Study on the Early Warning for Flash Flood Based on Random Rainfall Pattern

Flash floods cause great harm to people's lives and property safety. Rainfall is the key factor which induces flash floods, and critical rainfall (CR) is the most widely used indicator in flash flood early warning systems. Due to the randomness of rainfall, the CR has great uncertainty, which causes missed alarms when predicting flash floods. To improve the early warning accuracy for flash floods, a random rainfall pattern (RRP) generation method based on control parameters, including the comprehensive peak position coefficient (CPPC) and comprehensive peak ratio (CPR), is proposed and an early warning model with dynamic correction based on RRP identification is established. The rainfall-runoff process is simulated by the HEC-HMS hydrological model, and the CR threshold space corresponding to the RRP set is calculated based on the trial algorithm. Xinxian, a small watershed located in Henan Province, China, is taken as the case study. The results show that the method for generating the RRP is practical and simple, and it effectively reflects the CR uncertainty caused by the rainfall pattern randomness. All the Nash-Sutcliffe efficiencies are greater than 0.8, which proves that the HEC-HMS model has good application performance in the small watershed. Through sensitivity analysis, (0.5, b(max)), (r, b(max) < 0.5), and (r, b(max) > 0.5) are identified as key, safe, and dangerous rainfall patterns, respectively. The proposed early warning model is effective, which increases the forecast lead time and reduce the omissions rate of flash flood early earning.

Automated summary

Flash floods pose significant risks to lives and properties. A new early warning model based on random rainfall pattern generation has been proposed to improve prediction accuracy and reduce missed alarms. The model proved to be effective in increasing forecast lead time and reducing omissions rate for flash flood early warning.