Pollen from 50 lake surface samples in Brandenburg, Germany, in 2009

Past changes in plant and landscape diversity can be evaluated through pollen analysis, however, pollen based diversity indexes are potentially biased by differential pollen production and deposition. Studies examining the relationship between pollen and landscape diversity are therefore needed. The aim of this study is to evaluate how different pollen based indexes capture aspects of landscape diversity. Pollen counts were obtained from surface samples of 50 small to medium sized lakes in Brandenburg (Northeast Germany) and compiled into two sets, with one containing all pollen counts from terrestrial plants and the second restricted to wind-pollinated taxa. Both sets were adjusted for the pollen production/dispersal bias using the REVEALS model. A high resolution biotope map was used to extract the density of total biotopes and different biotopes per area as parameters describing landscape diversity. In addition tree species diversity was obtained from forest inventory data. The Shannon index and the number of taxa in a sample of 10 pollen grains are highly correlated and provide a useful measure of pollen type diversity which corresponds best to landscape diversity within one km of the lake and the proportion of non-forested area within seven km. Adjustments of the pollen production/dispersal bias only slightly improve the relationships between pollen diversity and landscape diversity for the restricted dataset as well as for the forest inventory data and corresponding pollen types. Using rarefaction analysis, we propose the following convention: pollen type diversity is represented by the number of types in a small sample (low count e.g. 10), pollen type richness is the number of types in a large sample (high count e.g. 500) and pollen sample evenness is characterized by the ratio of the two. Synthesis. Pollen type diversity is a robust index that captures vegetation structure and landscape diversity. It is ideally suited for between site comparisons as it does not require high pollen counts. In concert with pollen type richness and evenness, it helps evaluating the effect of climate change and human land use on vegetation structure on long timescales.