Phosphorus dynamics in Danish lakes and the implications for diatom ecology and palaeoecology

SUMMARY 1. Seasonal phosphorus variability may be considerable in eutrophic lakes and patterns are influenced by internal, as well as external, nutrient loading. The strong seasonality of planktonic diatoms, with main growth periods in the spring and also in the autumn in meso‐ to eutrophic lakes, means that the measure of total phosphorus (TP) that is most relevant for diatom ecology may not be adequately assessed by single or few measurements during the year. 2. The diatom species assemblage of surface sediments can be used to infer in‐lake nutrient concentrations. Weighted averaging (WA) and weighted averaging partial least squares (WAPLS) regression and calibration models for diatoms and annual mean TP were developed for a 29‐site data set of Danish lakes based on (a) all diatom species and (b) planktonic species only. Jack‐knifed error statistics were: =0.37, root mean squared error of prediction (RMSEP)=0.28 log₁₀μg TP L^?1 and mean bias=0.04 log₁₀μg TP L^?1 for the WAPLS 2‐component model based on all species; =0.23, RMSEP= 0.32log₁₀μg TP L^?1 and mean bias=0.07 log₁₀μg TP L^?1 for the WA model with tolerance downweighting based on planktonic species only. These are comparable with similar, published data sets. 3. A subset of 23 sites was used to develop models based on seasonal TP measurements. Mean spring TP concentrations gave only slightly improved RMSEP values for models based on all species and plankton‐only (0.24 log₁₀μg TP L^?1 and 0.29 log₁₀μg TP L^?1, respectively). 4. Weighted averaging models derive environmental optima, for individual species, which are not necessarily of ecological relevance. However, good water chemistry data are required to model species' responses adequately and to develop calibration data sets.