Prediction of River Pollution Under the Rainfall-Runoff Impact by Artificial Neural Network: A Case Study of Shiyan River, Shenzhen, China

Climate change and rapid urbanization have made it difficult to predict the risk of pollution in cities under different types rainfall. In this study, a data-driven approach quantify effects rainfall characteristics on river was proposed applied case study Shiyan River, Shenzhen, China. The results indicate that most important factor affecting is dry period followed by average intensity, maximum 10 min, total amount rainfall, initial runoff intensity. addition, an artificial neural network model developed event mean concentration (EMC) COD based correlations between EMC. Compared light rain (< mm/day), predicted EMC five times lower heavy (25–49.9 mm/day) two moderate (10–24.9 mm/day). By converting chemical oxygen demand river, load non-point-source estimated be 497.6 t/year (with accuracy 95.98%) River typical characteristics. can used guide urban rainwater utilization engineering design Shenzhen. findings also provide insights for predicting rainfall-runoff developing related policies other cities.