Spatial and temporal variation characteristics and frequency analysis of extreme precipitation from 1959 to 2017: A case study of the Longtan watershed, southwest China

Abstract

Abstract Extreme climate events represented by extreme precipitation seriously constrain the sustainable and healthy development of economy and society. In this study, based on 59-year precipitation data from 15 meteorological stations in the Longtan watershed, seven extreme precipitation indices (continuous drought index (CDD), consecutive wet days (CWD), maximum 1-day precipitation amount (RX1D), maximum consecutive 5-day precipitation amount (RX5D), precipitation on very wet days (R95p), simple daily intensity index (SDII) and heavy precipitation days (R50 mm)) were selected from three perspectives of duration, frequency and intensity, and then the spatio-temporal variation characteristics of Longtan watershed extreme precipitation were clarified. The results indicated that the indices related to the frequency or intensity of extreme precipitation basically had a non-significant increasing trend in the past 59 years, but the CWD showed a significant decrease, reflecting a trend of drought in the watershed as a whole. Spatially, three spatial variation types of extreme precipitation were observed by the empirical orthogonal function (EOF) analysis, that was, the variation of consistent in the whole region, southwest–northeast reverse phase, and partial-whole reverse phase, among which the first type was the most typical. Frequency analysis was also performed with nine distribution models, and from the results of the Kolmogorov–Smirnov (K–S) test, the generalized extreme value (GEV) model was found to fit each index series better. At different levels of return periods, the indices for RX1D, RX5D and R95p roughly showed an increasing trend from upstream to downstream, with large differences in spatial precipitation.