Sedimentary diatom concentrations and accumulation rates as predictors of lake trophic state

Diatom concentrations in surface sediments are positively correlated with limnetic chlorophyll a concentrations in Florida (USA) lakes. Using this relationship, I examine models that provide quantitative inferences for trophic state in historical applications.The best model predicts chlorophyll a trophic state index (TSI) values from log-transformed diatom concentrations and explains approximately half the variance in the dependent variable. Diatom accumulation rates are not better than sedimentary diatom concentrations as predictors of TSI. The entire diatom assemblage is as sensitive an indicator of TSI as are the planktonic diatoms alone. A model that considers the ecological preferences of specific taxa was found to be a better predictor than the model based on total diatom concentration.The sedimentary diatom concentration model provides a useful method for assessing historical changes in primary productivity, except in lakes where factors (e.g., silica limitation, blue-green bacterial inhibition) limited diatom production, or post-depositional changes removed sedimentary diatoms. TSI inferences are presented for sediment cores from two Florida lakes, one of which demonstrates a problematic application, and the other of which does not.