Detailed urban drainage data are important for urban nonpoint source (NPS) pollution prediction. However, the difficulties in collecting complete pipeline data usually interfere with urban NPS pollution studies, especially in large-scale study areas. In this study, NPS pollution models were constructed for a typical urban catchment using the SWMM, based on five drainage datasets with different resolution levels. The influence of the data resolution on the simulation results was examined. The calibration and validation results of the higher-resolution (HR) model indicated a satisfactory model performance with relatively detailed drainage data. However, the performances of the parameter-regionalized lower-resolution (LR) models were still affected by the drainage data scale. This scale effect was due not only to the pipe routing process but also to changes in the effective impervious area, which could be limited by a scale threshold. The runoff flow and NPS pollution responded differently to changes in scale, primarily because of the difference between buildup and washoff and the more significant decrease in pollutant infiltration loss and the much greater increase of pollutant flooding loss while scaling up. Additionally, scale effects were also affected by the rainfall type. Sub-area routing between impervious and pervious areas could improve the LR model performances to an extent, and this approach is recommended to offset the influence of spatial resolution deterioration.
