VUB final report IRMA-SPONGE 2 - Homepages Vub Ac - Vrije ...

IRMA-SPONGE 

Subproject 2 

Development of Flood Management Strategies for Rhine and 

Meuse Basins in the Context of Integrated River Management 

Discharge Coefficients for the Kikbeek Subbasin, 

a Brook Subbasin at the Belgian Side of 

the River Border Meuse (Grensmaas) 

P. van Rossum, O. Batelaan*, T. Asefa, and F. De Smedt 

VUB Contribution to the Final Report 

Vrije Universiteit Brussel (VUB) 

Department of Hydrology and Hydraulic Engineering 

Brussels, Belgium 

December 2001 

* Author to whom correspondence should be addressed: Vrije Universiteit Brussel (VUB), 

Department of Hydrology and Hydraulic Engineering, Pleinlaan 2, 1050 Brussels, Belgium. 

Tel: +32-2-629.30.39, Fax: +32-2-629.30.22, e-mail: batelaan@vub.ac.be

IRMA-SPONGE Subproject 2: VUB Contribution to the Final Report 

Preface 

This report and the IRMA-SPONGE Umbrella Program 

In recent years, several developments have contributed not only to an increased public interest in flood 

risk management issues, but also to a greater awareness of the need for improved knowledge 

supporting flood risk management. Important factors are: 

• Recent flooding events and the subsequently developed national action plans. 

• Socio-economic developments such as the increasing urbanisation of flood-prone areas. 

• Increased awareness of ecological and socio-economic effects of measures along rivers. 

• Increased likelihood of future changes in flood risks due to land use and climate changes. 

The study leading to this report aimed to fill one of the identified knowledge gaps with respect to flood 

risk management, and was therefore incorporated in the IRMA-SPONGE Umbrella Program. This 

program is financed partly by the European INTERREG Rhine-Meuse Activities (IRMA), and 

managed by the Netherlands Centre for River Studies (NCR). It is the largest and most comprehensive 

effort of its kind in Europe, bringing together more than 30 European scientific and management 

organisations in 13 scientific projects researching a wide range of flood risk management issues along 

the Rivers Rhine and Meuse. 

The main aim of IRMA-SPONGE is defined as: “The development of methodologies and tools to 

assess the impact of flood risk reduction measures and scenarios. This to support the spatial planning 

process in establishing alternative strategies for an optimal realisation of the hydraulic, economical 

and ecological functions of the Rhine and Meuse River Basins." A further important objective is to 

promote transboundary co-operation in flood risk management. Specific fields of interest are: 

• Flood risk assessment. 

• Efficiency of flood risk reduction measures. 

• Sustainable flood risk management. 

• Public participation in flood management issues. 

More detailed information on the IRMA-SPONGE Umbrella Program can be found on our website: 

www.irma-sponge.org. 

We would like to thank the authors of this report for their contribution to the program, and sincerely 

hope that the information presented here will help the reader to contribute to further developments in 

sustainable flood risk management. 

Ad van Os and Aljosja Hooijer 

(NCR Secretary and IRMA-SPONGE project manager) 

ii


Executive summary IRMA-SPONGE subproject 2 

Development of flood management strategies for the Rhine and Meuse 

basins in the context of integrated river management 

Uncertainties in water management 

Willem van Deursen, Carthago Consultancy 

Hans Middelkoop, Universiteit Utrecht 

The aim of this project is to develop a methodology to find integrated, robust water management 

strategies for the Rhine and Meuse basin. The formulation of water management strategies is complex 

due to uncertainties in future water management. These uncertainties exist in the future physical 

boundary conditions for water management, such as uncertainties in future climate change and sea 

level rise. Uncertainties are also introduced as a result of various yet unknown socio-economic and 

agro-economic developments that will affect water demand (such as population growth, industrial 

expansion) or the hydrological cycle (land use changes, urbanisation, use of different crop types). 

These uncertainties are related to future developments in the society, the client of water management. 

Added to these uncertainties are the uncertainties related to the models and concepts we use to 

formulate and analyse water management. Overall, all these uncertainties are related to assumptions 

and choices, relate to the perspectives of the parties involved in water management. 

This project deals with incorporating uncertainties in future flood management strategies in the Rhine 

and Meuse basins. This uncertain future gives the boundary condition for formulating strategies for 

water management. Even if we could agree on the set of future conditions for which we should 

develop strategies, we would still face a tremendous set of possible management options. Safety can 

be obtained by rising dikes and embankments or by widening the floodplains of the river or by 

increasing retention in the catchments in all its different forms. Different people appraise these options 

from different backgrounds and different perspectives, and there is no overall winner amongst the 

possible strategies. 

The objectives of this project are thus to provide an analysis of the uncertainties related to climate 

change and land use change and their hydrological response, and to provide a method to formulate 

robust strategies for flood management. 

Why robust strategies? Robust strategies are strategies that remain valid even if the assumptions on 

which they were based change. Robust strategies are flexible towards the future, changing conditions 

can be accommodated. Robust strategies explicitly deal with uncertainties, and incorporate the 

analysis of the uncertainties in the formulation of the strategy. 

iii


Perspectives on water management 

The Perspectives Method is used as a framework for a structured analysis of the uncertainties of future 

developments. The basic rationale of this method is that the uncertainties, which are associated with 

the assumptions, preferences and choices, provides the opportunity for numerous valid interpretations 

of how social, economic and environmental processes evolve. These subjective interpretations of 

uncertainties can be clustered into a limited number of viewpoints or perspectives. These perspectives 

reflect the choices concerning structural uncertainties throughout the whole cause-effect chain of 

social, economic and physical changes in a river basin as a result of human interventions. 

The perspectives can be characterised according to a typology of cultures or individual’s social context 

or ways of life. These ways of life are the Hierarchist (strong group boundaries and binding 

prescriptions), the Individualist (weak group boundaries and little prescribed roles) and the Egalitarian 

(strong group involvement and minimal regulation). The three perspectives are considered as 

extremes. The resulting spectrum, defined by these extremes, comprises a variety of rather hybrid, 

more moderate world views and management styles. 

A typology of perspectives 

In the Egalitarian perspective, it is assumed that people are, in principal, good, but that they can be 

influenced easily. These might be negative influences but humans can be guided positively by means 

of intimate relationships with other people and nature. Personal development can be obtained by 

spiritual growth rather than by consumption of goods. The Egalitarian world view implies an attitude 

of risk avoidance. The management style belonging to this can, therefore, be characterised as being a 

preventative strategy. The Egalitarian perspective advocates drastic and structural social, cultural and 

institutional changes in the current capitalistic economic system. Nature is considered extremely 

vulnerable and small disturbances can have catastrophic consequences. Human activities which affect 

the natural environment must therefore be avoided. 

In the Hierarchical perspective, people are sinful by nature. However, people can be controlled (and 

educated) by a proper government and institutions. Regulation, management and control must prevent 

large problems. This management style can be characterised by an attitude of accepting some risks. In 

this perspective, nature is robust within certain limits: nature is able to overcome small disturbances. 

However, crossing certain limits causes serious trouble for the way in which nature functions. The 

hierarchical perspective emphasises the relation between humans and nature where the mutual 

dependence and balance between both parties is important. In this perspective, an attempt is made to 

guarantee this balance. 

In the Individualistic perspective, human nature is egocentric and based on personal gains. In this 

perspective, people are considered as rational, self-assured actors trying to satisfy their material needs. 

Changes and uncertainties are interpreted as challenges and can, in principle, be solved. This 

perspective is characterised by a large belief in market mechanisms and technology. The management 

style can be characterised as being adaptive. Nature is assumed to be extremely robust and is able to 

survive a few disturbances. Anthropogenic influence, even if large, results in mild and harmless 

disruption. In this perspective people are considered the centre of the world and natural resources are 

at the service of people and can be exploited. 

iv


The Perspectives Method was not developed with a specific focus on water management, so the 

general descriptions of the recognised perspectives has to be specified for water management related 

descriptions. 

Interpreting Perspectives with respect to water management 

Perspective 

Heuristic rules 

Egalitarian Hierarchical Individualistic 

Focus Nature and the environment Control and a responsible 

government 

Heuristic rule 1 Nature is vulnerable and 

environmental risks are 

avoided; prevention is better 

than cure. 

Stability through regulation, 

hierarchy and standards; 

regulation of nature and the 

environment; acceptance of 

differences. 

Heuristic rule 2 Equity. Avoiding risks and against 

changes; easy does it, 

otherwise you’ll break the 

line. 

Heuristic rule 3 Economy as a means and not 

as an objective; conscious 

consumption. 

Heuristic rule 4 People have solidarity and 

behave as such; collective 

interest. 

Authority through expertise 

and experience. 

Power and esteem are the 

motives for action. 

Economy and the individual 

responsibility 

Free market mechanism and 

anti-regulation; economic 

growth and technical 

development equal progress. 

Individual development and 

material self-interest are 

motives for action; success is 

a personal responsibility. 

Problems can be solved; risks 

produce opportunities and 

challenges. 

We define a Perspective as a consistent and coherent description of how the world functions and how 

policy should be carried out. In this definition, a Perspective has two dimensions: a World View and a 

Management Style. The World View is a coherent description of how the world functions. The 

Management Style is a coherent set of preferred policy options. 

Utopias and Dystopias 

If the World View and Management Style coincide we speak of an Utopia. If this is not the case there 

is a Dystopia. Dystopias describe what could happen if the world functions according to a perspective 

different to the perspective on which the policy strategy is based. Or vice versa, if reality functions in 

line with one’s favoured world view, but opposite strategies are applied. Thus, in terms of scenario 

development and model experiments, dystopias are future pathways involving ‘mismatches’ between 

world view and management style. In reality, dystopias are often the result of interplay of forces 

between actors. 

v


MANAGEMENT 

STYLE 

Utopia and dystopia 

WORLD VIEW 

Egalitarian 

Hierarchist 

Individualist 

Egalitarian 

UTOPIA 

DYSTOPIA 

DYSTOPIA 

Hierarchist 

DYSTOPIA 

UTOPIA 

DYSTOPIA 

Individualist 

DYSTOPIA 

DYSTOPIA 

UTOPIA 

This project has defined cases for each of the entries in the above table. These cases were discussed 

with experts and stakeholders. The cases were quantified: for each case a scenario for land use and 

climate change and management style was developed. These scenarios were used as input for a set of 

hydrological simulation models. The results of these modelling sessions were evaluated for the 

implications of the management styles in Utopia and Dystopia situations. In addition, the current water 

management practises of The Netherlands, Germany and Belgium were evaluated according to the 

perspective method. 

Simulation models and sensitivity analyses for the Perspectives Method 

Climate Change scenarios for the defined cases at the Rhine basin scale were developed by KNMI. 

Institute of Hydraulic Engineering of the University of Stuttgart focused on the development of 

downscaling methods for regional climate change scenarios and a methodology to the establish 

atmospheric circulation patterns which can subsequently be associated with flood events. Land use 

change (and other socio-economic change) scenarios for the entire basin comes from literature review. 

For the German regional land use changes, patterns and trends, Potsdam Institute for Climate Research 

(PIK) developed a methodology to provide geographically distributed land use patterns based on 

current land use allocation. The scenarios were used by PIK and Vrije Universiteit Brussel as input for 

regional hydrological models (WASIM-ETH and WetSpass) to examine the hydrological sensitivity of 

the Kikbeek and the Lein catchments to climate change and land use change. For the entire basin, the 

Rhineflow and Meuseflow water balance models were used. The effects of the proposed measures 

under the developed climate change scenarios for the floodplains were evaluated with the Landscape 

Planning of the Rhine-DSS. 

vi


Current water management in perspective 

The main observations from the inventory of sixteen Dutch integrated studies, plans and visions that 

the Dutch water management definitely is a Hierarchistic management style. As in the archetype, 

studies and reports form an important part of policy making, and each aspect and detail is further 

explored. Social factors, sea level rise and climate change are considered to be the largest source of 

uncertainty, although one study mentions the limited knowledge of the complex river system and the 

methods used as important sources of uncertainty. 

The result of the literature review and the inventory of existing studies, visions and strategies for 

Germany and Belgium was relatively poor. This yields the conclusion that that comprehensive and 

recent German and Belgium surveys of long-term future developments of the Rhine and Meuse are not 

available. The survey of Belgium and German studies do offer scenario descriptions, but are 

essentially an exploration of the technical and physical limits of water management. Because of this 

limited number of integrated studies and their decentralised character for water management, both 

Germany and Belgium seem to have a Hierarchist/Individualist based water management. They are 

definitely less pronounced Hierarchist than the Netherlands. 

The modelling exercises and the analysis of their results, combined with the discussions and 

workshops with experts yielded a number of observations for each perspective. The project basically 

evaluated the three scenario families, based on the central statement ‘water management according to 

perspective X results in…’. By confronting each management style with different futures, both utopian 

and dystopian, overall conclusions could be derived for each perspective and associated management 

style. 

Egalitarian water management 

The Egalitarian strategy is focused on the causes of water-problems, instead of dealing symptoms and 

effects (Individualist) or focussing on actors (Hierarchist). The approach aims at a sustainable 

solutions and high resilience of the water system. The strategy involves major environmental and 

landscaping measures resulting in large spatial claims for restoration and expansion of nature. A 

positive side effects is that this leads to a higher quality of life. However, due to the scarcity of space 

in the Netherlands, the increasing demands for room for nature and water may increase the pressure on 

other nature reserves, such as the Veluwe. Many of the landscaping measures, such as lowering 

floodplains and transforming agriculture areas into nature, are irreversible. Furthermore, the 

implementation costs of the strategy are very high, and other functions (such as industrial and urban 

expansion, inland navigation, agriculture) are subordinate to the protection and expansion of water and 

nature. In dystopian situations, when the expected calamities do not occur, the drastic measures and 

large costs have been futile. However, the strategy is more flexible than the Individualist strategy. It 

allows a change to another water management strategy if time proves that the risks are smaller than 

perceived initially. The Egalitarian faith in a combination of economic austerity and a desire for 

psychological and socio-cultural well-being will result in a long-term stabilisation or even curbing of 

climate change, thereby reducing the long term flood risks. In other words, this management style 

suggests favourable futures if one does not mind high expenses. 

vii


Hierarchist water management 

The Hierarchist aims at so-called win-win situations. In doing this, the Hierarchist avoids making real 

choices. In general, this compromise strategy of ‘running with the hare and hunting with the hounds’ 

has the largest public support. The strategy needs regular adjustments of the policy, is time consuming 

and expensive. The risk is that it gets stuck in conferences and sluggish decision-making Only a few 

measures might actually implemented, most likely those that are acceptable but not effective. It does 

not yield firm safety guarantees but if climate change appears to be insignificant the costs have been 

for nothing. Some of the futures associated with this management strategy thus run the risk of 

becoming ‘loss-loss’ situations. Because the Hierarchist tries to serve all functions but is confronted 

with limited financial and land resources, it is likely that all functions suffer. Per situation and per 

point in time no stakeholder is fully satisfied. Stakeholder interests may change considerably in 

response to events and calamities such as droughts or floods and the strategy can be characterised as 

reactive and fully ‘controlled’ by external factors and incidents. This water management strategy is 

actually not a vision on water policy, but a vision on how to organise water management. Because the 

management style actually addresses uncertainty associated with future developments through 

incrementalism (versus drastic measures) it is flexible and allows for changes to other management 

styles. 

Individualist water management 

The Individualistic management style can be characterised as passive and displays a short-term vision 

for water management measures. The Individualist aims at reducing cost, stimulating economic 

benefits, thereby accepting a relatively high, calculated risk. Measures will be implemented as 

adaptations to changing conditions. Irreversible damage to natural systems in the floodplains and the 

occupation of potential retention areas render it difficult to change to another management style. Large 

adjustments to accommodate an unforeseen drastic climate change will not be possible. On the short 

term, compared to the others, this strategy is relatively cheap. However, one of the risks associated 

with the Individualists distopias is that economic growth induces further climate change. The world 

associated with Individualistic management is thus extremely vulnerable for calamities, i.e. low 

probability events happening. In case of extreme flooding, because of high economic value and 

damage potential along the rivers the economic impacts are large. The future associated with the 

Individualistic management strategy is characterised as economic wealth, but even in the utopian case 

results in a lower quality of life in the broader sense. Focussing on water management, it is obvious 

that the Individualistic approach to uncertainty should be characterised as risk-taking. Summarising: 

low short term costs, but high long-term risks. 

Robustness of management styles 

Evaluating the various management styles we conclude that the Egalitarian management style is the 

most robust one, mainly due to the aspirations associated with safety and nature. The undisputed price 

tag is that it involves high expenses and large spatial claims. 

The Hierarchistic management style fulfils the objectives of integrated (win-win) solutions with 

nature, safety and reversible measures in most cases. However, in dystopian situations, investment 

cost will be high without leading to safety and reduced flood risks. In case of a serious climate change 

(> 2 °C temperature rise in the next 50 years), there may be no possibilities left for finding win-win 

solutions for all water functions. Also, the complexity of the planning process may result in a slow 

viii


response in case of severe climate change. The Hierarchist management style is thus less robust in 

view of a changing environment and an uncertain future. 

The Individualistic management style is adequately characterised as high risk taking but cost-efficient, 

at least in the short term. Positive impacts for all relevant functions (except nature) do only materialise 

in case the external context develops according to the Individualist assumptions and when the river 

basins systems are as robust as the Individualist assumes. In dystopian cases, technical measures have 

to be applied to counteract the climate-induced changes, which may lead to high cost. The 

Indvidualistic management style can therefore be considered as the least robust in view of 

uncertainties associated with external context and functioning of the water system. 

No management style is superior over all others under all conditions evaluated on the whole set of 

evaluation criteria. A major difficulty in the ‘cost-benefit’ assessment is the weighing of advantages 

against disadvantages, because they are of a different kind. One of the key differences between the 

perspectives is their inherent choices on the implementation cost versus accepted risk. Finding the 

balance between risks and costs is indisputably a policy dilemma that cannot be solved by using an 

ingenious water management strategy. Political decisions on water management unavoidably involves 

trade-offs on normative grounds. 

The vast majority of current policy plans on water management in the Netherlands falls within the 

Hierarchist perspective for their management style. If this strategy is not superior to the others it 

becomes opportune to analyse whether the results of the present project would advocate a different 

water management style. 

The results do not provide enough arguments for this change of strategy. At present, a switch to the 

Individualistic management style is not advocated, because it would be unwise to neglect the 

possibility of serious climate change in view of the current level of uncertainty. An Individualistic 

water management strategy decreases the capability to cope with potential future climate change. In 

theory, however, if the Hierarchist management style would appear more risky than the Individualist, 

the Individualistic management strategy is preferred, because it is less costly, and leaves more room 

for other functions. 

From a safety point of view, it can be advocated to switch to the Egalitarian management style, 

because it is the most robust strategy. The main issue remains however, whether society is 

ready/willing to pay the costs in financial terms and in terms of spatial claims. If the current 

Hierarchist water management proves to be more expensive than the Egalitarian strategy, a shift to the 

Egalitarian management style is supported. In this case, the latter yields more safety and nature at 

lower costs. 

Similar comparisons may be made on the basis of, for example, resilience, ecological values, or the 

possibility of combining different functions at all places within the water systems. However, it is clear 

that it would be a bad policy to put all eggs in the Hierarchist basket. The Hierarchist water 

management strategy has to be continuously evaluated in terms of relative risk (compared to the 

Individualist water management strategy) and relative costs (compared to the Egalitarian water 

management strategy). 

ix


Conclusions and Recommendations 

1) At the scale of the entire Rhine basin, climate change impacts cannot be compensated by land use 

changes, as the influence of climate change on extreme floods is much stronger than the influence 

of land use measures. 

2) Flood risk management in the lower river deltas cannot be based on the assumption that extreme 

floods can be prevented by upstream measures. This is because it is not certain that upstream flood 

retention measures will be implemented and that they are as effective as anticipated, especially 

under very extreme flow conditions. 

3) The effects of landuse changes on peak discharges in small catchments are limited and strongly 

depend on the type of precipitation (convective vs. advective) and antecedent conditions, implying 

that … 

4) … future peak flows in small catchments depend on the changes in variability of precipitation. 

Estimates of changes in extreme precipitation and precipitation variability currently rely heavily 

on the results of downscaling methods of precipitation obtained from global climate models. 

5) Current Dutch flood risk management can be characterised as complying with a Hierarchist 

management style (cf Thompson)), while German and Belgian management styles have common 

characteristics with an Individualistic style. 

6) Under changing climate conditions, the Hierarchist type of management runs the risk of becoming 

an expensive attempt to fully control flood risk problems, without actually solving the problems in 

a long-term view. 

7) No flood risk management strategy is superior in all respects and in all circumstances. Flood risk 

management is not merely a technical optimisation problem: safety versus societal costs is really a 

policy dilemma. (Win-win situations cannot always be attained). 

8) The three Cultural Perspectives applied in the present study do not fully discriminate between all 

differences in water management when considering the international dimension. Additional 

dimensions for characterisation differences in national management styles are therefore needed. 

9) Considering the present-day and future uncertainties for water management in the Rhine and 

Meuse basins research should be more aimed at defining integrated and coherent scenarios that 

can underpin adequate water management strategies given the uncertainties. 

10) This should be done by combining social sciences with environmental sciences, and by combining 

physical/mathematical modelling tools with expert sessions and participatory stakeholder 

processes 

11) Integration of water management and spatial planning is essential, because spatial claims often 

collide with claims for water management, which is likely to result in higher risks and higher 

costs. 

