To better understand how ecosystems are changing, a multifaceted approach to measuring biodiversity that considers species richness and evolutionary history across spatial scales is needed. Here we compiled 162 datasets for fish, bird, and plant assemblages across the globe and measured how taxonomic and phylogenetic diversity changed at different spatial scales (within site α diversity and between sites spatial β diversity). Biodiversity change is measured from these datasets in three ways: across land use gradients, from species lists, and through sampling of the same locations across two time periods. We found that local species richness and phylogenetic α diversity (Faith’s PD) increased for all taxonomic groups. However, when measured with a metric that is independent of species richness (phylogenetic species variation, PSV), phylogenetic α diversity declined for all taxonomic groups. Land use datasets showed declines in species richness, Faith’s PD, and PSV. For all taxonomic groups and data types, spatial taxonomic and phylogenetic β diversity decreased when measured with Sorensen dissimilarity and phylogenetic Sorensen dissimilarity, respectively, providing strong evidence of global biotic homogenization. The decoupling of α and β diversity, as well as taxonomic and phylogenetic diversity, highlights the need for a broader perspective on contemporary biodiversity changes. Conservation and environmental policy decisions thus need to consider biodiversity beyond local species richness to protect biodiversity and ecosystem services.

Most column names are self-explaining. dat_1_n represents the number of sites of "time period" 1 for each dataset. `smd_logrr` represents the log response ratio of diversity, i.e. log(dat_2_ave/dat_1_ave). `signma_rr` represents the variance of log response ratio. See the paper for details.