Modern pollen-plant diversity relationship in open landscapes of Tibetan Plateau

Abstract

Past plant diversity dynamics can provide a historical perspective on biodiversity change and ecosystem dynamics as a guide to conservation management in the future. Pollen analysis is one of the most widely used approaches to reconstruct plant diversity change through time, but the representativeness of pollen diversity to plant diversity is not yet fully understood and especially in open landscapes. In this paper, we compare pollen assemblages from topsoil in the open landscapes of the Tibetan Plateau with the surrounding vegetation to investigate their compositional similarity and the correlations between their diversity (richness and evenness), and to evaluate the influence of climate variables, landscape characteristics, and human disturbance on the pollen diversity. The results show that pollen assemblages vary obviously in different vegetation types and they can represent their surrounding plant communities at low taxonomic levels. Pollen richness boarders a better match with plant richness than pollen evenness with plant evenness across the whole study area, but the correlation is weak (radj2 < 0.2, p < 0.001). Transforming plants into their pollen equivalents has little improvement for the correlations between pollen and plant richness/evenness. Pollen richness and evenness can capture well the trend in plant richness and evenness along the vegetation gradient, suggesting that they can be used as gamma (landscape scale) rather than alpha (local scale) diversity metrics in the study area. Mean annual precipitation (MAP), mean annual temperature (MAT), and patch richness (PatchS) are positively associated with pollen richness. MAP, PatchS, and human disturbance (HD) are significantly predictors for pollen evenness. These confirm that pollen richness and evenness can provide insights into past plant diversity dynamics and help evaluating the effects of climate and habitat changes and human disturbance on long timescales.