimental
conditions (Boury, 2001, Boury et al., in preparation).
The fishers themselves search for marked individuals each time they lift their nets. Recaptured
eels are immediately stored and recorded in the fishing logbooks. Controls are also carried out during
the catch periods to validate these declarations.
In parallel, daily catches are recorded in the fishing logbooks concerning each fishery. These
data, combined with fishing effort, enable the results to be expressed in catch-per-unit-effort (CPUE),
corresponding to the catch from one fishing night with one “guideau”.
Biometric measurements (size, weight) of landed individuals are carried out within the fisheries
during the season in order to characterise the migrating population and any temporal and spatial
trends that it exhibits. These measurements also enable the catch to be expressed in biomass, using
the average weight of individuals landed during each season.
In the case of a simple capture-recapture, the survey is carried out in two phases: a first phase
of capture and marking when nm fish are caught, marked and then freed; then, in a second phase, all
the fish caught below the marking site are examined. This phase can be modelled as x independent
repetitions of a two-mode-experiment m (marked) and ^m (unmarked). In practice, the aim is to
calculate the catch probability for the whole fishery during each downstream migration period through
independent marking studies. The Petersen method is used to calculate these catch probabilities and
the number of eels migrating downstream.
Table 9.2 -
Global estimates of fluxes across the seasons using overall recapture rates and
inspected post-marking catch.
Season 2001-2002 2002-2003 2003-2004 2004-2005
Fishing effort season (days*gears) 647 392 391 319
Fishing effort post-marking (days*gears) 545 241 250 185
N marked catch 1,649 1,202 1,072 781
N inspected catch 43,357 35,685 33,298 9,120
N recaptures 241 149 128 72
Recapture rate 0.1461 0.1239 0.1194 0.0922
Estimated seasonal flux 352,185 468,245 436,154 170,582
These experiments provide an estimate of the number of eels leaving the middle reach of the
Loire (upstream of Ancenis) each year. The population sex ratio is close to 100% female. They are 9
years old on average. These parameters did not vary significantly during the study period.
338

CPUE monitoring
CPUEs are calculated from commercial fishing logbooks, official catch declarations or
experimental monitoring targeting silver eels. They give some indication of abundance and of temporal
variation. Detailed information can be found in the methodological handbook concerning the use of
gear-based fishing data 3 .
For example, a silver eel abundance indicator was developed in the Loire river catchment from
commercial catches (Boisneau and Boisneau, unpublished data). Within the network of river
catchments which were included in the INDICANG framework, the Loire catchment is the only one with
a series of commercial fisheries targeting silver eels. There are fifteen barges all fitted with the same
“dideau” or “guideau” type of fishing gear (Duhamel du Monceau, 1772). This fishing gear resembles a
trawl, 9m wide, 5m high and 22 to 25m long, with a decreasing mesh size from 120 mm at the opening
to 20 mm at the end, which extends into a detachable bag where the silver eels are caught. But unlike
the trawl, the boat and the “guideau” do not move. They are held against the current by a strong
anchoring and cable system (figure 9.7) and it is therefore the current which opens the net where the
silver eels are caught. The technical characteristics and the dimensions of “guideaux” being fixed in
time, there is no variation in fishing power, either between boats or between fishing seasons for the
same boat. Only catchability varies with the river flow, because when flow, and therefore the wet
cross-section of the stream, increases the proportion of water filtered by the “guideau” decreases.
It was possible to collect data on fishing operations and the number of silver eels caught during
each operation from fishers using four “guideaux” in the same downstream sector of the Loire, These
data enable the construction of a yearly abundance index of silver eels migrating down the lower
reach of the Loire, resulting in a chronological series of yearly indices from which trends may be
identified (figure 9.8).
Figure 9.7 - Setting up a “guideau” on the Loire
The data used are thus:
the daily eel catches during the downstream migration season from 1 st October to 15 th February;
the daily fishing effort expressed in number of fishing trips;
the period analysed: 1987–2002.
3
See Chapter 6.
339

lndex calculation
The index for a year is the mean of the decimal logarithms of daily catches for all the fisheries.
Over the whole period, the trend in this index does not have a slope that is significantly different
from 0 (non-parametric seasonal Kendall test, rk=0.034 , P= 0.800) but this does not necessarily mean
that silver eel numbers are stable, especially after 2002.
Figure 9.8 - Temporal variation in the silver eel abundance index on the Loire.
It is possible to adapt this type of tool to other river catchments in order to generate a silver eel
abundance index. In this case, a single device of the “guideau” or “tézelle” type, located upstream of
the dynamic tide, may suffice. The dimensions can then be reduced (for example 6m wide by 3m
high). The cost of a "guideau" is estimated at €75,000. The cost of a smaller experimental fishery can
be estimated at €45,000. Staff costs must be added to equipment costs (two people per fishery).
Silver eels only leave fresh water in quite specific hydroclimatic conditions: floods, atmospheric
depression, dark night, drop in temperature … These difficult conditions are dangerous for commercial
and/or scientific fishing operations. Therefore, they require properly trained staff in order to ensure the
safety of people and equipment as well as appropriate sampling.
Category 2: monitoring the existing stock
A number of monitoring operations were undertaken in order to describe the status of the yellow
eel population. Such eels >200 mm in fact make up the sedentary fraction of the eel stock in each river
catchment (Laffaille et al., 2005), and some of these sedentary individuals begin their silvering
metamorphosis before starting to migrate downstream. Most eels have completed this metamorphosis
by the autumn (September-October) and await “favourable” hydroclimatic conditions in the river
catchment before adopting a genuine migratory behavioural pattern.
340

Three external, visible signs characterise silvering (Durif et al., 2005; Acou et al., 2005). The
most reliable method to identify a silver eel is to measure the ocular diameters (horizontal and vertical
widths of one eye on each eel) with a calliper. When precise ocular measurement is not possible,
silvering is assessed by simply recording the presence of the 3 external signs of silvering (colour,
lateral line and ocular hypertrophy).
It is theoretically possible to estimate the proportion of silver eels from the characteristics of the
sedentary stock and this can be used as an indirect measurement of silver eel production (Feunteun et
al., 2000; figure 9.9). However, the relationship between this index of silver eel abundance and the
true number of those leaving the river catchment must be established case by case.
Sampling-based size
structure of silver eels
Trap on downstream migration
Electrofishing & fyke net
Figure 9.9 - Comparing the size structure of the silver eel population estimated in the
sedentary stock (white bars) and caught on their journey downstream (black bars)
(from Feunteun et al., 2000), in the Frémur (France) in September 1996 (a) and in
September 1997 (b). White bars (electrofishing and fyke-net in the river
catchment): 68 and 47 eels respectively for (a) and (b); black bars (downstream
trap in the lower reach); 678 and 655 eels respectively for (a) and (b).
341

Estimating the potential size of the silver eel population
In some studies, attempts to estimate the size of eel populations were extrapolated to the whole
river catchment, using the water surface area (Feunteun et al., 2000), tag-recapture experiments
(Feunteun et al., Gabriel et al., 2004) or eel-habitat relationship modelling. The number and
characteristics of the silver eel population present in the river catchment are estimated from
information collected during electrofishing campaigns.
Quantification aims to provide abundance indicators of silver eels that are potential candidates
for migration in the coming year, taking into account the characteristics of the silver eel population
produced by the river catchment. Specific methodologies must be adapted to each catchment in order
to provide migratory potential indicators and in fine, taking into account silver stage mortality, silver eel
escapement indicators. These methods must also be implemented in line with the general objectives
described above. It is more difficult to develop such methods on large river catchments or in deep
habitats where sampling effort remains low, considering the size of the environment 4 .
Density of the existing population and CPUE.
In most cases, the sampling effort is far too low compared to the size of the system, in particular
in very large water surface areas (large rivers, lakes and lagoons). It then becomes unrealistic to
estimate the size of the migratory stock. However, it is possible to provide abundance indices, for
example density (number and weight of silver eels per unit of water surface area, Cucherousset et al.,
in press).
The Lobon-Cervia team carried out a study of the eel population (short-cycle, male-dominated)
on the Esva river (a small watercourse in the north-west of Spain) between 1990 and 2006. Given the
population’s longitudinal segregation, 9 sites were monitored on three tributaries selected according to
their distance to the tidal limit. On average, these sites were 3, 5, 18 and 26km from the sea. Densities
were estimated during electrofishing operations which enabled the use of both the stock depletion
method (3 consecutive runs) and the Zippin method to estimate abundances 5 . In order to eliminate
catch variability due to size, calculations were made separately for eels under 15cm and over 15cm.
Eels collected over this period provided the following size-age keys (Lobón-Cerviá et al., 1995).
< 13 cm ; less than one year in the river (age 1) ;
13-20 cm ; less than 2 years (age 2) ;
20-30 cm ; less than 3 years (age 3) ;
30-36 cm ; less than 4 years (age 4) ;
> 36 cm ; age 5 (only a few individuals).
The combined analysis of age structure and population density over a long period of time shows
that the Esva eel population consists of 5 age classes. Silvering starts in September for eels longer
than 29 cm, i.e. around 1,565 days. These silver eels appear in winter and disappear from the system
the following spring once downstream migration is completed.
4
Chapters 2 and 8.
5
Chapter 8.
342

Figure 9.10 shows the size structure of all eels caught in 1990 and 2006. When silvering begins,
males account for 17% of the catch with sizes between 30 and 38cm. Only 2 females were observed
and 2 individuals of indeterminate sex (40 - 43cm).
500
A
400
300
200
100
No. of Observations
0
200
0 5 10 15 20 25 30 35 40 45 50 55
B
150
100
50
0
30 33 35 38 40 43 45 48 51
Length (cm)
Figure 9.10- (A) Size structure of the silver eel population caught in September, based on
1,600 individuals collected between 1990 and 2006 in the Esva river (north-west of
Spain). (B) Males in the 30-38cm size range showed silvering characteristics.
During the study period, only 4 eels longer than 40cm were caught and were
potentially females.
Estimating the reproductive potential.
The objective is to determine the proportion of each sex and their average biomass in the
catchment’s silver eel production:
sex ratio: in order to facilitate field work, the sex ratio can be estimated from the size-class
frequency diagrams of downstream migrants. Generally, males do not exceed a total length of 450
mm, although this must be confirmed in each catchment;
size and biomass of males and females: the relationship between the size and weight of individuals
must be established for males and females in order to estimate the population’s average fecundity
and to obtain information on the quality of individuals 6 .
343

The Reproductive Potential indicator is based on the analysis of the eel density profile along the
main hydrographic linear axis (figure 9.11). The level of accuracy of this indicator depends on the
degree of available information on migratory individuals or on the population structure inside the
catchment. The objective of this indicator is to specify the type and level of silver eel production in a
catchment.
Short system
with no barrier
Short system
with barrier
Long system
with no barrier
Long system
with barriers
Figure 9.11- Variation in eel density profiles (CPUE, Catch per unit effort) as a function of the
length of the hydrographic system (short or long) and the presence or absence of
barriers. (from Robinet et al., 2008).
Long-term observations on the Esva river enable annual silver eel abundance to be monitored
and may provide an indicator of the population’s reproductive potential.
The results provide the average density of eels beginning their silvering process (eels aged 4 in
September). As regards older eels found in the river catchment, these graphs can be assimilated to
the reproductive potential, which is of 136 silver eels per ha (99% male), on average, during the 19-
year study. However, these values varied significantly; densities dropped noticeably between 1988
and 1996, and remained low until 2001, before returning to their initial level in 2006 (figure 9.12).
6
See Chapter 5.
344

300
Age-4 (ind ha -1 )
200
100
0
1986 1991 1996 2001 2006
Cohorts
Figure 9.12 - Evolution between 1998-2006 of the average yearly density of eels aged 4 (30 to
38cm) caught on 9 sites located along the Esva river (north-west of Spain)
In the ORIA (Basque Country), during the Indicang programme, electrofishing showed rapidly
decreasing abundance, which was over 2,500 ind/ha 10km from the sea (figure 9.13). The rapid
decline is clear, with eels disappearing from catches 45km from the sea. This type of regular (once or
twice a year) inventory undertaken over a long period reveals the trend in this feature. When the
population size increases, densities should increase and more of the length of the watercourse should
be colonised. This type of analysis which simply describes the yellow eel population (existing stock)
can, in the absence of any other data, provide a rough estimate of the spawner potential.
Figure 9.13 - Abundance profile (number/ha) of the sedentary eel population of the Oria river
catchment in 2005. Densities are derived from electrofishing (stock depletion,
Zippin estimator) (E. Diaz unpublished data).
In large rivers (over 10m wide and 1m deep), it is not possible to calculate densities using
capture-recapture or stock depletion methods. In this case, commercial or amateur fishing CPUE can
be monitored using protocols based on fishers’ declarations and catch characteristics. All this
345

information must be collected and validated by directly (checking of the declarations) or indirectly
(search for, and identification of, inconsistent data using statistical analysis). Experimental monitoring
methods may also be implemented, using gear-based fishing and electrofishing. In France, Onema
(formerly CSP) initiated monitoring of gear-based catches and electrofishing records some ten years
ago.
In order to have an idea of silver eel abundance, observations must focus on the one hand, on
eels between 300 and 450mm, including males close to silvering (and due to depart in the following
few months) and growing females and, on the other hand, on eels larger than 450mm which are
females, an (unknown) proportion of which is due to migrate downstream in the coming months or
years. This case is based on the hypothesis that relative eel abundance (expressed as Catch Per Unit
Effort) in each river catchment and in these two size ranges is proportional to the numbers of silver
eels leaving the river catchment each year 7 .
Given the way in which eel size structure and density is organised according to an estuaryupstream
gradient, it is essential that sampling is organised according to an upstream-downstream
gradient with stations distributed regularly along the watercourse.
In large watercourses, electrofishing must follow a protocol adapted to eels. Numerous studies
have shown that non-targeted fisheries underestimate abundance and distort the population size
structure 8 . Two types of methods have been tested for many years in large watercourses (such as the
Rhône): point-based abundance sampling or “EPA” (“Echantillonnages Ponctuels d’Abondance”) and
representative river section sampling (“méthode des pêches par ambiance”) both based on
electrofishing.
The “EPA” method was developed by Nelva et al (1979) on the Rhône. It relies on electrical
fishing gear releasing rectified alternative current (210-500 V; 7-11A). The unit is taken on board
a boat. The fishing operation i.e. “EPA” consists of throwing the anode towards the bank (at a
depth > 1 m) and of applying the current for 1 minute. All the fish are then caught with a hand net
during this fishing period. A minimum of 20 “EPA”, at least 5 to 10m apart, are carried out at a
given station. This method has been used since 1979 on the Rhône, between Lyon and Arles,
and now provides a series of almost 30 years of data (Feunteun et al., 1997).
The “pêches par ambiance” method (which samples a representative section of the river) consists
of sampling a sector continuously for around one hundred metres, mostly along the banks or in
habitats less than 1 metre deep. For example, the Aix en Provence Cemagref has developed this
method on the Rhône since 1995.
Another method consists of sampling only the small tributaries of large rivers where the stock
depletion method can be used. In this case, stations located near the confluence with the river are
selected on each small tributary. A series of small rivers are then selected, going from downstream to
upstream, and electrofishing is undertaken to provide densities along the longitudinal axis. This type of
sampling was used for example in Brittany on the Arguenon (Laffaille et al., 2001) and the Aulne
7
Chapters 2 and 8.
346

