A test of two models of different levels of complexity for predicting changes of phosphorus concentration in a lake's outflow

The objective of this investigation is to compare the ability of a one-box model and a two-box model to predict the temporal variations of total phosphorus in the outflow of several lakes. The one-box model considers only total phosphorus and treats the lake as well-mixed, all year round. The two-box model considers two phosphorus fractions and the epilimnion and hypolimnion. For five shallow Canadian Shield lakes, some of the temporal variability in lake phosphorus concentrations can be explained on the basis of changes in hydraulic inflows and phosphorus loadings. Both models give roughly similar predictions for shallow lakes with high hydraulic loadings (1–200 m/year) and small variations in total phosphorus concentrations. In comparing the settling and hydraulic outflow mechanisms, a sensitivity analysis shows that both models are more sensitive to the settling rate if the lake's hydraulic loadings are less than 1 m/year, and more sensitive to the outflow rate if the lake's hydraulic loadings are greater than 100 m/year. This analysis suggests a priority for concentrating economic and other resources in conducting field measurement programs for testing phosphorus models.