Reproduction of Rainfall-Tidal Scenarios and Flood Risk Analysis Based on Representative Historical Rainfall Profile

Abstract Coastal cities affected by tidal dynamics frequently experience heavy rainfall and pluvial flooding. Accurate assessment of urban flood risk and the enhancement of hydrological early warning capabilities are therefore critically important. This study proposes the identification of seven localized representative rainfall patterns based on the distribution of historical rainfall profiles, which, combined with regional empirical formulas for extreme precipitation, are used to generate urban design rainfall scenarios for flood risk analysis. Furthermore, a representative stochastic rainfall event database is constructed and superimposed with tidal processes to reproduce rainfall and tidal scenarios based on historical distribution characteristics. These scenarios are then used to drive an urban rainfall-runoff model to simulate flood responses and assess risks, while also verifying the feasibility of replacing future rainfall forecast inputs with the stochastic rainfall event database. Simulation results show that some river water levels are affected only by rainfall, while others are influenced by both rainfall and tidal interactions. Significant differences in urban river channel responses are observed under different rainfall patterns. In particular, short-duration intense rainfall events (within 180 minutes) require heightened attention. As rainfall duration increases, peak water depths exhibit a “high–low–high” variation trend. Analysis of 1,000 stochastic rainfall events found that 225 threshold exceedance incidents were mainly concentrated in rainfall patterns 1, 4, and 6, which accounted for more than 80% of exceedances, with pattern 1 having the highest exceedance rate and thus requiring prioritized focus. The results of this study provide more concrete technical support for flood risk assessment and for improving the efficiency and reliability of flood forecasting in coastal cities.