Network analysis to assess landscape connectivity trends: Application to European forests (1990-2000)

摘要

Landscape networks and ecosystems worldwide are undergoing changes that may impact in different ways relevant ecological processes such as gene flow, pollination, or wildlife dispersal. A myriad of indices have been developed to characterize landscape patterns, but not all of them are equally suited to evaluate temporal changes in landscape connectivity as is increasingly needed for biodiversity monitoring and operational indicator delivery. Relevant advancements in this direction have been recently proposed based on graph theoretical methods to analyze landscape network connectivity and on the measurement of habitat availability at the landscape scale. Building from these developments, we modify a recent index and present the equivalent connected area (ECA) index, defined as the size of a single patch (maximally connected) that would provide the same probability of connectivity than the actual habitat pattern in the landscape. The temporal changes in ECA can be directly compared with the changes in total habitat area. This allows for additional and straightforward insights on the degree to which the gains or losses in habitat amount can be beneficial or deleterious by affecting landscape elements that uphold connectivity in a wider landscape context. We provide a demonstrative example of application and interpretation of this index and approach to monitor changes in functional landscape connectivity. We focus on the trends in European forests at the province level in the period 1990-2000 from Corine land cover data, considering both changes in the forest spatial pattern and in the average permeability of the landscape matrix. The degree of connectivity was rather stable over most of the study area, with a slight overall increase in forest connectivity in Europe. However, a few countries and regions concentrated remarkably high changes in the analyzed period, particularly those with a low forest cover. The species traits also affected the responses to landscape pattern changes, which were more prominent for those species with limited dispersal abilities. We conclude discussing the potential of this approach for consistent indicator delivery, as well as the limitations and possibilities of application to a variety of situations, for which the required quantitative tools are freely available as open source projects.