A comparative analysis of four rainfall-runoff models on small watesheds near Tucson, Arizona

Abstract

The performance of four rainfall-runoff models on small watersheds near Tucson, Arizona is compared. The models are: "Flood Peak Estimator" of Pima County, Arizona, TR-20, HEC-1, and Hydrosoftware Inc.'s "RAINFLO". In sensitivity analyses, variations in parameters describing effective rainfall, time of concentration, and channel losses are compared to resulting changes in runoff volume and Qpeak. Volume was found most sensitive to the Curve Number while Qpeak was most sensitive to time of concentration. On five small watersheds, the models were applied using a 100-year rainfall. Qpeak was compared to Q100 developed by Log Pearson Type-III from 15-18 years of data on each watershed. On three, the models showed some agreement with LP-III. On two, the model results were far below LP-III, most likely due to inadequacy in methods for estimating time of concentration. Applicability and shortcomings of the models were discussed.