Diatom-environmental relationships in 64 alkaline southeastern Ontario (Canada) lakes: a diatom-based model for water quality reconstructions

Lake eutrophication is a problem in many areas of Ontario, although the his tory of nutrient enrichment is poorly documented. The aim of this study was to construct a diatom-based transfer function to infer past phosphorus lev els in Ontario lakes using paleolimnological analyses. The relationship bet ween diatom assemblages and limnological conditions was explored from a sur vey of diatoms preserved in the surface sediments of 64 Southern Ontario la kes, spanning a total phosphorus gradient of 0.004 to 0.054 mg L-1. Over 42 0 diatom taxa were identified, 98 of which were sufficiently common to be c onsidered in statistical analyses. Canonical correspondence analysis (CCA) determined that pH, ammonium, aluminium, spring total phosphorus(TP), stron tium, total nitrogen (TN), maximum depth (MaxZ), chlorophyll a (Chla) and m ean depth were significant variables in explaining the variance in the diat om species data. The environmental optima of common diatom taxa for the lim nologically important variables (TP, pH, TN, MaxZ, Chla) were calculated us ing weighted averaging (WA) regression and calibration techniques, and tran sfer functions were generated. The diatom inference model for spring TP pro vided a robust reconstructive relationship (r(2) = 0.637; RMSE = 0.007 mg L -1; r(boot)(2) = 0.466; RMSEboot = 0.010 mg L-1). Other variables, includin g pH (r(2) = 0.702; RMSE = 0.208; r(boot)(2) = 0.485; RMSEboot = 0.234), TN (r(2) = 0.574; RMSE = 0.0899 mg L-1; r(boot)(2) = 0.380; RMSEboot = 0.127 mg L-1) and MaxZ (r(2) = 0.554; RMSE = 1.05 m; r(boot)(2) = 0.380; RMSEboot = 1.490 m), were also strong, indicating that they may also be reconstruct ed from fossil diatom communities. This study shows that it is possible to reliably infer lakewater TP and other limnological variables in alkaline So uthern Ontario lakes using the WA technique. This method has the potential to aid rehabilitation programs, as it can provide water quality managers wi th the means to estimate pre-enrichment phosphorus concentrations and an in dication of the onset and development of nutrient enrichment in a lake.