**5. Different approaches on the habitat identification through RS in the context of Europe**

The lack of a simple and direct relationship between habitats and any biophysical feature detected by RS restricts the possibilities for automated image classification processes to habitat identification. In this sense, the current wide range of remote sensing techniques and products have supported many suggestions at different scales and using different approaches. The rationale underlying for all of them is the idea of selecting key variables and algorithms to the identification of the habitat entity, integrating knowledge from ancillary data sources. Some of these approaches are mentioned and briefly described in the next paragraphs. Also we propose a new methodology (based on a previous model proposed in Martinez et al., 2010(Martínez et al. 2010)) which presents some key concepts to be consider in a future standardized process.

#### **5.1 Decision rules implemented through a Geographical Information System (GIS): the example of the European PEENHAB project (Mücher et al. 2004; Mücher et al. 2009)**

The overall objective of the European PEENHAB project was to develop a methodology to identify spatially all major habitats in Europe according to the Annex I of the Habitats Directive (231 habitats, (EuropeanCommission 2007). This should result in a European Habitat Map with a spatial scale of 1: 2,5M and a minimum mapping unit of 100km2 with a minimum width of 2,5km. It was expected that this European Habitat Map was the main data layer in the design of the Pan-European Ecological Network (PEEN), which is widely recognized as an important policy initiative in support of protected *Natura 2000* sites.

PEENHAB proposed a new methodology to allow the spatial identification of individual habitats to European scale, based on specific expert knowledge and the design of decision rules on the basis of their description in Annex I. Habitats were identified by a combination of spatial data layers implemented in a GIS decision rule. The methodology was implemented following five steps: i) the selection of appropriate spatial data sets; ii) the definition of knowledge rules using the descriptions of Annex I habitats; iii) the use of additional ecological expert knowledge; iv) the implementation of the models for the individual habitats; v) validation (Mücher et al. 2009).

The spatial datasets used as ancillary data were: CORINE land cover database, biogeographic regions, distribution maps of individual plant species, digital elevation models, soil databases and other geographic and topographic data.

It is relevant that at Spain level CLC map shows important inconsistencies with bogs and mires (7110\*, 7120, 7130, 7230), water courses (3260), alluvial forests (91E0\*) or coastal

Total, very high and high gaps should be considered as important inconsistencies which enhance the limited capacity of the CLC map for representing natural and semi-natural habitats and reveal the inappropriate use of the CLC map as a biodiversity proxy, both at European and regional level. In some cases gaps can be explained because of the CLC methodology which makes not possible to identify habitats with less than 25ha or linear features below 100m in width. Also, discrepancies among countries could be attributed to

Then, though theoretically possible (Groom et al. 2006; Hansen et al. 2004), the use of some components of the complex habitat entity, such land covers, as a surrogate parameter of a

The lack of a simple and direct relationship between habitats and any biophysical feature detected by RS restricts the possibilities for automated image classification processes to habitat identification. In this sense, the current wide range of remote sensing techniques and products have supported many suggestions at different scales and using different approaches. The rationale underlying for all of them is the idea of selecting key variables and algorithms to the identification of the habitat entity, integrating knowledge from ancillary data sources. Some of these approaches are mentioned and briefly described in the next paragraphs. Also we propose a new methodology (based on a previous model proposed in Martinez et al., 2010(Martínez et al. 2010)) which presents some key concepts to

**5.1 Decision rules implemented through a Geographical Information System (GIS): the example of the European PEENHAB project (Mücher et al. 2004; Mücher et al. 2009)**  The overall objective of the European PEENHAB project was to develop a methodology to identify spatially all major habitats in Europe according to the Annex I of the Habitats Directive (231 habitats, (EuropeanCommission 2007). This should result in a European Habitat Map with a spatial scale of 1: 2,5M and a minimum mapping unit of 100km2 with a minimum width of 2,5km. It was expected that this European Habitat Map was the main data layer in the design of the Pan-European Ecological Network (PEEN), which is widely recognized as an important policy initiative in support of protected *Natura 2000* sites. PEENHAB proposed a new methodology to allow the spatial identification of individual habitats to European scale, based on specific expert knowledge and the design of decision rules on the basis of their description in Annex I. Habitats were identified by a combination of spatial data layers implemented in a GIS decision rule. The methodology was implemented following five steps: i) the selection of appropriate spatial data sets; ii) the definition of knowledge rules using the descriptions of Annex I habitats; iii) the use of additional ecological expert knowledge; iv) the implementation of the models for the

The spatial datasets used as ancillary data were: CORINE land cover database, biogeographic regions, distribution maps of individual plant species, digital elevation

**5. Different approaches on the habitat identification through RS in the** 

differences among the skills and expert knowledge of image interpreters.

particular habitat is uncertain and it should be previously evaluated.

lagoons (1150\*).

**context of Europe** 

be consider in a future standardized process.

individual habitats; v) validation (Mücher et al. 2009).

models, soil databases and other geographic and topographic data.

For example, for the Annex I habitat "Calcareous Beech Forest (code 9150)", first a rule was defined that selects the broadleaf forests from the CORINE land cover database, then a second rule was used to select the beech distribution map from the Atlas Florae Europaeae, and a third rule identified the calcareous soils from the European soil database. The combination of these three filters will form the decision rule that delimitates the spatial extension of calcareous beech forest.

The main advantage of this approach is the suggestion of using specific knowledge, implemented as a GIS decision rule, to identify individual habitat maps as they are described in Annex I of Habitats Directive. The approach use remote sensing data in an indirect way (through the use of CORINE land cover and other input variables) along with other suitable ancillary data. Results are appropriate at European scale in order to set guidelines for the strategic design of the Pan-European Ecological Network.
