Conectividad ecológica entre áreas transfronterizas: aplicación de parámetros del paisaje en el cuadrante noroccidental de la Península Ibérica (Zamora, León, Ourense y Bragança)

The north-eastern border of Portugal with Spain is a high biodiversity region, containing many protected areas but is also strongly affected by the construction of large communication infrastructures (high-speed road and rail) and (hydro and wind) power stations. The nature conservation strategies are not coordinated or coincident between the two countries. Atlas of species occurrence based on 10x10 km UTM grid is the only available information on the diversity of species, particularly those with greater value. Also cartographic recognition of changes in land use is not coincident for the two countries, and other systems at European level such as the CORINE LAND COVER, lacks detail or are not suitable to deal the processes time scales. This work presents the first results on developing a methodology for modelling habitat distribution of the most important Iberian fauna values – Iberian wolf (Canis lupus signatus), Eurasian otter (Lutra lutra), Pyrenean desman (Galemys pyrenaicus) and Iberian emerald lizard (Lacerta schreiberi) – relating data from 10 x 10 km atlas grids with land-cover patterns derived from a remote sensing unsupervised classification. Categorical information resulting from the classification of satellite imagery was quantified spatially by the prior establishment of a radius of moving window for evaluation of parameters or landscape metrics. Land cover patterns are described by class proportion, patch dimension and complexity, edge density, adjacencies and nearby classes. Relative Operating Characteristic (ROC) was used to select the pattern descriptors used as indicators to model the habitat of each species by logistic regression. Models based only on landscape composition presented a higher predictive value than the models that also include metrics inherent to the configuration, association and diversity of the landscape. Potential habitat models with greater predictive value established concern cases in more detail, i.e. larger number of classes and smaller area of minimum patch. The established models identify regions needed of actions and measures to prevent, minimize and compensate the disturbance inherent in the construction of infrastructures, including the highways A4 (Trásos- Montes - Alto Douro) and A-52 (Rias Baixas highway), the Olmedo-Zamora-Santiago de Compostela High Speed Railway corridor, in its passage along the Serra de Nogueira, Serra de Bornes, Serra do Canizo, Sierra de la Gamoneda and the valleys of the rivers Manzanas, Tuela and Tera. Habitat models show how strongly conditioned are those four species by human disturbance and they identify the most sensitive sites needing additional protection actions and measures.