Different Effects of Regional Species Pool on Plant Diversity between Forest and Grassland Biomes in Arid Northwest China

Species pool hypothesis is broadly known and frequently tested in various regions and vegetation types. However it has not been tested in the arid Xinjiang region of China due to lack of data. Here with systematic data from references and field survey, we comprehensively examined species pool hypothesis in this region. Took species richness in 0.1° × 0.1° grid cells as regional species richness (RSR) which were obtained from the distribution maps of vascular plant species, and took species diversity of 190 and 103 plots in forest and grassland biomes across Xinjiang as local species richness (LSR), together with the digitalized soil pH and climate data, we tested the species pool hypothesis in this region. We found that: (1) the average RSR was higher in mountains than that in basins and it was negatively correlated with soil pH in mountains while positively correlated with soil pH in basins in Xinjiang; (2) RSR showed a positive correlation with mean annual precipitation (MAP) while showed a hump-shaped pattern with mean annual temperature (MAT); and the changing patterns of LSR were different for forest and grassland along the geographical and climate gradients; (3) LSR of forest was more affected by RSR than by climate, while on the contrary, LSR of grassland was more affected by climate than by RSR. Our results validated the species pool hypothesis in revealing that RSR had a significant role in shaping LSR patterns in addition to climate. We concluded that the relative effects of climate vs. RSR on LSR differed markedly between the forest and grassland communities across Xinjiang. Our results also showed that RSR revealed a contrasting relationship with soil pH in mountains and in basins, which might reflect differences in evolutionary processes of various habitats. In summary, our research systematically analyzed the correlation of species richness in regional and local scales in Xinjiang which provides more insights into the understanding of species pool hypothesis.