(Laffaille and Lafage, 2003). Annual monitoring using an adapted and long-term sampling programme
(yearly or twice yearly) can provide average abundance variations in eel populations, and stock and
size structure variation of the eels caught can also be obtained if information concerning the size and
the ocular diameter is also collected.
During the Indicang programme (unpublished data), experimental fishing was carried out in the
Garonne river catchment using specific eel protocols. Cemagref and MIGADO 9 worked jointly in
tributaries going from upstream to downstream the lower reach of the Garonne. Electrofishing
inventories with 2 or 3 runs were made in shallow zones (riffles). These sought to catch small eels in
order to estimate colonisation fronts 10 . However, larger individuals can be caught and their numbers
estimated using the Zippin method. Hence, the densities of eels 30 to 45cm, and above, could be
represented according to the distance to the dynamic tide limit in the estuary (figure 9.14). This shows
that 30 to 45cm eels (growing males and females) are distributed as a normal curve and that eels
above 45cm (females) are distributed evenly up to 140km from the dynamic tide.
This representation gives an imprecise estimate of spawning stock abundance in the lower
reach of the Garonne because the validity of the results is highly dependent on the sampling protocol
and the targeted life stages. Given that young eels were targeted here, the results must be interpreted
with great caution and be complemented by experimental fishing and protocols focusing on older life
stages.
Density (no individuals/100m²)
25
20
15
10
5
0
Individuals 30-35cm
Individuals > 45cm
Polynomial (Individuals 30-35cm)
Polynomial (Individuals > 45cm)
R 2 = 0,1759
P=0,05
0 20 40 60 80 100 120 140 160 180
-5
Distance to the dynamic tide (km)
Figure 9.14 - Abundance indicator of the silver eel population in the Garonne
lower reach. The shape of the curves represents the number of eels larger than
45cm (females) and between 30 and 45cm (growing males and females). These
size groups include silvering eels about to migrate in the coming weeks or
months.
8
See Chapter 8.
9
An association which aims to restore and protect migratory fish in the Garonne and Dordogne catchments.
10
See Chapter 8.
347

9.3.2. Escapement potential
9.3.2.1. Context and objective
The reproductive potential represents the number of silver eels migrating downstream in
favourable climatic conditions and without any anthropogenic mortality but does not represent the
number of eels which actually escape from the catchment. Hence, in order to estimate the
Escapement Potential, a specific silver stage mortality rate must be applied to the reproductive
potential. This mortality is directly related to the catchment context and concerns principally: the
presence of hydroelectric turbines, of downstream migratory eel fisheries, of drinking-water reservoirs,
and the characteristics of reserved flow pipes.
9.3.2.2. Data exploitation and relationship with other indicators
Estimating the escapement from a catchment means that all the silver stage mortality factors
must be subtracted from the reproductive potential. If F’ = the mortality rate due to anthropogenic
action:
then F’= F + H + T + P + p
where F is the mortality related to fisheries
H is the mortality related to habitat (density–dependent high death rate due to barriers)
T is the mortality related to hydroelectrical turbines
P is the mortality related to pollution
and p is the mortality related to parasitism
Escapement Potential = Reproductive Potential x F’. * downstream migration intensity If
F’ = 0, mortality is entirely natural and the Escapement Potential is equal to the Reproductive
Potential.
Table 9.3 summarises all the observations and descriptors required to qualify and quantify the
potential escapement and the reproductive potential in a given river catchment. It includes the notion
of minimum and optimum management charts introduced at the beginning of this chapter.
348

Table 9.3 – Synopsis of necessary and complementary information required to qualify or
quantify silver eel escapement and the reproductive potential of the population
produced in a given river catchment.
Requirement /
Indicang
General
objective
Detailed
objectives
Period of
the year
Optimal
frequency
Descriptors
Compulsory
Monitoring by
habitat sector
Characterise
habitat sectors
Only once
River length and habitat surface area
by stream order
Compulsory
Monitoring the
trend in
reproductive
potential in the
catchmentss
Estimate
relative
abundance
and the size
structure of
the existing
stock by
habitat sector
As late as
possible
during the
period of low
water level
(Autumn)
One survey per
annum
Individual size and weight of all eels, or
of a 50-individual sub-sample by
fishing operation (specific or nonspecific
monitoring, fisheries)
One survey per
annum
Parameters of fishing operations (in
order to calculate the effort unit
corresponding to catches, expressed in
numbers and biomass per 100m 2 or by
effort unit), location (habitat sector,
distance to the sea).
Characterise
the fraction of
pre-migratory
individuals by
habitat sector
Size and sexratio
structure
As late as
possible
during the
period of low
water level
(Autumn)
One survey per
annum
Ocular diameter measurements
(horizontal and vertical) on individuals
> 30cm
Silvering diagnosis (lateral line, colour)
Compulsory
Estimate
mortality of
human origin
in the silver
stage.
Mortality due
to barriers to
migration
Continuously
Initial assessment
+ each time a
construction or
protection device
is modified
upstream
Location, characteristics of reservoir
dams, reserved flow systems and
protection devices upstream (if any)
Mortality due
to each
hydroelectric
turbine
Continuously
Initial assessment
+ each time
turbines or a
protection device
is modified
upstream
Location, characteristics of each
turbine (type, rotor diameter, velocity
according to the unit) and protection
device upstream
Commercial
and amateur
fisheries
harvests of
downstream
migratory eels
Continuously
Data collection
once a year
Silver eel catch data, characteristics of
fishing gear (type, number), number of
fishing days / year / gear, catch
characteristics (number, biomass,
CPUE, size structure)
Massive
occasional
mortalities
surveillance surveillance Frequency and intensity of massive eel
mortalities (ONEMA surveillance)
349

Requireme
nt /
Indicang
General
objective
Detailed
objectives
Period of the year
Optimal
frequency
Descriptors
Optional
Monitoring
downstream
migratory
eels
Quantify / semiquantify
fluxes
by catchment
All year long ~ Every day Size, weight and date of silver eel
catches in the downstream traps
(usually exhaustive quantification)
During downstream
migratory peaks
Several times
each winter
Escapement quantification (markrecapture)
During the whole season
(commercial) or all year long
(experimental fishery)
Regularly (e.g.
every new moon
or more
frequently)
Number and weight of silver eels by
fishing operation (commercial or
experimental fisheries)
Fishing operation parameters (in order
to calculate the effort unit related to
catches), location (habitat sector,
distance to the sea).
Characterise
silver eel
parameters
During each monitoring operation of eels
migrating downstream, on a representative
sample (30 to 50 individuals)
Size and weight of silver eels migrating
downstream
Ocular diameter measurements
(horizontal and vertical)
Optional
Know
spawner
quality
Define the
physiological
state of silver
eels
During
downstream
migration
Initial assessment on a
representative sample (20 to 30
individuals by sex), then
monitoring the trend every 4 to 5
years Any catch method appears
to be suitable
Lipid dosage by individual (migration in
the water column)
Know the age at
downstream
migration
Proportions of the different lipid types in
muscles and gonads
Average ovocyte size after atresia,
gonad weight, digestive tract weight,
female fecundity
Otolith reading (avoid the burning &
cracking method)
Characterise
past exposure to
pollutants and/or
the
contamination
level
Dosage of toxic equivalents by
contaminant family (pesticides, dioxinlike,
PCBs, PAHs, metals)
Activity of exposure biomarkers (EROD,
metallothioneins etc.)
Great structural impairment of silver eel
hepatocytes and ovaries/testes
Characterise the
health condition
Serology and/or hematocrit level (EVEX
etc.)
Condition of the swim bladder (present
or past presence of Anguillicola
crassus)
350

References
.
Acou A., Boury P., Laffaille P., Crivelli A.J., Feunteun E., 2005. Towards a standardized characterization
of the potentially migrating silver European eel (Anguilla anguilla, L.). Arch. Hydrobiol. 164: 237-255.
Acou A., Gabriel G., Laffaille P., Feunteun E., in press. Differential production and condition indices of
premigrant eels (Anguilla anguilla) in two small Atlantic coastal catchments of France. Am. Fish. Soc.
Symp.
Acou A., Lefebvre F., Contournet P., Poizat G., Panfili J., Crivelli A.J., 2003. Silvering of females eels
(Anguilla anguilla) in two sub-populations of the Rhône Delta. Bull. Fr. Pêche Piscic., 368 : 55 – 68.
Acou A., Poizat G., Crivelli A.J., 2006. Errors in ocular index measurements in European eel (Anguilla
anguilla). Ecology of Freswater Fish, 15: 578-582.Ancona U., 1921. Effeti dell’innizione sul tubo
degerente dell’anguilla. Mem. R. Comit. Talass. Ital., Venise, I.
Adam G., 1997. L’anguille européenne (Anguilla anguilla): dynamique de la sous-population du lac de
Grand-Lieu en relation avec les facteurs environnementaux et anthropiques. Thèse de Doctorat à
l’Université Paul Sabatier, Toulouse II, spécialité hydrobiologie, 353 pages.
Adam G. et Elie P., 1994. Mise en évidence des déplacements d’anguilles sédentaires (Anguilla anguilla
L.) en relation avec le cycle lunaire dans le lac de Grand-Lieu (Loire-Atlantique). Bulletin Français de la
Pêche et de la Pisciculture, 335: 123-132.
Ancona U., 1927. Richerche sull’ingrandimento dell’occhio dell’Anguilla in rapporto alla maturita
sessuale e considerazioni sul suo significata biologico. R. C. Acad. Nineci. 5: 360-363.
Anonyme, 2002. Guidance for Choosing Sampling Design for Environmental Data. United States
Environmental Protection Agency- Office of Environmental Information, Final Report EPA/QA/G-5S,
180pp.
Anonyme, 2003. Réseau de surveillance de l’anguille européenne dans les réseaux de canaux du
Marais poitevin. Rapport Parc Interrégional Marais poitevin - Cemagref Bordeaux - CSP, 80 p.
Anonyme, 2004. De l’utilité de la petite pêche. Samudra Mars 2004 : 25-27.
Antunes C., 1994. The seasonal occurrence of glass eels (Anguilla anguilla L.) in the Rio Minho
between 1991 and 1993 (North of Portugal). Int. Revue ges. Hydrobiol. 79 (2):287-294
Antunes C., 1996. A pesca da enguia no rio Minho-que futuro? I Simpósio Ibérico sobre a Bacia
Hidrográfica do rio Minho. Actas (79-83)
Antunes C. et Tesch F., 1997. A critical consideration of the metamorphosis zone when identifying daily
rings in otoliths of European eel, Anguilla anguilla (L.) Ecology of Freshwater Fish 6:2, 102–107
Aprahamian M.W., 1986. Eel (A. anguilla) production in the River Severn, England. Polskie Archiwum
Hydrobiologii, i
INDICANG – Handbook - References 351

Aprahamian M.W., 1988. Age, structure of the eel population (Anguilla anguilla) in the rivers Severn
(England) and Dee (Wales). Aquacult. Fish. Manage., 19: 365-376.
Aprahamian M.W, 2000. The growth rate of eel in tributaries of the lower River Severn, England, and its
relationships with stock size. J. Fish Biol., 56: 223-227.
Arias A.M. et Drake P., 1985. Estructura de la poblacion y regimen alimentario de A. anguilla en los
esteros de San Fernando (Cadiz). Invest. Pesq., 49: 475-491.
Ardizzone, G. D. et Corsi, F., 1985. Eel population structure, dynamics and fishing yield in a
mediterranean coastal lagoon. Oebalia, 11(2) : 547-560.
Avise J.C., Helfman G.S., Saunders N.C., Hales L.S., 1986. Mitochondrial DNA differentiation in North
Atlantic eels: population genetic consequences of an unusual life history pattern. Proceedings of the
National Academy of Science of the USA, 83: 4350-4354.
Avise J.C., Nelson W.S., Arnold J., Koehn R.K., Williams G.C., Thorsteinsson V., 1990. The
evolutionary genetic status of Icelandic eels. Evolution, 44(5) : 1254–1262.
Bailey N.T.S., 1951. On estimating the size of mobile populations from recapture data. Biometrika, 38 :
293-306.
Baisez A., 2001. Optimisation des suivis des indices d’abondance et des structures de taille de l’anguille
européenne (Anguilla anguilla) dans un marais endigué de la côte atlantique. Relations « Espècehabitats
». Thèse Biologie, Université Toulouse III - Cemagref Bordeaux, 396 p.
Baisez A. et Laffaille P., 2005. Un outil d’aide à la gestion de l’anguille : le tableau de bord anguille du
bassin Loire. Bulletin Français de la Pêche et de la Pisciculture, 378-379: 115-130.
Baisez A., Rigaud C., Feunteun E., 2000. Hétérogénéité de répartition de l'anguille européenne
(Anguilla anguilla, L.) observée par pêche électrique dans un marais endigué de la côte atlantique
française (Marais breton). Cybium, 24: 23-32.
Barak N.A.E. et Masson C.F., 1992. Population density, growth and diet of eels, Anguilla anguilla L., in
two rivers in eastern England. Aquaculture and Fisheries Management, 23: 59-70.
Baras E., Jeandrain D., Serouge B., Philippart J.C., 1998. Seasonal variations in time and space
utilization by radio-tagged yellow eels (A. anguilla) in a small stream. Hydrobiologia, 371/372: 187-198.
Baras E., Philippart J.C., Salmon B., 1996. Estimation of migrant yellow eel stock in large rivers through
the survey of fish passes: a preliminary investigation in the River Meuse (Belgium). In Cowx I.G. Stock
assessment in inland fisheries Fishing news books: 82-92.
Baras E., Salmon B., Philippart J.C., 1994. Evaluation de l'efficacité d'une méthode d'échantillonnage
par nasses des anguilles jaunes (Anguilla anguilla L.) en migration dans la Meuse. Bulletin Français de
la Pêche et de la Pisciculture, 335: 7-16.
Barbin G.P. et McCleave J.D., 1997. Fecundity of the American eel Anguilla rostrata at 45°N in Maine,
USA. Journal of Fish Biology 51:840-847.
Bardonnet A., Bolliet V., Belon V., 2005a. Recruitment abundance estimation. Role of glass eels
(Anguilla anguilla) response to light. J. Exp. Mar. Biol. Ecol., 321: 181-190.
INDICANG – Handbook - References 352