12) Perspective-based flood risk management scenarios should not only consider the temporal 

dimension with different lines of future development, but also should take into account differences 

in management styles within the river basin. 

x


Abstract of VUB contribution to subproject 2 

The activities of the Vrije Universiteit Brussel (VUB) within the framework of the IRMA-SPONGE 

subproject 2 focussed on the hydrological aspects of the river Border Meuse (Grensmaas in Dutch) 

and one of its Belgian subbasins: the Kikbeek subbasin. The first part of the research concerned a 

literature research on present Belgian (Flemish) governmental and non-governmental policies and 

proposals for the river and water management on the Belgian side of the Border Meuse. This literature 

research included literature on related nature restoration and development programmes in the area. The 

second part of the research concerned the hydrological characteristics of the Kikbeek subbasin, which 

served as an example for the Belgian subbasins of the Border Meuse. 

The hydrological characteristics of the Kikbeek subbasin were analysed by way of a hydrological 

modelling with WetSpass, a GIS based, spatially distributed hydrological model that has been 

developed at the VUB. Based on geographical input data, such as topographic data (slopes), soil types, 

land use, groundwater tables, and average seasonal meteorological data (temperature, precipitation, 

potential evapotranspiration, and wind speed), and sets of model parameters that are based on physical 

and empirical relationships, the WetSpass model calculated the spatially distributed yearly and 

seasonal evapotranspiration, groundwater recharge, and surface runoff in the Kikbeek subbasin. 

Subsequently, fast and slow discharge coefficients could be calculated from these spatial data, as these 

discharge coefficients are related to the total surface runoff and groundwater recharge, respectively. 

The sensitivity of the discharge coefficients of the Kikbeek subbasin towards climate and land use 

changes was analysed by the modelling of a number of (combined) climate and land use scenarios. For 

this modelling, the WetSpass model of the actual hydrological situation in the Kikbeek subbasin was 

used as a starting point, and only the geographical input data had to be adjusted. The climate scenarios 

were, on one side, realistic (wet) greenhouse scenarios in which the temperature and precipitation will 

increase proportionally. On the other side, climate scenarios were modelled in which the climate will 

change due to a sudden change in the North Atlantic oceanic circulation or in which the temperature 

and precipitation changes are not coupled (dry scenarios). In the land use scenarios, the actual land use 

data were adjusted in such a way that the effects of extreme, but possible, land use changes could be 

analysed. For instance, all non-urban areas were turned into agricultural land, or all agricultural land 

and meadows were turned into deciduous forests or urban areas. 

1


Table of contents 

Preface ...............................................................................................................................................ii 

Executive summary IRMA-SPONGE subproject 2................................................................................iii 

Abstract of VUB contribution to subproject 2.........................................................................................1 

List of abbreviations ................................................................................................................................3 

1. Introduction ......................................................................................................................................4 

2. River and water management on the Belgian side of the Border Meuse..........................................6 

2.1. Sources of literature ..................................................................................................................6 

2.2. Results of literature search ........................................................................................................6 

3. Hydrological characteristics of the Kikbeek subbasin .....................................................................7 

3.1. Introduction...............................................................................................................................7 

3.2. Description of the WetSpass model ..........................................................................................7 

3.3. The Kikbeek subbasin.............................................................................................................12 

3.4. Geographical input data ..........................................................................................................13 

3.5. WetSpass modelling of the hydrological situation in the Kikbeek subbasin ..........................22 

4. Sensitivity of the hydrological characteristics of the Kikbeek subbasin towards climate and land 

use changes.....................................................................................................................................25 

4.1. Introduction.............................................................................................................................25 

4.2. Climate scenarios ....................................................................................................................25 

4.3. Land use scenarios ..................................................................................................................37 

4.4. Combined climate and land use scenarios...............................................................................43 

5. Discussion ......................................................................................................................................47 

6. Conclusions ....................................................................................................................................48 

7. References ......................................................................................................................................50 

Appendix. Hydrological Atlas for the Kikbeek subbasin......................................................................58 

2


List of abbreviations 

AMINAL = Administratie Milieu-, Natuur-, Land- en Waterbeheer [Administration of Environmental, 

Nature, Land, and Water Management], Ministerie van de Vlaamse Gemeenschap [Ministry of 

Flanders], Departement Leefmilieu en Infrastructuur [Department of Environment and Infrastructure], 

Brussels, Belgium. 

AWZ = Administratie Waterwegen en Zeewezen [Administration of Waterways and Marine Affairs], 

Ministerie van de Vlaamse Gemeenschap [Ministry of Flanders], Departement Leefmilieu en 

Infrastructuur [Department of Environment and Infrastructure], Brussels, Belgium. 

BBLV = Bond Beter Leefmilieu Vlaanderen [Society for a Better Environment in Flanders], Brussels, 

Belgium. 

DTM = Digital topography model 

FELNET = Flanders Environmental Library Network 

IBW = Instituut voor Bosbouw en Wildbeheer [Institute of Forestry and Wildlife], Geraardsbergen, 

Belgium. 

ICIS = International Centre for Integrative Studies, Maastricht, The Netherlands. 

INB = Instituut voor Natuurbehoud [Institute of Nature Conservation], Brussels, Belgium. 

GIS = Geographical information system 

KMI = Koninklijk Meteorologisch Instituut [Royal Meteorological Institute], Brussels, Belgium. 

KNMI = Koninklijk Nederlands Meteorologisch Instituut [Royal Dutch Meteorological Institute], De 

Bilt, The Netherlands. 

KU Leuven = Katholieke Universiteit Leuven [Catholic University Leuven], Leuven, Belgium. 

PET = Potential evapotranspiration 

NGI = Nationaal Geografisch Instituut [National Geographical Institute], Brussels, Belgium. 

PIME = Provinciaal Instituut voor Milieu-Educatie [Provincial Institute of Environmental Education 

(of the Province of Antwerp)], Lier, Belgium. 

RIZA = Rijksinstituut voor Integraal Zoetwaterbeheer en Afvalwaterbehandeling [Dutch Institute for 

Inland Water Management and Waste Water Treatment], Lelystad, The Netherlands. 

TAW = Tweede Algemene Waterpassing [Belgian standard elevation datum] 

VLINA = Vlaams Impulsprogramma voor Natuurontwikkeling [Flemish Governmental Impulse Programme 

for Nature Development] 

VLM = Vlaamse Landmaatschappij [Flemish Land Society], Brussels, Belgium. 

VMM = Vlaamse Milieumaatschappij [Flemish Environmental Society], Erembodegem, Belgium. 

VUB = Vrije Universiteit Brussel [Free University Brussels], Vakgroep Hydrologie en 

Waterbouwkunde [Department of Hydrology and Hydraulic Engineering], Brussels, Belgium. 

WetSpass = Hydrological model for water and energy transfer between soil, plants, and atmosphere 

under quasi-steady state conditions 

WNF = Wereld Natuurfonds Vlaanderen [World Wildlife Fund Flanders], Brussels, Belgium. 

3


1. Introduction 

The report presented here summarizes the research that was carried out by the Department of 

Hydrology and Hydraulic Engineering of the Vrije Universiteit Brussel (VUB) within the framework 

of the IRMA-SPONGE subproject 2. Within this subproject on strategies for Rhine and Meuse basins 

in the context of integrated river management, the activities of the VUB focussed on hydrological 

aspects of the river Border Meuse (Grensmaas in Dutch) and one of its subbasins. On one side, 

general attention was paid to Belgian literature and policies concerning the river and water 

management of the Border Meuse; on the other side, a detailed study was carried out of the 

hydrological characteristics of the selected subbasin. 

The Border Meuse is that part of the river Meuse (Maas in Dutch) where it forms the international 

border (grens in Dutch) between Belgium and the Netherlands (see Figure 1). More specific, it forms 

the major part of the border between the Belgian Province of Limburg in the west and the Dutch 

Province of Limburg in the east. On the Belgian side, the Border Meuse is located between the 

villages of Kessenich in the north and Smeermaas in the south. On the Dutch side, it is located 

between Maasbracht and Borgharen. The subbasin of the Border Meuse to which particular attention 

was paid was that of the Kikbeek, which is one of the generally small brooks on the Belgian side of 

the river. Via the Kikbeek and the regional groundwater flow system in the subbasin, the Kikbeek 

subbasin discharges into the Border Meuse near the city of Maasmechelen (see Figure 1). 

Figure 1. Location of the river Border Meuse and the Kikbeek subbasin 

The objectives of the VUB contribution to the IRMA-SPONGE subproject 2 were the following: 

• Identification of scientific literature (articles, reports, etc.) and policies and proposals of Belgian 

(Flemish) governmental organisations (AMINAL, AWZ, INB, IBW, VLM, VMM, etc.) and nongovernmental 

organisations (BBLV, Greenpeace, Natuurreservaten, WNF) on the integrated river 

and water management on the Belgian side of the Border Meuse, including related nature restoration 

and development programmes. 

• Analysis of the hydrological characteristics of a Belgian subbasin of the Border Meuse (i.e. the 

Kikbeek subbasin), or more precise: determination of the yearly and seasonal discharge coefficients 

that are related to the total surface runoff and groundwater flow. 

• Analysis of the sensitivity of the hydrological characteristics towards climate and land use changes. 

The part of the research concerning the river and water management on the Belgian side of the Border 

Meuse will be discussed in Chapter 2, and the parts concerning the hydrological characteristics of the 

4


Kikbeek subbasin and their sensitivity towards climate and land use changes will be discussed in 

Chapter 3 and Chapter 4, respectively. Discussion and general conclusions will be given in 

respectively Chapter 5 and 6. 

5


2. River and water management on the Belgian side of the Border Meuse 

The first part of the research of the VUB within the framework of the IRMA-SPONGE subproject 2, 

concerned the river and water management on the Belgian side of the Border Meuse. Of particular 

interest were existing scientific literature and Belgian governmental and non-governmental proposals 

and policies. An overview of them was required by the ICIS and the Dutch partners in IRMA- 

SPONGE subproject 2 for the development of integrated water management scenarios. Besides a 

comprehensive search for documentation on the river and water management on the Belgian side of 

the Border Meuse, the VUB carried out a literature search on related nature restoration and 

development programmes and proposals of the Belgian (Flemish) authorities for the area in concern. 

The latter are predominantly developed in the framework of the bilateral Belgian-Dutch 'Living 

Border Meuse' Project [Project ‘Levende Grensmaas’ in Dutch]. The aim of this large-scale integrated 

water management project is to give the Border Meuse a more natural course in order to stimulate a 

wide variety of natural river processes, to promote nature development and nature-friendly recreation, 

and to protect urban and other areas of economical interest from flooding. To complete the literature 

search, more general, related studies concerning the hydrology, ecology, and land use of the Border 

Meuse basin were included as well. 

2.1. Sources of literature 

For the collection of data on literature that was not available at the VUB itself, visits has been brought 

to the Afdeling Maas & Albertkanaal [Meuse and Albert Canal Division] of the AWZ in Hasselt, 

Belgium, and to the INB. From the side of the Belgian (Flemish) authorities, these organisations, the 

first in particular, are primarily in concern with the 'Living Border Meuse' Project (Gielen, 2000; and 

Vanacker & Van Looy, 2000). In addition, a literature search has been carried out on the FELNET 

website (www.felnet.org). FELNET is the Flanders Environmental Library Network and appeared to 

be a powerful tool for environmental literature search. It is connected to the library catalogues of the 

major environmental information centres in the Flemish part of Belgium, such as AMINAL, BBLV, 

IBW, INB, PIME, VLM, and VMM. Another advantage of FELNET is that it easily reveals where 

particular literature is available. 

2.2. Results of literature search 

The results of the literature search are listed as references in Chapter 7, where they are indicated with 

an asterisk. Most of the literature has been written in Dutch. In that case, an English translation of the 

title is included in the reference list. Frequently, however, an abstract in English is available as well. In 

the reference list, it is indicated as well where in Belgium the literature can be obtained. 

Recommended overviews of the 'Living Border Meuse' Project from the Belgian (Flemish) side, 

including the state of affairs and the Belgian (Flemish) preferential alternative, are given by Nagels et 

al. (1999), Odou (1998), M. De Coster (1998), and Toebat et al. (2000). These overviews give a good 

impression of the policies and proposals of the Belgian authorities on the aspired integrated river and 

water management on the Belgian side of the Border Meuse. 

6


3. Hydrological characteristics of the Kikbeek subbasin 

3.1. Introduction 

The second part of the research concerned the analysis of the hydrological characteristics of a subbasin 

of the Border Meuse on the Belgian side of the river. In particular, the aim of this part of the research 

was the determination of average yearly and seasonal, fast and slow discharge coefficients of the 

subbasin in concern. These discharge coefficients were necessary for the verification of the discharge 

parameters that were used in the MEUSEFLOW model of the Dutch partners in IRMA-SPONGE 

subproject 2. As the discharge coefficients are related to the total surface runoff (fast discharge) and 

the groundwater flow system (slow discharge) of the subbasin, they could be determined by way of 

spatially distributed hydrological WetSpass modelling. This type of modelling will be delineated in 

Section 3.2. 

The subbasin of the Border Meuse that was analysed by way of WetSpass modelling was the Kikbeek 

subbasin, which will be described in Section 3.3. For the Kikbeek subbasin, average yearly and 

seasonal, fast and slow discharge coefficients were calculated from the average values of the surface 

runoff and groundwater recharge, which were determined by the WetSpass modelling. The 

geographical input data for the WetSpass modelling of the Kikbeek subbasin are presented in Section 

3.4, and results that were obtained by the WetSpass modelling and the subsequent calculations are 

presented in Section 3.5. The obtained hydrological parameters of the Kikbeek subbasin were used, 

subsequently, as a basis for the analysis of their sensitivity towards climate and land use changes, 

which will be discussed in Chapter 4. 

3.2. Description of the WetSpass model 

Introduction 

This section is based on the description of the WetSpass model in Batelaan & De Smedt (2001). The 

GIS based, spatially distributed hydrological WetSpass model calculates the spatially distributed 

yearly and seasonal evapotranspiration, surface runoff (fast discharge), and groundwater recharge 

(slow discharge). As input for this calculation, it uses spatially distributed data concerning topography, 

geomorphology, soil types, land use, hydrology (groundwater table), and meteorology (temperature, 

precipitation, potential evapotranspiration (PET), and wind speed). WetSpass is an acronym for “water 

and energy transfer between soil, plants and atmosphere under quasi-steady state conditions”. It was 

built upon the foundations of the time dependent spatially distributed water balance model WetSpa 

(Batelaan et al., 1996; Wang et al., 1997 and De Smedt et al., 2000). 

In particular, WetSpass was developed as a physically based methodology for the estimation of longterm 

spatial patterns of the groundwater recharge (and also actual evapotranspiration and surface 

runoff) that could be used as input in regional groundwater flow models, which are used for the 

analysis of regional groundwater flow systems (infiltration-discharge relations). As these groundwater 

flow models are often quasi-steady state, they usually need such long-term average groundwater 

recharge data as input. As WetSpass is a spatially distributed model, it also accounts for the spatial 

variation in the groundwater recharge, which is the result of distributed land use, soil type, slope, etc. 

As this variation can be significant, which is demonstrated clearly by Chapman (1999), this is an 

important feature of WetSpass. 

The definition of groundwater recharge, on which the WetSpass model is based, is taken from Freeze 

(1969): “recharge is the entry into the saturated zone of water made available at the water table 

surface, together with the associated flow away from the water table within the saturated zone”. 

7


Model description 

The total water balance for a grid cell in a spatially distributed grid (or raster) is split up in 

independent water balances for vegetated, bare-soil, open-water and impervious parts of the grid cell 

(see Figure 2). This allows accounting for the non-uniformity of the land use per grid cell, which 

depends on the resolution of the grid cell. The processes in each part of a grid cell are set in a 

cascading way. This means an order of occurrence of the processes, after the precipitation event, is 

assumed. Defining such an order is a prerequisite for the seasonal time scale with which the processes 

will be quantified. A number of physical and empirical relationships is used to describe the processes. 

The extent of each process is consequently limited by a number of constraints. 

Water balance components 

Transpiration Interception 

Impervious 

runoff 

Vegetated 

recharge 

Bare-soil 

Vegetated 

Open water 

runoff 

runoff 

runoff 

Groundwater recharge 

Groundwater 

Figure 2. Schematic water balance of a hypothetical grid cell 

(from: Batelaan & De Smedt, 2001) 

The water balance for vegetated surfaces is given by: 

Impervious 

evaporation 

Bare-soil 

evaporation 

Open water 

evaporation 

Impervious 

recharge 

Bare-soil 

recharge 

Root zone 

Transmission zone 

Saturated zone 

Precipitation 

Slope 

Impervious 

evaporation 

P = I + Sv + Tv + Rv (1) 

where P is the average seasonal precipitation [LT −1 ], I is the interception by vegetation [LT −1 ], Sv is 

runoff over land surface beneath vegetation [LT −1 ], Tv is the actual transpiration [LT −1 ], and Rv is 

groundwater recharge [LT −1 ]. The term actual evapotranspiration, ETv, is used here for the sum of the 

8


transpiration, Tv, and the evaporation from the bare-soil between the vegetation, Es. ETtot, the total 

actual evapotranspiration is the sum of the interception evaporation of the vegetation, I, and the actual 

evapotranspiration, ETv. 

The interception fraction has been shown to be reasonably constant at a given annual precipitation 

value and exhibits a consistent decrease with increasing annual total rainfall (Roberts, 1983). 

Therefore, the intercepted value is parameterised as a constant percentage from precipitation, 

dependent on the vegetation type (Calder, 1979; and Nonhebel, 1987). 

The surface runoff, Sv, is calculated in two stages. In the first stage, the potential surface runoff, Sv,pot, 

is calculated as a coefficient times the precipitation minus the interception: 

Sv,pot = Cs,v (P – I) (2) 

where Cs,v is a surface runoff coefficient for vegetated infiltration areas, based on the rational formula 

(Smedema & Rycroft, 1988; Pilgrim & Cordery, 1992; and Chow et al., 1988). Cs,v is a function of 

vegetation type, soil type, and slope. In groundwater discharge areas saturated surface runoff is 

occurring. Here, the surface runoff coefficient is very high and assumed to be constant, due to its 

reduced dependency on soil and vegetation type and the generally near to river position of the runoff 

producing areas. In the second stage, the potential surface runoff is actualised by taking into account 

differences in precipitation intensities in relation to soil infiltration capacities. Rubin (1966) showed 

that in this case Hortonian overland flow is rare. 

Sv = CHor Sv,pot (3) 

CHor is a coefficient, which parameterises the part of the seasonal precipitation that is actually 

contributing to the (Hortonian) surface runoff. In groundwater discharge areas, all intensities of 

precipitation contribute to surface runoff (i.e. CHor equals 1). In infiltration areas, only high intensity 

storms will generate surface runoff. For the precipitation station at Uccle (Brussels, Belgium), an 

analysis has been made of 10-minute precipitation data for the period 1948-1998. For each season, the 

cumulative precipitation falling with an intensity bigger than 1, 2, 3 mm/hr, etc. is determined. 

Clearly, it appeared that the cumulative high intensity precipitation amount is much bigger in summer 

than in winter. For each soil class, the infiltration rate (Rawls et al., 1992; and Saxton et al., 1986) has 

been related to the precipitation intensity. The Hortonian surface runoff could now be determined as 

the fraction of the seasonal precipitation with a higher intensity than the infiltration capacity. 

In order to come to a seasonal distributed evapotranspiration value, WetSpass proposes to convert the 

open-water evaporation value, as commonly available from the Penman equation, to a reference 

transpiration value (Federer, 1979) based on a vegetation coefficient. 

Tr,v = c Eo (4) 

where Tr,v is the reference transpiration of a vegetated surface [LT −1 ], Eo is the potential evaporation of 

open water [LT −1 ], and c is the vegetation coefficient [-]. The vegetation coefficient can be determined 

as the quotient of the reference vegetation transpiration, as given by the Penman-Monteith equation, 

and the potential open-water evaporation, as given by the Penman equation, resulting in: 

c = (1 + γ/∆) / (1 + (1 + rc/ra) γ/∆) (5) 

where the proportionality constant Δ is the slope of the first derivative of the saturated vapour pressure 

curve [ML −1 T −2 C −1 ], γ is the psychrometric constant [ML −1 T −2 C −1 ], rc is the canopy resistance [TL −1 ], 

and ra is the aerodynamic resistance [TL −1 ]. Dingman (1994) derived a similar equation as equation 

(5), but included the soil moisture dependent canopy resistance function from Stewart (1988). 

Obviously, in vegetated groundwater discharge areas, the actual transpiration, Tv, is equal to the 

reference transpiration, Tr,v, since soil water availability is not limiting. 

Tv = Tr,v if Gd – ht ≤ Rd (6) 

9


where, Gd, is the groundwater depth [L], ht is the tension saturated height [L], and Rd is the rooting 

depth [L]. The actual transpiration for vegetated areas, where the groundwater level is below the root 

zone, is calculated as: 

Tv = f(θ) Tr,v if Gd – ht > Rd (7) 

where ƒ(θ) is a function of the water content. In WetSpass, for a time invariant situation, the 

methodology developed by Vandewiele et al. (1991) is used for defining ƒ(θ): 

f(θ) = 1 – a1 w/Tr,v with w = P + (θfc – θpwp) Rd (8) 

where a1 is a calibrated parameter related to the sand content of a soil type (Van der Beken & 

Huybrechts, 1990), w is the available water content for transpiration [LT −1 ], and (θfc – θpwp) is the plant 

available water content. 

