Abstract
Between 1951 and 1979, total phosphorous concentrations in Lake Constance increased from 7 to 87 μg L −1 . Following wastewater treatment, phosphorus levels were brought under control, returning to 7.6 μg L −1 by spring 2007. The biological and chemical data from 1980 to 2004 were first modelled by seasonal time series analyses and then used to create a general model. Excluding collinear variables allowed the data set to be condensed to six variables that could be fitted into a general linear model that explained ∼75% of the observed annual variation in chlorophyll a . A clear seasonal influence was apparent, with chlorophyll a tracking trends in temperature and the progress of spring. A nonseasonal influence was also observed in the interaction of two biological components, the proportion of phytoplankton biomass available to Daphnia (i.e. the percentage of ingestible size a levels. In contrast, the concentration of soluble reactive phosphorus (SRP) correlated positively with chlorophyll a . The effect of SRP showed a significant seasonal component, as it was more abundant in spring than at other times of year. In general, the model predicts a negative exponential response of chlorophyll a to further depletion of SRP in Lake Constance, while the temperature trends predicted by current global warming scenarios will result in a moderate increase in productivity. Data from 2005 to 2007 were used to verify the model. The modelled chlorophyll a values were nonbiased and showed a close match to the measured values ( r 2 : 75%). Thus the applicability, reliability, and informative value of the model for pelagic Lake Constance was confirmed. The approach might easily be applied to other waters.