Testing the reliability of pollen-based diversity estimates

Rarefaction analysis is a common tool for estimating pollen richness. Using modern and fossil pollen data from the Canadian Arctic and Greenland, we examine the effects of pollen concentration (grains/cc) and evenness (the distribution of species abundances) on palynological richness. Our results show that pollen richness and concentration have a strong negative correlation at low pollen concentrations. There is a positive correlation between pollen evenness and richness, although the strength of this relationship is difficult to determine. Rarefaction analysis on samples of low concentration or high evenness is likely to lead to pollen richness being less underestimated than on samples of high concentration or low evenness. These findings corroborate theoretical research on these issues.