An intelligent SWMM calibration method and identification of urban runoff generation patterns

The accuracy of urban runoff simulation using the Storm Water Management Model (SWMM) largely depends on parameter calibration. This study proposes a universal and effective method to enhance model accuracy by optimizing parameter value ranges through an unsupervised intelligent clustering algorithm. Simulation scenarios with varying proportions of pervious and impervious areas are established, and sensitivity analysis is conducted to rank key parameters and identify dominant runoff generation patterns. The results show that when the impervious area is less than 10%, the most sensitive parameters are Zero.Imperv, N.Imperv, and Dstore-Imperv, indicating that runoff primarily originates from pervious surfaces. As the impervious area increases, runoff generation shifts to impervious areas, where the Unit Hydrograph Model, with fewer parameters and a simpler calibration process, leads to higher simulation accuracy. These findings improve the reliability of SWMM calibration and provide a reference for setting accuracy requirements under different urban surface conditions.