Bardonnet A., Rigaud C., Labonne J., 2005b. Etude expérimentale du comportement des civelles
(Anguilla anguilla) en faciès lentique. Influence de la densité et de la disponibilité en abris. Bull. Fr.
Pêche Pisc., 378-379: 47-65.
Barse A.M., Secor D.H., 1999. An exotic nematode parasite of the American eel. Fisheries 24(2) : 6-10.
Beaulaton L. et Castelnaud G., 2005. The efficiency of the selective stream tidal transport in glass eels
entering the Gironde (France). Bull. Fr. Pêche Pisci., 378-379: 5-21.
Beaulaton L. et Castelnaud G., 2005. Abundance trends of glass eels (Anguilla anguilla) between 1978
and 1999 from fisheries data in the Gironde basin, France. In: Casselman JM and Cairns DK, eds. Eels
at the edge: Science, Status, and Conservation Concerns. Proceedings of the 2003 International Eel
Symposium. Bethesda, Maryland: American Fisheries Society. Sous presse.
Beaulaton L., Taverny C., Castelnaud G., 2006. Status of the European stock of sea lamprey
(Petromyzon marinus): abundance, exploitation and biological characteristics of adult migrants entering
the Gironde (France). Soumis Fisheries Research.
Beaulaton L., 2007. Indicateurs halieutiques et diagnostic sur l’abondance des espèces amphihalines et
des crustacés dans le bassin de la Gironde. Propositions d’amélioration des suivis statistiques des
pêches fluvio-estuariennes françaises. Thèse Doctorat de l'Institut National Polytechnique de Toulouse.
Fonctionnement des écosystèmes et des Agrosystèmes. En préparation. (com pers ?)
Beverton R.J.H. et Holt S.J., 1957. On the dynamics of exploited fish populations. Her majesty's
stationery Office, London : 533 p.
Beaumont M.A., 1999. Detecting population expansion and decline using microsatellites. Genetics, 153,
2013-2029.
Bellet C., Daniel P., De Casamajor M.N., Sanchez F., Prouzet P., Charron M.H., Martinet J.P., Vignes
J.C., 1998. Infestation de l’anguille européenne sur le bassin de l’Adour et sur certains cours d’eau
côtiers des départements des Landes et des Pyrénées Atlantiques. Doc. Migradour / G.D.S.A.A., 29 p.
Belpaire C., 1987. Possibilities of culturing glass-eels (A. anguilla) in earthern pond in Melgium. EIFAC.
Working group on eel (Bristol). 10 p.
Belpaire C., Van Driessche H., Gao F.Y., Ollevier F., 1992. Food and feeding activity of glass-eels A.
anguilla stocked in earthern ponds. Irish Fish. Invest. Ser A (Freshwater), 36: 43-54.
Berg R., 1989. The growth of eels. A critical assesment of data from open waters. EIFAC Working party
on eels, Porto, 1989. Rome Eifac , Fao, 10 pp.
Berg R., 1990. The assessment of size class proportions and fisheries mortality of eel using various
catching equipment. Int. Revue ges. Hydrobiol., 75: 775-780.
Bergersen R. et Klemetsen A., 1988. Freshwater eel Anguilla anguilla (L.) from North Norway with
emphasis on occurrence, food, age and downstream migration. Nordic Journal. Freshwater Res., 64:
54-66.
Bertin L., 1951 Les anguilles 2ème édition, Payot (Ed), Paris, 191 pages.
Beullens K., Ding E.H., Ollevier F., Komen J., Richter C.J.J., 1997. Sex differentiation, changes in
length, weight and eye size before and after metamorphosis of European eel (Anguilla anguilla L.)
maintained in captivity. Aquacuture, 153: 151-162.
INDICANG – Handbook - References 353

Biais G., 2003. Expertise et régulation des pêches : regard issu d’une démarche systémique. Nature
Sciences Société, 11 : 195-201.
Bienarz K., Epler P., Malczewski B., Passakas T., 1981. Development of European eel (A. anguilla)
gonads in artificial conditions. Aquaculture, 22: 53-66.
Blanc G., 1989. L'anguillicolose. Répartition géographique. Mise au point d'une méthode de diagnostic
par échotomographie. Moyens de lutte. Thèse de Docteur Vétériaire., Université de Nantes. 194 p.
Blondel J., Ferry C., Frochot B., 1970. La méthode des indices ponctuels d’abondance (IPA) ou des
relevés d’avifaune par “stations d’écoute”. Alauda, 38: 55-71.
Boëtius J., 1980. Anguilla anguilla. A presentation of old and new data of total numbers of vertebrae
with special reference to the occurrence of Anguilla rostrata in Europe. Dana, 1: 93-112.
Boëtius I. et Boëtius J., 1980. Experimental maturation of female silver eels, Anguilla anguilla. Estimates
of fecundity and energy reserves for migration and spawning. Dana 1: 1-28.
Boëtius I. et Boëtius J., 1985. Lipid and protein content in Anguilla anguilla during growth and starvation.
Dana 4:1-17.
Boëtius J. et Harding E.F., 1985. A re-examination of Johannes Schmidt’s Atlantic eel investigations.
Dana 4: 129-162.
Bohlin T. et Cowx I.G., 1990. Implications of unequal probability of capture by electric fishing on the
estimation of population size. In Developments in electric fishing, Cowx Eds, Fishing News Books: 145-
155.
Bohlin T. et Sundström B., 1977. Influence of unequal catchability on population estimates using the
Lincoln index and the removal method applied to electro-fishing. Oikos, 28: 123-129.
Boisneau P. et Mennesson-Boisneau C., 2001. La faisabilité d’un indice d’abondance de l’anguille
d’avalaison en Loire ». Premières rencontres scientifiques « L'anguille en Loire », 19 mars 2001,
Angers. Diaporama : 19p.
Bolger T. et Connolly P.L., 1989. The selection of suitable indices for the measurement and analysis of
fish condition. Journal of Fish Biology, 34: 171-182.
Bolopion J., Forest A., Sourd L.J., 2000. Rapport sur l'exercice de la pêche dans la zone côtière de la
France. Ministère de l'Agriculture et de la Pêche : 113 p + figures.
Boubée J., Mitchell C.P., Chisnall B.L., West D.W., Bowman E.J., Haro A., 2001. Factors regulating the
downstream migration of mature eels (Anguilla spp.) at Aniwhenua Dam, Bay of Plenty, New Zealand.
New Z. J. Mar. Freshw. Res. 35:121-134.
Boussouar A. et Prouzet P., 2007. Le modèle comportemental civelle : un outil de gestion.
Communication orale – Journées de restitution des travaux du GDR Adour. Ifremer-UPPA.
Boussouar A., Prouzet P., Sanchez F., De Casamajor M.-N., 2005. Comportement migratoire des
civelles (Anguilla anguilla) en estuaire : modélisation et apllication à la vitesse de dplacement du flux et
à sa disponiblité pour sa pêcherie. In Colloque Défi golfe de Gascogne (Poster).
INDICANG – Handbook - References 354

Bordajandi L.R., Gomez G., Fernandez M.A., Abad E., Rivera J., Gonzalez M.J., 2003. Study on PCBs,
PCDD/FS organochlorine pesticides, heavy metals and arsenic contents in freshwater species from the
River Turia (Spain). Chemosphere, 53: 163-171.
Bourg D., 1993. Hans Jonas et l’écologie. La Recherche 256 : 886-890.
Bouvet J.-C., Prouzet P., Bru N., 2006. Quantification de la biomase saisonnière de civelles (Anguilla
anguilla) dans l’estuaire de l’Adour et estimation du taux d’exploitation saisonnier de la pêche
professionnelle de 1998 à 2005. Rapport Ifremer – projet Indicang. http://www.ifremer.fr/indicang/sitesthematiques/pdf/flux-saison.pdf
Bozeman, E.L., Helfman G.S., Richerdson T., 1985. Population size and home range of American eels
in a Georgia tidal creeks. Trans. Amer. Fish. Soc., 114: 821-825.
Branchi I., Capone F., Alleva E., Costa L.G., 2003. Polybrominated diphenylethers : neurobehavioural
effects following developmental exposure. Neurotoxicology, 28: 449-462.
Brandt A.V., 1971. An account of the various kinds of traps employed for the capture of eels. Eifac
Technical Paper, 14: 69-79.
Brethes J.C., 1990. Le processus d’évaluation des stocks. In : Brethes J.C., O’Boyle R.N. (eds.),
Méthodes d'évaluation des stocks halieutiques, Centre international d'exploitation des océans, Québec :
17-30.
Brethes J.C., 1999. Spatialiser la gestion dans l’approche de précaution : mortalité globale et mortalité
locale. In : Les espaces de l’halieutique, 4 ème forum halieumétrique, Rennes, 29/06-1/7/99 : 205-211.
Briand C. et Boussion D., 1997. Suivi des migrations de poissons sur la Vilaine. Bilan de la première
année de fonctionnement des passes à poissons d'Arzal. Rapport Mission Institution d'Aménagement
de la Vilaine : 59 p.
Briand C., Fatin D., Fontenelle G., Feunteun E., 2002. Effect of re-opening of a migratory axis for eel at
a watershed scale (Vilaine river, Southern Bittany). American Fisheries Society Symposium.
Briand C., Fatin D., Fontenelle G., Feunteun E., 2003. Estuarine and fluvial recruitment of European
glass eels in an exploited Atlantic estuary. Fisheries Management and Ecology, 10: 377-384.
Briand C., Fatin D., Fontenelle G., Feunteun E., 2005a. Effect of re-opening of a migratory pathway for
eel (A. anguilla) at a watershed scale. Bulletin Français de la Pêche et de la Pisciculture, 378-379: 67-
86.
Briand C., Fatin D., Fontenelle G., Feunteun E., 2005b. Estimating the stock of glass eels in an estuary
by mark-recapture experiments using vital dyes. Bulletin Français de la Pêche et de la Pisciculture, 378-
379: 23-46.
Briand C., Fatin D., Lambert P., 2004. Effets de la température et de la salinité sur le développement de
la pigmentation des civelles (A. anguilla). Cybium, 28: 247-254.
Broad T.L., Townsend C.R., Arbuckle C.J., Jellyman D.J., 2001. Microhabitat use by longfinned eels in
New Zealand streams with contrasting riparian vegetation. Journal. Fish Biol., 59: 1385-1400.
Brosse S., Laffaille P., Gabas S., Lek S., 2001. Is scuba sampling a relevant method to study
microhabitat in lakes? Examples and comparisons for three European species. Ecology of Freshwater
Fish, 10: 138-146.
INDICANG – Handbook - References 355

Bru N, Lejeune M., Prouzet P. , 2004. Estimations journalière et saisonnière de l’abondance d’un flux de
civelles en estuaire ouvert : Application à l’Adour. (Poster) IXème Colloque International
d’Océanographie du golfe de Gascogne.
Bru N., Bouvet J.-C., Prouzet P., 2006. Quantification par une méthode analytique de la biomase
saisonnière de civelles (Anguilla anguilla) dans l’estuaire de l’Adour et estimation du taux d’exploitation
saisonnier de la pêche professionnelle de 1998 à 2005. Rapport Ifremer – projet Indicang
http://www.ifremer.fr/indicang/sites-thematiques/pdf/estim-civelle-1998-2005.pdf
Bruslé J., 1990. Effects of heavy metals on eel, Anguilla sp. Aquat. Living Resour., 3 : 131-141.
Bruslé J., 1994. L’anguille européenne Anguilla anguilla, un poisson sensible aux stress
environnementaux et vulnérable à diverses atteintes pathogènes. Bull. Fr. Pêche Piscic., 335 : 237-260.
Caddy J., 1998. A short review of precautionary reference points and some proposals for their use in
data-poor situations. FAO Fisheries Technical Paper, 379: 30 p.
Caddy J.F. et Mahon R., 1996. Points de référence en aménagement des pêcheries. FAO Document
technique sur les pêches, 347 : 89 p.
Caddy, J.F. et Griffiths R.C., 1995. Living marine resources and their sustainable development: some
environmental and institutional perspectives FAO Fisheries Technical Paper, 353 : 167 p.
Cagnon C., Lauga B., Karama S., Mouchès C., 2004. Diversité génétique des civelles en migration dans
l’estuaire de l’Adour. Communication orale , IXème Colloque International d’Océanographie du golfe de
Gascogne, Pau France, Juin 2004.
Caill-Milly N., 2002. Site atelier du bassin d’Arcachon. Rapport Final Contrat PECOSUDE
« Caractéristiques des petites pêches côtières et estuariennes de la côte atlantique du sud de
l’Europe », 42 p.
Caill-Milly N. et Prouzet P., 1999. Pêches basques et Sud-Landaises : Ressources, flottille, économie,
commercialisation. Repères Océan, 17, 121 p.
Callaghan R.M. et McCarthy T.K., 1992. Population and growth of eels in Dunkellin River. Irish Fish.
Inv., Ser. A, 36: 59-65.
Cantrelle I., 1984. Le marquage par coloration appliqué à l’étude des migrations de civelles (A.
anguilla). Cybium, 8: 69-78.
Carle F.L. et Strub M.R., 1978. A new method for estimating population size from removal data.
Biometrics, 34: 621-630.
Carrs D.N., Elston D.A., Nelson K.C., Kruuk H., 1999. Spatial and temporal trends in unexploited yellow
eel stocks in two shallow lakes and associated streams. J. Fish Biol., 55: 636-654.
Carry L. et Delpeyroux J-M., 2003. Etude des rythmes de migration des espèces amphibiotiques et
holobiotiques de la Garonne au niveau de la station de contrôle de Golfech au cours de l’année 2002.
Rapport MIGADO G18-03-RT : 26 p. et annexes.
Castelnaud G. et Babin D., 1992. La pêche professionnelle fluviale et lacustre en France. Enquête au fil
de l'eau. Collection Etudes, Ressources en eau, CEMAGREF/Min. Environnement : 291 p.
INDICANG – Handbook - References 356