The groundwater recharge can be calculated, as a rest term, from the water balance: 

Rv = P – Sv – ETv – Es – I (9) 

The methodology here described will result in the estimation of the spatially distributed groundwater 

recharge in function of vegetation, soil type, slope, groundwater depth, precipitation regime, and other 

climatic variables. Even in groundwater discharge areas, some recharge will be calculated. This is in 

agreement with the conceptual picture that also in discharge areas a thin unsaturated zone is present, 

allowing some recharge. However, in summer the calculated recharge in discharge areas will be often 

negative as a result of the potential transpiration of the vegetation. High winter recharge will in some 

cases compensate the negative recharge. 

Change in storage is brought into the model, on a seasonal basis, in two ways. First of all, it is possible 

to have different groundwater levels for the winter and summer situation. Secondly, it is assumed that 

during winter the plant available soil moisture is filled up and that during summer this reservoir can be 

depleted. 

A similar procedure as for the vegetated surfaces is applied to the bare-soil, open-water and 

impervious surfaces. 

Water balance per grid cell 

The total water balance, per grid cell and season, can now be calculated with the previously described 

water balance components for the vegetated, bare-soil, open-water and impervious parts of the grid 

cell: 

ETcell = av ETv + as ETs + ao ETo + ai ETi (10) 

Scell = av Sv + as Ss + ao So + ai Si (11) 

Rcell = av Rv + as Rs + ai Ri (12) 

where ETcell, Scell, and Rcell are, respectively, the total evapotranspiration, surface runoff and recharge in 

a grid cell. av, as, ao, and ai are, respectively, the vegetated, bare-soil, open-water and impervious area 

fractions of a grid cell (Batelaan & De Smedt, 2001). 

GIS implementation 

As the spatially distributed WetSpass model was coded in Avenue, it could be integrated completely in 

ArcView GIS, a geographical information system (GIS). In ArcView GIS, WetSpass had to be loaded 

as a text script and to be compiled, before it could be used. In addition, an input file (see Figure 3) had 

to be created that directs WetSpass, while running, to the ArcView GIS grids that should be used as 

input in the modelling. These grids comprise spatially distributed and seasonal data of the slope, soil 

type, land use, groundwater table, and meteorology (temperature, precipitation, PET, and wind speed). 

Moreover, the input file directs WetSpass to a number of parameter files (i.e. the data base files with 

extension “.dbf”, see Figure 3) that contain comprehensive sets of parameters, which are related to the 

10


soil and land use types and that are required for the model simulations in WetSpass. For instance, the 

parameter file “runoffcoef10.dbf” relates a runoff coefficient to the soil and land use type and to the 

slope. The use of separate parameter files allows for easy definition of new land use or soil types, as 

well as changes in the parameter values. Moreover, the use of a separate input file allows for easy 

replacement of input grids by adjusted land use or meteorological grids to model, for instance, land 

use or climate scenarios. 

WetSpass input file 

path\ 

path\landuse_param_w4.dbf 

path\ 

path\landuse_param_z4.dbf 

path\ 

path\soil_param12.dbf 

path\ 

path\ 

path\ 

path\ 

path\ 

path\ 

path\ 

path\ 

path\ 

path\ 

path\ 

path\runoffcoef10.dbf 

Figure 3. General structure of a WetSpass input file 

The soil type and land use grids used in the WetSpass modelling are classified grids. The soil 

classification system used by WetSpass is based on that of the U.S. Department of Agriculture 

(USDA, 1951). The land use classification system follows that of the land use map of Flanders, which 

is based on classified Landsat 5 satellite images, which were resampled to a 50 by 50 m resolution 

(VLM, 1998a). 

Using the distributed groundwater recharge from the WetSpass model in a steady-state groundwater 

flow model will improve the prediction of the modelled groundwater table, and of the groundwater 

discharge and recharge areas. However, the groundwater table is used as input to the WetSpass 

modelling. Therefore, the groundwater flow model and the WetSpass model have to be performed one 

after the other, for a number of times, while exchanging the groundwater recharge and groundwater 

table data. This iteration process will usually lead to a stable solution for the spatially distributed 

groundwater recharge and groundwater table after a few iterations. 

In this report, it will be shown that the WetSpass model could also be applied to obtain fast and slow 

discharge coefficients for a particular (sub)basin. How these coefficients can be calculated from the 

WetSpass output will be outlined in Section 3.5. 

Application in previous projects 

WetSpass was used for a land planning project in the Grote Nete basin, Belgium (Batelaan et al., 

2000a). In this study the effects of land use changes on the groundwater discharge areas were 

analysed. An estimation of the spatially distributed groundwater recharge was therefore essential. 

11


Therefore, a WetSpass model was derived from a larger WetSpass model that was developed for the 

basins of the rivers Dijle, Demer, and Nete (Batelaan, in prep.). The larger WetSpass model was 

calibrated for these basins on basis of discharge measurements of 17 gauging stations, of which two 

are located in the Grote Nete basin. Total discharge, as well as surface runoff and base flow, 

determined by two different discharge separation techniques, were used for the calibration of the water 

balance components in the WetSpass model. A groundwater flow model, with groundwater recharge 

from the WetSpass model of the Grote Nete basin as input, was calibrated in conjunction with the 

calibration of the WetSpass model. Groundwater discharge areas calculated by the groundwater model 

were for the Grote Nete area also verified by comparison with field maps of the phreatophytic 

vegetation (Batelaan et al., 2000a). 

In a similar way, WetSpass was applied in the groundwater flow modelling of three Belgian wetland 

ecosystems (Doode Bemde, Vorsdonkbos, and the Valley of the Zwarte Beek) that are situated in the 

valleys along the rivers Dijle, Demer, and Zwarte Beek. Here, modelled groundwater tables were 

compared to the observed ones in the wetland ecosystems, and modelled groundwater discharge areas 

were compared to those that were indicated by the phreatophytic vegetation (Batelaan et al., 2000b, 

Huybrechts et al., 2000). 

3.3. The Kikbeek subbasin 

Selection of the subbasin 

This section gives a description of the Kikbeek subbasin, which was selected for the hydrological 

modelling in this project. The major criterions for this selection were the following: The subbasin had 

to be situated entirely at the Belgian side of the Border Meuse and it had to discharge into the Border 

Meuse. The major stream had to follow a more or less natural course and it had to be large enough to 

be of interest. Furthermore, it would be useful if the hydrological modelling of the subbasin in concern 

would be of interest for other projects in the region, in particular those related to nature restoration and 

development. Finally, the availability of hydrological data could be decisive. Such data are of great 

importance for the reliability of the hydrological model. 

Three possible candidates for the selection of the subbasin were the subbasins of the brooks Bosbeek, 

Kikbeek, and Ziepbeek. These brooks were all large enough to be of interest and discharge all from 

the Belgian side into the Border Meuse. The more northerly situated brooks Abeek and Itterbeek were 

no candidates, as they discharge in the Dutch part of the river Meuse and have their downstream parts 

in the Netherlands as well. The more southerly situated brook Asbeek was no candidate neither, as it 

does not discharge into the Border Meuse, but artificially into the canal that connects the canals 

Albertkanaal and Zuid-Willemsvaart. 

A serious disadvantage of the Bosbeek, Kikbeek, and Ziepbeek subbasins was, however, the lack of 

hydrological data. No time series were available of frequent (hourly or daily) water level and 

discharge measurements at the discharge points of these brooks. Such data were only available for the 

upstream part of the Bosbeek from the measurement station at Opoeteren (AWZ, 2000b). Still, the 

Bosbeek was not a good candidate for the hydrological modelling. The problem was that the course of 

its downstream part had been manipulated too much. Near Aldeneik, the Bosbeek discharges into the 

Border Meuse indirectly via a network of reservoirs (Gielen, 2000). 

The choice had, therefore, to be made between the Kikbeek and the Ziepbeek subbasins. They both 

discharge near Maasmechelen into the Border Meuse, the Kikbeek just a km downstream from the 

Ziepbeek. The courses of both brooks are fairly natural, although they are disturbed at some sites. 

Both brooks cross the canal Zuid-Willemsvaart. As they cross it via a siphon (or culvert), there is, 

however, no hydrological contact between the brooks and the canal and the disturbance is only 

limited. The course of the Kikbeek is probably more disturbed by a diversion along the junction of 

high road A2 and main road N78, south of Maasmechelen. The source area of the Kikbeek, in the 

12


forest Peensbos (see Map 1 in the Appendix), is affected too. Due to sand winnings, dewatering 

problems are observed here (Nagels, 2001). 

For two reasons, a hydrological model of the Kikbeek was, nevertheless, thought to be more 

interesting for the stakeholders than one of the Ziepbeek. The first concerned an expected 

improvement of the water quality. At present, the water in the Kikbeek is highly polluted with 

untreated waste water that is still discharged by the waste water treatment plant of Maasmechelen. In 

the coming years, this pollution will diminish because of an increase of the capacity of the waste water 

treatment plant (Gielen, 2000). Understanding of the hydrological characteristics of the Kikbeek 

subbasin can help to estimate the effects on the water quality of the Kikbeek and the Border Meuse, 

into which the presently polluted water discharges. 

The second reason concerned the plans for the Kikbeek in the framework of the 'Living Border Meuse' 

Project. In these plans, the bed of the Kikbeek will be lowered between the siphon and the Border 

Meuse. As a result, the Kikbeek can discharge in a more natural way into the Border Meuse. Now, the 

water just drops into the Border Meuse from a concrete bar. There are also plans to divert the lowered 

Kikbeek towards a side channel of the Border Meuse that might be formed in the river bed in the 

framework of the 'Living Border Meuse' Project. In that case, this side channel will be fed with water 

from the Kikbeek at normal water levels in the Border Meuse, and also by water from the Border 

Meuse at high water levels (Nagels et al., 1999; Toebat et al., 2000; and Gielen, 2000). Knowledge of 

the total discharge of the Kikbeek subbasin, and its variations in time, can help to determine the 

feasibility of these plans. 

Description of the Kikbeek subbasin 

The Kikbeek is about 10 km long. Its major source area is situated in the forest Peensbos, at the east 

side of the Kempish Plateau, at a level of about 90 m above TAW (see Map 1 and Map 2 in the 

Appendix). From there, the Kikbeek declines first rapidly to a level of about 50 m above TAW, and 

then slowly to a level of about 40 m above TAW at its discharge point (see Map 2). At about 6 km 

downstream, at Opgrimbie, the brook Groenstraatbeek joins the Kikbeek. Another km further 

downstream starts the 1½ km long diversion along the road junction. At a km from its discharge point, 

the Kikbeek crosses the Zuid-Willemsvaart via the siphon, and just before the diverted Lograafbeek 

joins the Kikbeek as well (see Map 1). The upstream part of the Kikbeek (the part until Opgrimbie) 

and that of the Groenstraatbeek (the part until Bovenwezet) are bedded predominantly in sand soils. 

The downstream parts of the Kikbeek and the Groenstraatbeek are bedded predominantly in sandy 

loam and silt soils (see Map 6 in the Appendix). The upstream part of the Kikbeek subbasin is 

dominated by mixed forests, the downstream part by agricultural land and urban areas (see Map 7 in 

the Appendix). 

3.4. Geographical input data 

Topography 

The basis of a spatially distributed, hydrological WetSpass model is the topography of the area in 

concern, as the topography determines most of the hydrological processes. To enter the spatially 

distributed, topographical data into WetSpass, a digital topography model (DTM) has to be developed. 

This DTM is a digital, spatially distributed grid, in which an average surface elevation has been 

allocated to each rectangular grid cell. 

In this project, a DTM has been developed for the wider region of the Kikbeek subbasin. This DTM is 

based on digital topographic maps, scale 1:10,000, from the Belgian National Geographic Institute 

(NGI, 1995). Besides common information, such as cities, villages, roads, and rivers, these 

topographic maps also include elevation contour lines. The elevation interval of the contour lines is 

2.5 m. 

13


The topographic maps used were imported as a georeferenced picture (tiff format) into ArcView GIS. 

Although the maps were georeferenced, no further digital geographical information was supplied with 

them. Therefore, the elevation contour lines had to be digitised. This was done within ArcView GIS 

using the line drawing mode and the general snapping tool (using a view scale of 1:5,000). The 

digitised contour lines were saved together in an ArcView GIS shapefile. The elevations in concern 

were allocated to them in the attribute table of the shapefile. The resulting map with the digitised 

contour lines is shown in Map 2 in the Appendix. 

To obtain a DTM, the ArcView GIS shapefile of the elevation contour lines had to be converted to a 

spatially distributed grid with an interpolated average surface elevation for each grid cell. For this 

purpose, the topogrid function in ArcInfo GIS could be used. First, however, the ArcView GIS 

shapefile of the elevation contour lines had to be imported into ArcInfo GIS, and to be converted to an 

arc file. The same had to be done with ArcView GIS shapefiles of the natural streams in the area, i.e. 

the Border Meuse and its tributaries on the Belgian side of the river, such as the Kikbeek. The 

ArcView GIS shapefiles of the (unnatural) canals in the area, such as the Zuid-Willemsvaart, were not 

used, as these canals have not influenced the surrounding topography. 

With the topogrid function in ArcInfo GIS, the arc file of the elevation contour lines was interpolated, 

taking the natural streams into account. For this interpolation, a 16 by 16 km grid was defined with 

grid cells of 20 by 20 m, starting at X = 233 km and Y = 173 km (Belgian Lambert co-ordinates). The 

Dutch part of the grid, generally the part east of the Border Meuse, was excluded from the DTM by 

the use of a mask in the interpolation process. The resulting interpolated grid was the DTM for the 

wider region of the Kikbeek subbasin (see Map 3 in the Appendix). In this grid, the value of each grid 

cell on the Belgian side of the 16 by 16 km grid was the average elevation for the 20 by 20 m grid cell. 

The value of each grid cell on the Dutch side was set to no data. 

The used mask was an ArcInfo GIS polygon file that comprised the Belgian part of the 16 by 16 km 

area (i.e. the wider region of the Kikbeek subbasin). To obtain it, a shapefile comprising a polygon of 

the region outline was constructed in ArcView GIS first. The east side of the polygon was formed by 

the shapefile of the Border Meuse (between Y = 173 and 189 km), the north side by the Y = 189 km 

line, the west side by the X = 233 km line, and the south side by the Y = 173 km line. Subsequently, 

this ArcView GIS shapefile was imported in ArcInfo GIS, where it was converted to the ArcInfo GIS 

polygon file that was used as a mask in the interpolation process. With the obtained DTM, a mask grid 

was produced for the wider region of the Kikbeek subbasin by dividing the DTM in ArcInfo GIS by 

itself. This resulted in a mask grid in which the value of each grid cell was 1 for the Belgian part of the 

grid, and set to no data for the Dutch part. 

For hydrological modelling, the DTM should not contain any sinks. A sink is a grid cell (or group of 

grid cells) in which the value in concern (here the elevation) is lower than the corresponding values in 

all its (or their) neighbour grid cells. As sinks in the DTM cause difficulties in modelling surface 

runoff, they need to be located and filled up to the level of the lowest neighbour grid cell. In ArcInfo, 

this is possible with the fill function (option sinks). The DTM appeared not to have any sinks. 

Outline of the Kikbeek subbasin 

Based on the DTM, the outline of the Kikbeek subbasin was determined in ArcInfo GIS. First, a flow 

direction grid was constructed from the DTM, using the flowdirection function. For each grid cell in 

this grid, an integer value indicates which of the eight neighbour grid cells has the lowest elevation 

and is the steepest downhill. The predominant flow direction of any surface runoff will be towards this 

grid cell. Subsequently, a flow accumulation grid was constructed from the flow direction grid, using 

the flowaccumulation function. For each grid cell in this grid, an integer value indicates the total 

number of connected grid cells from which surface runoff would flow towards the grid cell. The next 

step was the construction of a stream network grid. For this purpose the con (conditional) function and 

the flow accumulation grid were used. In fact, if the integer value in a grid cell of the flow 

accumulation grid exceeded 100, the grid cell was considered to be part of a stream. 

14


The stream network grid was used to allocate the most downstream grid cell of the stream that could 

be regarded as the Kikbeek. This grid cell (situated at X = 244.96 km and Y = 184.46 km) was the 

outlet grid cell where the Kikbeek subbasin would discharge into the Border Meuse. Using the 

selectpoint function in ArcInfo GIS, a point grid was made, in which the outlet grid cell only has a 

value of 1. Using this point grid, the flow direction grid, and the watershed function, a subbasin grid 

was formed, in which each grid cell that discharges through the outlet grid cell in concern has a value 

of 1, and all other grid cells have a value of 0. 

As a control, the obtained stream network and subbasin grids were compared in ArcView GIS with the 

shapefiles of the Border Meuse and its tributaries on the Belgian side of the river. No major 

discrepancies were observed. The only exception was the stream network of the Heiwickbeek, a brook 

that is a tributary of the Ziepbeek (see Map 1 in the Appendix). In the stream network grid, the 

Heiwickbeek stream network turned out to discharge into the Kikbeek, and not into the nearby 

Ziepbeek. This discrepancy, however, could be the result of local uncertainties in the DTM at the site 

of the outlet grid cell of the Heiwickbeek stream network. 

As a consequence, the Kikbeek subbasin grid had to be corrected for the incorrect inclusion of the 

Heiwickbeek subbasin. Additional topographic data to improve the DTM at the site of the 

Heiwickbeek outlet grid cell, which could have altered the local flow directions, were, however, not 

available. Therefore, the decision was made to exclude simply the grid cells that could be ascribed to 

the Heiwickbeek subbasin from the Kikbeek subbasin grid. For this purpose, a second subbasin grid 

was constructed for the Heiwickbeek subbasin, in a similar way as described above for the Kikbeek 

subbasin (the outlet grid cell was situated around X = 239.36 km and Y = 181.74 km). Subsequently, 

the obtained Heiwickbeek subbasin grid was subtracted in ArcInfo GIS from the Kikbeek subbasin 

grid. The resulting grid was supposed to comprise the area of the entire Kikbeek subbasin (see Map 4 

in the Appendix, which also shows the stream net of the Kikbeek subbasin). 

From the obtained Kikbeek subbasin grid, in ArcInfo GIS, a grid was constructed that could be used as 

a mask grid for any other grid of the Kikbeek subbasin used in the project. In this mask grid, the grid 

cells within the Kikbeek subbasin had a value of 1, and those outside the subbasin a value that was set 

to no data. The mask grid for the Kikbeek subbasin was made for the area between X = 236 and 246 

km, and between Y = 179 and 187 km. Finally, it was converted in ArcInfo GIS to a polygon file and 

subsequently to an ArcView GIS shapefile. This shapefile comprised a polygon of the outline of the 

Kikbeek subbasin. 

A last point to be discussed here is the subbasin of the Lograafbeek, a brook north of the Kikbeek (see 

Map 1 in the Appendix). This brook is blocked by the Zuid-Willemsvaart and diverted artificially to 

the Kikbeek. Nevertheless, this subbasin has not been included in the Kikbeek subbasin, as it does not 

belong to it naturally. The purpose of the study was to quantify the runoff of a natural subbasin on the 

Belgian side of the Border Meuse. Therefore, the Lograafbeek subbasin was left out of consideration. 

Geomorphology 

Besides the topography, the geomorphology of the Kikbeek subbasin had to be taken into 

consideration as well. Therefore, on basis of the DTM, a digital slope map was constructed showing 

the average steepness in each grid cell. For this, the slope function in ArcInfo GIS was used. The 

slopes were expressed as percentages (this unit is required for the WetSpass model). Slope maps were 

made for both the Kikbeek subbasin (see Map 5 in the Appendix) and the wider region, using the 

DTM and the two mask grids. 

Soil types 

The next step in the collection of input data for the WetSpass model was the construction of a digital 

soil map of the area of the Kikbeek subbasin. For this purpose, the digital soil map of Flanders was 

used (VLM, 1998b). This, however, is a spatially distributed ArcInfo GIS polygon file, and not a grid 

that can be read in WetSpass. In addition, the soil types are classified according to the Belgian soil 

15


code system, which is not compatible with the code system used by WetSpass. In the Belgian soil code 

system, the main soil type is indicated by a single capital (e.g. Z stands for sand and L for loam). 

Additional characters indicate further characteristics of the soil in concern (e.g. Zbf stands for dry sand 

with hardly any humus and/or iron B horizon). In WetSpass, on the other hand, the soil type is 

characterised by an integer (1 to 12). Therefore, the polygon file had to be converted to a grid, and the 

Belgian soil codes to WetSpass soil codes. 

For the conversion of the polygon file of the soil map of Flanders to a grid for the area of interest, the 

polygrid function in ArcInfo GIS was used. The grid was adjusted to the same grid structure as the 

DTM and limited to the same area by using the mask grid of the wider region of the Kikbeek subbasin 

as a mask in the conversion process. Before the conversion of the polygon file to a grid, the soil types 

in the polygon file were reclassified according to the WetSpa code system. For this reclassification, an 

ArcInfo GIS aml file was used, which made use of a reclassification table that was prepared in ArcInfo 

GIS as well. After the cancelling of the additional characters in the Belgian soil codes, the leading 

characters were changed into integers (from 1 to 13) according to the reclassification table (see Table 

1). Hence, the grid cell values in the produced soil grid for the wider region of the Kikbeek subbasin 

were integers from 1 to 13. 

In contrast with the WetSpa model, the WetSpass model does not deal with an impervious soil type 

(soil code 13) but with an impervious landuse class, such as roads, squares, parking sites, roofs, etc. 

Therefore, the soil grid had to undergo a final conversion step in which every soil code 13 was 

replaced by the (pervious) soil code in adjacent grid cells (see Table 1). Therefore, grid cells with a 

value of 13 were changed into no data grid cells first. Subsequently, the no data grid cells were filled 

with the value of the nearest neighbour grid cell by using the eucallocation function in ArcInfo GIS. 

