ON THE PROCESS OF SUMMER MONSOON ONSET OVER EAST ASIA

Vol.15 No. 4ACTA METEOROLOGICA SINICA2001ON THE PROCESS OF SUMMER MONSOON ONSET OVEREAST ASIA'CHEN Longxun (陈隆勋). LI Wei (李薇). ZHAO Ping (赵平)Chinese Academy of Meteorological Sciences. Beiing 100081and TAO Shiyan (陶诗言)Institute of Atmospheric Physics, Chinese Acadermy of Sciences. Bejjng 100029Received May 10, 2001; revised August 14. 2001ABSTRACTUsing daily observational rainfall data covered 194 stations of China from 1961 to 1995 andNCEP model analyzed pentad precipitation data of global grid point from 1979 to 1997， thedistribution of onset date of rainy season over Asian area fron spring to summer is studied in thispaper. The analyzed results show that therc exist two stages of rainy season onset over East Asianregion from spring to summer rainy season onset accompanying subtropical monsoon and tropicalmonsoon respectively. The former rain belt is mainly formed by the convergence of cold air and therecurved southwesterly flow from western part of subtropical high and westerly flow from the so-called western trough of subtropical region occurring during winter to spring over South Asia. Thelatter is formed in the process of subtropical monsoon rain belt over inshore regions of South ChinaSea originally coming from south of Changiang (Yangtze) River Basin advancing with northwardshift of subtropical high after the onset of tropical monsoon over South China Sea. The pre- floodrainy season over South China region then came into mature period and the second peak of rainfallappeared. Meiyu, the rainy season over Changjiang Huaibe River Basin and North China thenformed consequently. The process of sutmer tropical monsoon onset over South China Sea in 1998is also discussed in this paper. It indicated that the monsoon during summer tropical monsoononset over South China Sea is the result of the westerly flow over middle part of Soutb China Sca,which is from the new generated cyclone formed in north subtropical high entering into SouthChina Sea， converged with the tropical southweslerly flow recurved by the intensified cross-equatorial flow.Key words: subtropical monsoon rainy season, tropical monsoon rainy scason, summer nonsoononset, East AsiaI. INTRODUCTIONIn 1987, Tao and Chen (1987) put forward a figure of distribution of the onset date ofthe Asian summer monsoon by synthesizing the previous result of Chinese and Indianscholars. According to the result shown in the figure. the earliest onset of the summermonsoon started over the South China Sea (SCS). The result was recarded as one of themain bases for the South China Sea Monsoon Exp中国煤化工ich was carriedTYHCNMHG， This paper is supported by the State Ministry of Science and Technology under the key project A ofSouth China Sea Monsoon Experiment.CHEN Longxun. LI Wei, ZHAO Ping and TAO Shiyan437out in 1998 (PRC-US scientific Working Group 1995). Later， the issue of the seasonalvariation of the Asian summer monsoon has been further studied by many scholars (Lauand Yang 1997; Matsumoto 1997; Webster et al. 1998). They indicated their own onsetdate of the rainy season of the summer monsoon over East Asia according to different datasuch as OLR，wind and rainfall amount, and conforma bly indicated that the earliest onsetof the rainy season started over the Indo-China Peninsula, while the onset date of the rainyseason over the Indo-China Peninsula are different from each other with the onset datechanging from 28 April < Matsumoto 1997) t0 10 May (Lau and Yang 1997; Webster etal.，1998). In another respect. the climatological onset date of the summer monsoon overSouth China Sea has been studied by many Chinese scholars (Xie et al. 1996: Yan 1997;Chen et al. 1999; Liang and Wu 1999). Nearly conformably. the onset date of thesummer monsoon over SCs was defined in the fourth pentad of May although the methodand data were different. Therefore， the onset of the summer monsoon over Indo-ChinaPeninsula was almost 10- 20 days earlier than that over the SCS although the two regionsare different only 5一10 degrees in longitude. the reason is necessary to be explained.Chen et al. (1991) indicated that there existcd tropical monsoon and subtropical one in themonsoon system over East Asia， and that the rainfall showed a pattern of bimodaldistribution over the region from Hainan Island to South China，on an average, the firstpeak occurred in mid-April while the earliest could occur in March. this is the time of thestarting period of pre-flooding rainy season over South China or so-called "Meiyu" inTaiwan region. Till the onset of the tropical summer monsoon, the rainy season of thepre-looding season over South China reached peak time and proceeded northward. Thuswe should study whether the onset of the rainy season over the Indo-China Peninsula andthe onset of the rainy season of April over South China belong to subtropical rainy season,and whether the rain belt extended from South China to Indo-China Peninsula as well aswhether it was caused by the convergence of cold air and the recurved southwesterly flowfrom western part of subtropical high. The result of the paper showed that there existedtropical and subtropical rainy season onset over East Asian region. The