Background/Question/Methods The estimates of species diversity obtained by raw occurrence data in comparison with distribution maps lead to unrealistic or biased estimation of its components, with patterns changing with the scale of analysis. Raw data underestimates the local species richness (alpha diversity) and overestimates the measure that relates the local with the regional species richness (multiplicative beta diversity), as functions of area.

Results/Conclusions In this work we analyze the correlation of several environmental variables with alpha and beta diversities, at several spatial scales (1/32 degree − 1/2). We do this following a linear-nested and a random sampling scenarios, and using two different measures of spatial heterogeneity in environmental variables (variance and variation). Our results suggest that the three diversity measures correlate differently with environmental variables, depending on the scale of analysis, but clearly supporting the idea that there is a natural decomposition of factors into an area-invariant component (related to alpha), and an area dependent factor (related to beta).