Castelnaud G. et Gascuel D., 1984. Exploitation et suivi de la ressource Anguilla anguilla. Inventaire et
analyse des travaux français, réflexions et bilan. Groupe National Anguille, novembre 1983 : 116-142.
Castelnaud G. et Cauvin G., 2002. Site atelier de la Gironde. Caractéristiques des petites pêches
côtières et estuariennes de la côte Atlantique du sud de l'Europe. Rapport final, contrat européen
PECOSUDE n°99/024 ED/DG FISH (DGXIV). Cemagref, Groupement de Bordeaux : 58 p.
http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/site-atelier-gironde-rapport-3-5.pdf
Castelnaud G., Rochard E., Le Gat Y., 2001. Analyse de la tendance de l’abondance de l’alose, Alosa
alosa en Gironde à partir de l’estimation d’indicateurs halieutiques sur la période 1977-1998. Bull Fr.
Pêche et Piscic., 362/363 : 989-1015.
Castelnaud G., Coutancier B., Cerezuelle D., Guchan A., 1985. La pêche des migrateurs en Gironde :
2ème partie : Enquête socioprofessionnelle et propositions de gestion. CEMAGREF de Bordeaux,
Division ALA/MSHA/Ministère de l'Environnement/Département de la Gironde : 110 p.
Castignolles N., 1995. Automatisation du comptage et de la reconnaissance des espèces dans les
passes à poisons par l’analyse de séquence d’images. Thèse Doctorat INP Toulouse, 167 p.
Casselman J.M., Marcogliese L.A., Hodson P.V., 1997. Recruitment index for the upper Saint Lawrence
River and Lake Ontario eel stock: a re-examination of eel passage at the R.H. Saunders hydroelectric
generating station at Cornwall, Ontario, 1974-1995 : 161 – 169. In Peterson, R.H. (Ed.). The American
eel in eastern Canada: stock status and management strategies. Proceedings of Eel Workshop,
January 13 – 14, 1997, Québec City, QC. Canadian Technical Report of Fisheries and Aquatic
Sciences, 2196.
Catanzano J. et Rey H. 1997. La recherche halieutique entre science et action : réflexions sur fond de
crise. Nature Sciences Société, 5 : 19-30.
Cauvin G. et Paquignon G., 2001. Monographie sur la pêche maritime dans l’estuaire de la Gironde.
IMA de Biarritz, CLPMEM de Bordeaux : 69 p. + annexes.
Chadwick M. et O’Boyle R.N., 1990. L’analyse des données de capture et d’effort. In : Brethes J.C.,
O’Boyle R.N. (eds.), Méthodes d'évaluation des stocks halieutiques, Centre international d'exploitation
des océans, Québec : 77-101.
Changeux T., 2003. Evaluation des captures d'anguilles par les membres d'AAPPMA dans le bassin de
la Loire. LOGRAMI/CSP, Orléans.
Chapman D.G., 1952. Inverse, multiple and sequantial sample censures. Biometrics, 8: 286-306.
Chapman D.G., 1990. Statistical problems in fisheries. In : Hunter J. (ed.), Writing for fishery journals,
American Fisheries Society, Bethesda : 81-92.
Chaput G., Locke A., Cairns D., 1997. Status of American eel (Anguilla rostrata) from the southern Gulf
of Saint Lawrence, 69 – 93. In Peterson, R.H. (Ed.). The American eel in eastern Canada: stock status
and management strategies. Proceedings of Eel Workshop, January 13 – 14, 1997, Québec City, QC.
Canadian Technical Report of Fisheries and Aquatic Sciences, 2196.
Charlon N. et Blanc J.M., 1982. Etude des civelles d’Anguilla anguilla dans la région du bassin de
l’Adour. 1- Caractéristiques biométriques de longueur et de poids en fonction de la pigmentation. Arch.
Hydrobiol., 93 (2) : 238-255.
INDICANG – Handbook - References 357

Chisnall B.L. et Hayes J.W., 2001. Age and growth of shortfinned eels (Anguilla australis) in the lower
Waikato basin, North Island, New Zealand. New Z. J. Mar. Freshw. Res. 25 : 71-80.
Chisnall B.L. et West D.W., 1996. Design and trials of a large fine-meshed fyke-net for eel capture and
factors affecting size distribution of catches. N-Z Journal Mar. Freshwat. Res., 30: 355-364.
Clauss T. et Meunier F., 1998. L’alevinage de « Sandoz » en demi-teinte. Eaux libres, 24 : 34-35
Cochran W.G., 1977. Sampling techniques (3 ème éd.). Wiley and sons, New-york: 428 p.
Colombo G. et Grandi G., 1996. Histological study of the development and sex differenciation of the
gonad in the European eel. Journal of Fish biology, 48: 493-512.
Colombo G., Grandi G., Rossi R., 1984. Gonad differenciation and body growth in Anguilla anguilla L. J.
Fish. Biol., 24: 215-228.
Commission des Communautés Européennes, 1997. Communication sur les impôts, taxes et
redevances environnementaux dans le marché unique. COM(97) 9 final 26.03.1997.
Copp G.H., 1990. Shifts in the microhabitat of larval and juvenile roach (Rutilus rutilus) in a floodplain
channel. J. Fish Biol., 36: 683-692.
Corsi F. et Ardizzone G.D., 1985. Some environnemental conditions affecting yellow eels catchability.
Oebalia, XI-2: 561-571.
Costa J. L., Assis C.A., Almeida P.R., Moreira F.M., Costa M.J., 1992. On the food of the European eel,
Anguilla anguilla (L.), in the upper zone of the Tagus estuary, Portugal. Journal of Fish Biology, 41: 841-
850.
Couillard C.M., Hodson P.V., Castonguay M., 1997. Correlations between pathological changes and
chemical contamination in American Eels, A. rostrata, from the St Lawrence River. Can. J. Fish. Aquat.
Sci., 54.
Couliou, J.-R., 2002. Site atelier de l’estuaire de la Loire. Rapport Final Contrat PECOSUDE
« Caractéristiques des petites pêches côtières et estuariennes de la côte atlantique du sud de
l’Europe », 59 pages. http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/site-atelier-loire-rapport-
1-5.pdf
Cowx I.G., 1983. Review of the methods for estimating fish population size from survey removal data.
Fish. Mgmt, 14: 67-82.
Cross D.G., Stott B., 1975. The effect of electric fishing on the subsequent capture of fish. J. Fish. Biol.,
7: 349-357.
Cucherousset J., Paillisson J.M., Carpentier A., Thoby V., Damien J.P., Eybert M.C., Feunteun E.,
Robinet T., in press. Freshwater protected areas : an effective measure to reconcile conservation and
exploitation of the threatened European eels (Anguilla anguilla)? Ecology of Freshwater Fish.
Cuende F.-X., Caill-Milly N., Prouzet P., 2002. Site atelier de l’estuaire de l’Adour. Rapport Final Contrat
PECOSUDE « Caractéristiques des petites pêches côtières et estuariennes de la côte atlantique du sud
de l’Europe », 43 pages. http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/site-atelier-adourrapport-5-5.pdf
INDICANG – Handbook - References 358

Cuende F.-X. et Marty S., 2005. Rapport d’étape Indicang bassin de l’Adour. Institution
Adour/Migradour. Novembre 2005: 32 p. http://www.ifremer.fr/indicang/boite-bassinsversants/pdf/rapport-etape-bv-adour-2005.pdf
Dannewitz J., Maes G.E., Johansson L., Wickström H., Volckaert F., Järvi T., 2005. Panmixia in the
European eel : a matter of time…, Proc. R. Soc B., 272, 1129–1137.
Dave G., Johannsson-Sjöbeck M.L., Larsson Å., Lewander K., Lidman U., 1979. Effects of cortisol on
the fatty acid composition of the total blood plasma lipids in the European eel, Anguilla anguilla L.
Comp. Biochem. Physiol. 64A: 37-40.
Daverat F., Limburg K.E., Thibault I., Shiao J.C., Dodson J.J., Caron F., Tzeng W.N., Iizuka Y.,
Wickström H., 2006. Phenotypic plasticity of habitat use by three temperate species, A. anguilla, A.
japonica, A. rostrata. Mar. Ecol. Prog. Ser., 308: 231-241.
Daverat F., Tomas J., Lahaye M., Palmer M., Elie P., 2005 Tracking continental habitat shifts of eels
using otolith Sr/Ca ratios : validation and application to the coastal, estuarine and riverine eels of the
Gironde-Garonne-Dordogne watershed. Marine and Freshwater Research 56 (5) 619-627.
Daverat F. et Tomas J., 2006. Tactics and demographic attributes in the European eel (A. anguilla) in
the Gironde watershed, SW France. Marine Ecology Progress Serie, 307: 247-257.
Davey A.J.H. et Jellyman D.J., 2005. Sex determination in freshwater eels and management options for
manipulation of sex. Reviews in Fish Biology and Fisheries, 15: 37-52.
Debenay B., Lambert P., Rochard E., 1995. Estimation de la quantité totale de civelles susceptibles de
transiter dans le circuit de refroidissement du CPNB durant la saison de migration 1994-1995.
Cemagref, groupement de Bordeaux. 42 p.
De Boer J. et Hagel P., 1994. Spatial differences and temporal trends of chlorobiphenyls in yellow eels
(A. anguilla) from inland waters of the Netherlands. Science of Total Environment, 141: 155-174.
De Boer J., Van der Valk F., Kerkhoff M.A.T., Hagel P., Brinkman U.A.T., 1994. 8-year study on the
elimination of PCBs and other organochlorine compounds from eel (A. anguilla) under natural
conditions. Environment Science & Technology, 28: 2242-2248.
De Casamajor, M.-N., Bru N., Prouzet P., 1999. Influence de la luminosité nocturne et de la turbidité sur
le comportement vertical de migration de la civelle d'anguille (Anguilla anguilla L.) dans l'estuaire de
l'Adour. Bull. Fr. Piscic., 355: 327-347.
De Casamajor, M.-N., Lecomte-Finiger R., Prouzet P., 2001. Détermination de l'état d'amaigrissement
des civelles (Anguilla anguilla) en migration en zones côtière et estuarienne. C.R. Acad. Sciences de la
Vie, 324: 345-353.
De Casamajor, M.-N., Lecomte-Finiger R., Prouzet P., 2003. Caractéristiques biologiques des civelles
(Anguilla anguilla) lors de la transition en estuaire. Bull. Fr. Piscic., 368:109-124.
De Casamajor, M.-N., Prouzet P., Lazure P., 2000. Identification des flux de civelles (Anguilla anguilla)
à partir des relations d'allométrie en fonction des conditions hydroclimatiques de l'estuaire de l'Adour.
Aquat. Living Resourc., 13: 411-420.
De Leo G.A. et Gatto M., 1995. A size and age-structured model of the European eel (Anguilla anguilla
L.). Canadian Journal of Fisheries and Aquatic Sciences, 52: 1351-1367.
INDICANG – Handbook - References 359

De Nie H.W., 1987. Food, feeding periodicity and consumption of the eel (A. anguilla) in the shallow
eutrophic Tjeukemeer (The Netherlands). Arch. Hydrobiol., 109: 421-443.
Deelder C.L., 1954. Factors affecting the migration of the silver eel in Dutch inland waters. J. Cons.
perm. int. Explor. Mer 20: 177-185.
Deelder C.L., 1958. On the behaviour of elvers (Anguilla vulgaris, Turt.) migrating from the sea into
fresh water. J. Cons. Int. Expl. Mer, 24: 135-146.
Deelder C.L., 1970. Synopsis of biological data on the eel (A. anguilla). FAO Fisheries Synopsis N°80,
75 p.
Deelder C.L., 1984. Exposé synoptique des données biologiques sur l’anguille, Anguilla anguilla
(Linnaeus, 1758). FAO Fisheries Synopsis 80, Revision 1: 73 p.
Dekker W., 2000a. A Procrustean assessment of the European eel stock. ICES Journal of Marine
Science 57: 938-947.
Dekker, W. 2000b. The fractal geometry of the European eel stock. ICES Journal of Marine Science 57:
109-121.
Dekker W., 2000c. Impact of yellow eel exploitation on spawner production in Lake Ijsselmeer, The
Netherlands. Dana 12: 17-32.
Dekker W., Van Os B., Van Willigen J., 1998. Minimal and maximal size of eel. Bulletin Français de la
Pêche et de la Pisciculture, 349: 15-198.
Dönni W., Maier K. J., Vicenti H., 2001. Bestandsentwicklund des Aals (Anguilla anguilla) im Hochrein.
Mitt. Zur Fischerei, Buwal, Bern, 27, 99 pp. In ICES, 2003.
Duplessy J.-C., 1999. Climate and the Gulf Stream. Nature, 402, 593 – 595.
Dufour S., 1985. La fonction gonadotrope de l’anguille européenne, Anguilla anguilla, au stade argenté
(au moment du départ pour la migration de reproduction): les mécanismes de son blocage et sa
stimulation expérimentale. Thèse de 3 ème cycle, Université de Paris VI.
Duquesne E., 2007. Analyse du comportement des civelles dans l’Isle et des pratiques de pêche dans
l’Isle et la Dordogne en fonction des conditions hydroclimatiques :Application à l’évaluation de la
biomasse saisonnière et des taux d’exploitation de 1996 à 2007. Mémoire de Master 2 professionnel :-
Biodiversité des écosystèmes continentaux et marins. Rapport de stage Ifremer – ULCO et USTL, 55p.
Durif C., 2003. La migration d’avalaison de l’anguille européenne Anguilla anguilla : caractérisation des
fractions dévalantes, phénomène de migration et franchissement d’obstacles. Thèse de doctorat,
Université de Toulouse III, 350 p.
Durif C., Dufour S., Elie P., 2005. The silvering process of Anguilla anguilla: a new classification from
the yellow resident to the silver migrating stage. Journal of Fish Biology, 66: 1025-1043.
Durif C., Elie P., Dufour S., Marchelidon J., Vidal B., 2000. Analyse des paramètres morphologiques et
physiologiques lors de la préparation à la migration de dévalaison chez l’anguille européenne (A.
anguilla) du Lac de Grand-Lieu (Loire-Atlantique). Cybium, 24: 63-74.
Dutil J-D., Giroux A., Kemp A., Lavoie G., Dallaire J-P., 1988. Tidal influence on movements and on
daily cycle of activity of American eels. Transaction of the American Fisheries Society, 117: 488-494.
INDICANG – Handbook - References 360

Edeline E., 2005. Facteurs de contrôle de la dispersion continentale chez l’anguille. Thèse Université
Toulouse III - Cemagref , 144 p.
Edeline E., Dufour S., Briand C., Fatin D., Elie P., 2004. Thyroïd status is related to migratory behaviour
in Anguilla anguilla glass eels. Mar. Ecol. Prog. Ser., 282: 261-270.
Edeline E., Lambert P., Rigaud C., Elie P., 2006. Effects of body condition and water temperature on
Anguilla anguilla glass eel migratory behavior. J. Exp. Marine Biol. Ecol., 331: 217-225.
EELREP, 2005. Estimation of the reproduction capacity of European eel. Final report, EU contract
Q5RS-2001-01836, 272p. Web site link: http://www.fishbiology.net/eelrepsum.html
Egusa S., 1979. Notes on the culture of the European eel (Anguilla anguilla L.) in Japanese eel-farming
ponds. Rapp. P.-v. Réun. Cons. Int. Explor. Mer, 174: 51-58.
Elie P., Edeline C., 1984. Etude de la population d'anguille (A. anguilla) de l'estuaire et du bassin
versant de la Vilaine: pêche, biologie, écologie. Examen particulier de l'impact du barrage d'Arzal sur la
migration anadrome. Université Rennes 1 - Cemagref – URFDAAPP : 175 p.
Elie P., Lecomte Finiger R., Cantrelle I., Charlon N., 1982. Définition des limites des différents stades
pigmentaires durant la phase civelle d’anguille Anguilla sp. (poisson téléostéen anguilliforme). Vie et
milieu, 32 (3) : 149-157.
Elie P. et Rochard E., 1994. Migration des civelles d’anguilles dans les estuaires, modalités du
phénomène et caractéristiques des individus. Bull. Fr. Pêche Pisc., 335: 81-98.
Epstein F.H., Cynamon M., Mc Kay W., 1971. Endocrine control of Na-K-ATPase and seawater
adaptation in Anguilla rostrata. Gen. Comp. Endocrinol. 16: 323-28.
Evans D. et Grainger R., 2002. Gathering data for resource monitoring and fisheries management. In
Handbook of fish biology and fisheries. Hart P.J.B. and Reynold J.D. (eds) Oxford, United Kingdom:
Blackwell : 84-102.
FAO, 1993. Reference points for fishery management : their potential application to straddling and
highly migratory resources. FAO Fisheries Circular, N° 864 : 52 p.
FAO, 1995. Code de conduite pour une pêche responsable. Rome, FAO : 46 p.
FAO, 1996a. Precautionary approach to fisheries. Part 2 : scientific papers. Prepared for the Technical
consultation on the precautionary approach to capture fisheries. Lysekil, Sweden, 6-13 June 1995. FAO
Fisheries technical paper, 350(2) : 210 p.
FAO, 1996b. Examen par sous-régions et par sous-secteurs de la situation de l’Aquaculture et des
pêches dans les eaux intérieures des pays de la CECPI. FAO Rapport sur les pêches N° 509, suppl. 1 :
91 p.
FAO, 1998. Pêches continentales. Directives techniques pour une pêche responsable, N°6. Rome,
FAO : 52 p.
FAO, 1999. Directives pour la collecte régulière de données sur les pêches de capture. Établies à la
Consultation d’experts FAO/DANIDA. Bangkok, Thaïlande, 18-30 mai 1998. FAO document technique
sur les pêches, N° 382. Rome, FAO : 123 p.
FAO, 2002. La situation mondiale des pêches et de l’aquaculture. FAO. Rome : 152 p.
INDICANG – Handbook - References 361

