A fundamental question at the interface of ecology, evolution, and biogeography is whether the mechanisms of community assembly vary along latitudinal gradients. Turnover in species composition (β-diversity) through space tends to increase towards the tropics, a pattern which can reflect at least three non-mutually exclusive processes. First, the size of the regional species pool (γ-diversity) influences β-diversity through stochastic sampling effects: as γ-diversity increases towards the tropics, “spillover” of species among localities is expected to increase from sampling alone. Second, differences in β-diversity between temperate and tropical regions may reflect variation in the strength of deterministic processes such as niche selection along environmental gradients. Third, in high-diversity regions with many rare species, spatial (e.g., dispersal limitation) and stochastic (e.g., ecological drift) processes can strongly influence β-diversity. We evaluate these hypotheses by comparing patterns of β-diversity between temperate and tropical forests that span a 10-fold difference in γ-diversity. Using standardized methods, we compare the composition of woody plants (>2.5-cm DBH) in Missouri (Ozark Ecoregion, USA; γ-diversity ~ 50 species) and Bolivia (Madidi National Park; γ-diversity ~ 600 species) using >36 0.1-ha plots in each region that span similar environmental and spatial gradients.
Results/Conclusions
Observed β-diversity was higher in Bolivia than in Missouri, but this trend disappeared after accounting for differences in the size of the regional species pool using a null model. Moreover, we found positive and similar deviations of observed β-diversity from the null model, indicating higher turnover than expected by chance in both temperate and tropical forests. To compare potential mechanisms influencing null-model deviations, we used variance-partitioning methods to examine the relative importance of environmental (soil, topography, climate), spatial (e.g., geographic distance), and spatially structured environmental variables influencing β-diversity. We found that deterministic processes had a similar influence on β-diversity in Bolivia and Missouri: pure environmental and spatially structured environmental variables combined explained ~40% of the total variation in β-diversity in both regions. Pure spatial processes had a similar but weaker influence than environmental processes in both regions, explaining 7-10% of the total variation. By explicitly accounting for stochastic sampling effects owing to differences in regional species pools, our study suggests that local deterministic processes may have similar influences on community assembly in temperate and tropical forests.