Evaluating relationships between sediment chemistry and anoxic phosphorus and iron release across three different water bodies

This study evaluated phosphorus (P) and iron (Fe) release from anoxic sediment cores in 3 different water bodies, Lake Winnipeg in Manitoba and Lake Simcoe and Hamilton Harbour in Ontario, and developed empirical models to predict P and Fe release rates as functions of sediment and water chemistry. Maximum P and Fe sediment release rates, which occurred shortly after sulfate (SO4) was depleted in most cores, showed unique positive linear relationships for each lake. High P release rates only occurred from Lake Simcoe and Lake Winnipeg sediments with molar ratios <= 8.1 of NaOH-extractable aluminum (A1) to reducible Fe (Fe extractable with bicarbonate-dithionite [BD]). High Fe release rates only occurred in Lake Winnipeg sediments, perhaps due to low SO4 reduction rates (<115 mg m(-2) d(-1)). Three approaches were used to develop linear regression models for maximum P and Fe release rates using data from all lakes: only Lake Winnipeg, only Lake Simcoe, and low (<= 8.1) and high molar ratios (>20) of NaOH-extractable Al to BD-extractable Fe. Regression models differed depending on the study sites used, and only 3 of the 5 significant regression models for P release had high predictive value (R-2 > 0.7). Four of the 5 significant Fe release models had R-2 > 0.7, but we were unable to find a significant Fe model for Lake Simcoe. These results are important because they delineate the limitations of this approach to modeling release using operationally defined fractions.