FAO, 2003. Review of the state of world fishery resources: inland fisheries. FAO fisheries circular N°
942, Rev. Rome, FAO : 60 p.
Ferchaud S., 2002. Site atelier de l’Ile d’Yeu. Rapport - Final Contrat PECOSUDE « Caractéristiques
des petites pêches côtières et estuariennes de la côte atlantique du sud de l’Europe », 36 pages.
Fernandez-Delgado C., Hernando J.A., Herrera M., Bellido M., 1989. Age and growth of yellow eels,
Anguilla anguilla, in the estuary of the Guadalquivir River (S W Spain). J. Fish Biol., 34: 561-570.
Feunteun E., 1994. Le peuplement ichtyologique du marais littoral endigué de Bourgneuf-Machecoul
(France, Loire-Atlantique). Approche méthodologique pour une analyse quantitative de la distribution
spatiale du peuplement ichtyologique et de la dynamique de certaines populations. Thèse de
l’Université de Rennes 1.
Feunteun E., 2001. Movements in inland habitats : adaptive value for anguillid eels. Advances in eel
biology, Tokyo 28-30 septembre 2001.
Feunteun E., 2002. Management and restoration of European eel population (Anguilla anguilla): an
impossible bargain. Ecological Engineering, 18: 575-591.
Feunteun E., Acou A., Guillouet J., Lafaille P., Legault A., 1998. Spatial distribution of an eel population
(A. anguilla) in a small coastal catchment of Northern Brittany (France). Consequences of hydraulic
works. Bull. Fr. Pêche. Pisc., 349: 129-139.
Feunteun E., Acou A., Lafaille P., Legault A., 2000. European eel (A. anguilla): prediction of spawner
escapment from continental population parameters. Canadian Journal of Fisheries and Aquatic
Sciences, 57: 1627-1635.
Feunteun E., Laffaille P., Robinet T., Briand C., Baisez A., Olivier J.M., Acou A., 2003. A Review of
Upstream Migration and Movements in Inland Waters by Anguillid eels. Toward a General Theory. In
Aida K, K Tsukamoto & K Yamauchi (Eds). Eel Biology, Springer-Verlag Tokyo, Chap. 14 : 191-213.
Feunteun E., Rigaud C., Elie P., Lefeuvre J-C., 1999. Les peuplements piscicoles des marais littoraux
endigués atlantiques : un patrimoine à gérer ? Le cas du marais de Bourgneuf-Machecoul (Loire-
Atlantique, France). Bull. Fr. Pêche Piscic., 352: 63-79.
Fontaine M., 1975., Physiological mechanisms in the migration of marine and amphihaline fish. Adv.
Mar. Biol. 13: 241-355.
Fontaine M. et Olivereau M., 1962. Nutrition et sexualité chez les poissons. Ann. Nutr. Alim. 16: 125-
152.
Fontaine Y.A., 1989. Les anguilles : migration et reproduction. Oceanis 15: 197-206.
Fontaine Y.A., 1994. L'argenture de l'anguille : métamorphose, anticipation, adaptation. Bull. Fr. Pêche
Piscic. 335: 171-85.
Fontaine Y.A., Pisam M., Le Moal C., Rambourg A., 1995. Silvering and gill « mitochodrial-rich » cells in
the eel, Anguilla anguilla. Cell Tissue Res. 281: 465-471.
Fontenelle G., 1991. Age et longueur moyenne des anguilles (A. anguilla) en Europe, une revue
critique. EIFAC working party on eel, Dublin. 16 p.
INDICANG – Handbook - References 362

Fontenelle G., Gascuel D., Mounaix B., 1990. Characteristics of an eel (Anguilla anguilla) fyke-net
fishery within a small estuary (River Blavet, France). Rev. Gesamt. Hydrobiol., 75: 797-806.
Fricke H. et Kaese R., 1995. Tracking of artificially matured eels (Anguilla anguilla) in the Sargasso Sea
and the problem of the Eel’s Spawning Site. Naturwissenschaften 82: 32-36.
Fries, L.T., Williams D.J., Johnson S.K., 1996. Occurrence of Anguillicola crassus, an exotic parasitic
swim bladder nematode of eels, in the southeastern United States. Transactions of the American
Fisheries Society. 125 : 794–797.
Frontier S., 1983. Stratégies d’échantillonnage en écologie. Masson, Paris. 494 p.
Frost, W. E. (1950). The eel fisheries of the river Bann, Northern Ireland, and observations on the age of
the silver eels. J. Cons. Perm. Int. Explor. Mer 16, 358-383.
Garcia S., 1994. The precautionary principles: its implications in capture fisheries management. Ocean
& Coastal Management, 22 : 99-125.
Garcia S. et Reveret J.-P., 1991. Recherche et structure des pêches artisanales : paradigmes et
méthodes de recherche. Une introduction. In : La Recherche face à la Pêche Artisanale, Symp. Int.
Orstom-Ifremer, Montpellier, France, 3-7 Juillet 1989, J.-R. Durand, J. Lemoalle et J. Weber (eds).
Paris, Orstom, 1991(1) : 253-268.
Garcia S.M. et De Leiva Moreno J.I., 2005. Evolution of the state of fish stocks in the Northeast Atlantic
within a precautionary framework, 1970-2003: a synoptic evaluation. ICES J. Mar. Sci. 62:1603-1608.
Gascuel D., 1986. Flow carried and active swimming migration of the glass-eel (A. anguilla) in the tidal
area of a small estuary on the French Atlantic coat. Helgol. Meeresunters, 40: 321-326.
Gascuel D., 1995. Efforts et puissances de pêche : redéfinition des concepts et exemple d'application.
In Gascuel D., Durand J.-L. and Fonteneau A. (eds). Premier forum halieumétrique. Les recherches
françaises en évaluation quantitative et modélisation des ressources et des systèmes halieutiques.
Rennes: ORSTOM Edition, Paris : 159-181.
Gerdeaux D., 1987. Revue des méthodes d’estimation de l’effectif d’une population par pêches
successives avec retrait. Programme d’estimation d’effectif par la méthode de Carle et Strub. Bull. Fr.
Pêche Pisc., 304: 13-21.
Gillis D.M., Peterman R.M., 1998. Implications of interference among fishing vessels and the ideal free
distribution to the interpretation of CPUE. Can. J. Fish. Aquat. Sci., 55 : 37-46.
Gillis D.M., Peterman R.M., Tyler A.V., 1993. Movement dynamics in a fishery : application of the ideal
free distribution to spatial allocation of effort. Can. J. Fish. Aquat. Sci., 50 : 323-333.
Girard P., 1998. Le poisson sentinelle des milieux aquatiques : pertinence et optimisation des
indicateurs sanitaires. Bull. Fr. Pêche Pisci : 300-351, 429-443
Girard P. et Elie P., 2007. Manuel d’identification des principales lésions anatomo-morphologiques et
des principaux parasites externes des anguilles. Etude Cemagref Bordeaux n° 110 : 81 p. Projet
Indicang : http://www.ifremer.fr/indicang/documentation/pdf/liste-annexes-guide-method.pdf
Girardin M, Castelnaud G, Beaulaton L., 2006. Surveillance halieutique de l'estuaire de la Gironde :
Suivi des captures 2004. Etude de la faune circulante 2005. Cemagref Cestas: 211.
INDICANG – Handbook - References 363

Glova G.J., 2001. Effects of the presence of subadult longfinned eels (A. dieffenbachii) on cover
preferences of juvenile eels (A. spp) in replicate channels. New Zealand Journal of Marine and
Freshwater Research, 35: 221-233.
Glova G.J., 2002. Density effects on juvenile shortfinned eel (A. australis) cover preferences in replicate
channels. New Zealand Journal of Marine and Freshwater Research, 36: 483-490.
Glova G.J., Jellyman D.J., Bonnett M.L., 2001. Spatio-temporal variation in the distribution of eel
(Anguilla spp) populations in three New Zealand lowland streams. Ecology of Freshwater fishes, 10:
147-153.
Goemans G. et Belpaire C., 2002. New Belgian standard for fish and fisheries products leads to catch
and release obligation for eel (Anguilla anguilla) for fishermen in Flanders. Workshop on PCBs in Brno
(Tchech republic), 7 – 10 may 2002.
Gosset C, Travade F., Durif C., Rives J., Garaicoechea C., 2000. Etude des dispositifs de dévalaison
pour l’anguille argentée : Test de deux exutoires de dévalaison à la centrale hydroélectrique de Halsou
(Nive- 64). Rapport de contrat INRA/EDF, 35 pages + annexes.
GRISAM, 2004. Compte-rendu de la réunion des 5 et 6 octobre 2004 à Bordeaux, 23 p.
GRISAM, 2006. Boite à outils pour la mise en place d’un plan de gestion national pour l’anguille
européenne. Rapport GRISAM, juin 2006.
Gueguen J., 1992. La gestion des ressources, un mal nécessaire. In « Cols bleus » n° 2196 du 19
décembre 1992 : 6-13.
Guérault D., Désaunay Y., Lecomte Finiger R., Grellier P. et Beillois P., 1995. Biometry and
otolithometry of Anguilla anguilla (L.) glass eels : towards a model for seasonal variation. In VIII th SEI
Congress, OVIEDO.
Guigue T., 1994. Etude socio-économique de la pêche professionnelle en eau douce dans le
département de la Gironde et proposition d’un programme d’appui technique à la profession. Mémoire
DAA, Spécialisation Halieutique. ENSA de Rennes : 114 p.
Gulland J.A., 1969. Manual of methods for fish stock assessment. Part 1 : Fish population analysis.
FAO Manuals in Fisheries science, n°4, 154 p.
Haddering R.H., 1982. Experimental reduction of fish impingement by artificial illumination at Bergum
power station. Int. Rev. Ges. Hydrobiol., 67 : 887-900. In Bruslé, 1994.
Haddering R.H., Van Der Stoep J.W., Habraken J.M., 1992. Deflecting eels from water inlets of power
stations with light. Irish. Fish. Biol., 27 : 139-149. In Bruslé, 1994.
Haro A., 2003. Downstream Migration of Silver-Phase Anguillid Eels. In: Aida, K., K. Tsukamoto & K.
Yamauchi (Eds). Eel Biology, Springer Verlag (Tokyo), 215-222.
Helfman G.S., Facey D.E., Hales L.S., Bozeman E.L., 1987. Reproductive ecology of the American eel.
Am. Fish. Soc. Symp. 1:42-56.
Higgins P.J., 1985. An interactive computer program for population estimation using the Zippin method.
Aqua. Fish. Manage., 1: 287-295.
INDICANG – Handbook - References 364

Hodson P.V., Castonguay M., Couillard C.M., Desjardins C., Pelletier E., McLeaod R., 1994. Spatial and
temporal variations in chemical contamination of American eels, Anguilla rostrata, captured in the
estuary of the St Lawrence River. Can. J. Fish. Aquat. Sci., 51: 464-478.
Holmgren K., 1996. Effect of water temperature and growth variation on the sex ratio of experimentally
reared eels. Ecology of Freshwater Fish, 5: 203-212.
Holmgren K., Mosegaard H., 1996. Implications of individual growth status on the future sex of the
European eel. J. Fish Biol., 49: 910-925.
Holmgren K., Wickström H., Clevestam P., 1997. Sex-related growth of the European eel (A. anguilla)
with focus on median silver eel age. Can. J. Fish. Aquat. Sci., 54: 2775-2781.
Hussein S.A., 1981. The population density, growth and food of eels Anguilla anguilla L. in some
tributaries of the river Tweed. In: Proceedings of the second British Freshwater Fisheries Conference,
University of Liverpool: 120-128.
Harley S.J., Myers R.A., Dunn A., 2001. Is catch-per-unit-effort proportional to abundance? Canadian
Journal of Fisheries and Aquatic Science, 58 : 1760-1772.
Hilborn R., 1985. Fleet dynamics and individual variation: why some people catch more fish than others.
Can. J. Fish. Aquat. Sci., 42 : 2-13.
Holden M.J., Rait D.F.S., 1994. Manuel de sciences halieutiques. Deuxième partie- Méthodes de
recherches sur les ressources et leur application. FAO document technique sur les pêches. N° 115 Rev.
1. Rome, FAO : 223 p.
Ibbotson A., Smith J., Scarlett P, Aprahamian P., 2002. Colonisation of freshwater habitats by the
European eel Anguilla anguilla. Freshwater biology 47 : 1696 – 1706.
ICES/EIFAC, Working group on eel. ICES CM 2002/ACFM : 03.
ICES/EIFAC, Working group on eel. ICES CM 2003/ACFM : 06.
ICES/EIFAC, Working group on eel. ICES CM 2004/ACFM : 09.
ICES/EIFAC, Working group on eel. ICES CM 2005/I : 01, Ref G, ACFM
ICES/EIFAC, Working group on eel. ICES CM 2006/ACFM : 16.
Jellyman D.J. et Ryan C.M., 1983. Seasonal migration of elvers (Anguilla spp.) into lake Ponui, New
Zealand, 1974-1978. New Zealand Journal of marine and Freshwater research, 17: 1-15.
Jellyman D.J. et Sykes J.R.E., 2003. Diel and seasonal movements of radio-tagged freshwater eels,
Anguilla spp., in two New Zealand streams. Environmental biology of fishes, 66: 143-166.
Jessop B.M., 1987. Migrating American Eels in Nova Scotia. Trans. Am. Fish. Soc. 116:161-170.
Kah O., Dufour S., Baloche S., Breton B., 1989. The GnRH systems in the brain and pituitary of normal
and hCG treated European silver eels. Fish Physiol. Biochem. 6:279-284.
Kendall W., 1999. Robustness of closed capture-recapture methods to violations of the closure
assumption. Ecology, 80: 2517-2525.
INDICANG – Handbook - References 365

