Risk Assessment of Dynamic Diffusion of Urban Non-Point Source Pollution Under Extreme Rainfall

With the acceleration of urbanization, the diffusion mechanism of urban non-point source (NPS) pollution caused by extreme rainfall is not clear, which leads to high cost and difficulty in water environment treatment. In view of the shortcomings of dynamic diffusion simulations of mesoscale pollution, this paper proposes a simulation framework based on cellular automata, GIS geographic technology, and a two-dimensional shallow water model. Taking the 500 m × 500 m grid as the unit, we explore the migration laws of nitrogen and phosphorus pollutants and the response relationship between pollutant diffusion and land use under extreme rainfall scenarios. The results show that (i) the pollution risk increases significantly with diffusion, with the maximum pollution load in high-risk areas increasing by 181%, and the diffusion rate is positively correlated with the rate of change in rainfall intensity; (ii) forest land has the highest grid pollution load loss rate, whereas the water grid has the highest accumulation rate; (iii) this method can accurately identify the hot spots of pollution diffusion, providing a basis for the precise control of high-risk areas. This study can support the targeted governance of pollution sources and land planning optimization in urban storm and flood management, and help reduce environmental health risks in extreme climates.