Habitat fragmentation: From spatially explicit models to the analysis of patterns in real ecosystems

• Jordi Bascompte

Abstract
Destruction and fragmentation of habitats are a major threat to biological diversity. A theoretical framework aimed at understanding and predicting communities' responses to these destructive processes is still lacking, however. In this proposal I describe an integrative approach aimed at gaining better insight on the response of ecosystems to habitat destruction. It is based on an interaction between theoretical models and the analysis of patterns in real ecosystems using data bases. Spatially explicit models can be derived to account for the role of spatial structure on metapopulation abundance and species diversity. There are critical thresholds of destruction, derived from percolation theory, at which landscape structural properties and the subsequent metapopulation regional abundance change in a non-linear way. This non-linearity has important implications for conservation biology and landscape management. Some of these predictions can be tested using data bases. Information provided by the models can help to explain the patterns observed in real ecosystems. In turn such data is expect to improve the models.