Reconstructing past eutrophication trends from diatoms and biogenic silica in the sediment and the pelagic zone of Lake Constance, Germany

In Lake Constance, phosphorus concentrations and the seasonal development of phytoplankton communities in water samples from the pelagic zone were regularly recorded since the 1950's. Before the 1950's, there were occasional investigations of plankton communities since 1896. We compared these data with the sedimentary record in two sediment cores. Then, the eutrophication history of Lake Constance was inferred from diatoms. The record of biogenic silica in the cores is discussed with respect to diatom biomass increase.Diatom assemblages in the sediment cores precisely reflected the pelagic diatom development for the period 1971--1992. Both sediment cores and the water samples have a high interannual variability of diatom assemblages. Below a sediment depth of 27 cm (AD 1920), more than 50% of the diatoms were partly corroded, and we limited the reconstruction of trophic state changes to the interval of 1920--1993. Oligotrophic conditions of Lake Constance were indicated by the dominance of various Cyclotella taxa from 1920 to 1940. Since 1939/1940, increasing abundance of it Tabellaria fenestrata showed oligotrophic to mesotrophic conditions. Between 1953 and 1956, increasing Stephanodiscus hantzschii and disappearing Cyclotella indicated advanced eutrophication and total phosphorus values ranged between 8--10 mg m-3 during turnover in late winter. Further eutrophication was shown by disappearing T. fenestrata and increasing S. minutulus in 1963. Maximum TP concentrations of 87 mg m-3 occurred in 1979/80 and was accompanied by increasing abundances of Aulacoseira granulata. From 1986 to 1992, reoccurrence of Tabellaria fenestrata and Cyclotella indicate some recovery of Lake Constance.Biogenic silica and diatom abundances were similar among cores but indicate a 3--4 fold increase of diatom biomass only. This was far below the estimate of biomass increase from sedimentary pigment data (25 fold) and the estimate of phytoplankton data from the literature (70 fold).