Using habitat extent and composition to predict the occurrence of woodland birds in fragmented landscapes

摘要: The removal, alteration and fragmentation of habitat are key causes biodiversity decline worldwide. In Australia, temperate woodlands have been disproportionately cleared following European settlement. Biodiversity in such systems may be reversed by restoration native vegetation on agricultural land. However, rebuilding functioning will require understanding the determinants species distributions existing habitat. We used logistic regression bird occurrence data from 240 sites across northern Victoria, to determine probability 29 woodland-dependent species. modelled as a function variables that characterise both extent (amount) composition surrounding sites. Our specific goal was whether predictive performance models is improved accounting for compared with alone. For nearly all species, addition weighting distance explanatory power models, explaining average 5.4 % (range 0–27.6 %) residual uncertainty accounted Models incorporate variation can not only provide more accurate predictions occurrence, but also guide appropriate restoration. They highlight need fertile soils support productive types. These woodland birds inform spatially-explicit optimisation model restoring cover land this region, achieving gains while minimizing loss production.