Evaluation of CLIGEN Precipitation Parameters and Their Implication on WEPP Runoff and Erosion Prediction

The quality of synthesized daily weather data directly affects the output of hydrological and agricultural response models. The objectives of this study were to evaluate the ability of the CLIGEN model to reproduce daily, monthly, and annual precipitation amounts, extremes, and internal storm patterns (i.e., storm duration, relative peak intensity, and time to peak) and to assess further the impact of generated storm patterns on WEPP runoff and erosion prediction. Four Oklahoma stations with more than 50 years of daily precipitation data and eight other sites across the U.S. with an average record of 10 years of measured storm patterns were used. Mean absolute relative errors for simulating daily, monthly, and annual precipitation across the four Oklahoma stations were 4.7%, 1.7%, and 1.5% for the means and 3.7%, 6.7%, and 15% for the standard deviations, respectively. Mean absolute relative errors for the all–time maxima of daily, monthly, and yearly precipitation were 17.7%, 8.9%, and 6.5%, respectively. Storm pattern generation, especially storm duration, was determined to need improvement for better prediction of runoff and soil erosion. The measured storm patterns showed positive linear correlations between precipitation, duration, and relative peak intensity, but little correlation was shown for generated storm patterns. The CLIGEN–generated durations were generally too long for small storms and too short for large storms. Inaccurate storm pattern generation led to WEPP prediction errors as high as 35% for average annual runoff and 47% for annual sediment yield on the test sites. To improve WEPP runoff and erosion prediction, storm duration generation should be reconsidered, and a distribution–free approach may be used to induce proper correlations between the input storm variables.