Now a soil grid was obtained for the wider region of the Kikbeek subbasin with soil types according to 

the WetSpass soil code system (integers from 1 to 12). To obtain a WetSpass soil grid for the Kikbeek 

subbasin only (see Map 6 in the Appendix), the soil grid for the wider region of the Kikbeek subbasin 

was multiplied with the mask grid for the Kikbeek subbasin, within the frame of this mask grid (the 

area between X = 236 and 246 km, and between Y = 179 and 187 km). 

Table 1. Conversion table for Belgian, WetSpa, and WetSpass soil codes 

Type of soil 1 st character 

in Belgian 

soil code 

WetSpa 

soil code 

WetSpass 

soil code 

Sand Z 1 1 

Loamy sand S 2 2 

Sandy loam P 3 3 

Loam L 4 4 

Silty loam - 5 5 

Silt A 6 6 

Sandy clay-loam - 7 7 

Silty clay-loam - 8 8 

Clay-loam E 9 9 

Sandy clay - 10 10 

Silty clay - 11 11 

Clay U 12 12 

Impervious O 13 1 to 12 * 

* Soil code of nearest neighbour 

16


A statistical overview of the distribution of soil types in the Kikbeek subbasin is presented in Table 2. 

From this table, it is clear that the soil is dominated by sand (57.5 % of the surface) and secondly by 

loamy sand to sandy loam (37.9 %). In the conversion from WetSpa codes to WetSpass codes, the soil 

code 13 for impervious soils was changed to other soil codes for 5836 grid cells, which is 10.0 % of 

the grid cells in the Kikbeek subbasin. 

Type of soil 

Table 2. Overview of spatially distributed soil data used 

for the WetSpass modelling of the Kikbeek subbasin 

WetSpass soil 

code 

Number of 

grid cells 

Total area 

(ha)* 

Percentage of 

subbasin area 

Sand 1 33565 1343 57.5 

Loamy sand 2 5131 205 8.8 

Sandy loam 3 16981 679 29.1 

Loam 4 1334 53 2.3 

Silt 6 1350 54 2.3 

* 1 ha = 1 hm 2 = 0.01 km 2 

Land use 

Total 58361 2334 100.0 

Another important input in the WetSpass model is the type of land use. For this, the land use map of 

Flanders was used, which is available as an ArcInfo GIS land use grid (VLM, 1998a). Using the mask 

grid for the wider region of the Kikbeek subbasin, a land use grid was extracted from it for the same 

area as the DTM, and with the same grid cell size (20 by 20 m). With the mask grid for the Kikbeek 

subbasin, a land use grid for the Kikbeek subbasin only was produced subsequently from the produced 

land use grid for the wider region of the subbasin. The land use codes in the two produced land use 

grids were the same as the land use codes (integers from 1 to 202) that were used in the land use map 

of Flanders (see Table 3). 

The land use grid for the Kikbeek subbasin was corrected for the presence of the Zuid-Willemsvaart. 

As this canal was left out of consideration to model as much as possible a natural situation (and 

therefore not included in the interpolation process for the construction of the DTM), it needed to be 

left out of the land use grid as well. It was, however, clearly present. Therefore, the grid cells with a 

value of 51 (the land use code for a navigable river or canal) and a situation that corresponds with that 

of the Zuid-Willemsvaart had to be given another value. These 163 grid cells (0.28 % of the subbasin 

area) were selected and set to no data grid cells first, and filled with the values of the nearest 

neighbour grid cells subsequently. For the latter, the eucallocation function in ArcInfo GIS was used. 

The resulting land use grid for the Kikbeek subbasin is shown in Map 7 in the Appendix. 

A statistical overview of the types of land use in the Kikbeek subbasin is presented in Table 3. 

Agricultural land and forests dominate equally the land use in the subbasin (both 31 % of the area). 

Urban areas and infrastructure fill up about 23 % of the subbasin area. The rest is filled up with (wet) 

meadows (9 %), heathers (3 %), mud flats (2 %), and surface water (1 %, predominantly lakes). 

Groundwater table 

A shallow groundwater table or a groundwater table at the surface influences the evapotranspiration, 

surface runoff and recharge. However, this situation occurs in most catchments only in a small part of 

the valley, here the recharge is low. Deeper groundwater tables do not influence evapotranspiration, 

surface runoff or recharge. As no data were available concerning the groundwater table, it was set, 

throughout the Kikbeek subbasin, at an arbitrary level of 1 m below the land surface. It was assumed 

17


that this level would result in realistic a recharge condition for most of the catchment, while still some 

evapotranspiration from the groundwater might occur in the valley. Hence, in ArcInfo GIS, a grid with 

arbitrary groundwater tables for the Kikbeek subbasin was produced, by subtracting 1 m from the 

elevations in the DTM. 

Meteorological data 

Until this point, the discussion focussed on the conditional, more or less invariable input data. The 

leading, variable input data for the WetSpass model are, however, the meteorological data. Based on 

all the conditions set above (topography, soil types, land use, etc.), they determine the final outcome of 

the WetSpass model. The amount of precipitation determines how much water can be divided between 

evapotranspiration, groundwater recharge, and surface runoff. The potential evapotranspiration (PET), 

temperature, and wind speed are important factors in the calculation of the part of precipitation that 

will evapotranspirate. 

18


Table 3. Overview of spatially distributed land use data used 

for the WetSpass modelling of the Kikbeek subbasin 

Type of land use Land use 

code 

Urban areas: 

Number of 

grid cells 

Total area 

(ha)* 

Percentage of 

subbasin area 

- city center 1 3454 138 5.9 

- built up 2 4450 178 7.6 

- industry 3 260 10 0.4 

- open built up 10 3222 129 5.5 

Infrastructure: 

subtotal 11386 455 19.5 

- highway 201 297 12 0.5 

- district road 202 425 17 0.7 

- other infrastructure 4 1256 50 2.2 

Agriculture: 

subtotal 1978 79 3.4 

- agricultural land 21 12584 503 21.6 

- field with maize or other tuberous plants 27 5450 218 9.3 

Meadows: 

subtotal 18034 721 30.9 

- meadow 23 5288 212 9.1 

- wet meadow 28 217 9 0.4 

Forests: 

subtotal 5505 220 9.4 

- deciduous forest 31 3011 120 5.2 

- coniferous forest 32 7426 297 12.7 

- mixed forest 33 7397 296 12.7 

Heathers and shrubs: 

subtotal 17834 713 30.6 

- heather 35 1962 78 3.4 

- shrub 36 18 1 0.0 

subtotal 1980 79 3.4 

Mud flats 44 1179 47 2.0 

Surface waters: 

- navigable river or canal 51 113 5 0.2 

- lake 52 352 14 0.6 

* 1 ha = 1 hm 2 = 0.01 km 2 

subtotal 465 19 0.8 

Total 58361 2334 100.0 

19


Precipitation 

A digitised map with contour lines of the time-averaged yearly precipitation in Belgium (Dupriez & 

Sneyers, 1979; and Huybrechts et al., 2000), saved as an ArcInfo arc file, was used to construct 

precipitation grids for the WetSpass model. In the digitised map, the interval between the contour lines 

is 50 mm/year (= 50 dm 3 /m 2 year). With the topogrid function in ArcInfo GIS, the contour lines were 

interpolated to obtain a time-averaged yearly precipitation grid. For this interpolation, a 60 by 82 km 

grid was defined with grid cells of 20 by 20 m, starting at X = 200 km and Y = 150 km, respectively. 

The interpolation area has been chosen much wider than the area of the DTM in order to include a 

representative number of contour lines and to avoid errors caused by irregularities at the borders of the 

interpolation area. In addition, the interpolation area was extended to the 60 by 82 km area in order to 

produce a time-averaged yearly precipitation grid for the entire Belgian part of the Border Meuse 

region. This would make it applicable in other projects as well. 

With the use of the mask grid for the Kikbeek subbasin, a precipitation grid for the subbasin was 

extracted from the produced precipitation grid. In the resulting time-averaged yearly precipitation grid 

for the Kikbeek subbasin (see Map 17 in the Appendix), the spatial average of the time-averaged 

yearly precipitation was 770 mm/year, the minimum 756 mm/year, and the maximum 780 mm/year 

(see Table 4). 

The WetSpass model does, however, not take the yearly precipitation into account, but the total 

summer and winter precipitation. Hence, the time-averaged yearly precipitation grid for the Kikbeek 

subbasin had to be split up in a summer and winter precipitation grid. For the area in concern, there 

were, however, no data available to distinguish between summer and winter precipitation. Therefore, it 

was assumed that the ratio between summer and winter precipitation would correspond to the timeaveraged 

ratio observed at the meteorological station at Uccle (Brussels). At this station, the total 

precipitation in a year was 780 mm on a long-term time average. In summer (April to September) it 

was 400 mm, and in winter (October to March) 380 mm (Sneyers & Vandiepenbeeck, 1995). Hence, 

51.3 % on average of the yearly precipitation fell in summer, and 48.7 % in winter. In ArcInfo GIS, 

time-averaged total summer and winter precipitation grids for the Kikbeek subbasin were produced, by 

multiplying the time-averaged yearly precipitation grid with these percentages. See Table 4 for the 

ranges in the summer and winter precipitation. 

Potential evapotranspiration (PET) 

For the potential evapotranspiration (PET), a digitised map (saved as an ArcInfo GIS arc file) with 

contour lines of the time-averaged yearly PET in Belgium was available as well (Gellens- 

Meulenberghs & Gellens, 1992; and Huybrechts et al., 2000). But, although the interval between the 

contour lines in this map was 12.5 mm/year, the number of contour lines in the Belgian part of the 

Border Meuse region was too small to be used in a proper interpolation. Therefore, it was assumed 

that the time-averaged yearly PET was constant throughout the region. As a single contour line passed 

the entire region from north to south, the time-averaged yearly PET has been given the value of this 

contour line: 675 mm/year. Hence, a time-averaged yearly PET grid was produced for the same 60 by 

82 km area as the time-averaged yearly precipitation grid for the Belgian part of the Border Meuse 

region, with the same grid cell size of 20 by 20 m, in which each grid cell had a value of 675 mm/year. 

From this grid, a time-averaged yearly PET grid for the Kikbeek subbasin was extracted, using the 

mask grid for the subbasin. In this grid too, each grid cell had a value of 675 mm/year (see Table 4). 

As was the case for the precipitation (see above), the WetSpass model takes into account only the total 

summer and winter PET. Hence, the time-averaged yearly PET grid for the Kikbeek subbasin had to 

be split up. Again, for the area in concern, no data were available to distinguish between summer and 

winter PET. Hence, it was assumed here too that the ratio between summer and winter PET was equal 

to the time-averaged ratio observed at Uccle. At this meteorological station, the time-averaged total 

PET in a year was 657 mm. In summer it was 543 mm, and in winter 114 mm (Sneyers & Vandiepenbeeck, 

1995). Hence, it was assumed that in the Belgian part of the Border Meuse region too, the total 

20


PET was in summer 82.6 % of the yearly PET (675 mm), and in winter 17.4 %. In ArcInfo GIS, timeaveraged 

total summer and winter PET grids for the Kikbeek subbasin were produced, by multiplying 

the time-averaged yearly PET grid with these percentages. In the resulting summer and winter PET 

grids, all grid cells had values of (82.6 % × 675 mm/year =) 557.8 mm/year and (17.4 % × 675 

mm/year =) 117.2 mm/year, respectively (see Table 4). 

Temperature 

Regarding the temperature of the Belgian part of the Border Meuse region, no applicable data were 

available. Therefore, the long-term, time-averaged summer and winter temperatures for the 

meteorological station at Uccle were used in the WetSpass model. These were 14.1 and 5.0 °C, 

respectively (Sneyers & Vandiepenbeeck, 1995). As described above for the PET, summer and winter 

temperature grids were produced for the Belgian part of the Border Meuse region, and for the Kikbeek 

subbasin in particular. In these grids, all grid cells had values of 14.1 °C and 5.0 °C, respectively (see 

Table 4). 

The applicability of the seasonal temperature data from Uccle for the Kikbeek subbasin was verified 

by an examination of seasonal temperature maps in the Atlas of Belgium (Poncelet, 1973). According 

to these maps, the average temperature in July was at most 0.25 °C higher in the area of the Kikbeek 

subbasin in comparison with Uccle, and in January it was about 0.75 °C lower. In April and October, 

there was hardly any difference between the average temperature in the area of the Kikbeek subbasin 

and that at Uccle. The average yearly temperature in the Kikbeek subbasin was, however, about 0.25 

to 0.75 °C lower. This examination gave the impression that the average winter temperature used in 

the WetSpass model was possibly about 0.5 °C too high. 

In addition, long-term (1906-2000) temperature data for the Maastricht Aachen Airport at Beek, the 

Netherlands, were examined. As this airport is situated nearby the Kikbeek subbasin (about 3 km 

south and 5 km east of it on the Dutch side of the Border Meuse), the temperature data for the airport 

could be representative for the subbasin as well. Therefore, the temperature data were downloaded 

from the KNMI website (KNMI, 2001) and used in the calculation of the long term, average summer 

and winter temperatures. It turned out that at Beek, between 1961 and 2000, the average summer 

temperature was 14.3 °C (0.2 °C higher than at Uccle), and the average winter temperature 5.0 °C (the 

same as at Uccle). As the differences with the seasonal temperatures at Uccle were rather small, it was 

concluded that the use of these seasonal temperatures for the Kikbeek subbasin was fairly reasonable. 

Wind 

The last meteorological parameter to be used as input in the WetSpass model was the average wind 

speed in summer and winter. Here too, data form the meteorological station at Uccle had to be used, as 

no applicable regional data were available. At Uccle, the time-averaged wind speed is 3.27 m/s in 

summer and 3.84 m/s in winter (Sneyers & Vandiepenbeeck, 1995). In a similar way as described for 

the PET and the temperature grids, summer and winter wind grids were produced for the Belgian part 

of the Border Meuse region and for the Kikbeek subbasin. In these grids, all grid cells had values of 

3.27 m/s and 3.84 m/s, respectively (see Table 4). 

As for the temperature, long-term (1961-2000) wind speed data were available from the KNMI 

website (KNMI, 2001), but they could be used only for comparison. These wind speed data comprised 

a long-term hourly time series of the potential wind speed at Beek from August 1961 to December 

2000. This potential wind speed (PWS) is related to the measured wind speed (MWS) at 10 m height 

(above the surface), according to the following equation: 

PWS = { (log RH – log Rz0) (log BH – log Lz0) / (log BH – log Rz0) (log RH – log Lz0) } MWS 

In this equation, RH is the reference height of 10 m, BH is the blending height of 60 m, Rz0 is the 

reference roughness length of 0.03 m, and Lz0 is the local roughness length. Around Beek, Lz0 = 0.47 

m (KNMI, 2001). Hence, 

21


PWS = 1.21 MWS at Beek 

At Beek, the average potential wind speed (PWS), calculated from the time series for the period from 

August 1961 to December 2000, is 4.08 m/s in summer and 5.07 m/s in winter (KNMI, 2001). Hence, 

the average wind speed at Beek (at 10 m) is 3.37 m/s in summer and 4.18 m/s in winter. These 

seasonal average wind speeds do not differ considerably from those observed at Uccle. As it is also not 

known how comparable the wind conditions in the Kikbeek subbasin are with those at Beek, it was 

decided to use the average seasonal wind speed data from Uccle, which were considered to be a 

reasonable approximation. 

Table 4. Overview of meteorological input data for the WetSpass modelling of the Kikbeek 

subbasin 

Parameter Unit Average 

value a 

Precipitation: 

Standard 

deviation 

Minimum 

value b 

Maximum 

value b 

- per year mm 770.4 5.4 756.3 780.0 

- in summer mm 395.1 2.8 387.8 400.0 

- in winter mm 375.3 2.6 368.5 380.0 

Potential evapotranspiration (PET): 

- per year mm 675.0 constant value n/a n/a 

- in summer mm 557.8 constant value n/a n/a 

- in winter mm 117.2 constant value n/a n/a 

Temperature: 

- per year °C 9.6 constant value n/a n/a 

- in summer °C 14.1 constant value n/a n/a 

- in winter °C 5.0 constant value n/a n/a 

Wind speed: 

- per year m/s 3.56 constant value n/a n/a 

- in summer m/s 3.27 constant value n/a n/a 

- in winter m/s 3.84 constant value n/a n/a 

a. Total volume for subbasin / area of subbasin 

b. Total volume for grid cell in concern / grid cell area 

3.5. WetSpass modelling of the hydrological situation in the Kikbeek subbasin 

With the geographical input data presented in the previous section, the actual hydrological situation in 

the Kikbeek subbasin was modelled in WetSpass. For this purpose, a WetSpass input file, which 

referred to the required grids, was created first; see Figure 3 in Section 3.2). In the WetSpass 

modelling, the hydrological effects of the (unnatural) canal Zuid-Willemsvaart were ignored. It was 

assumed that these effects hardly influence the average discharge of the entire subbasin. Probably, 

they only delay the fast discharge a little bit. The modelling results of the actual situation in the 

Kikbeek subbasin were used as a reference for the climate and land use scenarios, which will be 

discussed in Chapter 4. 

22


The output of the WetSpass modelling was a set of grids comprising the calculated average yearly, 

summer, and winter evapotranspiration, groundwater recharge, and surface runoff per grid cell (see 

Map 18 to Map 26 in the Appendix). For each calculated parameter, the spatial average value for the 

actual hydrological situation in the Kikbeek subbasin is tabulated in Table 5. The minimum and 

maximum grid cell values and the standard deviation are given too. Multiplication of the average grid 

cell values by the subbasin area (2334 ha = 23.34 km 2 ) will give the total volumes for the subbasin. 

For instance, the total groundwater recharge per year for the Kikbeek subbasin is (257.8 mm/yr × 

23.34 km 2 =) 6.02 × 10 6 m 3 /yr (= 686 m 3 /hr), and the total surface runoff per year (53.1 mm/yr × 23.34 

km 2 =) 1.24 × 10 6 m 3 /yr (= 141 m 3 /hr). The sum of both, 7.26 × 10 6 m 3 /yr (= 828 m 3 /hr), is the total 

volume of water that yearly discharges from the Kikbeek subbasin into the Border Meuse. 

In order to control the consistency of the model output, the water balance has been calculated as well. 

This was done, by subtracting the average evapotranspiration, groundwater recharge, and surface 

runoff from the average precipitation (see Table 5). The difference between them (i.e. the error in the 

water balance) should be (close to) zero, and is less than 1 % of the average precipitation (see Table 

5). Therefore, the model output seems to be fairly consistent. 

From the results of the WetSpass modelling, discharge coefficients for the Kikbeek subbasin could be 

calculated as well. As the groundwater recharge finally reaches the Border Meuse and is assumed to 

discharge into it, it is regarded as the slow discharge from the Kikbeek subbasin. On the other hand, 

the surface runoff that streams via the surface and the Kikbeek into the Border Meuse is regarded as 

fast discharge. Hence, by dividing the average groundwater recharge and the average surface runoff 

with the average precipitation, the average slow and fast discharge coefficients for the Kikbeek 

subbasin can be obtained, respectively. In Table 5, the calculated average yearly and seasonal 

discharge coefficients are presented as well. 

The modelling and calculation results in Table 5 demonstrate clearly the seasonal differences in the 

hydrology of the Kikbeek subbasin. In summer, the major part of the precipitation (86.9 %) 

evapotranspirates (directly or via the vegetation). The rest of the summer precipitation is distributed 

almost equally between groundwater recharge and surface runoff (6.8 and 6.9 %, respectively). In 

winter, the situation is completely different. Due to the lower temperatures and the lack of green 

vegetation, then only 31.6 % of the slightly smaller amount of precipitation evapotranspirates. The 

surface runoff, however, hardly changes (6.8 %), but the groundwater recharge increases dramatically 

(up to 61.6 %). This implicates that 89.7 % of the groundwater recharge takes place in winter, and 

only 10.3 % in summer. It also implicates that most part (82.6 %) of the total discharge from the 

Kikbeek subbasin into the Border Meuse has precipitated once in winter. The seasonal variability in 

the discharge itself will, of course, depend on the seasonal fluctuations in the groundwater tables and 

the dynamic intensity of the surface runoff process. 

23


Table 5. Results of WetSpass modelling for the actual situation 


value a 

Per year: 

Standard 

deviation 

Minimum 

value b 

Maximum 

value b 

Precipitation (Pr) mm 770.4 5.4 756.3 780.0 

Evapotranspiration (Et) mm 462.0 59.6 212.0 675.0 

Groundwater recharge (Re) mm 257.8 68.0 0.0 387.9 

Surface runoff (Ro) mm 53.1 92.2 1.2 484.6 

Water balance (WB = Pr–Et–Re–Ro) mm –2.4 N/a n/a n/a 

Error in water balance (WB/Pr) % –0.3 N/a n/a n/a 

Slow discharge coefficient (Re/Pr) % 33.5 N/a n/a n/a 

Fast discharge coefficient (Ro/Pr) % 6.9 N/a n/a n/a 

In summer: 



Groundwater recharge (Re) mm 26.7 34.0 –116.2 121.6 


Water balance (WB = Pr–Et–Re–Ro) mm –2.4 N/a n/a n/a 

Error in water balance (WB/Pr) % –0.6 N/a n/a n/a 



In winter: 





Water balance (WB = Pr–Et–Re–Ro) mm 0.0 N/a n/a n/a 

Error in water balance (WB/Pr) % 0.0 N/a n/a n/a 





24


4. Sensitivity of the hydrological characteristics of the Kikbeek subbasin 

towards climate and land use changes 

4.1. Introduction 

In order to analyse the sensitivity of the actual discharge coefficients (see Section 3.5) towards climate 

and land use changes, the discharge coefficients have been calculated for a number of likely and more 

extreme climate and land use scenarios. This was carried out by adjustment of the input parameters in 

the developed WetSpass model (see Chapter 3). In two series of fourteen climate scenarios, only the 

meteorological parameters were adjusted; in a series of nine land use scenarios, only the land use grid 

was changed; and in two series of five combined climate and land use scenarios both the 

meteorological parameters and the land use grid were changed. In this chapter, the description of the 

scenarios and the results of their modelling will be discussed in this order. 

