Quantifying background nutrient concentrations in coastal waters: a case study from an urban embayment of the Baltic Sea.

Successful management of damaged coastal ecosystems requires reliable scientific evidence of their past state. Here we demonstrate that the sediment record of biotic indicators can be used to quantitatively reconstruct nutrient concentrations preceding the short time span covered by monitoring records. We generated a diatom-based weighted-averaging partial least squares transfer function model for total dissolved nitrogen with a prediction accuracy of 0.09 microg L(-1) (log10 units). The model was applied to sediment core data from Laajalahti Bay, an urban embayment in Helsinki, Finland, where its performance was validated against a approximately 30 yr record of water-quality data and known land-use changes in the watershed. The model tracked well the trends in the nutrient record, although it underestimated very high nutrient concentrations in this highly impacted embayment. The generally good agreement between the actual and predicted values implies that the approach has considerable potential in assessing background nutrient concentrations in coastal waters.