Abstract: Latitudinal patterns of tree β-diversity reveal important insights into the biogeographical processes that influence forest ecosystems. Although previous studies have extensively documented β-diversity within relatively small spatial extents, the potential drivers of β-diversity along latitudinal gradients are still not well understood at larger spatial extents. In this study, we determined whether tree β-diversity is correlated with latitude in forests of southeastern China, and if so, what ecological processes contribute to these patterns of tree β-diversity. We specifically aimed to disentangle the relative contributions from interspecific aggregation and environmental filtering across various spatial extents. We delineated regional communities comprising multiple nearby national forest inventory (NFI) plots around random focal plots. The number of NFI plots in a regional community served as a surrogate for spatial extent. We also used a null model to simulate randomly assembled communities and quantify the deviation (β-deviation) between observed and expected β-diversity. We found that β-diversity decreased along a latitudinal gradient and that this pattern was clearer at larger spatial extents. In addition, latitudinal patterns of β-deviation were explained by the degree of species spatial aggregation. We also identified environmental factors that drive β-deviation in these forests, including precipitation, seasonality, and temperature variation. At larger spatial extents, these environmental variables explained up to 84% of the β-deviation. Our results reinforce that ecological processes are scale-dependent and collectively contribute to the β-gradient in subtropical forests. We recommend that conservation efforts maintain diverse forests and heterogeneous environments at multiple spatial extents to mitigate the adverse effects of climate change.

Key words: Environmental filtering, Latitudinal pattern, Forest beta diversity, Spatial extent, Species spatial aggregation, Regional community