Impact of Uncertain Knowledge of Model Parameters on Estimated Runoff and Phosphorus Loads in the Lake Okeechobee Basin

Uncertain knowledge of the values for parameters of the CREAMS-WT model used for estimating runoff volumes and phosphorus loads into streams draining into Lake Okeechobee in south Florida leads to uncertainty in these estimates. In this study, first order analysis and Monte Carlo simulation were used to quantify this uncertainty in terms of probability distribution functions and confidence intervals on model outputs. Sensitivity coefficients were used to identify the parameters that were the most important for this analysis. In the case of runoff volume, one parameter accounted for 73% of the uncertainty in terms of variance. Thus improving knowledge of this single parameter can have a significant impact on reducing the uncertainty in the predicted runoff.