4.2. Climate scenarios 

Scenario descriptions 

The climate scenarios used in this study are similar to those described by Können (2001). Three of 

them (climate scenarios with codes 1 to 3, see Table 6) are the same as previously formulated 

greenhouse scenarios (Können et al. 1997; and Kors et al., 2000), which are assumed to be valid for 

the entire Rhine/Meuse basin, including the Kikbeek subbasin. The greenhouse scenarios are based on 

three classical estimates of the temperature change due to the global warming: a low, central, and high 

estimate. In the central estimate, the mean temperature will increase with 2 °C until the year 2100; in 

the low and high estimates, with 1 and 4 °C, respectively (see Table 8). The temperature changes will 

evolve linearly in time. Therefore, until the year 2050, the temperature increase will be half of that 

until the year 2100 (see Table 7; Können, 2001). 

In the greenhouse scenarios, the precipitation will change proportionally to the temperature change. 

This proportionality is based on empirical relations between observed mean temperatures and the 

precipitation in the past. As a consequence, the average yearly precipitation will have increased by 6 

% in 2100 in scenarios according to the central estimate of the temperature increase (climate scenarios 

with code 2, see Table 6 and Table 8). This increase will even be 12 % in scenarios according to the 

high estimate, but only 3 % in those according to the low estimate (see Table 8). Again, in 2050, the 

increase of the average yearly precipitation will be half of that in 2100 (see Table 7). In contrast with 

the temperature change, the precipitation change depends on the season. In winter the average 

precipitation change is twice as much as the average yearly precipitation change, in summer it is only 

one third (see Table 7 and Table 8; Können, 2001). An impression of the implications of a temperature 

and precipitation increase for the water management in the Rhine/Meuse basin is presented by 

Kwadijk (2000). 

The evaporation is assumed to remain linear with temperature. Hence, in the greenhouse scenarios, it 

will increase with an increase in temperature (in the high estimate even by 16 % in 2100). This 

proportionality is based on a study by Brandsma (1995), and the evaporation changes for 2050 and 

2100 presented in Table 7 and Table 8, respectively, are based on a study by Haasnoot et al. (1999). 

As the increase in temperature, the increase in evaporation is assumed to be season independent 

(Können, 2001). 

The average wind speed is not expected to change much in the greenhouse scenarios. A margin of ± 5 

% seems to be sufficient, as such a margin corresponds to the observed decadal variability in the past 

century (Können, 2001). Therefore, each of the climate scenarios that are similar to the previously 

formulated greenhouse scenarios (i.e. the scenarios with codes 1 to 3; Können et al. 1997; and Kors et 

25


al., 2000) comprises three wind speed variants. In variant a, the average wind speed is 5 % higher; in 

variant b, it is the same; and in variant c, it is 5 % lower (see Table 7 and Table 8). 

Besides the greenhouse scenarios, a dry scenario is presented (climate scenario with code 5, see Table 

6). In this scenario, the temperature and precipitation changes are uncoupled. The temperature and 

evaporation changes are the same as in the greenhouse scenarios that were based on the high estimate 

of the temperature change (climate scenarios with code 3, see Table 6, and also Table 7 and Table 8). 

The average precipitation, however, decreases by 10 % in summer and winter, in both the scenarios 

for 2050 and 2100 (see Table 7 and Table 8). As the probability for higher wind speeds decreases, the 

average wind speed decreases by 0 to 10 %, both in 2050 and 2100 (see Table 7 and Table 8). In both 

the scenarios for 2050 and 2100, the average wind speed decreases by 0 % in variant b, 5 % in variant 

c, and 10 % in variant d (see Table 6; Können, 2001). 

Another different group of scenarios comprises scenarios in which the climate change is induced by a 

sudden change in the North Atlantic thermohaline circulation. According to a study by Klein Tank & 

Können (1997), such a change can result in a cooling of the ocean, and consequently in a cooling of 

the atmosphere. Under the assumption of an unchanged atmospheric circulation, a worst-case cooling 

of 4 °C of the ocean will result in a season independent atmospheric cooling of about 2 °C in the 

Netherlands (the same will probably be true for adjacent areas, including the Kikbeek subbasin). Such 

a sudden change in the oceanic circulation is not bound to a predictable moment. It could happen at 

any moment and it will not evolve linearly in time. The estimation is that it can be completed in about 

5 to 10 years. As a consequence, its effects should be superposed on the climate at the moment of the 

event (Können, 2001). Therefore, two scenarios were examined: One in which only a climate change 

takes place that is induced by a sudden change in the North Atlantic thermohaline circulation (climate 

scenario with code 4, see Table 6), and another in which this climate scenario is superposed on the 

greenhouse scenario that is based on the central estimate of the temperature change due to global 

warming (climate scenario with code 4/2, see Table 6). 

With respect to the precipitation, it is unclear what will happen in case of a change in the oceanic 

circulation. The best assumption is probably that the temperature and precipitation changes remain 

coupled. This means that, if the event would occur now (or later without another climate change), the 

precipitation changes would be equal to that of the greenhouse scenario that is based on the central 

estimate of the temperature change in 2100 (+ 2 °C), but of opposite sign. Hence, the average yearly 

precipitation would decrease by 6 %, the average summer precipitation by 2 %, and the average winter 

precipitation by 12 % (see Table 7 and Table 8). Similarly, it is assumed that the relationship between 

temperature and evaporation will remain linear. Hence, if the mean temperature decreases by 2 °C, the 

evaporation will decrease proportionally by 8 % (see Table 7 and Table 8). Finally, the wind speed is 

not expected to be a subject of change, because of the assumption of an unchanged atmospheric 

circulation (Können, 2001). In combination with the greenhouse scenarios that are based on the central 

temperature estimate (climate scenarios 2050-2b and 2100-2b), the climate changes caused by the 

global warming will compensate in the year 2050 half the effects of a change in the oceanic circulation 

and neutralise them completely in the year 2100 (see Table 7 and Table 8). 

26


Code* Climate scenario description 

actual situation 

Table 6. Description of climate scenarios 

1 greenhouse scenario (wet), low estimate of the temperature change 

2 greenhouse scenario (wet), central estimate of the temperature change 

3 greenhouse scenario (wet), high estimate of the temperature change 

4 sudden change in the North Atlantic circulation, no further climate changes 

4/2 combination of climate scenarios 4 and 2 

5 dry scenario based on the high estimate of the temperature change 

Wind subcode* Change in wind 

a more wind (average wind speed 5 % higher) 

b no change 

c less wind (average wind speed 5 % lower) 

d much less wind (average wind speed 10 % lower) 

* The full code for a climate scenario comprises the year in concern, the climate scenario code and the wind 

subcode. For instance, 2050-2b is the code for the climate scenario for the year 2050, in which the climate 

has changed according to the central estimate (wet) without a change in average wind speed. 

27


Table 7. Climate changes in the year 2050 according to the proposed climate scenarios 

Climate scenario 2050-1 2050-2 2050-3 2050-4* 2050-4/2* 2050-5 

Description Low 

estimate 

wet 

Central 

estimate 

wet 

High 

estimate 

wet 

Change 

North Atlan. 

circ. 

Scen. 

2050-4 and 

2050-2 

High 

estimate 

dry 

Change in temperature +0.5 °C +1 °C +2 °C –2 °C –1 °C +2 °C 

Change in precipitation: 

- per year +1.5 % +3 % +6 % –6 % –3 % –10 % 

- in summer +0.5 % +1 % +2 % –2 % –1 % –10 % 

- in winter +3 % +6 % +12 % –12 % –6 % –10 % 

Change in evaporation: 

- per year +4 % +4 % +8 % –8 % –4 % +8 % 

- in summer +4 % +4 % +8 % –8 % –4 % +8 % 

- in winter +4 % +4 % +8 % –8 % –4 % +8 % 

Change in wind ±5 % ±5 % ±5 % no change ±5 % –10 to 0 % 

* In these scenarios, it is assumed that a change in the thermohaline circulation of the Northern Atlantic 

Ocean takes place between now and the year 2050. 

Table 8. Climate changes in the year 2100 according to the proposed climate scenarios 

Climate scenario 2100-1 2100-2 2100-3 2100-4* 2100-4/2 2100-5 

Description Low 

estimate 

wet 

Central 

estimate 

wet 

High 

estimate 

wet 

Change 

North 

Atlan.circ. 

Scen. 

2050-4 and 

2050-2 

High 

estimate 

dry 

Change in temperature +1 °C +2 °C +4 °C –2 °C no change +4 °C 

Change in precipitation: 

- per year +3 % +6 % +12 % –6 % no change –10 % 

- in summer +1 % +2 % +4 % –2 % no change –10 % 

- in winter +6 % +12 % +25 % –12 % no change –10 % 

Change in evaporation: 

- per year +4 % +8 % +16 % –8 % no change +16 % 

- in summer +4 % +8 % +16 % –8 % no change +16 % 

- in winter +4 % +8 % +16 % –8 % no change +16 % 

Change in wind ±5 % ±5 % ±5 % no change ±5 % –10 to 0 % 

* In these scenarios, it is assumed that a change in the thermohaline circulation of the Northern Atlantic 

Ocean takes place between now and the year 2100. 

28


For each investigated climate scenario, the grids of the meteorological parameters concerning the 

actual situation in the Kikbeek subbasin (see Section 3.4) were adjusted in accordance with the figures 

in Table 7 or Table 8. The temperature grids for summer and winter were adjusted by adding in 

ArcInfo GIS the tabulated temperature change to the grid cell values. The precipitation, PET, and wind 

speed grids for summer and winter were, on the other hand, adjusted by multiplying the grid cell 

values in ArcInfo GIS by a factor that corresponds to the tabulated change in terms of percentage. For 

instance, for the modelling of scenario 2100-2b, all grid cell values in the summer precipitation grid 

were multiplied by a factor of 1.02, and in the winter precipitation grid by a factor of 1.12. After the 

adjustment of the meteorological grids, also the input files for the WetSpass modelling were adjusted 

for the climate scenarios. An overview of all modelled climate scenarios is given in Table 9. 

Table 9. Overview of modelled climate scenarios 

Modelled climate scenarios 

2050-1a 2050-2a 2050-3a 2050-4b 2050-5b 

2050-1b 2050-2b 2050-3b 2050-4/2b 2050-5c 

2050-1c 2050-2c 2050-3c 2050-5d 

2100-1a 2100-2a 2100-3a 2100-4b 2100-5b 

2100-1b 2100-2b 2100-3b 2100-4/2b 2100-5c 

2100-1c 2100-2c 2100-3c 2100-5d 

Results of the WetSpass modelling 

For the greenhouse scenario 2100-2b that is based on the central temperature estimate for the year 

2100, the output of the WetSpass modelling is presented in Table 10 (and in Map 27 to Map 32 in the 

Appendix). As for the actual situation in the Kikbeek subbasin (see Table 5 in Section 3.5), for climate 

scenario 2100-2b, the average, minimum, and maximum grid cell values as well as the calculated 

standard deviations are tabulated in Table 10 for the average yearly, summer, and winter 

evapotranspiration, groundwater recharge, and surface runoff map. As in Table 5, the average yearly, 

summer, and winter precipitation are given for comparison, and the calculated water balance and the 

error within it in order to control the consistency of the model output. Finally, the calculated average 

yearly and seasonal discharge coefficients are presented as well. 

29


Table 10. Results of the WetSpass modelling for the (wet) greenhouse scenario 2100-2b, 

which is based on the central temperature estimate for the year 2100 


value a 

Per year: 

Standard 

deviation 

Minimum 

value b 

Maximum 

value b 





Water balance (WB = Pr–Et–Re–Ro) mm –3.1 n/a n/a n/a 

Error in water balance (WB/Pr) % –0.4 n/a n/a n/a 

Slow discharge coefficient (Re/Pr) % 33.4 n/a n/a n/a 

Fast discharge coefficient (Ro/Pr) % 7.1 n/a n/a n/a 

In summer: 



Groundwater recharge (Re) mm 12.5 37.0 –150.6 112.6 


Water balance (WB = Pr–Et–Re–Ro) mm –3.1 n/a n/a n/a 

Error in water balance (WB/Pr) % –0.8 n/a n/a n/a 



In winter: 





Water balance (WB = Pr–Et–Re–Ro) mm 0.0 n/a n/a n/a 

Error in water balance (WB/Pr) % 0.0 n/a n/a n/a 





30


From comparison of Table 10 with Table 5 in Section 3.5, it is clear that according to climate scenario 

2100-2b the average yearly evapotranspiration, groundwater recharge, and surface runoff will have 

increased all in 2100 due to the increase of the mean temperature (+ 2 °C) and the average yearly 

precipitation (+ 52.9 mm). Most part of the additional yearly precipitation will just evapotranspi rate 

(+ 30.3 mm) or lead to an increase of the average yearly groundwater recharge (+ 17.6 mm). The 

average yearly surface runoff will increase with only 5.7 mm. Despite the overall increase of the 

average yearly groundwater recharge, the average groundwater recharge in summer will decrease 

significantly from 26.7 mm to 12.5 mm. Hence, according to climate scenario 2100-2b, in 2100, the 

groundwater recharge will take place predominantly in winter as well, but even more than it already 

does (262.8 mm in comparison with 231.1 mm). With the exception of the slow discharge coefficient 

in summer, the yearly and seasonal discharge coefficients will not change dramatically with respect to 

the actual situation. The slow discharge coefficient in summer, however, will be more than halved (to 

3.1 %), but was already relatively low (6.8 %). 

For more relative comparisons, the changes of the average yearly and seasonal precipitation, 

evapotranspiration, groundwater recharge, and surface runoff with respect to the actual situation are 

given, in Table 11, in terms of percentage. In this table, the relative WetSpass output for the other 

modelled climate scenarios is given as well. The percentages were obtained by division of the average 

grid cell value of the parameter in concern for the climate scenario by the average grid cell value of the 

same parameter in the actual situation (see Table 5). 

31


Parameter 

Per year: 

Table 11. Results of the WetSpass modelling for the climate scenarios, 

expressed as changes with respect to the actual situation (average values) 

Scenario 

2050-1a 

% 

2050-1b 

% 

2050-1c 

% 

2050-2a 

% 

2050-2b 

% 

2050-2c 

% 

2050-4/2b 

% 

Precipitation (Pr) 1.7 1.7 1.7 3.4 3.4 3.4 –3.4 

Evapotranspiration (Et) 2.0 2.9 3.8 2.4 3.3 4.2 –3.3 

Groundwater recharge (Re) 1.2 –0.4 –2.0 5.0 3.4 1.8 –3.4 

Surface runoff (Ro) 2.6 2.6 2.6 5.4 5.4 5.4 –5.4 

In summer: 



Groundwater recharge (Re) –12.0 –25.6 –39.9 –12.9 –26.6 –40.8 26.4 


In winter: 



Groundwater recharge (Re) 2.7 2.5 2.4 7.0 6.9 6.7 –6.8 


Per year: 

Scenario 2100-1a 

% 

2100-1b 

% 

2100-1c 

% 

2100-2a 

% 

2100-2b 

% 

2100-2c 

% 

2100-4/2b 

% 

Precipitation (Pr) 3.4 3.4 3.4 6.9 6.9 6.9 0.0 

Evapotranspiration (Et) 2.4 3.3 4.2 5.7 6.6 7.5 0.0 

Groundwater recharge (Re) 5.0 3.4 1.8 8.4 6.8 5.2 0.0 

Surface runoff (Ro) 5.4 5.4 5.4 10.8 10.8 10.8 0.0 

In summer: 



Groundwater recharge (Re) –12.9 –26.6 –40.8 –39.3 –53.1 –67.6 0.0 


In winter: 



Groundwater recharge (Re) 7.0 6.9 6.7 13.9 13.7 13.6 0.0 


to be continued on next page 

32


parameter 

Per year: 

Table 11 (continued). Results of the WetSpass modelling for the climate scenarios, 


Scenario 

2050-3a 

% 

2050-3b 

% 

2050-3c 

% 

2050-4b 

% 

2050-5b 

% 

2050-5c 

% 

2050-5d 

% 

Precipitation (Pr) 6.9 6.9 6.9 –6.9 –10.0 –10.0 –10.0 

Evapotranspiration (Et) 5.7 6.6 7.5 –6.6 2.2 3.0 3.8 

Groundwater recharge (Re) 8.4 6.8 5.2 –6.8 –30.9 –32.3 –33.8 

Surface runoff (Ro) 10.8 10.8 10.8 –10.8 –10.6 –10.6 –10.6 

In summer: 



Groundwater recharge (Re) –39.3 –53.1 –67.6 52.6 – – –165.4 


In winter: 





Per year: 

Scenario 

2100-3a 

% 

2100-3b 

% 

2100-3c 

% 

2100-4b 

% 

2100-5b 

% 

2100-5c 

% 

2100-5d 

% 





In summer: 



Groundwater recharge (Re) –91.7 – – 52.6 – – –218.1 


In winter: 





In Table 11, it can be seen that for all greenhouse scenarios (climate scenarios with codes 1 to 3), in 

general, similar hydrological changes were found, though more or less sound. In nearly all greenhouse 

scenarios, all the average yearly and seasonal values of the hydrological parameters in concern will 

increase, with the exception of the average groundwater recharge in summer. In two greenhouse 

scenarios only (2050-1b and 2050-1c), also the average yearly groundwater recharge will decrease, but 

33


only slightly. In comparison with the central temperature estimate scenarios (code 2), the hydrological 

changes (of positive or negative sign) will clearly be larger in the high temperature estimate scenarios 

(code 3), and considerably smaller in the low temperature estimate scenarios (code 1). Also, it is clear 

that the hydrological effects of the climate changes in 2050 will be about half of those that are 

expected for 2100. Moreover, the hydrological changes will generally be a little larger in scenarios 

with less wind (subcode c) and a little smaller in scenarios with more wind (subcode a). Consequently, 

the greenhouse scenarios with the most dramatic hydrological changes were scenarios 2100-3b and 

2100-3c. In these two scenarios, the groundwater recharge in summer will decrease by even more than 

100 %. This means that according to these two scenarios in summer a net groundwater 

evapotranspiration will start to occur somewhere between the years 2050 and 2100. 

More dramatic results were obtained for the variants of the dry scenario (climate scenarios with code 

5). In these dry scenarios, the groundwater recharge in summer will have decreased by about 150 % in 

2050, and by about 200 % in 2100 (see Table 11). Hence, in these scenarios too, a net 

evapotranspiration of groundwater will occur in summer, being, however, much more pronounced in 

summer and starting earlier (already before 2050). In contrast with the wet greenhouse scenarios 

(codes 1 to 3), the groundwater recharge will also decrease in winter. This decrease will be, however, 

not more than about 22 % of 231.1 mm (see Table 5). Hence, a net evapotranspiration of groundwater 

in winter is not to be expected before 2100 or soon afterwards. As a result of the reversed or 

diminished groundwater recharge in summer and winter, the average yearly groundwater recharge will 

diminish too in the dry scenarios (by about 30 % in 2050 and about 40 % in 2100, see Table 11). Also 

the surface runoff will decrease in the dry scenarios, but only by about 10 to 12 % (more or less season 

independent, see Table 11). The average evapotranspiration, on the other hand, will increase, 

particularly in winter, but also in summer (see Table 11). As the average precipitation will decrease in 

the dry scenarios, the increase of the evapotranspiration (due to the rise of temperature) will clearly 

take place on the expense of the groundwater recharge and the surface runoff (which will both 

decrease). 

For the scenarios in which only climate changes occur that are induced by a sudden change in the 

North Atlantic thermohaline circulation (the identical scenarios 2050-4b and 2100-4b), the changes in 

the output parameters of the WetSpass modelling (with respect to the actual situation) are practically 

completely opposite of those for greenhouse scenario 2100-2b (which is equal to scenario 2050-3b). 

Furthermore, they are twice as big as the changes of the average yearly and seasonal 

evapotranspiration, groundwater recharge, and surface runoff in greenhouse scenario 2050-2b, but also 

opposite of sign (see Table 11). This is not a surprise, as the changes in the input parameters for the 

compared scenarios are related similarly (see Table 7 and Table 8). 

In combination with the greenhouse scenarios 2050-2b and 2100-2b, the climate changes caused by 

the global warming would compensate in the year 2050 half the effects of a change in the oceanic 

circulation and neutralise them completely in the year 2100. Actually, the changes in the input 

parameters for the WetSpass modelling (with respect to the actual situation) were for scenario 2050- 

4/2b half of those for scenario 2050-4b. For scenario 2100-4/2b, these changes were all zero (see 

Table 7 and Table 8). Exactly the same trends were seen in the changes of the output parameters of the 

WetSpass modelling (with respect to the actual situation). 

