气候变化背景下海河流域极端降水特征及不同重现期降水量估计OA北大核心

With the ongoing intensification of global warming,extreme precipitation events have become more frequent,exerting dramatic impact on socio-economic development and poses a severe risks to lives and property in the Haihe River basin.This study analyzes the spatiotemporal variability and statistical characteristics of extreme precipitation in the basin under climate change,utilizing daily precipitation observation from 1961 to 2022.Extreme precipitation events are defined using the annual maximum precipitation(AM)series and peak-over-threshold(POT)series.The results show that the spatial distribution patterns of the multi-year average AM and POT series are similar,with maximum daily precipitation primarily concentrated west of the Taihang Mountains and south of the Yanshan Mountains.Variability analysis reveals that annual maximum daily precipitation at most meteorological stations ranged between 0 and 50 mm,with the highest variability along the Taihang and Yanshan Mountains ranges,where the standard deviation reaches approximately 40 to 50 mm.Furthermore,trends in the an-nual maximum precipitation show spatial heterogeneity:among the 133 meteorological stations analyzed,54 exhibit increasing trend,whereas 79 stations showed a decreasing trend over the study periods(1961-2022).On a station-specific scale,the majority of extreme daily precipitation events occurred during the flood season(July to August)in the 1960s and 1970s. To model extreme precipitation,various extreme value distribution function,including the Generalized Ex-treme Value(GEV),Generalized Pareto(GP),and Gamma distributions,were evaluated using the L-moment method and the Kolmogorov-Smirnov(K-S)test.The findings demonstrated that the GEV distribution effectively models the AM series,while the GP distributions provides an optimal fit for POT series.Furthermore,comparative analysis of precipitation estimates across different return periods suggests that the AM series,in conjunction with the GEV distribution provide a better representation of extreme precipitation in the Haihe River basin.To validate the model performance,three historical extreme rainfall events,each independently assessed as exceeding a 100-year return period threshold,were selected.The GEV based AM extreme exhibited smaller relative error(ranging from 6％and 11％)compared to the GP based POT estimates,further confirming the superior performance of the GEV distribution in simulating extreme precipitation.These findings have important implications for disaster risk assessment,flood mitigation strategies,and sustainable socio-economic development in the Haihe River basin.