Grade Of Membership Models Reveal Geographical And Environmental Correlates Of Floristic Structure In A Temperate Biodiversity Hotspot

Identifying the contours and correlates of species turnover is central to understanding the nature of biogeographical regions. The Hengduan Mountains region of south-central China (HMR) is well known for its high diversity of plants, but its boundaries and internal floristic structure are poorly understood, especially in relation to geographical and environmental factors.With data on occurrences and elevational ranges of seed plants across the HMR and adjacent areas of the greater Qinghai-Tibet Plateau, we identified motifs (distinct species assemblages) by Grade of Membership models, and characterized relative contributions of geography, elevation, and climate to their spatial patterns.Motifs segregate primarily by latitude, elevation, and correlated environmental variables, most sharply across the tropical-temperate divide. Secondarily, they segregate by longitude and geographical features, and reveal a novel divide across the Jinsha River. A core set of motifs corresponds to previous delineations of the HMR.The HMR biodiversity hotspot is more a mosaic of floristic elements than a cohesive entity. Grade of Membership models effectively reveal the geographical contours of biotic structure, and are a valuable new tool for biogeographical analysis.