Spatial distribution of combined risk of precipitation characteristics in Kunming city based on Copula functions

Climate change has led to increased frequency of extreme rainfall events, posing significant threat to urban water security． To assess the spatial distribution of combined risk based on different precipitation characteristics can provide important scientific evidence for prevention and mitigation of extreme rainfall disasters and urban flood control． In this paper the 5-minute rainfall data from 56 rain gauges in urban Kunming from 2019 to 2021 are applied to overdetermined method for precipitation event sampling, to calculate rainfall duration, peak rainfall, and total rainfall． Combined risk probability analysis of precipitation characteristics is done by selecting optimal marginal distribution functions and Copula functions, to quantitatively assess combined risk probability of precipitation characteristics under different representative conditions and their spatial distribution characteristics． Variable of each station in Kunming city largely follows Pearson Type Ⅲ (PE3) distribution, variable follows log-normal (lognorm) distribution, variable P follows Generalized Extreme Value (GEV) distribution, Generalized Logistic (GLO) distribution and Generalized Pareto (GPO) distribution． Peak rainfall P and rainfall duration of all stations are negatively correlated or independent, all following Frank Copula and Independent Copula functions． Clayton Copula and Frank Copula functions can better describe joint distribution of variable combinations (Q₀, Q) and (T, Q) for most stations． In mountainous areas in the eastern part of urban Kunming, rainfall events with longer duration, higher peak rainfall, or greater total rainfall are likely to occur, while the central part of the main urban area is more prone to rainfall events with long duration, high peak rainfall, and large total rainfall．