Undersampling Correction Methods To Control Gamma-Dependence For Comparing Beta-Diversity Between Regions

Measures of beta-diversity are known to be biased by differences in gamma-diversity (i.e., gamma-dependence), making it challenging to compare beta-diversity across regions. Undersampling corrections have been designed to reduce effects of gamma-dependence on beta-diversity arising from the problem of incomplete sampling. However, no study has systematically tested the effectiveness of these corrections or examined how well they reflect beta-diversity patterns across ecological gradients. Here, we conduct these tests by comparing two undersampling corrections with the widely used individual-based null model approach, using both simulated communities along an ecological gradient and empirical data across a wide range of gamma-diversity and sample sizes. We found that undersampling corrections using diversity accumulation curves were more effective than the null-model approaches in removing gamma-dependence. In particular, the corrected beta-Shannon diversity index was least dependent on gamma-diversity, and was the most reflective of the beta-diversity pattern along a simulated ecological gradient. Moreover, a corrected Jaccard-Chao index applied to null model results removed gamma-dependence more effectively than either the correction alone or the null model alone. Undersampling corrections are effective tools for removing gamma-dependence bias, thus facilitating comparisons of beta-diversity across regions.