A land similarity approach to modeling complex ecological networks

In response to the limited success of isolated nature protection areas as a way to protect endangered habitats and species as well as biodiversity, there has been a policy shift toward the implementation of ecological networks, in Europe and in the rest of the world. Network behavior can be an emerging feature due to the natural configuration of habitat patches, or it can be induced by the establishment of green infrastructure. Graph models of ecological networks have become a field of application of complex network analysis, so that researchers could improve the understanding of the network behavior in wide areas. The most common approach, however, is still that of choosing a single species, whether to be representative of a number of other species or because it is particularly important to conserve, and study the network behavior with respect to that species. This paper proposes an approach based on the analysis of land use data and the computation of similarity scores between sites, as a way to provide a high-level understanding of features of sites that are part of the European "Natura 2000" ecological network.