Weak evidence of long-term extinction debt in Pannonian dry sand grasslands

Habitat loss is one of the major drivers of the reduction in biological diversity worldwide. European dry grasslands are particularly endangered. However, the persistence of populations can temporarily, mitigate species loss - a process referred to as 'extinction debt'. We test this hypothesis using historical and present day habitat maps and current plant biodiversity data collected in the forest-steppe zone of Europe. In 16 5 km x 5 km study sites, representing the landscape heterogeneity of the Kiskunsag region (Hungary), 86 20 m x 20 m vegetation plots were surveyed in open and closed calcareous sand grasslands. Grassland diversity was measured as the number of specialist species, defined by statistical fidelity measures using primary and secondary grassland plots. Landscape context was quantified using the areal extent of semi-natural forest-steppe vegetation in a 300 m neighborhood of the plots, based on recent and historical maps (1783, 1860, 1950s, 1987-1989 and 2005). The number of specialist species was estimated with Poisson generalized linear models using the present landscape context, climatic conditions, and a proxy of soil type as covariates. To test for the effect of historical legacies, Pearson residuals from the present models were tested for significant relationships between the residuals and the historical landscape contexts using linear models. We found that the present landscape context had no significant relationship with the specialist species richness of the primary grassland fragments. However, we found a significant relationship between the historical landscape context of the 19th century and the residuals of the present model. Even though the extent of natural vegetation in the 20th century showed more drastic changes, the landscape context in 1950s and 1987-1989 exhibited no significant statistical relationship with the residuals. This delay of species loss is consistent with the extinction debt hypothesis. (C) 2013 Elsevier B.V. All rights reserved.