Abstract: To solve the problem in urban waterlogging simulations caused by the difficulty in obtaining urban sewer network data, this paper uses different generalization methods to construct numerical models for areas with no sewer network data, based on the SWMM model, a two-dimensional TELEMAC-2D model, and the coupled model TSWM. Using a case study of the Liede River basin in Guangzhou, we compare the results with the scenario of sewer network data available, and discuss the applicability of different generalization methods. The results show that the flows of one-dimensional simulations and the submergence depths of two-dimensional simulations by the models with a sewer network basically agree with the measurements, demonstrating a good reliability of the models for the study area. Our simulations of the scenario with the sewer network for different return periods of rainfall are correlated satisfactorily with the inundation extents simulated by the rainfall discount method, the infiltration increase method, and the pipe network generalizing method. Of the three methods, the infiltration increase method has the largest correlation coefficient and its inundation extent and depth are the closest to the scenario with the sewer network, showing its better applicability to highly urbanized areas with no sewer network data.

Key words: urban stormwater, no sewer network data, TSWM model, infiltration increase method, Liede River basin