Synoptic Weather Typing And Precipitation Projection Of Guangzhou Area

As one of the essential economic development areas of South China,Zengcheng-Guangzhou,located in the Pearl River Delta,has been significantly affected by monsoon throughout the year and is severely troubled by extreme rainfall and urban waterlogging.Studies in this paper were based on five kinds of data:1)hourly station observation of Zengcheng from January 1990 to September 2013,2)JRA55 reanalysis data at surface,925hPa,850hPa,700hPa,600hPa and 500hPa,3)the China Meteorological Administration tropical cyclone database,4)surface weather chart from the Japan Meteorological Agency,5)model data from IPSL-CM5A,including historical,RCP2.6,RCP4.5 and RCP8.5 daily outputs.Based on synoptic weather typing and rainfall simulation,rainfall characteristics and regional climate change were analyzed.To begin with,based on principal component analysis,cluster analysis,discriminant analysis,logistic regression and nonlinear regression,CA(Cluster Analysis)synoptic weather typing and daily rainfall simulation were developed.There were 38 weather types in total,which included six typhoon related weather types,corresponding to six typhoon situations,as typhoon located to the east,center,south,north,center(common on Spring and Autumn)of Zengcheng and multiple weather systems.In terms of simulating total precipitation and typhoon related precipitation,daily rainfall simulation was proved to be quite effective of light and medium rainfall,while of heavy rainfall,it still needed to be improved.Then,based on self-organizing map,SOM synoptic weather typing,weather types' climate implication and major rainfall categories were investigated.There were 20 weather types in 4x5 SOM map,which could represent classical winter and summer monsoon patterns,annual cycle,as an anti-clockwise rotation of common weather types along SOM map's edges in different seasons,and three major precipitation patterns,including front category in first-rainy season,typhoon category in second-rainy season and continental cold high in non-rainy season.Afterwards,based on statistical methods,IPSL-CM5A historical,RCP2.6,RCP4.5 and RCP8.5 daily regional-scale modeling data was downscaled into hourly station-scale data.This downscaling method was suitable for weather elements of Zengcheng,having the ability to accurately reflect each weather elements'daily and annual cycles,and to provide possible regional climate changes of three RCPs in 21 century.In RCP2.6 scenario,surface temperature increases about 2? in Zengcheng area and annual frequency of high-temperature events increases to about 60days in the first 50 years of 21 century.In the next 50 years,surface temperature has little changes.During the entire 21 century,humidity condition does not have significant changes.In RCP4.5 scenario,temperature increases and humidity decreases.The changing degree of RCP4.5 is more intense than RCP2.6 and less intense than RCP8.5.In RCP8.5 scenario,Zengcheng experiences the strongest increasing in temperature and decreasing in humidity,which brings more high-temperature days in the end of 21 century.Besides,10m north wind decreases and the Prevailing Westerlies intensity also decreases.Finally,based on the previous downscaled data and two synoptic weather typing methods,possible regional climate changes in precipitation were forecasted.CA weather typing and SOM weather typing provided a similar result,which was in all three RCP scenarios,there would be a hotter and drier tendency,decreasing in rainfall frequencies and typhoon frequencies with less typhoon related rainfall and no significant change in daily rainfall amount.This might cause higher frequencies of extreme rainfall events and disaster events of typhoon and extreme weather.RCP8.5 scenario has the highest possibility that rainfall would occur in hotter and drier weather conditions,followed by RCP4.5 and RCP2,6.From the studies of SOM weather typing rainfall analyses,at the end of 21 century,major rainfall events would most common in hot dry weather type(1,1)rather than warm humid weather type(4,1)under RCP8.5 scenario.Same tendency was also indicated in RCP4.5 scenario with less intense changing degree,while RCP2.6 scenario maintains the previous rainfall characteristic that major rainfall events most common in type(4,1).