Dissolved organic carbon reconstructions from diatom assemblages in PIRLA project lakes, North America

Diatom-based palaeolimnological reconstructions of dissolved organic carbon (DOC) are presented for four regional data sets of the North American ‘Paleoecological Investigation of Recent Lake Acidification (PIRLA)’ project, and for a combined, three-region set. Species optima and tolerances along the DOC gradient were estimated by using maximum likelihood and weighted-averaging regression. Weighted-averaging regression appears to be the most robust and tractable technique for estimating optima, and the apparent error (mean standard error of the relation) was as good for weighted-averaging calibration as for maximum likelihood calibration. Calculated species optima are not entirely consistent among regions and the best ‘indicators’ for DOC in the PIRLA data-sets are not in good agreement with those found in the literature. Example reconstructions demonstrate that DOC changes are often less than 100 µmol l-1, and that the DOC declines in some recently acidified lakes parallel reconstructed pH declines.