Landscapes Forest and Global Change - ESA - Escola Superior ...

C. Carvalho-Santos 2010. Fine-scale mapping of High Nature Value farmlands
184
3.1. Problems with existing methodologies
Ecological, historical and cultural differences in farming landscapes among countries require
region-specific rules to identify HNVf. This paper addresses this problematic and presets a new
methodology to map HNVf at a local level, considering the importance of this identification to
the improvement of rural natural and economic environment.
The standard procedures for mapping HNVf in Europe include the use of land use data (CLC -
Corine Land Cover), with classes based on an Environmental Stratification (Metzger et al.
2005)When available, the methodology also suggests the use of complementary information on
farming practices, altitude and latitude, soil quality, climatic condition, steepness of slope at
national level to improve the cartography (Paracchini et al. 2006). However, the resulting maps
cannot be used to draw conclusions on the presence of HNV farmland at the local level, but
only at the regional level (Paracchini et al. 2006). In fact, scale is very important when trying to
map HNVf, because different agro-ecological processes operate at different scales that must be
taken into account.
For HNVf identification at the local scale the application of a downscaling exercise using a
bigger scale land use map seemed to be a good option. In Portugal, local-scale land cover/use
analysis is based on the COS products (Portuguese land cover map, 1:25.000), obtained from
the interpretation of aerial photos. However, there is no satisfactory direct relationship between
the CLC and COS classifications (table 1), and an HNVf map based on the application of the
regional-scale methodology to a COS map would exhibit more than twice the extent of HNVf
area than a map obtained using the CLC dataset. So, differences in land cover classifications in
maps with different scales may result in very different maps for a same area. Overall, this
implies that the methodology for identifying and mapping HNVf should be revised in order to
adapt it to multi-scalar exercises.
3.2. Local scale HNVf mapping – proposal of a new methodology
In order to best consider those areas that could be excluded when applying the CLC
methodology, a new refined methodology has been developed to identify HNVf at local scales.
The national land cover dataset (COS) is the central dataset in this novel methodology.
The first step is to define the “total farmland area”, considering not only the pure agricultural
and agro-forestry areas, but also small patches of neighbouring forest and semi-natural areas
spatially and functionally linked with croplands. We defined the maximum areas of 5ha for
forests and 1ha for semi-natural areas. Herewith, we are placing the farmland not as fragments
with restricted boundaries, but in its context as a continuous place where biodiversity circulates
among habitats.
Even if the methodology considers two different levels of analysis, the patch level and the civil
parish level, the final HNVf map should be presented at the less detailed scale (the parish level),
in order not to lose information in the transition among scales. Moreover, as a landscape
concept, HNVf should not be mapped directly at the individual patch of COS, but using context
attributes such as landscape metrics and other features of the farming landscape and of the
territory itself (Figure 1). Landscape attributes include those related to landscape composition
and to landscape structure. Available data on farming (from agrarian censuses) were also added
at the parish level, to identify the importance of primary sector of activity in each parish. Finally,
natural value was taken into account, using available data on regional biodiversity and
ecosystems from a previous project (FCUP 2009).
Mean values were calculated for all parameters and for each parish, based on “total farmland
area”. To isolate any surrogacy among variables a correlation analysis (e.g. using the Spearman
index) should be carried out. Finally, a global HNV score is obtained for each patch or parish by
reclassifying the selected parameters into five classes using equal breaks, and then by averaging
their values. The final scale ranges from 1 (low nature value farmland) to 5 (high nature value
Forest Landscapes and Global Change-New Frontiers in Management, Conservation and Restoration. Proceedings of the IUFRO Landscape Ecology
Working Group International Conference, September 21-27, 2010, Bragança, Portugal. J.C. Azevedo, M. Feliciano, J. Castro & M.A. Pinto (eds.)
2010, Instituto Politécnico de Bragança, Bragança, Portugal.