The processes that determine community composition may vary enormously across scales. The degree to which predictors of broad-scale ecological patterns scale down to the scale of local communities, where actual interactions among individuals occur, remains unclear. Elucidating the relationship between regional richness and local richness can shed light on the processes controlling local community assembly from the corresponding species pool. In this study, we contrast the local richness of 330 extensively surveyed terrestrial vertebrate communities worldwide with their respective regional richness at varying scales.
Results/Conclusions
We find that the amount of variation in species richness explained by environmental predictors increased consistently with increased spatial scale. Some transitions in the identity of environmental predictors across scales were detected. Proportional local richness in relation to regional richness decreased with increased regional richness. Consequently, the discrepancy between local and range-map diversity increased predictably along environmental and latitudinal gradients. However, once gradients in regional diversity were statistically accounted for, little residual variation in local diversity was explained by local environment. The amount of residual variation increased with the scale used to define the regional species pool. Local net primary productivity and elevation heterogeneity were the best predictors of residual variation between local and range map diversity.
The interaction between the size of the regional species pool and the strength of local species assembly processes will ultimately determine local diversity. Our results suggest that after accounting for regional differences, local richness is relatively insensitive to environmental gradients. This highlights the stochastic nature of local assembly processes. From a conservation perspective, the degree to which regional estimates of richness, e.g. based on range-maps, may provide useful estimates of local richness in e.g. yet unsurveyed reserves varies among taxa and their typical range sizes. However, considerable uncertainty still exists and these estimates will not be considerably improved by using additional environmental data.