Pollen richness : a reflection of vegetation diversity or pollen-specific parameters?

Our study aims to increase the understanding about the impacts of potential drivers of pollen richness by using a pollen-vegetation modelling approach. We used the Sutton-Prentice dispersal model implemented in the HUMPOL software suite to explore the effects of factors commonly associated with pollen richness: vegetation diversity (plant richness and evenness), land-cover characteristics (patch richness, evenness and size, and basin size) and pollen-related parameters (productivity and fall speed). The impact of the factors was tested using modelling scenarios involving all combinations of the included factors and their values. All tested factors had a statistically significant impact on modelled pollen richness. Pollen-type based plant richness was the dominant determinant of pollen richness; however, pollen productivity became co-dominant at higher plant richness levels, effectively limiting the detection of species with low pollen productivity. The fall speed of pollen and sedimentary basin size had moderate impact, but gained importance in simulations with elevated plant richness. In patchy land-cover, patch richness was the most significant determinant of pollen richness, while patch evenness and size were the least important factors. Our modelling approach provides insights into the significance of common factors in determining pollen richness and its connection to plant diversity, as well as a theoretical basis for understanding the substantial variation among earlier empirical pollen-plant richness studies. The estimation of plant richness from pollen richness in past vegetation diversity studies could be improved by the separation of taxa with low and high pollen productivity and an awareness of the pollen source area