Discharge coefficients 

For all examined climate scenarios, average yearly and seasonal, fast and slow discharge coefficients 

have been calculated for the entire Kikbeek subbasin from the results of the WetSpass modelling. An 

overview of them is presented in Table 12. In the greenhouse scenarios (codes 1 to 3), similar trends 

came forward as have been seen already in greenhouse scenario 2100-2b (see above): With the 

exception of the slow discharge coefficient in summer, the discharge coefficients will not change 

dramatically with respect to the actual situation. For the slow discharge coefficient in summer, 

however, a significant decrease is observed. This decrease is the largest in the greenhouse scenarios 

that are based on the high temperature estimate, and the smallest in those that are based on the low 

34


one. The decrease of the slow discharge coefficient in summer will also be larger in 2100 than in 2050, 

which is of course not a surprise. Finally, it is also larger in the less windy greenhouse scenarios 

(scenarios with subcode c), and smaller in the windier ones (scenarios with subcode a). In two 

greenhouse scenarios (2100-3b and 2100-3c), the slow discharge coefficient in summer will become 

negative (–0.4 and –1.3 %, respectively). It was in these two scenarios that a net evapotranspiration of 

groundwater in summer was simulated (see above). 

In contrast to the greenhouse scenarios, the climate scenarios that include a sudden change in the 

North Atlantic thermohaline circulation (scenarios with code 4) foresee an increase of the average 

slow discharge coefficient in summer (up to 10.5 % in the identical scenarios 2050-4b and 2100-4b). 

As in the greenhouse scenarios, the other average, yearly and seasonal, fast and slow discharge 

coefficients will not change dramatically. In scenario 2100-4/2b, of course, not any change with 

respect to the actual situation is observed (see Table 12). 

The dry scenarios (with code 5) show significant decreases of the average slow discharge coefficients 

for both seasons. As a result, the average yearly slow discharge coefficient will decrease as well. The 

average yearly and seasonal fast discharge coefficients, on the other hand, will not change noticeably. 

The decrease of the average slow discharge coefficients will be larger in summer than in winter. In 

summer, the average slow discharge coefficients will even be negative, both in 2050 and 2100. Its 

values range between –3.1 and –8.9 %, which is even more negative than in the greenhouse scenarios 

2100-3b and 2100-3c. In summer and winter, the decrease of the average slow discharge coefficient 

will be larger in the less windy dry scenarios (subcodes c and d), and larger in 2100 in comparison 

with 2050 (see Table 12). 

35


Scenario Average 

per year 

Table 12. Discharge coefficients for the Kikbeek subbasin 

in the climate scenarios for 2050 and 2100 

Fast discharge coefficients a Slow discharge coefficients b 

Average 

in summer 

Average 

in winter 

Average 

per year 

Average 

in summer 

Average 

in winter 

% % % % % % 

Actual situation 6.9 6.9 6.8 33.5 6.8 61.6 

For 2050: 

2050-1a 6.9 7.1 6.8 33.3 5.9 61.4 

2050-1b 6.9 7.1 6.8 32.8 5.0 61.3 

2050-1c 6.9 7.1 6.8 32.2 4.0 61.2 

2050-2a 7.0 7.2 6.9 33.9 5.8 62.2 

2050-2b 7.0 7.2 6.9 33.5 4.9 62.1 

2050-2c 7.0 7.2 6.9 32.9 4.0 62.0 

2050-3a 7.1 7.4 6.9 33.9 4.0 62.6 

2050-3b 7.1 7.4 6.9 33.4 3.1 62.5 

2050-3c 7.1 7.4 6.9 32.9 2.1 62.4 

2050-4b 6.6 6.4 6.8 33.5 10.5 60.4 

2050-4/2b 6.7 6.7 6.8 33.5 8.6 61.0 

2050-5b 6.8 6.9 6.7 25.7 –3.1 56.0 

2050-5c 6.8 6.9 6.7 25.2 –4.0 55.9 

2050-5d 6.8 6.9 6.7 24.6 –4.9 55.7 

For 2100: 

2100-1a 7.0 7.2 6.9 33.9 5.8 62.2 

2100-1b 7.0 7.2 6.9 33.5 4.9 62.1 

2100-1c 7.0 7.2 6.9 32.9 4.0 62.0 

2100-2a 7.1 7.4 6.9 33.9 4.0 62.6 

2100-2b 7.1 7.4 6.9 33.4 3.1 62.5 

2100-2c 7.1 7.4 6.9 32.9 2.1 62.4 

2100-3a 7.4 8.0 6.9 34.2 0.5 63.6 

2100-3b 7.4 8.0 6.9 33.7 –0.4 63.5 

2100-3c 7.4 8.0 6.9 33.2 –1.3 63.5 

2100-4b 6.6 6.4 6.8 33.5 10.5 60.4 

2100-4/2b 6.9 6.9 6.8 33.5 6.8 61.6 

2100-5b 6.8 6.9 6.7 22.6 –7.0 53.6 

2100-5c 6.8 6.9 6.7 22.0 –7.9 53.5 

2100-5d 6.8 6.9 6.7 21.4 –8.9 53.3 

a. Total surface runoff / total precipitation for the Kikbeek subbasin 

b. Total groundwater recharge / total precipitation for the Kikbeek subbasin 

36


4.3. Land use scenarios 


In the land use scenarios here presented (see Table 13), only the land use grid that was used as input in 

the WetSpass modelling of the actual situation in the Kikbeek subbasin (see Section 3.4, and 

particularly Table 3) was adjusted. It was adjusted in such a way that the effects of extreme, but 

possible land use changes could be examined. In five cases (land use scenarios with codes A to E), the 

land use will develop unidirectionally towards uniformity. In these cases, the non-urban areas will be 

turned entirely into deciduous forests (scenario A), meadows (scenario B), agricultural land other than 

maize crops (scenario C), maize crops (scenario D), or open urban areas (scenario E). Besides the 

actual urban areas, only the actual non-urban infrastructure, surface waters, and mud flats will remain 

as they are. In scenario B, the wet meadows will, of course, remain wet. 

In four other cases (land use scenarios with codes F to K, see Table 13), only particular types of land 

use will be turned completely into other types. In one of them, the agricultural land (including maize 

crops) and the (wet) meadows will become deciduous forests (scenario F); in another the agricultural 

land only (scenario G). In the third case, the agricultural land (including maize crops) will become 

meadows (scenario H), and in the fourth case the agricultural land (including maize crops) and the 

(wet) meadows will become open urban areas (scenario K). From the present point of view, these four 

cases are probably more realistic than the first five. They do not affect the present more or less natural 

areas with forests, heathers, and shrubs. As these scenarios are still rather extreme, it is, however, 

more realistic to expect them to occur only partly and/or in combination with other land use scenarios. 

Nevertheless, their hydrological modelling with WetSpass can give a good impression of how the 

hydrology of the Kikbeek subbasin and its discharge coefficients will react on future land use changes. 

37


Land use 

scenario code a 

Table 13. Description of land use scenarios 

Description of land use changes 

actual situation 

A non-urban areas b (except infrastructure, surface waters, and mud flats) replaced by 

deciduous forests 

B non-urban areas b (except infrastructure, surface waters, mud flats, and wet meadows) 

replaced by meadows 

C non-urban areas b (except infrastructure, surface waters, and mud flats) replaced by 

agricultural land c 

D non-urban areas b (except infrastructure, surface waters, and mud flats) replaced by maize 

crops 

E non-urban areas b (except infrastructure, surface waters, and mud flats) replaced by open 

urban areas 

F agricultural land (including maize crops) and (wet) meadows replaced by deciduous forests 

G agricultural land (including maize crops) replaced by deciduous forests 

H agricultural land (including maize crops) replaced by meadows 

K agricultural land (including maize crops) and (wet) meadows replaced by open urban areas 

a. The full code for a land use scenario comprises the year in concern and the land use scenario code. For 

instance, 2001-A is the code for the land use scenario in which the land use would change (now), without 

further (climate) changes, according to land use scenario A. 

b. Non-urban areas include agricultural land (including maize crops), (wet) meadows, deciduous forest, 

coniferous forest, mixed forest, heather, shrubs, mud flats, surface waters, and non-urban infrastructure 

c. All types of agricultural land, except maize crops 

For the adjustment of the land use grid for the land use scenarios, the grid cell values in the land use 

grid for the actual situation in the Kikbeek subbasin (see Table 3 in Section 3.4) had to be adjusted. In 

other words, for each land use scenario, the land use codes (the grid cell values in the land use grid) 

had to be changed for the land use types that were involved. These changes were carried out in 

ArcInfo GIS, and an overview of them is presented in Table 14 (and in Map 8 to Map 16 in the 

Appendix). An overview of the modelled land use scenarios is given in Table 15. As in these land use 

scenarios the land use will change only, and all other (mostly meteorological) input parameters for the 

WetSpass modelling will remain the same as in the actual situation, the land use scenarios have been 

given codes from 2001-A to 2001-K, in which the year 2001 corresponds to the actual 

(meteorological) situation. 

38


Land use scenario 

code a 

Table 14. Adjusted land use codes in the land use grids for the land use scenarios 

Changed land use codes b New land use 

code b 

Number of grid cells 

in concern 

A 21, 27, 23, 28, 32, 33, 35, and 36 31 40342 

B 21, 27, 31, 32, 33, 35, and 36 c 23 37848 

C 27, 23, 28, 31, 32, 33, 35, and 36 21 30769 

D 21, 23, 28, 31, 32, 33, 35, and 36 27 37903 

E 21, 27, 23, 28, 31, 32, 33, 35, and 36 10 43353 

F 21, 27, 23, and 28 31 23539 

G 21 and 27 31 18034 

H 21 and 27 23 18034 

K 21, 27, 23, and 28 10 23539 

a. For descriptions of the land use scenarios, see Table 13. 

b. For an explanation of the land use codes, see Table 3. 

c. In grid cells with land use code 28 (wet meadows), the grid cell value was not changed 

Table 15. Overview of modelled land use scenarios 

Modelled land use scenarios 

2001-A 2001-B 2001-C 2001-D 2001-E 

2001-F 2001-G 2001-H 2001-K 


From the output of the WetSpass modelling of the land use scenarios, the changes of the average 

yearly and seasonal evapotranspiration, groundwater recharge, and surface runoff with respect to the 

actual situation were calculated. The results of these calculations are presented in Table 16. The 

changes in terms of percentage were obtained by division of the average grid cell value of the 

parameter in concern for the land use scenario by the average grid cell value of the same parameter in 

the actual situation (see Table 5). The changes of the average yearly and seasonal precipitation, which 

are given for comparison (as in Table 11 for the climate scenarios), are, of course, all 0 %, since there 

were no meteorological changes included in the land use scenarios. 

For most of the land use scenarios, the changes in the WetSpass output are not more dramatic than in 

the climate scenarios. Exceptions are the land use scenarios that result in more maize crops (scenario 

2001-D) or in more urban areas (scenarios 2001-E and 2001-K). In the latter, urban scenarios, a 

strong, almost season independent increase of the surface runoff is observed, on the expense of 

evapotranspiration and groundwater recharge. Probably, this is the result of the increased total area of 

built-up and impermeable soils in the urban scenarios. In the maize scenario (2001-D), a surprisingly 

strong increase (by 197.6 %) of the average groundwater recharge in summer is observed. It should, 

39


however, be recalled that the groundwater recharge in summer was not very high (26.7 mm in the 

actual situation, see Table 5). Hence, an increase of it can easily become a strong increase in terms of 

percentage. In the maize scenario, also a relatively strong increase (by 15.5 %) of the average surface 

runoff in summer is observed. This increase is observed as well in the land use scenario that results in 

more agricultural land other than maize crops (agricultural scenario 2001-C). In contrast, a decrease 

(by 3.7 to 14.2 %) of the average surface runoff in summer is observed in the forest and meadow 

scenarios, which result in more deciduous forests (land use scenarios 2001-A, 2001-F, and 2001-G) or 

in more meadows (land use scenarios 2001-B and 2001-H), respectively. For the forest scenarios, a 

decrease of the average groundwater recharge in summer is observed as well. Similar changes will be 

found when the agricultural land (including maize crops) will be replaced by forests (scenario 2001-G) 

or by meadows (scenario 2001-H). In both cases, the average groundwater recharge and surface runoff 

in summer will decrease in favour of the evapotranspiration. 

From the results of the WetSpass modelling of the five unidirectional land use scenarios (2001-A to 

2001-E), the following conclusions can be drawn: the presence of deciduous forests diminishes the 

groundwater recharge and surface runoff in summer, and the evapotranspiration in winter (see the 

results for scenario 2001-A); the presence of meadows predominantly promotes the groundwater 

recharge in summer (see scenario 2001-B); the presence of agricultural land promotes the groundwater 

recharge and surface runoff in summer, the first especially in presence of maize crops (see scenarios 

2001-C and 2001-D); and the presence of (open) urban areas strongly promotes the surface runoff 

throughout the seasons (see scenario E). 

40


Table 16. Results of the WetSpass modelling for the land use scenarios, 


Parameter 

Per year: 

Scenario 

2001-A 

% 

2001-B 

% 

2001-C 

% 

2001-D 

% 

2001-E 

% 

Precipitation (Pr) 0.0 0.0 0.0 0.0 0.0 

Evapotranspiration (Et) –0.7 –2.1 –2.5 –13.3 –6.1 

Groundwater recharge (Re) 2.7 4.3 2.7 22.1 –10.8 

Surface runoff (Ro) –7.0 –3.0 8.8 8.8 105.5 

In summer: 


Evapotranspiration (Et) 3.7 –1.3 –2.0 –16.6 –7.6 

Groundwater recharge (Re) –33.2 21.0 10.2 197.6 –5.1 

Surface runoff (Ro) –14.2 –3.7 15.5 15.5 100.5 

In winter: 


Evapotranspiration (Et) –13.5 –4.2 –4.0 –4.0 –1.7 


Surface runoff (Ro) 0.7 –2.2 1.7 1.7 110.7 

Per year: 

Scenario 2001-F 

% 

2001-G 

% 

2001-H 

% 

2001-K 

% 

Precipitation (Pr) 0.0 0.0 0.0 0.0 

Evapotranspiration (Et) 1.2 1.0 1.9 –0.5 

Groundwater recharge (Re) –0.6 –0.4 –2.3 –9.7 

Surface runoff (Ro) –7.3 –6.8 –5.1 51.4 

In summer: 


Evapotranspiration (Et) 3.3 2.8 2.7 –1.0 

Groundwater recharge (Re) –28.4 –22.5 –26.6 –34.0 


In winter: 


Evapotranspiration (Et) –5.2 –4.0 –0.5 0.9 

Groundwater recharge (Re) 2.7 2.1 0.5 –6.8 


41



As for the climate scenarios, average yearly and seasonal, fast and slow discharge coefficients have 

been calculated for the examined land use scenarios. An overview of them is presented in Table 17. 

For the discharge coefficients in the land use scenarios too, the changes with respect to the actual 

situation were generally not more dramatic than in the climate scenarios. Exceptions were again the 

land use scenarios that result in more maize crops (scenario 2001-D) or in more urban areas (scenarios 

2001-E and 2001-K). 

In the urban scenarios (2001-E and 2001-K), the average fast discharge coefficient will increase 

strongly in summer and winter (up to 13.9 and 14.4 %, respectively). The average slow discharge 

coefficient, on the other hand, will decrease significantly, especially in winter (by about 7 % in 

scenario 2001-E). In the maize scenario (2001-D), the average slow discharge coefficient shows a 

surprisingly strong increase in summer (up to 20.1 %), and for the same season a clear (but less strong) 

increase of the average fast discharge coefficient is seen as well (up to 8.0 %). A similar increase of 

the average fast discharge coefficient in summer is seen in the agricultural scenario (2001-C). In 

contrast, a comparable decrease of the average fast discharge coefficient in summer (down to about 6 

%) is observed in the forest scenarios (2001-A, 2001-F, and 2001-G), and a smaller one in the meadow 

scenarios (2001-B and 2001-H). In the forest scenarios, a strong decrease of the average slow 

discharge coefficient in summer (down to about 5 %) is observed as well. The same is observed for the 

meadow scenario in which the agricultural land (including maize crops) will be replaced by meadows 

(scenario 2001-H), but the opposite for the meadow scenario in which all non-urban areas will be 

replaced by meadows (scenario 2001-B). Finally, a significant increase of the average slow discharge 

coefficient in winter (up to 65.8 %) is observed in the forest scenarios. 

From the average discharge coefficients for the five unidirectional land use scenarios (2001-A to 

2001-E), it can be concluded that urbanisation of the Kikbeek subbasin will result, season 

independent, in strong increases of the fast discharge coefficients, and in possibly strong decreases of 

the slow discharge coefficients. An increase in agricultural land (including maize crops) will result in 

an increase of the fast and slow discharge coefficients in summer, the latter particularly in case of an 

increase of the maize crops. An increase of deciduous forest will result, on the other hand, in a 

decrease of the fast and slow discharge coefficients in summer, but in an important increase of the 

slow discharge coefficient in winter. Finally, more meadows will result in larger slow discharge 

coefficients, in summer and winter. 

42


Table 17. Discharge coefficients for the Kikbeek subbasin in the land use scenarios 


per year 

Fast discharge coefficients a Slow discharge coefficients b 

Average 

in summer 

Average 

in winter 

Average 

per year 

Average 

in summer 

Average 

in winter 

% % % % % % 


2001-A 6.4 5.9 6.9 34.4 4.5 65.8 

2001-B 6.7 6.7 6.7 34.9 8.2 63.1 

2001-C 7.5 8.0 7.0 34.4 7.4 62.7 

2001-D 7.5 8.0 7.0 40.9 20.1 62.7 

2001-E 14.2 13.9 14.4 29.9 6.4 54.5 

2001-F 6.4 5.9 6.8 33.3 4.8 63.2 

2001-G 6.4 6.0 6.8 33.3 5.2 62.9 

2001-H 6.5 6.4 6.7 32.7 5.0 61.8 

2001-K 10.4 10.1 10.8 30.2 4.5 57.4 

a. Total surface runoff / total precipitation for the Kikbeek subbasin 

b. Total groundwater recharge / total precipitation for the Kikbeek subbasin 

4.4. Combined climate and land use scenarios 


In two series of five combined climate and land use scenarios, the possible effects of land use changes 

and global warming on the hydrological characteristics of the Kikbeek subbasin were examined. For 

the WetSpass modelling of these scenarios, not only adjusted land use grids were used, but also 

adjusted grids of the meteorological parameters. The adjusted land use grids that were used were those 

that were constructed for the rather extreme, but possible, unidirectional land use scenarios described 

in the previous section (scenarios 2001-A to 2001-E, see Table 13 and Table 14, and also Table 3 in 

Section 3.4). The adjusted grids of the meteorological parameters were those that were used in the 

WetSpass modelling of the central (wet) greenhouse scenarios 2050-2b and 2100-2b. These 

greenhouse scenarios were based on the central estimate of the temperature increase in 2050 and 2100 

and on the assumption that the average wind speed would not change (see Table 6, Table 7, and Table 

8 in Section 4.2). An overview of the modelled scenarios is given in Table 18. 

43


Table 18. Overview of modelled combined climate and land use scenarios 

Modelled combined climate and land use scenarios* 

2050-A2b 2050-B2b 2050-C2b 2050-D2b 2050-E2b 

2100-A2b 2100-B2b 2100-C2b 2100-D2b 2100-E2b 

* The full code for a combined climate and land use scenario comprises the year in 

concern, the land use scenario code (see Table 13), the climate scenario code (see 

Table 6), and the wind subcode (see also Table 6). For instance, 2050-A2b is the 

code for the scenario in which the land use would be changed in 2050 according to 

land use scenario A, and the climate according to central wet greenhouse scenario 

2 without a change in the average wind speed (subcode b). 


The results of the WetSpass modelling of the combined climate and land use scenarios are presented in 

Table 19. Comparison of these results with those given in Table 11 and Table 16 (see previous 

sections), reveals that the observed trends in the modelling results for the combined scenarios are, as 

expected, generally a combination of the trends observed in the modelling results for the climate 

scenarios and the trends observed in the modelling results for the land use scenarios. For a number of 

output parameters, the combination of the climate and land use scenarios has enforced the effects of 

the climate and land use changes observed in the separate scenarios. For these parameters, the change 

with respect to the actual situation is larger (in absolute terms) than the corresponding changes in the 

corresponding climate and land use scenarios (in Table 19, these accumulative changes are printed 

bold). For the other parameters, this change is situated between the corresponding changes in the 

corresponding climate and land use scenarios. From Table 19, it can be seen that the surface runoff is 

enforced in the combined central greenhouse and agricultural, maize, or urban scenarios (2050/2100- 

C/D/E2b), particularly in summer. It can also be seen that in winter the groundwater recharge will be 

enforced in all scenarios, except the urban ones (2050/2100-E2b). In the combined central greenhouse 

and forest scenarios (2050/2100-A2b), however, the groundwater recharge will more strongly decrease 

in summer in favour of the evapotranspiration. In general, it can be concluded that the modelling 

results for the combined climate and land use scenarios can give a reasonable impression of the extent 

into which the hydrological characteristics of the Kikbeek subbasin can change due to possible climate 

and land use changes in the coming century. 

44


Table 19. Results of the WetSpass modelling for the combined climate and land use 

scenarios, expressed as changes with respect to the actual situation (average values)* 

Parameter 

Per year: 

Scenario 

2050-A2b 

% 

2050-B2b 

% 

2050-C2b 

% 

2050-D2b 

% 

2050-E2b 

% 


Evapotranspiration (Et) 3.2 0.9 0.3 –10.6 –4.1 


Surface runoff (Ro) –1.6 2.4 14.3 14.3 117.3 

In summer: 



Groundwater recharge (Re) –72.6 2.2 –9.5 179.2 –17.4 


In winter: 


Evapotranspiration (Et) –9.9 0.6 –0.1 –0.1 2.2 


Surface runoff (Ro) 6.9 3.9 7.9 7.9 123.5 

Per year: 

Scenario 2100-A2b 

% 

2100-B2b 

% 

2100-C2b 

% 

2100-D2b 

% 

2100-E2b 

% 





In summer: 



Groundwater recharge (Re) –112.9 –15.7 –28.5 161.1 –28.6 


In winter: 


Evapotranspiration (Et) –6.3 5.5 3.8 3.8 6.2 

Groundwater recharge (Re) 21.3 15.6 16.0 16.0 1.1 


* The changes that are printed bold are larger (in absolute terms) than those in the 

corresponding climate and land use scenarios. 

