Effects of landscape structure on multiple dimensions of bat biodiversity

Background/Question/Methods

Scientific understanding of the spatiotemporal dynamics of biodiversity is primarily based on considerations of the taxonomic dimension. Such approaches are insensitive to ecological or evolutionary differences among species, which may play dominant roles during species assembly into local communities along environmental gradients. Few studies have simultaneously assessed taxonomic, functional, and phylogenetic dimensions of biodiversity and even fewer studies have done so along disturbance gradients. We assessed taxonomic biodiversity (Simpson’s diversity), as well as phylogenetic and functional biodiversity (Rao’s quadratic entropy) for phyllostomid bats — important seed dispersal and pollination agents in the tropics— along an anthropogenic disturbance gradient represented by 15 sites in northeastern Costa Rica. Wet and dry seasons were assessed separately, as resources available to bats are season specific. Functional diversity was quantified using 10 ecological (e.g., diet, foraging location and strategy) and 10 ecomorphological (e.g., size, cranial dimensions, and wing dimensions) traits. At each site, four compositional (metrics sensitive to proportions of land-cover) and four configurational (metrics sensitive to spatial arrangement of land-cover) characteristics of landscape structure were quantified at each of three spatial scales (1, 3, 5 km radius circles). Hierarchical partitioning identified the best explanatory landscape characteristics for each dimension of biodiversity at each spatial scale.

Results/Conclusions

The landscape characteristics that accounted for significant variation in biodiversity were dependent on dimension of biodiversity and season. During the wet season, landscape composition most explained variation in functional diversity (i.e., forest area at 3 km scale) and Simpson’s diversity (i.e., forest area and patch density at 3 km scale), whereas, phylogenetic diversity was associated with configurational characteristics (i.e., shape index at 3 and 5 km scales). During the dry season, functional and phylogenetic diversities were associated with landscape composition (i.e., patch density at 1 and 5 km scales, and forest area at 1 and 3 km scales), whereas Simpson’s diversity was associated with landscape configuration (i.e., nearest neighbor distance at 3 and 5 km scales). Differences between dimensions demonstrate the complex ways in which biodiversity responds to anthropogenic disturbance and that one dimension of biodiversity may not be a good surrogate for another. Based on taxonomic diversity, forest area most affects biodiversity when resources were abundant, whereas distance between forest patches was more important when resources were rare. Incorporation of ecological and evolutionary variation in biodiversity assessment suggests that conservation of greater numbers of compact forest patches may conserve more ecological and evolutionary variation, potentially preserving ecosystem services.