Tropical forest beta diversity across multiple taxa and scales

Background/Question/Methods

Numerous studies of different taxa have demonstrated that tropical forests exhibit pronounced turnover in species composition (beta diversity) across a wide range of spatial scales. However, few studies have simultaneously explored the patterns and underlying processes of beta diversity of multiple co-occurring taxa. Thus, little is known about the degree to which different taxa exhibit similar or different patterns of beta diversity and whether or not deterministic processes of community assembly differ between taxa. To fill this research gap, we used a replicated nested sampling design to simultaneously sample seven taxa at six sites across Peninsular Malaysia: trees, ants, bats, dung beetles, stingless bees, bats, and birds.  For each taxon, we partitioned diversity at two spatial scales and compared relative turnover rates across taxa and scales. We also used a null model approach to identify the component of turnover expected from stochasticity alone (random establishment and sampling error). Finally, we investigated the influence of spatial distance, topography, and logging history on observed turnover patterns.

Results/Conclusions

Observed relative turnover rates were consistently higher than null model expectations; this pattern held across all taxa and both scales. We found no clear relationship between taxon-level dispersal ability and observed local or regional turnover (e.g. moths and birds exhibited greater turnover than trees and ants). However, after accounting for expected stochastic turnover, we found that the least mobile taxa (trees and ants) exhibited the highest turnover due to deterministic processes (including dispersal limitation). Cross-scale comparisons revealed that observed local turnover was generally higher than observed regional turnover, but our null model revealed that deterministic turnover was higher at the regional scale for the five most mobile taxa (but not for trees and ants). Despite clear inter-site differentiation, we found no evidence of distance decay at the regional scale for any taxon. At the local scale, compositional patterns for several taxa were associated with physical distance, topography, and/or logging history. Our findings demonstrate that directly observed turnover patterns do not necessarily align with deterministic turnover, and confirm long-standing (but largely untested) hypotheses about dispersal ability and beta diversity in tropical forests.