Contribution of environmental filtering and dispersal limitation to species turnover of temperate deciduous broad‐leaved forests in China

Abstract Questions How and to what extent do environmental filtering and dispersal limitation affect species turnover among sites in the temperate deciduous broad‐leaved forests of C hina? Do different life forms respond to environmental filtering and dispersal limitation in the same way? Location Seventeen mountains (31°8′–44°22′N, to 104°42′–128°8′E) in C hina. Methods We measured the plant species composition of 495 plots. We calculated the Jaccard's similarity ( S Jac ) of species composition, geographic distance ( D g ) and environmental divergence ( D e ) for each plot pair for trees, shrubs and herbs. Three plot pair groups were distinguished: (1) plot pairs with both plots from the same mountain; (2) plot pairs with two plots from different mountains but with a similar environment; (3) plot pairs with two plots from different mountains and with a heterogeneous environment. We analysed the relationship between S Jac , D g and D e for all three groups, and performed partial redundancy analysis. We also calculated beta‐diversity of mountains, and explored latitudinal and climatic patterns of beta‐diversity. Results S Jac decreases significantly with both D e and D g for all three plot pair groups. Environmental and geographic variables together explain 13.8%, 11.6% and 11.1% of variance in species composition of trees, shrubs and herbs, respectively. For all life forms, environment explains more variance than geography, and environment and geography have strong co‐varying effects. The beta‐diversity for each mountain decreases with latitude, but increases with mean annual temperature ( MAT ). Conclusion Environmental filtering rather than dispersal limitation predominantly regulates patterns of species turnover in the temperate deciduous broad‐leaved forests of C hina. Furthermore, the increases of beta‐diversity along MAT also support the effects of environment on species turnover. These results indicate that environment might influence the distribution of most species in temperate forests of C hina. However, dispersal limitation still independently accounts for a small amount of variance in species turnover. We therefore conclude that environmental filtering and dispersal limitation have both combined and independent effects, but environmental filtering is more important in shaping the community assembly of temperate deciduous broad‐leaved forests of C hina.