Multi-time Scale Characteristics And Regionalization Of Monthly Precipitation In The Yangtze River Basin

Precipitation is a complex and non-stationary hydro-climatic variable,which is affected by many factors such as geographical location,surface conditions and atmospheric circulation.Therefore,studying the multi-scale spatial and temporal characteristics of precipitation,identifying uniform precipitation regions,and analyzing the possible influence factors of precipitation are essential for prediction of precipitation and management of water resources.In this study,the ensemble empirical mode decomposition(EEMD)is applied to analyze the periodic characteristics and long-term trend of monthly precipitation based on the monthly precipitation observations at 189 meteorological stations during 1963-2016 in the Yangtze River Basin.Next,a multi-scale information entropy method(EME)based on EEMD is proposed to quantify the stochastic characteristics of monthly precipitation across the different time scales.Then,the spatial categorization of meteorological stations is performed using the fuzzy C-means clustering(FCM)algorithm in the basin.Finally,the lag-time correlation analysis and stepwise variable selection are employed to identify the significant climate factors impacting the periodic oscillations of monthly precipitation in each sub-region.And using the identified large-scale climate factors as the forecasting variables of monthly precipitation,themultivariate linear regression model is established at each sub-region of the basin for predicting monthly precipitation.The results are as follows:(1)The information entropy values of the monthly precipitation in the Yangtze River Basin exist significant heterogeneity along the latitudinal direction,which shows that the entropy values of monthly precipitation increase gradually from west to east.But presents quasi uniformity along the longitudinal direction,which means that the difference of the information entropy values in stations on the same longitude line is very small.At the same time,there is a strong negative linear correlation between the information entropy value and the variation coefficient and elevation,while there is a significant positive linear correlation between the information entropy value and the annual precipitation and the annual precipitation days.(2)The IMF1,IMF2 and IMF3 are the dominant characteristic components of monthly precipitation series of the Yangtze River Basin in the past 54 years,which reflected the seasonal changes with quasi-0.24-year,quasi-0.62-year and quasi-1.02-year,respectively,and are the main cycles of the oscillation period of the monthly precipitation sequence in the basin.The stochastic characteristics of IMF1,IMF2 and IMF3 components of monthly precipitation increase gradually from west to east along the latitudinal direction,and decrease gradually from south to north along the longitudinal direction.In particular,the contribution ofIMF2 to the randomness of monthly precipitation in the basin is the largest,while the contribution of IMF3 to the randomness of monthly precipitation in the basin is relatively small.The IMF4,IMF5 and IMF6 reflect the interannual oscillations of quasi-2.4-year,quasi-4.7-year,and quasi-9.5-year,respectively,while the IMF7 and IMF8 reflect the interdecadal oscillations of quasi-22.1-year and quasi-50.8-year,respectively.The long-term change trend of the monthly precipitation in most of the stations in the basin is remarkable.The Western Sichuan Plateau,the Poyang Lake Basin,the Dongting Lake Basin and the Yangtze River Delta are covered mainly by a significantly increasing long-term trend of monthly precipitation,while a significantly decreasing long-term trend of monthly precipitation occurs mainly in the northern region of the Yangtze River Basin.(3)The monthly precipitation in the Yangtze River basin can be divided into six homogeneous sub-regions: the western plateau region,the Hengduan mountain region,the northern low mountains and basin region,the southern low mountains and hills region,the southeastern Poyang Lake plain region and the eastern Yangtze River Delta region.The monthly precipitation of each sub-region is influenced by the SST over East China Sea(ECS),the SST over Kuroshio(KC)and the SST over South China Sea(SCS)with different time lags.The built monthly precipitation prediction model based on the identified climate factors ineach sub-region can predict the periodic oscillation pattern and the long-term change trend of the monthly precipitation,and the prediction performance of the model is gradually improved from east to west in space.