Abstract: Using a 2-km-resolution WRF model, along with its single-layer urban canopy model and Thompson bulk-parameterization microphysics scheme, numerical simulation of a convective rainstorm is executed over Guangzhou city in the Pear River Delta (PRD) to investigate the impacts of urban environments (including urban land use changes and urban air pollution-enhanced cloud droplet number concentration) on the convective precipitation. Results indicate that urban heat island and dry island effect produced by the land surface can lead to an enhanced urban boundary layer, which is conducive to air confluence and unstable energy increase in the areas near the city. Simulation shows that to the north and south of Guangzhou city, caused by the urban land surface, there exists respectively a convergence zone with enhanced CAPE (Convective Available Potential Energy) values. The simulated radar echoes being initiated within these zones and consistent well with observations reflect that the urban land surface plays a more direct role in convection initiation and development. Once convection develops, sensitive experiments show that enhanced cloud droplet number concentration might lead to more precipitation. In case with high concentration (polluted cases), precipitation is increased by up to 20%. Diagnostic analysis suggests that the increased precipitation results from that more rain water and supercooled cloud water produce in the convective clouds. These rain and cloud water, carried upward by strong upward motion intensified by more latent heat release during this process, can be transported to higher altitudes, where ice formation is enhanced through freezing or through interaction with other already existing ice species, and that finally results in surface precipitation increase.