Kennedy C.R. et Fitch D.J, 1990. Colonization, larval survival and epidemiology of the nematode
Anguillicola crassus, parasitic in the eel, Anguilla anguilla, in Britain. Journal of Fish Biology, 36: 117-
131.
Klechner R.C. et Krueger W.H., 1981. Changes in swimbladder retial morphology in Anguilla rostrata
during premigration metamorphosis. Journal of Fish Biology, 18: 569-577.
Kleckner R.C. et McCleave J.D., 1985. Spatial and temporal distribution of American eel larvae in
relation to North Atlantic Ocean current system. Dana, 4: 67-92.
Kleckner R.C., J.D. McCleave, G.S. Wippelhauser, 1983. Spawning of American eel, Anguilla rostrata,
relative to thermal fronts in the Sargasso Sea. Environmental Biology of Fishes, 9(3/4): 289-293.
Kleiber P. et Perrin C., 1991. Catch-per-effort and stock status in the U.S. North Pacific albacore
fishery : reappraisal of both. Fishery Bulletin, 89 : 379-385.
Klein-Breteler J.G.P., 1992. Effect of provenance and density on growth and survival of glass eels (A.
anguilla) in mesocosm experiments. Irish Fish. Invest., ser A (Freshwater), 36: 15-22.
Knights B., 1987. Agonistic behaviour and growth in the European eel, A. anguilla L, in relation to warmwater
aquaculture. Journal of Fish Biology, 31: 265-276.
Knights B., 1997. Risk assessment and management of contamination of eels (Anguilla spp) by
persistent xenobiotic organo-chlorine compounds. Ecology and Chemistry, 14: 171-212.
Knights B., 2003. A review of the possible impacts of long-term oceanic and climate changes and fishing
mortality on recruitment of anguillid eels of the Northern Hemisphere. The Science of the Total
Environment 310: 237-244.
Knights B., White E., Naismith I.A., 1996. Stock assessment of Europeen eel Anguilla anguilla L. In:
Cowx I.G. Stock assessment in inland fisheries Fishing news books: 431-447.
Kuhlmann H., 1976. Influence of temperature, food, initial size and origin on the growth and sexual
differenciation of elvers (A. anguilla). ICES/EIFAC Symposium on Eel Research and Management, 12,
17 p.
Kushnirov, D. et Degani G., 1995. Sexual dimorphism in yellow European eels, Anguilla anguilla (L.).
Aquaculture Research, 25 : 409-414.
Krueger W.H. et Oliveira K. 1997. Sex, size, and gonad morphology of silver American eels Anguilla
rostrata. Copeia 1997 : 415–420.
Krueger W.H. et Oliveira K., 1999. Evidence for environmental sex determination in the American eel,
Anguilla rostrata. Environ. Biol. Fish. 55: 381-389.
LaBar G.W. et Facey D.E., 1983. Local movements and inshore population sizes of American eels in
Lake Champlain, Vermont. Transaction of the American Fisheries Society, 112: 111-116.
LaBar, G.W., Hernando-Casal J.A., Delgado C.F., 1987. Local movements and population size of
European Eels Anguilla anguilla, in a small lake in southwestern Spain. Environnemental Biology of
Fish, 19: 111-117.
Labourg P.J., 1976. Les réservoirs à poissons du Bassin d’Arcachon. La Pisciculture française, 45: 35-
50.
INDICANG – Handbook - References 366

Laffaille P., Acou A., Guillouet J., 2005a. The yellow European eel (A. anguilla) may adopt a sedentary
lifestyle in inland freshwaters. Ecology of freshwater Fish, 14: 191-196.
Laffaille P., Acou A., Guillouët J., Mounaix B., Legault A., 2006. Patterns of silver eel (Anguilla anguilla
L.) sex ratio in a catchment. Ecology of Freshwater Fish, 15: 583-588.
Laffaille P., Baisez A., Lasne E., 2004a. Freshwater habitats of European eel. ICES Annual Science
Conference, 22-26 September, Vigo, Espagne.
Laffaille p., Baisez A., Rigaud C., Feunteun E., 2004b. Habitat preferences of different European eel
size classes in a reclaimed marsh: a contribution to species and ecosystem conservation. Wetlands, 24,
3: 642-651.
Laffaille P., Briand C., Fatin D., Lafage D., Lasne E., 2005b. Point sampling the abundance of European
eel (Anguilla anguilla) in freshwater areas. Arch. Hydobiol., 162: 91-98.
Laffaille P., Caraguel J.M., Legault A., sous presse. Temporal patterns in the upstream migration of
European glass eels (A. anguilla) at the Couesnon estuarine dam. Estuarine, Coastal and Shelf
Science.
Laffaille P., Feunteun E., Acou A., Lefeuvre J.-C., 2000. Role of European eel (Anguilla anguilla L.) in
the transfer of organic matter between marine and freshwater systems. Verh. Internat. Verein. Limnol.,
27: 616-619.
Laffaille P., Feunteun E., Baisez A., Robinet T., Acou A., Legault A., LEK S., 2003. Spatial organization
of European eel (A. anguilla) in a small catchment. Ecology of freshwater fish, 12: 254-264.
Lamaison G., 2005. Variabilité de la croissance de l’anguille européenne (A. anguilla) au sein du bassin
Gironde-Garonne-Dordogne. Rapport Cemagref Bordeaux - Université Bordeaux 1 : 43 p.
Lamarque P., 1976. Types de courant électrique à utiliser pour la capture optimale des poissons
principalement les anguilles. La Pisciculture Française, 47: 8 p.
Lambert P., 1997. Analyse des données de captures d'anguilles par pêche électrique et cartographie
des structures en tailles à l'échelle de la France. Rapport d'étude CSP - Cemagref, 50 p.
Lambert P., 2005. Exploration multiscalaire des paradigmes de la dynamique de la population
d’anguilles européennes à l’aide d’outils de simulations. Thèse de doctorat, Université Bordeaux 1 –
Cemagref.
Lambert, P., Feunteun E., Rigaud C., 1994. Etude de l'anguille en marais d'eau douce. Première
analyse des probabilités de capture observées lors des inventaires par pêche électrique. Bull. Fr. Pêche
et Piscic., 335: 111-121.
Lambert P. et Rigaud C., 1999. Recherche d'éléments de gestion de la population d'anguilles sur la
base des données produites par la RHP. Rap. Cemagref – CSP, 49 : 63 p.
Lammens E.H.R.R. et Visser J.T, 1989. Variability of mouth width in European eel, Anguilla anguilla, in
relation to varying feeding conditions in three dutch lakes. Env. Biol. Fishes, 26: 63-75.
Lamothe P.J., Gallagher M., Chivers D.P., Moring J.R., 2000. Homing and movement of yellow-phase
American eels in freshwater ponds. Environ. Biol. Fish., 58: 393-399.
INDICANG – Handbook - References 367

Langon M. et Dartiguelongue J., 1997. La dévalaison des anguilles (Anguilla anguilla) argentées. Test
de deux exutoires de dévalaison à la centrale hydroélectrique E.D.F. d’Halsou (Nive-64). Rapp. SCEAA,
Toulouse, 100 p. + tableaux.
Larinier M., Chanseau M., Rigaud C., Steinbach P., 2006. Eléments d’aide à la définition d’une stratégie
de restauration des axes de migration de l’anguille. Note GHAAPPE RA 06-01, 23 p.
Larinier M. et Dartiguelongue J., 1989. La circulation des poissons migrateurs : le transit à travers les
turbines des installations hydroélectriques. Bull. Fr. Pêche et Piscic. 312. 94 p.
Larsen K., 1972. Studies on the biology of Danish stream fishes. On seasonal fluctuations in the stock
density of the yellow eel in shallow stream biotopes and their causes. Meddelelser Fra Danmark Fiskeri,
OG Havundersogelser, 7(2): 23-46.
Lasne E., Laffaille P., sous presse. Analysis of distribution patterns of yellow European eels in the Loire
catchment using logistic models based on presence-absence of different size-classes. Ecology of
Freshwater Fish.
Laurec A. et Le Guen J.C., 1981. Dynamique des populations marines exploitées. Tome I : concepts et
modèles. Rapports scientifiques et techniques CNEXO, 45 : 117 p.
Léauté J.-P. et Caill-Milly N., 2003. Caractéristiques des petites pêches côtières et estuariennes de la
côte Atlantique du sud de l'Europe. Synthèse du contrat européen PECOSUDE n°99/024 ED/DG FISH
(DGXIV). Ifremer : 66 p.
Lee E T.W., 1979. Dynamique des populations d'anguilles A. anguilla des lagunes du bassin
d'Arcachon. Thèse Université Montpellier : 218 p.
Lefebvre F., Contournet P., Crivelli A.J., 2002. The health state of the eel swimbladder as a measure of
parasite pressure by Anguillicola crassus. Parasitology, 124: 457-463.
Lefebvre F., Mounaix B., Poizat G., Crivelli A.J., 2004. Impacts of the swimbladder nematode
Anguillicola crassus on Anguilla anguilla: variations in liver and spleen masses. J. Fish Biology, 64: 435-
447.
Le Fur J., Cury P., Laloë F., Durand M.-H., Chaboud C. 1999. Co-viabilité des systèmes halieutiques.
Nature Sciences Société, 7 : 19-32.
Legault A., 1987. L’anguille dans le bassin de la Sèvre Niortaise. Biologie, Ecologie, Exploitation.
Publication Halieutique ENSAR : 305 p. + annexes.
Legault A., 1990. Gestion des barrages estuariens et migration des anguilles. Int. Rev. ges. Hydriol., 75:
819-825.
Legault A., 1992. Etude de quelques facteurs de sélectivité des passes à anguilles. Bull. Fr. Pêche
Pisci., 325: 83-91.
Legault A., 1994. Etude préliminaire du recrutement fluvial de l'anguille. Bulletin Français de la Pêche et
de la Pisciculture, 335: 33-41.
Legault A., 1996. Colonization dynamics of a catchment area by eel. Characterization of migrating
populations in a free access river. In: Leclerc et al (Eds). Proceedings of 2nd Int. Symp. on
Habitats/Hydraulics. Vol. A, INRS-Eau/Iahr, Québec, Canada: pp 89-98.
INDICANG – Handbook - References 368

Legault A. et Feunteun E., 1992. Etude de la migration estivale des anguilles au barrage de Golfech.
Rapport Fish Pass - MIGADO, 26 p.
Legault A., Laffaille P., Guillouët J., Acou A., 2004. Importance of specific fish passes for European eel
(Anguilla anguilla L.) recruitment. In: Proceedings of the Fifth International Symposium on
Ecohydraulics, Madrid. Aquatic Habitat: Analysis & Restoration. De Jalon Lastra, D.G., Martinez P.V.
(eds.). IAHR, Madrid: pp. 937-941.
Legay J. M., 1993. Ici et maintenant, plus tard et ailleurs. Natures Sciences Sociétés, 1 : 144-147.
Lissardy M., De Casamajor M.-N., Sanchez F., Prouzet P., Morandeau G., Cuende F.-X., 2004.
Caractérisation et abondance des captures professionnelles estuariennes dans le bassin de l’Adour en
2003. Rapport IFREMER-CERECA, 34 pages + annexes.
Lissardy M., De Casamajor M.-N., Sanchez F., Prouzet P., Morandeau G., Cuende F.-X., 2005.
Caractérisation et abondance des captures professionnelles estuariennes dans l’estuaire de l’Adour en
2004. Rapport Cereca-Ifremer-Institution Adour : 32 p. + annexes.
Lobon-Cervia J., Bernat Y., Rincon P.A., 1990. Effects of eels (A. anguilla) removals from selected sites
of a stream on its subsequent densities. Hydrobiologia, 206: 207-216.
Lobon-Cervia J. et Carrascal M., 1992. Seasonal timing of silver eels in a Cantabrian stream. Arch. of
Hydrobiologie, 125: 121-126.
Lobon-Cervia J., Utrilla C.G., Rincon P.A., 1995. Variations in the population dynamics of the European
eel Anguilla anguilla (L.) along the course of a Cantabrian river. Ecology of Freshwater Fish, 4: 17-27.
Luneau S., Changeux T., Mertens D., 2003. Guide des engins de pêche fluviale et lacustre en France
métropolitaine. Collection Mise au Point, Conseil Supérieur de la Pêche : 198 p.
McCarthy, T.K. et Cullen P., 2000. Eel fishing in the River Shannon: Eel population changes, fishery
management options and fishery conservation issues. A synthesis report on the River Shannon eel
Management Programme 1992-2000. Electricity Supply Board, Dublin. 21pp.
McCleave J. D., 1993. Physical and behavioural control on the oceanic distribution and migration of
leptocephali. J Fish Biol, 43, 243-273.
McCleave J.D., 2001. Simulation of the Impact of Dams and Fishing Weirs on Reproductive Potential of
Silver-Phase American Eels in the Kennebec River Basin, Maine. North American Journal of Fisheries
Management 21:592-605.
McCleave J.D. et Klechner R.C., 1982. Selective tidal stream transport in the estuarine migration of
glass eels of the American eel (A. rostrata). Journal Cons. Intern. Explor. Mer, 40: 262-271.
McCleave J.D., Kleckner R.C., Castonguay M., 1987. Reproductive sympatry of American and
European eels and implications for migration and taxonomy. American Fisheries Society Symposium, 1:
286-297.
McCleave J.D. et Wippelhauser G.S., 1987. Behavioural aspects of selective tidal stream transport in
juvenile American eels. Amer. Fish. Soc. Symp., 1: 138-150.
Mahon R., 1980. Accuracy of catch-effort methods for estimating fish density and biomass in streams.
Envir. Biol. Fish., 5: 343-360.
INDICANG – Handbook - References 369

