The study of local and regional mechanisms driving spatial patterns in beta diversity is essential to the understanding of biodiversity. We aimed to predict the roles of multiple mechanisms operating at regional to fine spatial scales in structuring the beta diversity in a mountainous semi-arid rangeland of the Khorassan-Kopet Dagh floristic province located in NE Iran. We evaluated the relative contributions of three main filters, at the regional (stochastic but spatially structured dispersal filter), local (abiotic filter), and fine (biotic filter) spatial scales on beta diversity across communities. We partitioned beta diversity constrained by spatial, environmental, and biotic variables in 23 communities studied here, and used simple RDA and RDA-based variation partitioning to assess the contributions of studied filters on beta diversity. Moreover, spatial autocorrelation analyses were used to test neutral theory predictions. The relative contributions of the studied ecological filters explained 55% of variation in beta diversity. Although differences in the explained variations between unique fractions are low, the abiotic and biotic filters (signifying the niche-based processes) represent stronger effects directly and indirectly (via impact on significantly of other ecological filters) than dispersal (signifying neutral processes) on beta diversity. In addition, 45% of the variation in beta diversity was not explained by the studied ecological filters. In conclusion, independent and shared impacts of processes at different spatial scales determine beta diversity in our plant communities. However, unexplained variation in beta diversity requires further study of other facets of biodiversity and community assembly processes.

Keywords