45



As for the climate and land use scenarios, average yearly and seasonal, fast and slow discharge 

coefficients have been calculated for the examined combined scenarios as well. An overview of them 

is presented in Table 20. The changes in the discharge coefficients with respect to those in the 

corresponding land use scenarios (see Table 17) were generally small. A major exception, however, 

were the changes in the average slow discharge coefficients in summer. These discharge coefficients 

were significantly lower in the combined scenarios than in the corresponding land use scenarios. This 

was most likely the result of the strongly decreased groundwater recharge in summer in the central 

greenhouse scenarios (see Table 11). 

A further comparison with the average discharge coefficients that were calculated for the separate 

climate and land use scenarios (see Table 12 and Table 17) reveals that a number of average discharge 

coefficients for the combined scenarios are larger or smaller than those in both the corresponding 

climate and land use scenarios (in Table 20, these discharge coefficients are printed bold). The average 

fast discharge coefficient in the combined central greenhouse and agricultural, maize, or urban 

scenarios (2050/2100-C/D/E2b), for instance, was higher than in the separate climate and land use 

scenarios, particularly in summer. The average slow discharge coefficient in winter was generally 

higher too, with the exception of the urban scenarios (2050/2100-E2b). In summer, the average slow 

discharge coefficient was, however, significantly lower in the combined central greenhouse and forest 

scenarios (2050/2100-A2b). In 2100, it would even become negative. The values of the rest of the 

average discharge coefficients for the combined scenarios are between those of the corresponding 

climate and land use scenarios. 

Table 20. Discharge coefficients for the Kikbeek subbasin 

in the combined climate and land use scenarios for 2050 and 2100 


per year 

Fast discharge coefficients a,b Slow discharge coefficients a,c 

Average 

in summer 

Average 

in winter 

Average 

per year 

Average 

in summer 

Average 

in winter 

% % % % % % 


For 2050: 

2050-A2b 6.6 6.2 6.9 34.0 1.8 66.2 

2050-B2b 6.8 6.9 6.7 35.0 6.8 63.3 

2050-C2b 7.6 8.2 7.0 34.6 6.0 63.3 

2050-D2b 7.6 8.2 7.0 40.9 18.7 63.3 

2050-E2b 14.5 14.5 14.4 30.3 5.5 55.1 

For 2100: 

2100-A2b 6.7 6.4 6.9 33.6 –0.9 66.7 

2100-B2b 6.9 7.2 6.7 35.2 5.6 63.5 

2100-C2b 7.7 8.5 7.0 34.9 4.7 63.8 

2100-D2b 7.7 8.5 7.0 41.0 17.3 63.8 

2100-E2b 14.8 15.1 14.5 30.7 4.7 55.6 

a. The average discharge coefficients that are printed bold are larger or smaller than those 

in both the corresponding climate and land use scenarios. 

b. Total surface runoff / total precipitation for the Kikbeek subbasin 

c. Total groundwater recharge / total precipitation for the Kikbeek subbasin 

46


5. Discussion 

The WetSpass modelling of the actual hydrological situation in the Kikbeek subbasin and the 

(combined) climate and land use scenarios seemed to give reasonable results. The obtained surface 

runoff and groundwater recharge data, the calculated discharge coefficients, and the observed trends in 

the scenarios seemed to be realistic and did not show any inexplicable irregularities. The assumption 

that these results are realistic is supported by one of the major characteristics of the WetSpass model. 

That is that it is based on physical and empirical relationships. In principal, no adjustable model 

parameters are involved. In practice, of course, some model parameters will have to be adjusted 

slightly to fit the model to observed situation. This adjustment (or fine tuning) has been carried out in 

the previous projects in which WetSpass was applied successfully (see the end of Section 3.2). The 

modelled river basins in these projects (e.g. the Grote Nete basin and the Valley of the Zwarte Beek) 

are situated probably not too far away from the Kikbeek subbasin. Therefore, it is expected that the 

model parameters for the Kikbeek subbasin will not be significantly different. In that case, the 

reliability of the WetSpass modelling will depend predominantly on the reliability of the input data. 

Regarding the input data, the major uncertainties can be expected in the geomorphological, the 

meteorological and groundwater table data. The quality of the first, in fact the slope grid, depends 

strongly on the quality of the DTM, which depends on the accuracy of the (digitised) elevation contour 

lines, the interval between them, and the errors made in the interpolation process. With respect to the 

used meteorological data, it will be clear that some errors must have been introduced to the WetSpass 

model. As not much meteorological data were available for the region of the Kikbeek subbasin, most 

of the required data were derived from meteorological data from the measurement station at Uccle, 

which is situated about 100 km west from the Kikbeek subbasin. Nevertheless, comparison with 

meteorological data from Beek in the Netherlands (a measurement station that is situated relatively 

nearby) revealed that the used meteorological data were probably not so bad. No information on the 

groundwater table was available, which is a limitation in the analysis. The effects of the changing 

recharge on the water table and hence on the surface runoff and evapotranspiration could therefore not 

be back coupled. It is expected that this feedback on the groundwater table would influence the 

absolute simulated values. However the relative order of the effects of the different scenarios is not 

expected to change. 

Although the results of the WetSpass modelling seemed to be realistic, and the way of modelling 

supports their reliability, the impossibility of their verification, due to the lack of hydrological data for 

the Kikbeek subbasin (i.e. hourly or daily time series of measured water levels and discharge volumes 

at the discharge point), made their reliability still questionable. Consequently, the modelling results, 

and the discharge coefficients calculated from them, should not be applied in other models or used by 

stakeholders without awareness of their possible uncertainty. Nevertheless, they can be regarded as 

suitable estimations, until they will be verified by comparison with observed data. Therefore, the 

major recommendation that comes forward from this study will be the start of regular measurements 

of water levels and discharge volumes at the discharge points (and preferably at upstream points too) 

of a number of brooks that discharge from the Belgian side into the Border Meuse. 

47


6. Conclusions 

• A good impression of the Belgian (Flemish) policies and proposals on the aspired integrated river 

and water management on the Belgian side of the Border Meuse is given by a number of overviews 

of the Belgian part in the Belgian-Dutch 'Living Border Meuse' Project. 

• Despite the lack of hydrological and meteorological data for the area, a probably reliable 

hydrological model for the Kikbeek subbasin (one of the Belgian subbasins of the Border Meuse) 

could be developed by means of WetSpass modelling. 

• From the results of the WetSpass modelling, yearly and seasonal, fast and slow discharge 

coefficients could be obtained for the entire Kikbeek subbasin, for which it was assumed that the 

fast discharge is equal to the total surface runoff, and the slow one to the total groundwater 

recharge. 

• The WetSpass modelling revealed that in summer the major part of the precipitation (86.9 %) in the 

Kikbeek subbasin evapotranspirates, whereas the evapotranspiration is significantly lower in winter 

(31.6 %). In winter, the groundwater recharge (or slow discharge) is significantly larger (61.6 %) 

than in summer (6.8 %). The surface runoff (or fast discharge) is relatively small and nearly season 

independent (6.8 % in winter and 6.9 % in summer). An interesting implication is that most part 

(82.6 %) of the total discharge of the Kikbeek subbasin originates from the precipitation in winter. 

• The WetSpass model appeared to be well applicable in the modelling of a number of (combined) 

climate and land use scenarios. The probably realistic climate scenarios were similar to the 

previously described ones that were developed at the KNMI, and the more extreme land use 

scenarios were developed at the VUB to examine the maximum effects of possible land use 

changes. 

• In the (wet) greenhouse scenarios, the discharge coefficients will not change dramatically. An 

exception is the average slow discharge coefficient in summer that will decrease significantly and 

might even become negative. 

• In contrast to the greenhouse scenarios, a strong increase of the average slow discharge coefficient 

in summer (up to 10.5 %) was seen in the climate scenarios that include a sudden change in the 

North Atlantic thermohaline circulation. As in the greenhouse scenarios, the other discharge 

coefficients will not change much. 

• The dry climate scenarios show significant decreases of the average slow discharge coefficients in 

both seasons. In summer, they will even be more negative (down to –8.9 %) than they would be in 

some of the greenhouse scenarios. The average fast discharge coefficients, on the other hand, will 

not change noticeably. 

• From the modelling results of the land use scenarios, it can be concluded that an expansion of the 

urban areas can lead to a strong increase of the average fast discharge coefficient in both seasons 

(up to 14.4 %) and to a strong decrease of the average slow discharge coefficients in winter (down 

to 54.5 %). An increase of agricultural land (and/or maize crops) will result in an increase of the 

fast and slow discharge coefficients in summer, the slow one particularly in case of an increase of 

maize crops (up to 20.1 %). An increase of deciduous forest will result, on the other hand, in a 

decrease of the fast and slow discharge coefficients in summer, but in an important increase of the 

slow discharge coefficient in winter (up to 65.8 %). Finally, more meadows will result in larger 

slow discharge coefficients, in summer and winter. 

• The changes of the discharge coefficients in the combined climate and land use scenarios with 

respect to the actual situation were generally a combination of the changes of the discharge 

48


coefficients in the separate climate and land use scenarios. For the discharge coefficients, the 

effects of the climate and land use changes could be accumulative or compensative. 

• The WetSpass model of the Kikbeek subbasin is thought to be realistic, since the WetSpass model 

is based on physical relationships and has been applied successfully in nearby river basins. A 

recommendation for improvement is coupling of the WetSpass simulation with a groundwater 

modelling in order to account for the feedback effects of the recharge on the groundwater table. 

The model reliability would increase if it could be verified. Therefore, it is recommended that the 

hydrological data that are needed for such verification would be gathered, not only for the Kikbeek 

subbasin, but also for other brook subbasins that discharge from the Belgian side into the Border 

Meuse. 

49


7. References 

References on the river and water management on the Belgian side of the river Border Meuse are 

indicated with an asterisk (*). Between square brackets is indicated where they are available. 

* Allemeersch, L. (1990): De Limburgse Grensmaas [The Limburg's Border Meuse, in Dutch], 

Leefmilieu 1990(3), p. 77-85. [at VLM and Stichting Leefmilieu, Antwerp, Belgium] 

* Allemeersch, L. (1994a): De Grensmaas: Landschapswandeling [The Border Meuse: Landscape 

walking tour, in Dutch], Provinciaal Natuurcentrum Limburg [Provincial Nature Centre Limburg], 

Hasselt, Belgium, ISBN: 90-6685-139-2. [at BBLV and INB and Stichting Leefmilieu, Antwerp, 

Belgium] 

* Allemeersch, L. (1994b): Planten op winterdijken langs de Grensmaas [Plants on the winter dikes 

along the Border Meuse, in Dutch], Likona Jaarboek 1993, p. 21-25. [at PIME] 

* Anonymous (1994): Grensoverschrijdend natuurontwikkelingsplan Grensmaas [Border-crossing 

nature-development plan for the Border Meuse, in Dutch], Lisec, Genk, Belgium. [at INB] 

* Anonymous (1995): De Grensmaas komt tot leven [The Border Meuse gets alive, in Dutch], 

Natuurbehoud 26(4), p. 18-23. [at PIME] 

* Anonymous (1996): Grensmaasproject - Stand van zaken [Border Meuse Project - State of affairs, in 

Dutch], Vlaams Parlement [Flemish Parliament], Brussels, Belgium, 1995-1996(14), p. 711-714 en 

719 (?). [at VLM] 

* Anonymous (1998): De "Verklaring van Namur" van 8 april 1998: De start van het "Actieplan 

Hoogwater Maas" [The "Declaration of Namur" of 8 April 1998: The start of the "Meuse Floods 

Action Plan", in Dutch], Water 99, p. 77-79. [at VUB] 

* Anonymous (2000): De Grensmaas [The Border Meuse, in Dutch], Aqua 1 (informatie sheet of 

Aquafin), p. 5-8. [at VMM] 

* AWZ (2000a): De Grensmaas [The Border Meuse, in Dutch], Waterspiegel 1(2), February 2000, 

AWZ. [at VUB] 

AWZ (2000b): Waterstand- en afvoergegevens in Vlaanderen; meetpuntgegevens van station 

Opoeteren [Water level and discharge data for Flanders; measurement data from station Opoeteren, 

in Dutch], website: www.lin.vlaanderen.be/awz/waterstanden, AWZ, Afdeling Waterbouwkundig 

Laboratorium en Hydrologisch Onderzoek [Hydraulics Laboratory and Hydrological Research 

Division], Hydrologisch Informatie Centrum [Hydrological Information Centre], Borgerhout, 

Belgium. 

Batelaan, O. (in prep.): Characterization of regional groundwater discharge areas, PhD thesis, VUB. 

Batelaan, O. & De Smedt, F. (2001): WetSpass: a flexible, GIS based, distributed recharge 

methodology for regional groundwater modelling, in Impact of human activity on groundwater 

dynamics, Gehrels, H., Peters, N.E., Hoehn, E., Karsten, J., Leibundgut, C., Griffioen, J., Webb, B. & 

Zaadnoordijk, W.J., editors, IAHS Publication 269, p. 11-17. Wallingford, Oxfordshire, UK. 

Batelaan, O., Wang, Z.M. & De Smedt, F. (1996): An adaptive GIS toolbox for hydrological 

modelling, in: Application of geographic information systems in hydrology and water resources 

management, Kovar, K. & Nachtnebel, H.P., editors, IAHS Publication 235, p. 3-9, Wallingford, 

Oxfordshire, UK. 

50


Batelaan, O., Asefa, T., Van Campenhout, A. & De Smedt, F. (2000a): Bepalen van de regionale 

grondwaterstroming naar een aantal kwelgebieden in het landinrichtingsproject Grote-Netegebied 

[Estimation of the regional groundwater flow to a number of groundwater discharge areas in the land 

development project Grote-Nete area, in Dutch], project report for the VLM, Division Herentals, 

VUB. 

Batelaan, O., Asefa, T., Van Rossum, P. & De Smedt, F. (2000b): Groundwater-flow modelling of 

three wetland ecosystems in river valleys in Flanders, Belgium, Conference Proceedings of the 

ERB2000 Conference, September 27-28, 2000, Ghent, Belgium. 

* Belgroma (1998): Effectenanalyse van het natuurontwikkelingsplan voor het Grensmaasgebied. 

Eindrapport [Effect analysis of the nature-development programme for the Border Meuse area. Final 

report, in Dutch], report for AMINAL, Afdeling Natuur [Nature Division], Brussels, and AWZ, 

Afdeling Maas & Albertkanaal [Meuse and Albert Canal Division], Hasselt, Belgium; Belgroma, 

Mechelen, Belgium. [at VUB] 

* Belpaire, C., De Charleroy, D., Gilson, P. & Beyens, J. (1994): Ontwikkelingsplan voor de visserij op 

de Grensmaas (Limburg) [Development programme for fishery on the Border Meuse (Limburg), in 

Dutch], IBW, Sectie Visserij [Section Fishery], Groenendaal, Belgium. [at IBW] 

* Beyens, J., Van Thuyne, G., Viaene, P. & Belpaire, C. (1996): Het visbestand in de 

grensoverschrijdende beken van het Maasbekken [Fish stocks in border-crossing brooks of the River 

Meuse basin, in Dutch], IBW, Sectie Visserij [Section Fishery], Groenendaal, Belgium. [at IBW] 

Brandsma, T. (1995): Hydrological impact of climate change: a sensitivity study for the Netherlands, 

PhD thesis, Technical University Delft. 

Calder, I.R. (1979): Do trees use more water than grass?, Water Services 1(January), p. 4. 

* Caspers, T. (2000): De Maas: een 'lieflijke' kus van het zuiden aan het noorden [The Meuse: a ‘sweet’ 

kiss of the south to the north, in Dutch], Natura 97(2), p.36-39. [at PIME] 

Chapman, T. (1999): A comparison of algorithms for stream flow recession and base flow separation, 

Hydrological Processes 13(5), p. 701-714. 

Chow, V.T., Maidment, D.R. & Mays, L.W. (1988): Applied hydrology, McGraw-Hill, New York, 

USA. 

* De Blust, G., Van Looy, K. & Vanacker, S. (1999): De oevers van de Grensmaas: Schatkamers voor 

de toekomst? [The River Meuse banks: Treasure-houses for the future?, in Dutch], Likona Jaarboek 

1998, p.42-51. [at PIME] 

* De Charleroy, D. & Belpaire, C. (1993): Visbestandsopname en natuurtechnische voorstellen voor de 

Oude Maas te Stokkem (Dilsen) [Fish stocks survey and nature-technical proposals for the Old Meuse 

at Stokkem (Dilsen), in Dutch], IBW, Sectie Visserij [Section Fishery], Groenendaal, Belgium. [at 

IBW] 

* De Coster, M. (1998): De levende Grensmaas. Ruimte voor een rivier over de grenzen heen [The 

living Border Meuse: Room for a river across the borders, in Dutch], brochure, Vlaams-Nederlandse 

Coördinatie-commissie Grensmaas [Flemish-Dutch Co-ordination Committee Border Meuse]; 

AMINAL, Afdeling Natuur, Dienst Limburg [Nature Division Limburg], and AWZ, Afdeling Maas & 

Albertkanaal [Meuse and Albert Canal Division], Hasselt, Belgium; and De Maaswerken [The Meuse 

Works], Afdeling Communicatie [Communication Division], Maastricht, The Netherlands. [at VUB] 

* De Coster, P. (1993): Overstroom: Natuurlijke overstromingsgebieden noodzakelijk [Over flow: 

Natural over-flow areas are necessary, in Dutch], Milieurama 11, p. 11-13. [at INB] 

51


* De Coster, P. (1998): Grind winnen of natuur herstellen? (Grensmaasproject) [Gravel winning or 

nature restoration? (Border Meuse Project), in Dutch], Milieurama 18(11), p. 8-10. [at BBLV and 

PIME] 

* De Jong, D. (2000): Samenwerking voor een grenzeloze natuur - Onbekendheid met elkaars 

opvattingen leidt tot moeilijkheden bij de ontwikkeling van grensoverschrijdend natuurbeleid. De 

Grensmaas als voorbeeld [Co-operation for a boundless nature – Unfamiliarity with each others 

opinions causes problems in the development of border-crossing nature management. The Border 

Meuse as an example, in Dutch], Rooilijn 6, p. 287-293. 

* De Ruiter, F. (2000): Het nieuwe struinen. Steeds meer ongereptheid langs de Grensmaas, maar ook 

elders [The new foraging. More and more unspoilt nature along the Border Meuse, and elsewhere, in 

Dutch], Natuurbehoud 31(2), p. 26-30. [at PIME] 

* De Smedt, F. & Van Vaerenbergh, W. (1991): Studie van de regionale grondwaterstroming in het 

Kempisch plateau en the Maasvallei [Study of regional groundwater-flow systems in the Campine 

plateau and the Meuse valley, in Dutch], in Het hydrologisch onderzoek in het grensgebied Luik- 

Maasbracht; Onderzoeksresultaten 1985-1990 [Hydrological research in the border region Liège- 

Maasbracht; Results 1985-1990, in Dutch], Rapporten en nota's 26, p. 71-83, Commissie voor 

Hydrologisch Onderzoek (Commission for Hydrological Research) TNO, Delft, The Netherlands. [at 

VUB] 

De Smedt, F., Liu, Y.B. & Gebremeskel, S. (2000): Hydrologic modelling on a catchment scale using 

GIS and remote sensed land use information, in Risk Analyses II, Brebbia, C.A., editor, p. 295-304, 

WIT Press, Southampton, UK, Boston, USA. 

Dingman, S.L. (1994): Physical hydrology, Macmillan, New York, USA. 

Dupriez, G.L. & Sneyers, R. (1979): De nieuwe pluviometrische kaarten van België [New 

pluviometrical maps of Belgium, in Dutch], KMI Publications, Serie A, nr 103, KMI. 

* Durwael, L. (1998): Grensmaas: Planten op het keienstrand. Het verslag van een zomers PWGonderzoek 

[Border Meuse: Plants on the boulders. Report of a summery PWG (?) research, in Dutch], 

Euglena 17(5), p. 33-36. [at PIME] 

Federer, C.A. (1979): A soil-plant-atmosphere model for transpiration and availability of soil water, 

Water Resources Research 15(3), p. 555-562. 

Freeze, R.A. (1969): The mechanism of natural ground-water recharge and discharge. 1. Onedimensional, 

vertical, unsteady, unsaturated flow above a recharging or discharging ground-water 

flow system, Water Resources Research 5(1), p. 153-171. 

Gellens-Meulenberghs, F. & Gellens, D. (1992): L’Evapotranspiration potentielle en Belgique: 

Variabilité spatiale et temporelle [Potential evapotranspiration in Belgium: Spatial and temporal 

variability, in French] KMI Publications, Serie A, nr 130, KMI. 

* Gielen, H. (2000): personal communications, AWZ, Afdeling Maas & Albertkanaal [Meuse and 

Albert Canal Division], Hasselt, Belgium. 

* Gilson, P., De Charleroy, D., Beyens, J. & Belpaire, C. (1994): Ontwikkelingsplan voor de visserij op 

de Grensmaas (Limburg) [Development programme for fishery on the Border Meuse (Limburg), in 

Dutch], IBW, Hoeilaart, Belgium. [at INB] 

Haasnoot, M., Vermulst, J.A.P.H. & Middelkoop, H. (1999): Impacts of climate change and land 

subsidence on the water systems in the Netherlands. Terrestrial areas, NRP project 952210, RIZA 

report 99.049, RIZA. 

