Resolution and Sensitivity Analysis of a Block-Scale Urban Drainage Model

The complexity of the urban watershed requires engineers to use computer models for water planning purposes. A widely-used industry standard model is the U.S. EPA Stormwater Management Model (SWMM) which predicts flow rates from watersheds characterized by the modeler in terms of both their hydrologic and hydraulic characteristics. To cost effectively model large urban areas within the time and budget constraints, engineers simplify the watershed domain of the model by aggregating subcatchments and conduit skeletonization. The process of selecting model inputs introduces uncertainty into SWMM predictions. Other sources of uncertainty include the mathematical equations used to represent the flow process, the measurement error associated with physical parameters in the field, and the determination of non-physical parameters. This study evaluates the sensitivity of user-specified subcatchment characteristics in predicting total volumetric runoff and peak runoff flow in SWMM. The model variables were adjusted according to a Monte Carlo simulation using a uniform distribution for the input parameters. The sensitivity of parameters will determine the criteria for calibrating the model and for representing low-impact development techniques.