Research On Temporal-Spatial Variations Of Extreme Precipitation In Central Asia And Its Return Period

Under global warming background,the global precipitation has increased.Extreme precipitation is a special case of precipitation,its frequency has increased and intensity has also intensified,which easily leaded secondary disasters such as floods,drought disasters,landslides,and debris flows,disrupted ecological balance,threatened people's lives and property safety,affected the sustainable development of economy and society.As the core area of the “One Belt and One Road”,Central Asia(CA)is very sensitive to changes in precipitation due to its special geographical location and natural economic environment,and is easily affected by extreme precipitation events.Therefore,it is necessary to reveal the temporal and spatial variation of precipitation especially extreme precipitation in CA,then to analyze and predict the characteristics of return level of extreme precipitation in different return periods,which is of great significance for the disaster prevention and mitigation work for the region,the rational use of water resources and the economic sustainable development of society.Based on the observations data and reanalysis data of precipitation in CA,the present study revealed the temporal-spatial characteristics of the annual and seasonal precipitation and extreme precipitation in Central Asia from 1979 to 2005.The applicability of two reanalysis datasets(ERA-Interim and NECP/CPC)in describing precipitation and extreme precipitation in CA was analyzed by calculating their correlation and root mean square error.Then the reanalysis dataset with better performance was used to reveal the temporal-spatial changes for annual and seasonal precipitation and extreme precipitation in CA in the past 36 years(1982-2017).The Extreme Value Theory was used to analyze the characteristics of the extreme precipitation return period in CA.The main conclusions are as follows:(1)The annual and seasonal(except for autumn)precipitation increased in CA during 1979-2005,and significantly in summer and winter.Seven extreme precipitation indices showed increasing trends including consecutive wet days(CWD),max 1-day precipitation amount(Rx1day),max 5-day precipitation amount(Rx5day),number of heavy precipitation days(R10),precipitation on very wet days(R95P),annual tot al wet-day precipitation(PRCPTOT)and simple daily precipitation intensity(SDII).CWD increased significantly,while the consecutive dry days(CDD)indicated decreasing trends.The spatial distribution of trends in annual precipitation and extreme precip itation seasonal series in CA showed "more in south-central and northern and less in two sides,and more in western but less in eastern".The high-value areas of precipitation and extreme precipitation were mainly located in the north of KG,TI and KZ,and the low-value areas were mainly located in XJ.(2)By analyzing the applicability of two reanalysis datasets(ERA-Interim and NECP/CPC),it indicated that the NCEP/CPC dataset had a higher correlation coefficient with the observation data regardless of the annual precipitation or extreme precipitation with smaller root mean square error and was better to express the characteristics of precipitation and extreme precipitation in CA.(3)The annual and seasonal precipitation series in CA showed increasing significant trends(except for autumn)of 14.40 mm/10 a,4.10 mm/10 a,4.80 mm/10 a,3.00 mm/10 a and 3.70 mm/10 a during 1982-2017,respectively.The tendency rates of CDD,CWD,Rx1 day,Rx5day,R10,R95 P,PRCPTOT and SDII in CA were-3.10 d/10 a,0.20 d/10 a,0.80 mm/10 a,1.90 mm/10 a,0.40 d/10 a,8.70 mm/10 a,14.60 mm/10 a and 0.40(mm/d)/10 a.CDD decreased while the other indicators increased significantly except for SDII increased insignificantly.(4)The annual,seasonal and extreme precipitation showed increasing trends in most parts during 1982-2017.For precipitation,the trends were significant in Xinjiang(XJ),northern Kyrgyzstan(KG),northern and northwest of Kazakhstan(KZ),while decreased in southern KG and Tajikistan(TI).For extreme precipitation,it increased faster in XJ,northern KZ and northern KG,while decreased significantly in southern KG and TI.(5)In the past 36 years,the areas where precipitation and extreme precipitation increased significantly(XJ)were the areas with less precipitation and less extreme precipitation.The areas where precipitation and extreme precipitation decreased significantly(KG and TI)were the areas with more precipitation and extreme precipitation.It indicated the fact of "dry gets wetter,wet gets drier" to a certain extent.(6)By comparing two methods with Extreme Value Theory used to fit the extreme precipitation,the generalized extreme value distribution(GEV)distribution function was more suitable in fitting the annual maximum(AM)series.There were significant spatial differences of precipitation extremes in different return periods generally with a "more in northwest and northern,while less in southeast" pattern.With the increasing of return periods,the daily extreme precipitation in CA also increased and the difference between high-value centers and low-value centers were widened.High-value centers mainly located in northwestern and northern of KZ and western of Turkmenistan(TX).And extreme precipitation in XJ was less.