Mann H., 1965. Über das Rückkehrvermogen verpflanzter Fluss-Aal. Arc. für Fishereiwissenschalft, 15:
177-185.
Mann R.H.K., Blackburn J.H., 1991. The biology of the eel (A. anguilla) in an English chalk stream and
interactions with juvenile trout Salmo trutta and salmon Salmo salar. Hydrobiologia, 218: 65-76.
Maes G.E., Pujolar J.M., Hellemans B. et Volckaert F.A.M., 2006. Evidence for isolation by time in the
European eel (Anguilla anguilla L.). Mol. Ecol., 15, 2095-2107.
Maes G.E. et Volckaert F.A.M., 2002. Clinal genetic variation and isolation by distance in the European
eel Anguilla anguilla (L.). Biol. J. Limn. Soc., 77, 509-521.
Masse J., Rigaud C., 1998. L'anguille et les marais littoraux. In: Hussenot et Buchet (eds) Colloque
« Marais maritimes et aquaculture » , Ifremer, Série Actes de colloques, n°19 : 141-154.
Matthews, M, Evans D., Rosell R. Moriarty C. et Marsh I., 2001. The Erne eel enhancement
programme, Final Report. Northern Regional Fisheries Board, Ballyshannon, Co Donegal, Ireland. EU
Peace and reconciliation Project No EU15.
Melia P., Bevacqua D., Crivelli A.J., De Leo G.A., Panfili J., Gatto M., 2006. Age and growth of Anguilla
anguilla in the Camargue lagoons. Journal of Fish Biology, 68: 876-890.
Meunier F.J., 1994. Données sur la croissance de l'anguille (A. anguilla) dans le cours moyen du Rhin,
région alsacienne. Bull. Fr. Pêche Piscic., 335: 133-146.
Michel P. et Oberdorff T., 1995. Feeding habits of fourteen European freshwater fish species. Cybium,
19: 5-46.
Mohr H., 1971. The effect of some behaviour patterns on the catching techniques for eel. EIFAC
Technical Paper, 14: 27-30.
Möller H., S. Holst, H. Lüchtenberg and F. Peterson, 1991. Infection of eel Anguilla anguilla from the
river Elbe estuary with two nematodes, Anguillicola crassus and Pseudoterranova decipiens. Diseases
of Aquatic Organisms 20 : 163 – 170.
Monten E., 1985. Fish ans turbines. Fish injuries during passage through power station turbines.
Vattenfall, Stockholm, 111 p.
Moran P.A.P., 1951. A mathematical theory of animal trapping. Biometrika, 38: 307-311.
Moreau G. et Barbeau C., 1982. Les métaux lourds comme indicateurs d'origine géographique de
l'anguille d'Amérique (A. rostrata). Can. Journal Fish. Aquat. Sci., 39: 1004-1011.
Moriarty C., 1973. Studies of the eel Anguilla anguilla in Ireland. 2. In Lough Conn, Lough Gill and North
Cavan Lakes. Irish Fisheries Investigations Ser. A, 13: 1-13.
Moriarty C., 1986a. Observations on the eels of Meelicks Bay, Lough Derg, 1981-1984. Vie et Milieu,
36: 279-283.
Moriarty C., 1986b. Riverine migration of young eels Anguilla anguilla (L.). Fisheries Research, 4: 43-58.
Moriarty, C., 1987. Factors influencing recruitment of the Atlantic species of anguillid eels. Amer. Fish.
Soc. Symp., 1: 483-491.
INDICANG – Handbook - References 370

Moriarty C., 1997. The European eel fishery in 1993 and 1994: First Report of a working group funded
by the European Union Concerted Action AIR A94-1939. Fisheries Bulletin (Dublin) 14, 52 pp.
Moriarty C., 2001. Maintenance and development of the Corrib Catchment eel fishery. A report
commissioned by the Western Regional Eel Fisherman's Association. 34 p.
Moriarty C. et Dekker W., 1997. Management of the European eel. Fish. Bull. (Dublin), 15: 110 p.
Morris S., Allchin C.R., Zegers B.N., Hafta J.J.H., Boon J.P., Belpaire C., Leonard P.E.G., van der
Leeuwen S.P.J., de Boer J., 2004. Distribution and fate of HBCD and TBBP-A brominated flame
retardants in North Sea estuaries and aquatic food webs. Environmental Science & Technology, 38:
549-504.
Morrison W.E. et Secor D.H., 2004. Abundance of yellow-phase American eels in the Hudson river
estuary. Transaction of the American Fisheries Society, 133: 896-910.
Mounaix B., 1992. Intercalibration et validation des méthodes d’estimation d’âge chez l’anguille
européenne (A. anguilla). Application au bassin versant de la Vilaine. Thèse ENSA Rennes : 146 p.
Mounaix B. et Fontenelle G., 1994. Anguilles estuariennes et fluviales : apport de l'otolithométrie. Bull.
Fr. Pêche Piscic., 335: 67-80.
Mullon C., Freon P., Cury P., 2005. The dynamics of collapse in world fisheries. Fish and Fisheries, 6 :
111-120.
Myers R.A., Hutchings J.A., Barrowman N.J., 1997. Why do fish stock collapse ? The example of cod in
Atlantic Canada. Ecological Applications, 7 : 91-106.
Naismith I.A. et Knights B., 1988. Migration of elvers and juvenile European eels, Anguilla anguilla L., in
the River Thames. Journal of Fish Biology, 33: 161-177.
Naismith I.A. et Knights B., 1990a. Studies of sampling methods and of techniques for estimating
populations of eels, Anguilla anguilla L. Aquaculture and Fisheries Management, 21: 357-367.
Naismith I.A. et Knights B., 1990b. Modelling of unexploited and exploited populations of eels (Anguilla
anguilla) in the Thames estuary. Journal of Fish Biology, 37: 975-986.
Naismith I.A. et Knights B., 1993. The distribution, density and growth of the European eel (A. anguilla)
in the freshwater catchment of the River Thames. J. Fish Biol., 42: 217-226.
Neis B., Schneider D.C., Felt L., Haedrich R.L., Fischer J., Hutchings J.A., 1999. Fisheries assessment:
what can be learned from interviewing ressource users. Can. J. Fish. Aquat. Sci., 56 : 1949-1963.
Nelva A., Persat H., Chessel D., 1979. Une nouvelle méthode d’étude des peuplements ichtyologiques
dans les grands cours d’eau par échantillonnage ponctuel d’abondance. C.R. Acad. Sci. Paris, 289:
1295-1298.
Neveu A., 1981a. Variations saisonnières et journalières de l’alimentation de l’anguille (Anguilla anguilla
L.) dans des conditions naturelles. Oecol. Applic., 2: 99-116.
Neveu A., 1981b. Densité et microrépartition de différentes espèces de poissons dans la Basse Nivelle,
petit fleuve côtier des Pyrénées Atlantiques. Bull. Fr. Pêche Pisci., 280: 86-102.
INDICANG – Handbook - References 371

Nielsen T. et Prouzet P., 2007. Capture-based aquaculture of the wild european eel (Anguilla anguilla).
Document FAO sous presse, 29p.
Nilo P. et Fortin R., 2001. Synthèse des connaissances et établissement d’une programmation de
recherche sur l’anguille d’Amérique (Anguilla rostrata). Univ. Du Québec à Montréal, Dépt Sciences
Biologiques pour la Société de la faune et des parcs du Québec, Direction de la recherche sur la faune,
Québec, 298 p.
Oberdorff T., Pont D., Hugueny B., Chessel D., 2001. A probabilistic model characterizing fish
assemblages of French rivers: a framework for environnemental assessment. Freshwater Biology, 46:
399-416.
Oliveira K., 1999. Life history characteristics and strategies of the American eel, Anguilla rostrata. Can.
J. Fish. Aquat. Sci. 56:795-802.
Oliveira K., McCleave J.D., Wippelhauser G.S., 2001. Regional variation and the effect of lake: river
area on sex distribution of American eels. Journal of Fish Biology, 58: 943-952.
Ooi, H.-K., Wang W.-S, Chang H.-Y., Wu C.-H., Lin C.-C., Hsieh M.-T., 1996. An epizootic of
anguillicolosis in cultured American eels in Taiwan. Journal of Aquatic Animal Health, 8 (2) : 163 – 166.
Opuszynski K., 1965. Observations on young eels (A. anguilla), age and biology. Roczniki Nauk.
Rolniczych, 86 B, 393-411.
Pallo S. et Travade F., 2001. Suivi du fonctionnement de la passe définitive à anguilles sur
l’aménagement hydroélectrique EDF de Tuilières (24). Contrôle de la migration et mise au point des
compteurs automatiques. Année 2001. Rapport MIGADO-EDF : 40 p et annexes.
Palstra A.O., Cohen E.G.H., Niemantsverdriet P.R.W., Van Ginneken V.J.T., Van Den Thillart,
G.E.E.J.M., 2005. Artificial maturation and reproduction of European silver eel : development of oocytes
during final maturation. Aquaculture 249 (1-4) : 533-547
Palstra A.P., Van Ginneken V.J.T., Murk A.J., Van Den Thillart, 2006. Are dioxin-like contaminants
responsible for the eel (Anguilla anguilla) drama ? ICES CM 2006/J : 34. 5 p.
Panfili J., 1993. Estimation de l’âge individuel des poissons : méthodologies et application à des
populations naturelles et tempérées. ORSTOM Paris, Trav. Doc. Microfich.: 456 p.
Panfili J. et Ximenes M.C., 1994. Evaluation de l'âge et croissance de l'anguille européenne (A.
anguilla) en milieu continental. Méthodologie, validation, application en Méditerranée, comparaison en
Europe. Bull. Fr. Pêche Pisc., 335 : 43-66.
Panfili J., Ximenes MC., Crivelli A.J., 1994. Sources of variation in growth of the European eel (Anguilla
anguilla) estimates from otoliths. Canadian Journal of Fish and Aquatic Sciences, 51: 506-515.
Pankhurst N.W., Lythgoe J.N., 1982. Structure and colour of the integument of the European eel
Anguilla anguilla (L.). J. Fish. Biol. 21: 279-296.
Pankhurst N.W., 1982a. Relation of visual changes to the onset of sexual maturation in the European
eel Anguilla anguilla (L.). J. Fish. Biol. 21: 127-140.
Pankhurst N.W., 1982b. Changes in body musculature with sexual maturation in the European eel,
Anguilla anguilla (L.). J. Fish. Biol. 21: 417-428.
INDICANG – Handbook - References 372

Pankhurst N.W., 1982c. Changes in the skin-scale complex with sexual maturation in the European eel,
Anguilla anguilla (L.). J. Fish. Biol. 21: 549-561.
Parsons J., Vickers K.U., Wardern Y., 1977. Relationship between elver recruitment and changes in the
sex ratio of silver eels Anguilla anguilla L. migrating from Lough Neagh, Northern Ireland. J. Fish. Biol.
10:211-229.
Parker G.A., 1992. The evolution of sexual size dimorphism in fish. Journal of Fish Biology, 41: 1-20.
Parker S.J., 1995. Homing ability and home range of yellow phase American eels in a tidally dominated
estuary. Journal of the Marine Biology Association of the UK, 75: 127-140.
Pauly D., 1995. Anecdotes and the shifting baseline syndrome of fisheries. Trends in Ecology &
Evolution, 10 : 430.
Peňáz M. et Tesch F.W., 1970. Geschlechtsverhältnis und Wachstum beim Aal (Anguilla anguilla) an
verschiedenen Lokalitäten von Nordsee und Elbe. Ber. Dt. wiss. Komm. Meeressforsch., 21: 290-310.
Persat H. et Copp G.H., 1990. Electric fishing and point abundance sampling for the ichtyology of large
rivers. In Developments in electric fishing, Cowx Eds, Fishing News Books: 197-209.
Philippart J-C., 1979. Observations concernant l’efficacité de la pêche à l’électricité dans une rivière de
la zone à barbeaux. Bull. Fr.. Pêche Pisc., 273: 157-171.
Peters G. et Hartmann F., 1986. Anguillicola, a parasite nematode of the swimbladder spreading among
eel populations in Europe. Diseases of Aquatic Organisms 1: 229-230.
Poinsard F. et Le Guen J.-C., 1975. Observations sur la définition d’une unité d’effort de pêche
applicable à la pêcherie de thon tropical africain. Rapp. PV réunion Cons. Int. Explor. Mer, 168, 39-43.
Poole W., Reynolds J.D., 1996. Growth rate and age at migration of Anguilla anguilla. Journal of Fish
Biology, 48: 633-642.
Poole W.R., Reynolds J.D., 1998. Variability in growth rate in European Eel Anguilla anguilla (L.) in a
western Irish catchment. Proc. Roy. Irish Acad. (Biol. & Environ.) 98B:141-145.
Prignon C., Micha J.C., Gillet A., 1998. Biological and environmental characteristics of fish passage at
the tailfer dam on the Meuse river, Belgium. In Jungwirth, M. Schmutz S. and Weiss S. ‚eds (1998). Fish
migration and fish bypasses. Fishing News Books, 69 – 84 – Tumpenny, A. W. H.
Prouzet P. (coordinateur) 2002. Historique des captures de civelles, intensité de leur exploitation,
variation de leur capturabilité par la pêche professionnelle maritime et indices de colonisation sur le
bassin versant de l’Adour - Projet EC/DG FISH 99/023 –Septembre 2002, 147 pages + annexes.
http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/historique-capture-civelle.pdf
Prouzet P. (coordinateur), 2003. Etude sur la civelle (Anguilla anguilla) dans l’estuaire de l’Adour :
Pêche, biologie, comportement. Modélisations hydrodynamique et comportementale. Estimation des
flux de civelles en estuaire. Rapport de contrat Ifremer/ DRV/RH-LA – Agence de l’Eau Adour-Garonne
–MEDAD-Institution Adour, 295p. http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/progamphihalins-civelle.pdf
Prouzet P. 2003. Les pêches et ressources estuariennes et continentales : Importance, évolution et
contraintes anthropiques. Un cas particulier, l’anguille. In Exploitation et surexploitation des ressources
marines vivantes. RST 17, 6.7, 233-248.
INDICANG – Handbook - References 373

Prouzet P., 2004. INDICANG: un projet pour synthétiser nos connaissances sur l’anguille et sur la
qualité de ses habitats par la mise en réseau d’opérations de suivis régionaux à une échelle de
dimension européenne. Chroniques Européennes, 54 : 50-51.
Prouzet P., Bouvet J.-C., Bru N., Boussouar A., 2008. Quantification de la biomasse de civelles
(Anguilla anguilla) dans l’estuaire de la Loire et estimation du taux d’exploitation saisonnier de la pêche
professionnelle pour les saisons de pêche 2003 à 205. Rapport Indicang
http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/rapport-final-pecherie-bv-gironde.pdf, 45p.
Pujolar J. M., Maes G. E. et Volkaert F. A. M., 2006. Genetic patchiness among recruits in the European
eel Anguilla anguilla. Mar Ecol Prog Ser, 307, 209-217.
Quensière J., 1993. De la modélisation halieutique à la gestion systémique des pêches. Nature
Sciences Société, 1 : 211-220.
Radke R.J. et Eckmann R., 1996. Piscivorous eels in Lake Constance; can they influence year class
strength of perch ? Ann. Zool. Fennici, 33: 489-494.
Rasmussen G. et Therkildsen B., 1979. Food, growth and production of Anguilla anguilla in a small
Danish stream. Rap. P-V Réun. CIEM, 174: 32-40.
Richkus W.A. et Whalen K.G., 1999. American Eel (Anguilla rostrata). Scoping study report. EPR, Palo
Alto, California TR- 111873.
Richards L.J. et Maguire J.J., 1998. Recent international agreements and the precautionary approach:
new directions for fisheries management science. Can. J. Fish. Aquat. Sci., 55 : 1545-1552.
Ricker W.E., 1980. Calcul et interprétation des statistiques biologiques des populations de poissons.
Ministère des pêches et des océans, Ottawa, Bulletin 191F : 409 p.
Rigaud C., Fontenelle G., Gascuel D., Legault A., 1988. Le franchissement des ouvrages hydrauliques
par les anguilles (A. anguilla). Présentation des dispositifs installés en Europe. Publication ENSA
Rennes, Biol-Halieutique, 9: 147 p et annexes.
Robins C.R., Cohen D.M., Robins C.H., 1979. The eels, Anguilla and Histobranchus, photographed on
the floor of the deep Atlantic in the Bahamas. Bulletin of Marine Science, 29: 401-45.
Robinet T. et Feunteun E., 2002. Sublethal effects of exposure to chemical compounds. A cause for
decline in Atlantic eels ? Ecotoxicology, 11: 265-277.
Rochard E., 1992. Mise au point d’une méthode de suivi de l’abondance des amphihalins dans le
système fluvio-estuarien de la Gironde. Application à l’étude écobiologique de l’esturgeon Acipenser
stutio. Thèse Biologie des populations et des écosystèmes, Université Rennes 1 – Cemagref: 295 p. et
annexes.
Rodriguez M.A., 2002. Restricted movement in stream fish: the paradigm is incomplete, not lost.
Ecology, 83: 1-13.
Roqueplo C., Lambert P., Gonthier P., 2000. Estimation de la mortalité des civelles de la Gironde après
leur passage dans le circuit de refroidissement du la centrale nucléaire du Blayais. Cemagref / EDF.
54p.
INDICANG – Handbook - References 374

