| In the current engineering practice and related research,most of the extreme value sequences are assumed to be stationary,and the stationary extreme value model with constant parameters is adopted by default.In recent years,extreme precipitation events have occurred frequently around the world,and the stable distribution model based on constant parameters has gradually been questioned.Studying the temporal and spatial evolution characteristics of extreme precipitation,and establishing a parameterized non-stationary extreme value model to analyze the non-stationary extreme precipitation frequency characteristics under changing climate conditions,which has profound theoretical and practical value for regional water resources management and flood control.In order to discuss the non-stationary characteristics of extreme precipitation in Henan from 1957 to 2018,this paper used linear trend regression,MK mutation test,Bayesian analysis,wavelet analysis,MK trend analysis,and empirical orthogonal function(EOF)to analyze the extreme precipitation series and the temporal and spatial evolution characteristics of extreme precipitation indicators.The main conclusions are as follows:(1)The extreme precipitation in southern,eastern,and northern Henan showed an increasing trend,while the central and western regions showed a decreasing trend.The mutation point was concentrated in the 1980s and 1990s,and spatially distributed in the east,south and west of Henan.The Henan extreme value sequence has short-period oscillations of 2-6a,medium-and long-period oscillations of 14-16a,and long-period oscillations of 28-31a.(2)The rainy-day precipitation intensity and the contribution rate of heavy precipitation increased significantly in eastern and southern Henan,indicating that extreme precipitation accounted for an increase in the proportion of total precipitation.And the situation gradually weakened from the east and south of Henan to the west.EOF analysis results of RX3D and R20 indicators shows that:The spatial distribution of the RX3D index can be decomposed into 2 modes:a uniform change pattern across the entire region,and a reverse phase pattern from the south to the northwest.There is no significant trend change in the time coefficient series of the two modes.R20 index can be decomposed into 3 modes:2 global consistent changing modes and a southern-northern anti-phase mode.The time coefficient sequence of the second mode with global consistent changes shows a significant upward trend,while the remaining modes have no significant trends.(3)In the probabilistic fitting of the non-stationary GEV model,the non-stationary GEV model obtained a better fitting effect than the stationary GEV distribution among the 13 station sequences of the 19 station sequences.In comparison with the goodness of fit of the P-III model,the three goodness of fit indicators of RMSE,R and Dn of the GEV model are better than those of P-? model.(4)The analysis conclusions of non-stationary frequency characteristics are as follows:?The spatial distribution of the recurrence levels in different regions of Henan was significantly different,and the overall distribution presented a spatial trend of decreasing from the south to the northwest.?In the spatial distribution of extreme precipitation under different return periods,the recurrence level of northern Henan under the large return period was quite different,and the nature of the thick tail was more significant.?Taking the recurrence level of the P-III model under the stationary assumption as the benchmark value,the extreme precipitation performance in 2019 under the climatic conditions showed that the increasing trend was concentrated in the northern and southern Henan,and the decreasing trend was concentrated in the Yellow River basin in western Henan.?In non-stationary sites with significant trends or points of change,the recurrence level of the GEV model was quite different from the baseline value.?The estimation of extreme precipitation under the larger return period level of the GEV model better reflected the fluctuating impact of climate change on the return level of extreme precipitation and solved the contradiction between safety considerations and cost considerations in design storm calculations under changing climate conditions... |
PDF Full Text Request