Research on the Critical Rainfall of Flash Floods in Small Watersheds Based on the Design of Characteristic Rainfall Patterns

Rainfall is one of the main causes of flash floods in mountainous watersheds. The critical rainfall, calculated by the rainfall patterns which represent typical rainfall processes, is an important index for the early warning of flash floods. However, due to the randomness and diversity of the rainfall process, the traditional single rainfall pattern is inconsistent with the actual diversified rainfall process, which brings great challenges to the early warning of flash floods. This paper proposes three characteristic parameters to describe the temporal distribution characteristics of typical rainfall processes. On this basis, the appropriate cumulative distribution functions (CDFs) are chosen to fit the cumulative rainfall-duration curves corresponding to typical rainfall processes, and the probability density functions (PDFs) can be used to represent the characteristic rainfall patterns. The HEC-HMS hydrological model is then used to simulate the rainfall-runoff process, and the critical rainfall corresponding to different characteristic rainfall patterns is calculated with a trial algorithm. The results demonstrate that: (1) the critical rainfall calculated by the designed characteristic rainfall patterns is more accurate than that calculated by the traditional rainfall pattern (TRP), proving that the design method of characteristic rainfall patterns proposed in this paper could increase the accuracy of early warning of flash floods in small watersheds. (2) The effects of the antecedent soil moisture condition (ASMC) and rainfall pattern on the critical rainfall are analyzed, which provides a valuable reference for the analysis and calculation of the critical rainfall of flash floods in small watersheds.