52


* Hermy, M. & Kuijken, E. (1987): Natuurtechnische aspecten bij aanleg, versterking en beheer van 

winterdijken langs de Maas [Nature-technical aspects in construction, reinforcement and management 

of winter dikes along the Meuse, in Dutch], INB, Hasselt (nowadays Brussels), Belgium. [at INB] 

* Hermy, M. & Kuijken, E. (1989): Landschappelijke, ecologische waarden van en een natuurplan voor 

de uiterwaarden van de Maas (Limburg): Een eerste aanzet [Landscape and ecological values of and 

a nature-development plan for the Meuse floodplains (Limburg): A first initiative, in Dutch], INB, 

Hasselt (nowadays Brussels), Belgium. [at INB] 

* Heylen, J. (1995): Hoge waterstanden Grensmaas, dec 93 en jan-feb 95, gerelateerd aan vroegere 

hoge waterstanden [Floods on the Border Meuse, December 1993 and January-February 1995, in 

relation to floods in the past, in Dutch], Verslag [Report], AWZ, Afdeling Waterbouwkundig 

Laboratorium en Hydrologisch Onderzoek [Hydraulic Laboratory and Hydrological Research 

Division], Dienst Hydrologisch Onderzoek [Hydrological Research Service]. [at VUB] 

* Heylen, J. (1997a): De hoogwaters op de Grensmaas in december 1993 en 13 maanden later in 

januari-februari 1995 [The floods on the Border Meuse at the Belgian-Dutch border in December 

1993 and 13 months later in January-February 1995, in Dutch], De Aardrijkskunde 1997(4), p. 13-26. 

[at VLM] 

* Heylen, J. (1997b): De hoogwaters op de Grensmaas in december 1993 en 13 maanden later in 


1993 and 13 months later in January-February 1995, in Dutch], Infrastructuur in het Leefmilieu 

1997(2), p. 81-92. [at PIME] 

* Heylen, J. (1998): De hoogwaters op de Grensmaas in december 1993 en 13 maanden later in 


1993 and 13 months later in January-February 1995, in Dutch], Water 99, p. 67-76. [at VUB] 

* Hoet, I., Plessers, L., Cardoen, F. & Nagels, K. (1998): Integraal rivierbeheer langs de Grensmaas. 

Een overzicht van het natuurontwikkelingsplan van het Grensmaasgebied [Integrated river 

management along the Border Meuse. An Overview of the nature development programme for the 

Border Meuse area, in Dutch], Water 99, p. 61-66. [at VUB] 

Huybrechts, W., Batelaan, O., De Becker, P., Joris, I. & Van Rossum, P. (2000): Ecohydrologisch 

onderzoek waterrijke vallei-ecosystemen [Ecohydrological research of wet valley ecosystems, in 

Dutch], Report VLINA project C96/03, INB. 

Klein Tank, A.M.G. & Können, G.P. (1997): Simple temperature scenario for a gulf stream induced 

climate change, Climatic Change 37, p. 505-512. 

KNMI (2001): Maandgemiddelde temperatuur en uurgemiddelde potentiele windsnelheid te 

Maastricht/Beek (KNMI-station 380) [Monthly mean temperature and hourly mean potential wind 

speed at station Maastricht/Beek (KNMI station 380)], website KNMI (voorlichting, klimatologische 

informatie, databestanden met tijdreeksen van Nederlandse stations): http://www.knmi.nl/voorl/weer/, 

KNMI. 

Können, G.P., Fransen, W. & Mureau, R. (1997): Meteorologie ten behoeve van de ‘Vierde Nota 

Waterhuishouding (NW4) [Meteorology for the ‘Fourth Note on Water Management (NW4)’, in 

Dutch], KNMI report, KNMI {cited in Können (2001)}. 

Können, G.P. (2001): Climate scenarios IRMA and perspectives, KNMI report, KNMI. 

Kors, A.G., Claessen, F.A.M., Wesseling, J.W. & Können, G.P. (2000): Scenario’s externe krachten 

voor WB21 [Scenarios concerning external forces for WB21, in Dutch], RIZA/WL and KNMI 

publication, KNMI {cited in Können (2001)}. 

Kwadijk, J. (2000): Nederland en het water [The Netherlands and the water, in Dutch], Aarde & Mens 

4(1), p. 19-23. 

53


* Meertens, H. (2001): De Grensmaas [The Border Meuse], Baobab ! De natuur verrast je 15, p.25-27. 

[at PIME] 

* Meertens, H. & Van Den Bosch, J. (2000): Het natuurpark van de 'Drie Eggen'. Een schoolvoorbeeld 

van samenwerking over de grens [The 'Drie Eggen' nature reserve. A school example of co-operation 

across the borders, in Dutch], Natuurhistorisch Maandblad 89, p. 166-171. [at INB and VUB] 

* Ministerie van de Vlaamse Gemeenschap (1999?,a): Project "Levende Grensmaas" [The 'Living 

Border Meuse' Project, in Dutch], brochure, AMINAL, Afdeling Natuur, Dienst Limburg [Nature 

Division Limburg], Hasselt, Belgium, and AWZ, Afdeling Maas & Albertkanaal [Meuse and Albert 

Canal Division], Hasselt, Belgium. [at VUB] 

* Ministerie van de Vlaamse Gemeenschap (1999?,b): The 'Living Border Meuse' Project, brochure, 

AMINAL, Afdeling Natuur, Dienst Limburg [Nature Division Limburg], Hasselt, Belgium, and AWZ, 

Afdeling Maas & Albertkanaal [Meuse and Albert Canal Division], Hasselt, Belgium. [at VUB] 

Nagels, K. (2001): personal communications, AMINAL, Afdeling Natuur, Dienst Limburg [Nature 

Division Limburg], Hasselt, Belgium. 

* Nagels, K. & Cardoen, F. (1997): Natuurontwikkeling berekend? Een tijdsbeeld van de 

effectenanalyse [Nature development calculated? A portrait of the era of effect analyses, in Dutch], De 

Aardrijkskunde 1997(4), p. 37-45. [at VLM] 

* Nagels, K., Hoet, I. & Van Looy, K. (1999): Project Levende Grensmaas. Stand van zaken en 

beschrijving Vlaams voorkeursalternatief [Living Border Meuse Project. State of affairs and 

description of the Flemish preferential alternative, in Dutch], AMINAL, Afdeling Natuur, Dienst 

Limburg [Nature Division Limburg], Hasselt, Belgium, and AWZ, Afdeling Maas & Albertkanaal 

[Meuse and Albert Canal Division], Hasselt, Belgium. [at VUB] 

NGI (1995): scanned topographic maps of Belgium, scale 1:10,000, sheets 26/1, 26/2, 26/3, 26/5, 

26/6, 26/7, 34/1, 34/2, and 34/3, CD-ROM, NGI. 

Nonhebel, S. (1987): Water use of Dutch forests: A modelling study, in Dutch, Forest and Water 

Commission, Report 7G, The Netherlands. 

* Odou, M. (1998): Het Vlaams-Nederlandse Grensmaasproject. Toelichting en stand van zaken [The 

Flemish-Dutch Border Meuse Project. Explanation and state of affairs, in Dutch], Water 99, p. 45-49. 

[at VUB] 

* Paulissen, E. (1973): De morfologie en de Kwartairstratigrafie van de Maasvallei in Belgisch Limburg 

[The morphology and Quaternary stratigraphy of the Meuse Valley in Belgian Limburg, in Dutch], 

Paleis der Academieën [Palace of the Academies], Brussels, Belgium. [at INB] 

* Peters, B. & Van Looy, K. (1996): Nieuwe kansen voor stroomdalgraslanden in het Zuidelijk Maasdal 

[New opportunities for channel grasslands in the Southern Meuse Valley, in Dutch], Natuurhistorisch 

maandblad 85, p. 120-126. [at INB] 

* Peters, B., Van Looy, K. & Kurstjens, G. (2000): Pioniervegetatie langs grindrivieren: De Allier en de 

Grensmaas [Pioneer vegetations along gravel rivers: The rivers Allier and the Border Meuse, in 

Dutch], Natuurhistorisch Maandblad 89, p. 123-136. [at INB and VUB] 

Pilgrim, D.H. & Cordery, I. (1992): Flood runoff, in Handbook of hydrology, Maidment, D.R., editor, 

p. 9.1-9.41, McGraw-Hill, New York, USA. 

* Plevoets, B., Vanbrabant, S., Van Dyck, J. & Willems, P. (1998): Risicoanalyse van dijken langs de 

Grensmaas [Risk analysis for dikes along the Border Meuse, in Dutch], Water 99, p. 50-60. [at VUB] 

54


Poncelet, L. (1973): Climatologie van België: Jaarlijkse en maandelijkse gemiddelden [Climatology of 

Belgium: Yearly and monthly averages, in Dutch and French], in Atlas van België, sheets 12 and 13, 

De Smet, R.E., Dumont, M.E., Dussart, F., Goossens, M. & Panier, U., editors, Nationaal Comité voor 

Geografie [National Committee for Geography], Brussels, Belgium. 

Rawls, W.J., Ahuja, L.R., Brakensiek, D.L. & Shirmohammadi, A. (1992): Infiltration and soil water 

movement, in Handbook of hydrology, Maidment, D.R., editor, p. 5.1-5.51, McGraw-Hill, New York, 

USA. 

Roberts, J. (1983): Forest transpiration: A conservative hydrological process?, Journal of Hydrology 

66, p. 133-141. 

Rubin, J. (1966): Theory of rainfall uptake by soils initially drier than their field capacity and its 

applications, Water Resources Research 2(4), p. 739-749. 

Saxton, K.E., Rawls, W.J., Romberger, J.S. & Papendick, R.I. (1986): Estimating generalized soilwater 

characteristics from texture, Soil Science Society of America Journal 50(4), p. 1031-1036. 

* Schulze, F.H., Salverda, P., Geilen, N. & Janssen, S. (2000): Ecologisch acceptabele 

afvoerfluctuaties: een norm voor de Grensmaas [Ecological acceptable discharge fluctuations: a 

standard for the Border Meuse, in Dutch], H2O 33(9), p. 30-32. [at PIME] 

* Sourbron, H. (1994): Grensmaas scheidt landen en geesten [Border Meuse divides countries and 

spirits, in Dutch], Leefmilieu 1994(2), p. 49. [at Stichting Leefmilieu, Antwerp, Belgium] 

Smedema, L.K. & Rycroft, D.W. (1988): Land drainage, Batsford, London, UK. 

Sneyers, R. & Vandiepenbeeck, M. (1995), Notice sur le climate de la Belgique [Notice on the Belgian 

climate, in French], Publication Scientific et Technique 002, 62, KMI. 

Stewart, J.B. (1988): Modelling surface conductance of pine forest, Agricultural and Forest 

Meteorology 43, p. 19-35. 

* Stubbe, L. (1985): Uiterwaarden van de Maas in Limburg [Meuse floodplains in Limburg, in Dutch], 

Natuurreservaten 7, p. 121-127. [at INB] 

* Toebat, J., Lantmeesters, K., Hoet, I. & Gielen, H. (2000): Het Vlaamse project "Levende Grensmaas" 

[The Flemish 'Living Border Meuse' Project, in Dutch], Natuurhistorisch Maandblad 89, p. 160-163. 

[at INB and VUB] 

* Tombeur, H. (1995): De volkenrechtelijke aspecten van de exclusieve verdragen inzake de 

bescherming van Maas en Schelde [International-law aspects of exclusive treaties concerning the 

protection of the Rivers Meuse and Scheldt, in Dutch], Water 85, p. 237-241. 

USDA (1951): Soil survey manual, Handbook 18, U.S. Department of Agriculture, Soil Survey Staff, 

USA. 

* Vanacker, S. (2000): Grindbanken: Soortenrijker dan je denkt? [Gravel banks: Richer in species than 

expected?, in Dutch], Natuurhistorisch Maandblad 89, p. 149-154. [at INB and VUB] 

* Vanacker, S. & Van Looy, K. (2000): personal communications, INB. 

* Vanacker, S., Van Looy, K. & De Blust, G. (1998): Typologie en habitatmodellering van de oevers 

van de Grensmaas [Classification and habitat modelling of the River Meuse banks, in Dutch], Report 

IN 98.04, INB. [at INB] 

* Vanacker, S., Van Looy, K. & De Blust, G. (1999): Typologie van de oevers langs de Grensmaas 

[Classification of the River Meuse banks, in Dutch], Nieuwsbrief @wel 1 – Water, p. 3 ff. 

(http://www.wel.be). [at VUB] 

Van der Beken, A. & Huybrechts, W. (1990): The water balance of the Flemish Community, in Dutch, 

Water 50, p. 88-92. 

55


Vandewiele, G.L., Xu, C.Y. & Ni-Lar-Win (1991): Methodology for constructing monthly water 

balance models on basin scale, VUB-Hydrologie 20, VUB. 

* Van Hoof, F. (2000): Jaarverslag 1999 - Maas, Albertkanaal en Kempische Kanalen [Annual report 

1999 - Meuse, Albertkanaal and Kempish canals, in Dutch], Antwerpse Waterwerken (AWW), 

Departement Waterkwaliteit, Antwerpen, Belgium. [at VMM] 

* Van Looy, K. (1996): Grensmaas wordt opnieuw een levende rivier: Grensoverschrijdende aanpak 

voor de Grensmaas [The Border Meuse will be a living river again: Border-crossing policy for the 

Border Meuse, in Dutch], Natuurreservaten 18(5), p. 10-13. [at PIME and VLM] 

* Van Looy, K. (1997a): Natuurontwikkelingsmogelijkheden aan de Vlaamse zijde van de Grensmaas 

[Possibilities for nature development at the Flemish side of the Border Meuse, in Dutch], De 

Aardrijkskunde 1997(4), p. 27-36. [at VLM] 

* Van Looy, K. (1997b): Natuurontwikkelingsmogelijkheden aan Vlaamse zijde van de Grensmaas 

[Possibilities for nature development at the Flemish side of the Border Meuse, in Dutch], Euglena 

16(5), p. 17-21. [at PIME] 

* Van Looy, K. (1999): Het watersysteem Grensmaas. Een analyse van de oppervlaktewateren in de 

Maasvallei [The Border Meuse water system. An analysis of surface waters in the Meuse valley, in 

Dutch], Report IN 99.17, INB. [at INB] 

* Van Looy, K. & Bijlsma, R.J. (2000): Structuur en ontwikkeling van het ooibos in Hochter Bampd 

[Structure and development of the floodplain forest of Hochter Bampd, in Dutch], Natuurhistorisch 

Maandblad 89, p. 143-148. [at INB and VUB] 

* Van Looy, K. & De Blust, G. (1995a): De Maas natuurlijk?! Aanzet tot een grootschalig 

natuurontwikkelingsproject in de Grensmaasvallei [The Meuse natural?! Initiative for a large-scale 

nature development project in the Border Meuse valley, in Dutch], Wetenschappelijke Mededeling van 

het INB (Scientific Reports of the INB) 1995(2), INB. [at INB and VUB] 

* Van Looy, K. & De Blust, G. (1995b): Ruimte voor de grindrivier. Natuurontwikkeling langs de 

Grensmaas [Room for a gravel river. Nature development along the Border Meuse, in Dutch], 

Infrastructuur in het Leefmilieu 1995(5), p. 287-293. [at INB] 

* Van Looy, K. & De Blust, G. (1998): Ecotopenstelsel Grensmaas: Een ecotopenindeling, 

referentiebeschrijving en vegetatietypering voor de Levende Grensmaas [The Border Meuse ecotope 

system: A classification of ecotopes, description of references, and vegetation characterization for the 

living Border Meuse, in Dutch], Report IN 98.25, INB. [at INB] 

* Van Looy, K. & De Blust, G. (1999a): Integraal monitoringmeetnet Grensmaas. Aanzet tot 

geïntegreerde monitoring vanuit bestaande meetnetten [Integrated monitoring network for the Border 

Meuse. Initiative for an integrated monitoring based on existing monitoring networks, in Dutch], 

Report IN 99.07, INB. [at INB] 

* Van Looy, K. & De Blust, G. (1999b): Stroomdalgraslanden op de Maasdijken. Een beheersvisie voor 

het Maasdijkenplan [Channel grasslands at the Meuse dikes. A management policy for the Meuse 

Dikes Plan, in Dutch], Report IN 99.11, INB. [at INB and VUB] 

* Van Looy, K. & Peters, B. (2000): Bosontwikkeling en morfodynamiek langs de Grensmaas [Riparian 

forest development and morphodynamics along the Border Meuse, in Dutch], Natuurhistorisch 

Maandblad 89, p. 137-142. [at INB and VUB] 

* Van Looy, K. & Van Ghelue, P. (1997): Wandelexcursie in de meander van Stokkem-Dilsen [Walking 

tour in the meander of Stokkem-Dilsen, in Dutch], De Aardrijkskunde 1997(4), p. 55-63. [at VLM] 

56


* Verpraet, L. (1998): Grensmaasplan. Natuurproject in stroomversnelling [Border Meuse Project. 

Nature development project in accelaration, in Dutch], N & S (Natuur- en Stedeschoon) 67(5), p. 18- 

20. [at PIME] 

* Viaene, P., Ollevier, F. & Hermy, M. (1996): Visserijbiologische evaluatie van enkele zijbeken van de 

Grensmaas [Fishery biological evaluation of some tributaries of the Border Meuse, in Dutch], thesis, 

KU Leuven, Faculteit Landbouwkundige en Toegepaste Biologische Wetenschappen [Faculty of 

Agricultural and Applied Biological Sciences]. [at IBW] 

VLM (1998a): Bodemgebruik, kleinschalig bodemgebruiksbestand van Vlaanderen en Brussel 

afgeleid uit Landsat TM satellietbeelden [Land use, small scale land use map of Flanders and 

Brussels, derived from Landsat TM satellite images, in Dutch], CD-ROM, Ondersteunend Centrum 

GIS Vlaanderen [Support Centre GIS Flanders], VLM. 

VLM (1998b): Bodemkaart van Vlaanderen en Brussel [Soil map of Flanders and Brussels, in Dutch], 

CD-ROM, Ondersteunend Centrum GIS Vlaanderen [Support Centre GIS Flanders], VLM. 

* VMM (2000): Ontwerp-Investeringsprogramma 2002 - 2005 (2006) voor de Bovengemeentelijke 

Zuiveringsinfrastructuur voor het Bekken van de Maas [Concept investment programme 2002 - 2005 

(2006) for the supercommunal water purification infrastructure of the Meuse basin, in Dutch], VMM, 

Afdeling Planning. [at VMM] 

Wang, Z.M., Batelaan, O. & De Smedt, F. (1997): A distributed model for water and energy transfer 

between soil, plants and atmosphere (WetSpa), Physical Chemistry of the Earth 21(3), p. 189-193. 

57


Appendix. Hydrological atlas for the Kikbeek subbasin 

Map 1. Location of the Kikbeek subbasin 

Map 2. Contour lines of the elevation in the wider region of the Kikbeek subbasin 

Map 3. Digital topography model (DTM) of the wider region of the Kikbeek subbasin 

Map 4. Calculated stream net and outline of the Kikbeek subbasin 

Map 5. Slope map of the Kikbeek subbasin 

Map 6. Soil map of the Kikbeek subbasin 

Map 7. Land use in the Kikbeek subbasin (actual situation) 

Map 8. Land use in the Kikbeek subbasin in scenario A 

Map 9. Land use in the Kikbeek subbasin in scenario B 

Map 10. Land use in the Kikbeek subbasin in scenario C 

Map 11. Land use in the Kikbeek subbasin in scenario D 

Map 12. Land use in the Kikbeek subbasin in scenario E 

Map 13. Land use in the Kikbeek subbasin in scenario F 

Map 14. Land use in the Kikbeek subbasin in scenario G 

Map 15. Land use in the Kikbeek subbasin in scenario H 

Map 16. Land use in the Kikbeek subbasin in scenario K 

Map 17. Precipitation per year in the Kikbeek subbasin (actual situation) 

Map 18. Groundwater recharge per year in the Kikbeek subbasin (actual situation) 

Map 19. Groundwater recharge in summer in the Kikbeek subbasin (actual situation) 

Map 20. Groundwater recharge in winter in the Kikbeek subbasin (actual situation) 

Map 21. Surface runoff per year in the Kikbeek subbasin (actual situation) 

Map 22. Surface runoff in summer in the Kikbeek subbasin (actual situation) 

Map 23. Surface runoff in winter in the Kikbeek subbasin (actual situation) 

Map 24. Evapotranspiration per year in the Kikbeek subbasin (actual situation) 

Map 25. Evapotranspiration in summer in the Kikbeek subbasin (actual situation) 

Map 26. Evapotranspiration in winter in the Kikbeek subbasin (actual situation) 

Map 27. Groundwater recharge per year in the Kikbeek subbasin in scenario 2100-2b 

Map 28. Groundwater recharge in summer in the Kikbeek subbasin in scenario 2100-2b 

Map 29. Groundwater recharge in winter in the Kikbeek subbasin in scenario 2100-2b 

Map 30. Surface runoff per year in the Kikbeek subbasin in scenario 2100-2b 

Map 31. Surface runoff in summer in the Kikbeek subbasin in scenario 2100-2b 

Map 32. Surface runoff in winter in the Kikbeek subbasin in scenario 2100-2b 

58

VUB final report IRMA-SPONGE 2 - Homepages Vub Ac - Vrije ...

Create successful ePaper yourself

Delete template?

Save as template?