Print PageThe future of tropical forest species depends in part on their ability to survive in human-modified landscapes. Theory suggests that species with particular traits may be greater risk of extinction as a consequence of habitat modification than others. While the search for mechanistic links between species traits and extinction risk have greatly expanded in recent years, the link between traits and persistence in human modified habitats remains poorly characterized for many ecologically important taxa, including those that provide key ecosystem functions. We tested the relationships between species traits and abundance response to modification of lowland tropical forest for 628 Scarabaeinae dung beetles species, a diverse and functionally important insect group. Using a global dataset compiled of 24 independent studies, we modeled the relationship between beetle abundance response to forest modification and three species traits posited to relate to beetle response to abiotic changes in modified forest (mean species body mass, diel activity and nesting strategy). We defined beetle abundance response as the log-ratio of per-trap abundance in intact forest to abundance in modified habitat, using pairs of ‘intact’ primary forest, and 1-3 modified forest types reported in each study. We used linear mixed models with and without the addition of genus as a covariate, to partially control for phylogenetic autocorrelation across the dataset.
Results/Conclusions
We found that body size and nesting strategy were significantly related to extinction risk, with larger species and those with a tunneling nesting strategy significantly more likely to decline after habitat modification. In general, fewer species and individuals persisted into the most modified areas. Diel activity was unrelated to the effects of habitat modification. Models with and without phylogenetic correction supported the same relationship between dung beetle species traits and abundance response in modified habitat. We discuss the implications of these results in the context of predictions of future persistence of biodiversity and functional values in human managed landscapes, and highlight how the current limitations of trait information for many mobile animal taxa constrains strong inference in the search of trait-based correlates of extinction proneness.