Rommel S.A. et McCleave J.D., 1973.Sensitivity of American eels (Anguilla rostrata) and Atlantic
salmon (Salmo salar) to weak electric and magnetic fields. Journal of Fisheries Research Board of
Canada 30: 657-663.
Rommel S.A. et Statsko A.B., 1973. Electronavigation by eels. Sea Frontiers, 19: 219-223.
Rosell, R., Evans D., Allen, M., 2005. The eel fishery in Lough Neagh, Northern Ireland. An example of
sustainable management ? Fisheries Management and Ecology, 12: 377-385.
Rossi, R. et Cannas, A., 1984. Eel fishing management in a hypersaline lagoon of Southern Sardinia.
Fish. Res., 2 : 285-298.
Rossi R. et Colombo G., 1976. Sex ratio, age and growth of silver eels in two brackish lagoons in the
Northern Adriatic Valli of Comacchio and Valle Nuova. Archo. Oceanogr. Limnol., 18: 327-341.
Rossi R. et Colombo G., 1979. Some observations on age, sex, and growth of silver eels (Anguilla
anguilla L.) in North Adriatic Lagoons. Rapp. PV. Reun. Cons. Int. Explor. Mer, 174: 64-69.
Rossi R. et Villani P., 1980. A biological analysis of eel catches, Anguilla anguilla, from the lagoons of
Lesina and Varano, Italy. Journal Fish. Biology, 16: 413-423.
Russell I.C. et Potter C.E., 2003. Implications of the precautionary approach for the management of the
European eel, Anguilla anguilla. Fisheries Management and Ecology, 10: 395-401.
Santillo D., Allsopp M., Walters A., Johnston P., Perivier H., 2006. Presence of perflorinated chemicals
in eels from 11 European countries. Green Peace Report, Note 07/2006: 31 p.
Scherrer B., 1984. Biostatistique. Gaëtan Morin Editeur, 850pp.
Schmidt J., 1906. Contributions to the life-history of the eel (Anguilla vulgaris, Flem.). Rapports et
Procès verbaux des Réunions du Conseil Permanent et International pour l’Exploration de la Mer. 5(4) :
137 – 264.
Schmidt J., 1922. The breeding places of the eel. Philosophical Transactions of the Royal Society of
London. Series B, 385(211): 179 – 208.
Schmidt J., 1925. The breeding places of the eel. Annual report of the Board of Regents of the
Smithsonian Institution, report for 1924: 279-316.
Schnute J., 1983. A new approach to estimating population by removal method. Can. J. Fish. Sci., 40:
2153-2169.
Schoth M. et Tesch F.-W., 1982. Spatial distribution of 0-group eel larvae (Anguilla sp.) in the Sargasso
Sea. Helgoländer Meeresuntersuchungen, 35: 309-320.
Schulze T., Kahl U., Radke R.J., Benndorf J., 2004. Consumption, abundance and habitat use of
Anguilla anguilla in a mesotrophic reservoir. Journal of Fish Biology, 65: 1543-1562.
Schwarz C.J. et Seber G.A.F., 1999. Estimating Animal Abundance : Review III. Statistical Science 14,
n°4, 427-456.
Seber G.A.F., Le Cren, E.D., 1967. Estimating population parameters from catches large relative to the
population. The Journal of Animal Ecology, 36: 631-643.
INDICANG – Handbook - References 375

Sinha U.R.P., Jones J.W., 1967. On the age and growth of the freshwater eel (Anguilla anguilla). J.
Zool. Cond., 153: 99-117.
Sinha, U.R.P. et Jones J.W., 1975. The European freshwater Eel. Liverpool Univ. Press. 146 p.
Sloane R.D., 1984. Preliminary observation of migrating adult freshwater eels (Anguilla australis
australis, Richardson) in Tasmania. Aust. J. Mar. Freshwat. Res., 35: 471-476.
Smith M.W. and J. . Saunders, 1955. The American eel in certain freshwaters of the maritime provinces
of Canada. Journal of Fisheries Research Board of Canada, 12(2): 238 – 269.
Smogor R.A., Aangermeier P.L., Gaylord C.K., 1995. Distribution and abundance of American Eels in
Virginia Streams: Tests of null models across spatial scales. Trans. Am. Fish. Soc., 124: 789-803.
Souchon Y., Andriamahéfa H., Cohen P., Breil P., Pella H., Lamouroux P., Malavoi J.R., Wasson J.G.
2000. Régionalisation de l’habitat aquatique dans le bassin de la Loire. Agence de l’Eau Loire Bretagne,
Cemagref Bely/LHQ. 297 p.
Stephenson R.L. et Lane D.E., 1995. Fisheries management science: a plea for conceptual change.
Canadian Journal of Fisheries & Aquatic Science, 52 : 2051-2056.
Steinmetz B., 1990. Electric fishing: some remarks on its use. In Developments in electric fishing, Cowx
Eds, Fishing News Books: 1-4.
Strubberg A., 1913. The metamoprhosis of elvers as influenced by outward conditions. Meddr. Komm.
Havunders, Ser. Fisk., 4: 1-11.
Susperregui N., (coordinateur) 2007. Rapport final Indicang : caractérisation de la pêcherie
professionnelle de civelles sur l’Isle – Bassin versant Gironde – Garonne – Dordogne.
http://www.ifremer.fr/indicang/boite-bassins-versants/pdf/rapport-final-pecherie-bv-gironde.pdf
Svedäng H., 1999. Vital population statistics of the exploited eel population on the Swedish west coast.
Fish. Research, 40: 251-265.
Tapie N., Budzinski H., Elie P., Gonthier P., 2006. Contamination en polychlorobiphényls (PCB) des
anguilles du système fluvio-estuarien de la Gironde. Rapport final, Novembre 2006, LPTC Bordeaux-
Cemagref Bordeaux, 58 p.
Tesch F.-W., 1977. The Eel. London Chapman et Hall.
Tesch F.-W. et G. Wegner, 1990. The distribution of small larvae of Anguilla sp. Related to hydrographic
conditions 1981 between Bermuda and Puerto Rico. Internationale Revue der gesamten Hydrobiologie,
6: 845-858.
Teyssier F., Gayou F., Filszinger P., Galiay E., Lemoine A., Taillebois L., 2002. Plan de gestion des
poissons migrateurs amphihalins Garonne – Dordogne – Charente - Leyre - Seudre. Situations et
propositions. Rapport CSP, Groupe anguille, 43 pages.
Tesch F.W., 1977. The eel. Biology and Management of Anguillid eels. Chapman & Hall, London, 434p.
Thillart G., Dufour S., Elie P., Volckaert F., Sebert P., Rankin C., Szekely C., Rijsingen J.V., 2005.
Estimation of the reproduction capacity of European eel. Rapport Q5RS-2001-01836. 272 p.
INDICANG – Handbook - References 376

Thompson A.J., Sargent J.R., 1978. Changes in the levels of chloride cells and (Na-K) dependant
ATPase in the gills of yellow and silver eels adapting to seawater. J. Exp. Zool. 200: 33-40.
Todd P.R., 1981a. Morphometric changes, gonad histology, and fecundity estimates in migrating New
Zealand freshwater eels (Anguilla spp.). New Z. J. Mar. Freshw. Res. 15: 155-170.
Todd P.R., 1981b. Timing and periodicity of migrating New Zealand freshwater eels (Anguilla spp.). New
Z. J. Mar. Freshw. Res. 15: 225-235.
Tongiorgi P., Balsamo M., Tosi L., 1986. Thermal preferences in upstream migrating glass-eels of
Anguilla anguilla (L.). J. Fish. Biol., 28: 501-510.
Troadec J.-P., 1986. Aménagement des pêches et des cultures marines en zones littorales:
perspectives et axes de recherche. La pêche Maritime, 1297 : 271-275.
Truong-Van B., Prouzet P., 2001. Méthodes de captures successives: synthèse, quelques contributions,
une application. Journal de la Société Française de Statstique, 142(3), 53-87.
Tzeng W.N., Cheng P.W., Lin F.Y., 1995. Relative abundance, sex ratio and population structure of the
Japanese eel Anguilla japonica in the Tanshui River system of northern Taiwan. J. Fish Biol., 46: 183-
201.
Tzeng W.N., Severin K.P., Wickström H., 1997. Use of otolith microchemistry to investigate the
environmental history of European eel (A. anguilla). Marine Ecol. Progress Series, 149: 73-81.
Van Baalen M. et Hochberg M.E., 2001. Dispersal in antagonistic interactions. In: Dispersal,.Clobert J.,
Danchin E;, Dhondt A.A., Nichols J.D., Eds, 299-310.
Van Den Thillart G., Van Ginneken V., Körner F., Heijmans R., Van Der Linden R., Gluvers A., 2004.
Endurance swimming of European eel. J. Fish. Biol. 65: 312-318.
Van Ginneken, V. et van den Thillart G., 2000. Eel fat stores are enough to reach the Sargasso. Nature
403:156-157.
Van Ginneken V., Antonissen E., Müller U.K., Booms R., Eding E., Verreth J., Van Den Thillart G.,
2005a. Eel migration to the Sargasso: remarkably high swimming efficiency and low energy costs. J.
Exp. Biol. 208: 1329-1335.
Van Ginneken V., Vianen G., Muusze B., Palstra A., Verschoor L., Lugten O., Onderwater M., Van
Schie S., Niemantsverdriet P., Vvan Heeswijk R., Eding E.P., Van Den Thillart G.. 2005b. Gonad
development and spawning behaviour of artificially-matured European eel (Anguilla anguilla L.). Animal
Biology 55: 203-218.
Van Ginneken V. Durif C., Balm S.P., Boot R., Verstegen M.W.A., Antonissem E., Van Den Thillart G.,
2007. Silvering of European eel (Anguilla anguilla L.): seasonal changes of morphological and metabolic
parameters. Animal Biology, 57: 63-77.
Venables W.N et Ripley B.D., 2002. Modern Applied Statistics with S. Fourth Edition, Springer Eds,
495p.
Vigier J-F, 1990. Les pathologies des anguilles. Coll. « Etudes » Cemagref, Série « Gestion des Milieux
Aquatiques », n°12 : 198 p.
INDICANG – Handbook - References 377

Voetgle B. et Larinier M., 2000. Etude des capacités de franchissement des civelles et anguillettes. Site
de Tuilières (Dordogne) et du barrage d’Arzal (estuaire de la Vilaine). Rapport GHAAPPE RA 00.05, 69
p et annexes.
Vollestad L.A., 1986. Growth and production of female yellow eels (Anguilla anguilla L.) from brackish
water in Norway. Vie & Milieu, 36: 267-271.
Vollestad L.A., 1988. Tagging experiment with yellow eel, Anguilla anguilla (L.) in brackish water in
Norway. Sarsia, 73: 157-161.
Vøllestad L.A., 1990. Optimal management of European eel in the Imsa River, Norway. J. Const. int.
Explor. Mer 46:115-120.
Vollestad L.A., 1992. Geographic variation in age and length at metamorphosis of maturing European
eel: environmental effects and phenotypic plasticity. Journal of Animal Ecology, 61: 41-48.
Vollestad L.A. et Jonsson B., 1988. A 13-year study of the population dynamics and growth of the
European eel (A. anguilla) in a Norvegian river: evidence for density-dependant mortality, and
development of a model for predicting yield. Journal of Animal Ecology, 57: 983-997.
Vøllestad L.A., Johnson B., Hvidsten N.-A., Naesje T.F., 1994. Experimental test of environmental
factors influencing the seaward migration of European silver eels. J. Fish Biol. 45:641-651.
White E.M. et Knights B., 1997a. Dynamics of upstream migration of the European eel (A. anguilla) in
the rivers Severn and Avon, England, with special reference to the effects of man-made barriers.
Fisheries Management and Ecology, 4: 311-324.
White E.M. et Knights B., 1997b. Environmental factors affecting migration of the European eel in the
Rivers Severn and Avon, England. Journal of Fish Biology, 50: 1104-1116.
Wiberg U.H., 1983. Sex determination in the European eel (Anguilla anguilla L.) : a hypothesis based on
cytogenetic results, correlated with findings of skewed sex ratio. Cytogen. Cell Gen. 36: 589–598.
Wickins J.F., 1985. Growth variability in individually confined elvers (A. anguilla). J. Fish Biol., 27: 469-
478.
Winn H.E., W.A. Richkus and L.K. Winn, 1975. Sexual dimorphism and natural movements of the
American eel (Anguilla rostrata) in Rhode Island streams and estuaries. Helgoländer Wissenschaftliche
Meeresuntersuchugen, 27: 156-166.
Wirth T. et Bernatchez L., 2001. Genetic evidence against panmixia in the European eel. Nature, 409,
1037–1039.
Wirth T. et Bernatchez L., 2003. Decline of North Atlantic eels: a fatal synergy ? Proc. R. Soc. Lond.
270, 681-688.
Ximenes M.C., 1986. L’anguille en Méditerranée française. Aspects écobiologiques et halieutiques.
Rapport Cemagref Montpellier, 166 p.
Yamaguchi N., Gazzard D., Scholey G., Mac Donald D.W., 2003. Concentrations and hazard
assessment of PCBs, organochlorine pesticides and mercury in fish species from the upper Thames.
Chemosphere, 50: 265–273.
INDICANG – Handbook - References 378

Yamamoto K. and Y. Yamauchi, 1974. Sexual maturation of Japanese eel and production of eel larvae
in the aquarium. Nature, 251: 220-222.
Zalewski M., Cowx I.G., 1990. Factors affecting the efficiency of electric fishing. In Fishing with
electricity. Eds Cowx & Lamarque, Fishing News Books.
Zippin C., 1956. An evolution of the removal method of estimating animal populations. Biometrics, 12:
163-189.
Zippin C., 1958. The removal method of population estimation. J. Wildlife Manag. 32: 325-339.
INDICANG – Handbook - References 379