Using pollen analysis to assess recent changes in floristic diversity, an example from the Swiss Plateau

Abstract. Assessing the diversity of pollen types in sedimentary deposits is one of the few available proxies for reconstructing past plant diversity. However, several biases complicate pollen based reconstructions. One of the main sources of bias is the overrepresentation of wind-pollinated relative to insect-pollinated plants. Increased abundance of wind-pollinated plants in the catchment changes both the size and the evenness of the total pollen pool, thereby distorting richness estimates based on a constant count sum. Several methods have been proposed to reduce this bias, including separating functional groups (wind- versus insect-pollinated), accumulation-based rarefaction, and increasing pollen count sizes. We evaluated these methods using two high-resolution pollen records from the Swiss Plateau, focusing on the effects of historical land-use changes. Our results demonstrate clear biases in standard diversity estimates and show that robust pollen-based diversity reconstructions require high pollen counts. Although, separating wind- and insect-pollinated taxa may provide additional ecological insights, it does not resolve the largest biases and can be potentially misleading. Accumulation-based rarefaction is conceptually and empirically the most robust. However, its practical applicability is limited as it requires precise chronological control, which is often not achievable in pollen records. Using running bins to increase count sizes and rarefaction to a constant sum without separating insect-pollinated taxa may therefore be the simplest way to improve the diversity signal for most pollen diagrams. The pollen records from the Seetal area in Switzerland show a decline in palynological richness that started already at the end of the 19^th century. By this time, major landscape reforms had started, leading to a shift from dynamic mixed farming systems to permanent grasslands for dairy farming. This change in land-use is consistent with the decline in palynological richness and an increases in pollen types from nitrophilous plants as well as a decline in wetland taxa. Further intensification occurring later during the 20th century did not cause a further decline of palynological richness. This may reflect the sensitivity of the pollen signal to the initial loss of common insect-pollinated taxa, capturing their initial decline, but not the further reductions of their abundances once they became rare.