Phylogenetic gradient analysis: environmental drivers of phylogenetic variation across ecological communities

Abstract

Differences in species composition across ecological communities are the result of multiple interacting mechanisms. Gradient analysis has been perhaps the most widely used statistical framework in describing how ecological communities vary in space, and aiding in determining the causes underlying these patterns. Direct gradient analysis allows the use of predictors such as environmental factors and spatial descriptors to directly estimate their contributions in explaining common and independent patterns of species distributions. In recent years, ecologists have started to explore how evolutionary history is associated with community patterns given the observation that species that share a common phylogenetic history tend also to have similar niches. Although ecological phylogenetics is among the fastest-growing fields in ecology, gradient analysis has not yet been fully integrated in this field. In this paper, we show and adapt the versatility of gradient analysis in describing and interpreting patterns of ecological communities based on their patterns of phylogenetic structure. Describing phylogenetic patterns across communities presents additional challenges regarding statistical inference in contrast to classic direct gradient analysis that are described and tackled here. We investigate the performance of our phylogenetic gradient analysis frameworks using simulations and provide a detailed example using a grassland community dataset.