Improving Swmm Predictions Using Dry and Wet Hydroclimatic Condition Parameter Sets Fit Using Automatic Calibration

Abstract

The Storm Water Management Model (SWMM) is a widely used urban watershed model. We compared the ability of SWMM to predict flows when independently calibrated to a dry and wet year, respectively, using OSTRICH-SWMM. The model developed using SWMM calibrated to a wet year performed better in the model assessment period. The water budgets differed markedly between the dry and wet years. The best fit estimates of SWMM parameters differed significantly between dry and wet years. For instance, Manning’s roughness coefficient for overland flow was higher in a dry year, as less runoff meant less flow on already wetted surfaces. Some parameters, e.g., % effective imperviousness, exhibited an expanded posterior probability distribution, increasing uncertainty of the parameter estimate. However, other parameters, such as Manning’s roughness coefficient for streams were well-defined. These changes in parameter sets for dry and wet hydroclimatic conditions affect the hydrological response of an urban watershed.

Keywords: Stormwater Management Model, OSTRICH-SWMM, automatic calibration, parameter estimate, posterior parameter distribution, probability distribution models