The mapping of earth's biodiversity has advanced our theoretical and empirical understanding of biodiversity and has thus guided conservation efforts. Yet, early biodiversity maps often relied on alpha diversity indices, while beta diversity has rarely been used for practical conservation actions. We used generalized dissimilarity modelling (GDM) and variance partitioning to map beta diversity patterns of Hainan Island, China, and explore its underlying factors based on a large dataset of 248,538 individual trees belonging to 1,016 species in 902 forest plots. We used principal component analysis and hierarchical clustering to visualize community similarity, and spatial overlap analysis to assess the ability of the current protected areas (PAs) to encompass beta diversity. The GDMs explained 27.65% and 26.58% of the variation in beta diversity at the genus and species levels, respectively. The community composition of tree species in Hainan presented a general east-to-west gradient, and three floristic regions were delineated. This biogeographical pattern is predominantly structured by mean annual precipitation. Environmental variables, rather than geographical distance, were the most important factors determining present beta diversity patterns. Currently, PAs of Hainan Island are concentrated on mountain forest areas, while the lowland forest has largely been ignored. Thus, we suggest that biodiversity mapping based only on alpha diversity is not enough to identify conservation gaps, and the inclusion of beta diversity in such maps constitutes a promising tool to maximize the biodiversity coverage of PAs. Our study provides empirical evidence that a spatially explicit analysis of beta diversity in a specific region can be used for conservation planning.
Keywords: Community composition; Floristic region; Generalized dissimilarity modelling; Hainan Island; Precipitation; Protected areas.
Copyright © 2019 Elsevier B.V. All rights reserved.