Variation in iron, aluminum and dissolved organic carbon mass transfer coefficients in lakes.

Steady-state mass balance models have been successfully used to predict annual loads and concentrations of numerous substances in lakes. A major limitation of the model is that the value of the mass transfer coefficient, nu(X), for some substance, X, is assumed constant and is taken as a mean value of a set of calibration lakes. The range in mean values between lakes for nu(X) is typically two- to three-fold. Thus, there is some potential for error in model predictions if the true but unknown value for a given lake differs from the mean, calibration value. Moreover, the use of a region-specific value for nu(X) for all lakes in a region means that when environmental conditions change, the model will have to be re-calibrated requiring many more years of monitoring. In this study, mass balances for two substances, total iron and aluminum, are presented along with simple empirical relationships that predict mass transfer coefficients for iron as a function of dissolved organic carbon (DOC) concentration and aluminum as a function of DOC and pH. The relationships, which are consistent with current understanding of iron and aluminum behavior, account for >90% of the variation in nu(X). Improvements in mass balance model predictions are expected when these coefficients are used in place of constant values.