Statistical Procedure for Evaluating Hydrologic/Water Quality Models

Input parameters are used in hydrologic/water quality models to describe specific situations. These input parameters are never known with certainty. The models, although largely physically based, are not capable of describing the exact hydrologic and chemical processes that take place under natural conditions. Evaluation of models by a comparison of observed and predicted results is fraught with ambiguities resulting from the large number of parameters that must be estimated and the inherent variability in natural systems. This article puts forth a model evaluation protocol based on a simulation procedure which transforms parameter uncertainty into prediction uncertainty using probability density functions. Confidence intervals are placed on model results and decisions regarding model acceptability are made based on the magnitude of measured data, confidence intervals, and performance criteria placed on the model.