山美水库流域极端降水时空变化及非平稳性特征

Temporal and spatial variation and non-stationary characteristics of extreme precipitation in the Shanmei reservoir basin

为探究极端降水的时空变化及其非平稳性特征,基于1972—2010年山美水库流域8个降水站点逐日降水数据,选取9个极端降水指数,利用Pre-Whitening Mann-Kendall(P-WM-K)方法分析流域极端降水的时空变化特征,并采用广义可加模型(generalized additive models for location,scale and shape,GAMLSS)检测极端降水的非平稳性特征。结果表明:在时间变化趋势上,极端降水频率指数中雨日数R_(10mm)和大雨日数R_(25mm)呈下降趋势,暴雨日数R_(50mm)呈上升趋势,且上升趋势达到0.05显著性水平;除降水总量P _(RCPTOT)外,其余强度指数呈上升趋势,且均达到0.05显著性水平,其中极端降水量R_(95p)线性倾向率达到30.5 mm/(10 a);在空间差异上,R_(50mm)和极端降水强度指数在流域东南部呈现上升趋势,且上升趋势显著,西北部P_(RCPTOT)下降较明显;R_(10mm)和R_(25mm)呈平稳特征,R_(50mm)全流域约50%的站点呈现非平稳特征,且以均值非平稳为主,除P_(RCPTOT)外,其余强度指数均以非平稳特征为主,且主要表现为均值非平稳。未来山美水库流域极端降水量和不确定性增加,灾害风险增大。

The probability distribution of extreme precipitation in the Shanmei reservoir basin is expected to change due to the dual influence of global climate change and human activities,showing non-stationary characteristics.Therefore,investigating the temporal and spatial trend characteristics and the non-stationarity of the extreme precipitation are valuable for policy decisions.Based on the daily precipitation data of 8 meteorological stations in the Shanmei reservoir basin from 1972 to 2010,9 extreme precipitation indices including 6 intensity indices and 3 frequency indices were used to describe the extreme precipitation characteristics.The Pre-Whitening Mann-Kendall(P-WM-K)method was adopted to analyze the temporal and spatial trend changes of extreme precipitation,and the generalized additive models for location,scale,and shape(GAMLSS)was employed to characterize the non-stationarities in the 9 indices in the Shanmei reservoir basin.The number of moderate precipitation days(R_(10mm))and the number of heavy rain days(R_(25mm))showed a downward trend while the number of very heavy rain days(R_(50mm))showed an upward trend with a 0.05 significant level.Except for the total precipitation(P_(RCPTOT)),the other intensity indices(daily intensity(S_(DII)),very wet day precipitation(R_(95P)),extremely wet day precipitation(R_(99P)),max 1-day precipitation(R_(X1day))and the max 5-day precipitation(R X5day))increased significantly.The linear trend rate of R_(95p)reached 30.5 mm/(10 a).The extreme precipitation indices had abrupt changes in the whole basin,and the mutation years mainly occurred in the late 1980s and early 1990s.In terms of spatial variation,R_(10mm) and R_(25mm) showed a download trend for all the stations,and reaches a 0.05 significant level in Dazhong and Zimei Stations.R_(25mm) showed a downward trend for the stations in the exception of the Dongguan and Shanmei Stations while the trend was not significant.The R_(50mm) and extreme precipitation intensity indices showed an upward trend in the southeastern part of the basin(Yongchun Station,Huyang Station,Dongguan Station,and Shanmei Station),and the trend was significant.The P_(RCPTOT) had decreased significantly in the northwest of the basin.The GAMLSS model was fitted well to each precipitation station,and the Filliben coefficient passed the 0.05 significance level.R_(10mm) and R_(25mm) showed stationary characteristics.The R_(50mm) showed non-stationary characteristics at 50%of the stations in the basin,and that was dominated by mean non-stationary.Except for P_(RCPTOT),the other intensity indices mainly showed non-stationary characteristics which mainly manifested as the mean non-stationary.Under the non-stationary conditions,the mean and variance of the extreme precipitation indices was mainly characterized by an increasing trend in the southeastern basin,and mainly showed a decreasing trend in the northwest.In light of climate variability and anthropogenic activities,the stable environment of extreme precipitation events in the Shanmei reservoir basin had been disturbed.The intensity of extreme precipitation and R_(50mm) increased significantly in the Shanmei reservoir basin,especially in the southeastern part of the basin.This showed non-stationary characteristics in the study area.The intensity of extreme precipitations as well as uncertainty would increase,which may lead to more related disasters in the future.