rainfall over manysubtropical regions showed a pattern of bimodal distribution. In early April, a rain beltwas formed by the convergence of cold air and the recurved southwesterly flow fromwestern part of subtropical high and westerly flow from the so-called western trough ofsubtropical region which occurred during winter 10 spring over South Asia. After the startof rainy season over north of South China or south of Changiang River Basin, the rain beltextended southwestward to Indo- -China Peninsula and southward to north of SCS, this wasthe first time of onset of the rainy season. but is over subtropical region. Then came theonset of the rainy season over the tropical region. The rain belt of the tropical monsoonmoved northward rapidly to 10- 15°N from equatorial region after the onset of the tropicalmonsoon over SCS. while the rain belt of subtropical monsoon over the Indo-ChinaPeninsula turned to a tropical one. The onset of thl中国煤化工h China movednorthward with the northward movement of subtro|YHCNMHGtheprocessofnorthward movement of the rain belt over South China after mid-May described in thefigure of Tao and Chen <1987). In addition. the onset of the summer monsoon in 1998 wasalso discussed in the paper.438ACTA METEOROLOGICA SINICAVol.15II. DATA AND DEFINITION OF ONSET DATE OF RAINY SEASONThe data used in this paper are as follows: (1) Daily rainfall data of 194 stationscovered China from 1951 to 1995 and year 1998; (2) TBB/GMS data from 1980 to 1998with resolution of 1°X1°; (3) NCEP global pentad grid rainfall data，with resolution of2.5°X2.5°，1979- 1997; and (4) NCEP/NCAR reanalysis 850 hPa wind and objectiveanalysis data of SCSMEX.The onset date of rainy season was defined as follows:Cv= (R(N)/(R(365)}.V65where relative rainfall coefficient GN denotes the ratio of daily mean rainfll during thecalculated period of N days to annual mean of daily rainfall. If the coefficients Cs. C1o, C15are all greater than Cx after a medium rain with daily rainfall ≥Rx, then the date whenmedium rain occurred is defined as rainy season onset date, whereR(5)R(365)C3=()/R(10) }C2o=10)/(365Cr={R(152)2)/(19China is a country with a vast territory, and the climate is very different. It is apparentlyunreasonable to define the onset of the rainy season over different regions by a uniformcriterion. Therefore in this paper， Rx,Cx are defined as follows for different kind ofterritory :(1) for western China: Rx=1.5,Cx=1.2;(2) for South China: R.=10. 0,Cx=1.5;(3) for North China and Northeast China: R,= 10.0,Cx=1.8;(4) other regions: Rx=3, 0,C;-1.5.The distribution of the onset of the rainy season over China (Fig. 1) defined by theabove method could generally show the feature of different climate. For the stationsbeyond boundary of China, only the pentad rainfall data are available, thus the onset ofthe rainy season over these regions is defined directly by the criterion as that the ratio ofpentad rainfall to annual rainfall is greater then 2%.II CLIMATE DISTRIBUTION OF THE ONSET DATE OF RAINY SEASON OVER ASIAIn order to compare the distribution of the onset date of rainy season over Asia ,especially over East Asia indicated by different scholars, we collected the results of themtogether in Figs. 1a- 1e, they are the results of: Tao and Chen (1987, Fig. 1a), Lau andYang (1997, Fig. 1b), Lin (1987, Fig. 1c), Matsumoto (1997, Fig. 1d) and Webster etal. (1998, Fig. 1e). It can be seen that the onset中国煤化工over the middlepart of Indo-China Peninsula defined by differentMYHCNMHG10May, 10-15May and 28 April (except for result of Lin 1987), while the onset of the rainy season overthe coastal region of South China is 10 May, 25 May, late May and 2 June respectively(except for the result of Matsumoto). Comparing the two regions, the onset date of theNo.4CHEN Longxun, LI Wei. ZHAO Ping and TAO Shiyan439(b)30 .20.I10.15%7s zn5060708090 10010120130140150E45^Np30°N40 t(d)、20 F35 t10 F30iBgi0➊:Mty 10 o:Juns 2, May 15 @:Juno 7o: May18❽: June 122510*S”0:May23 0: Juns .17l 0: Mty28\红7090110°EApr. May Jun. Jul. Aug. Sep Oct. Nov.30N(e)Fig. 1. Distribution of the onset date of the rainyseason over Asia given by differentscholars: (a) Tao and Chen (1987); (b)Lau and Yang (1997); (c) Lin (1987);(d) Matsumoto (1997); (e) Webster et528al. (1998); The value shown in Fig. 1c is .the ratio of pentad rainfall to annualrainfall of the Btations in the mainlandalong 110- 120°E and in Fig. le, the firstdigit denotes month, the second and thirddigit denotes day.508中国煤化工dMYHCNMHG100110E440ACTA METEOROLOGICA SINICAVol.15rainy season over Indo China Peninsula is about 10- 20 days earlier than that over thecoastal region of mainland. Over the Indo-China Peninsula, the earliest onset was locatedover east shore according to Lau, while over west shore according to Webster et al.(1998) and over middle part according to Matsumoto. All of Figs. 1b, 1d an le regardedthat the earliest onset started over Indo- China Peninsula and then propagated to SCS,however, seen from the curve of rainfall over China mainland that the rainfall of manystations of South China showed a feature of bimodal distribution from spring to summer ,actually, the rainfall over the Indo-China Peninsula shows the same feature as well (seeFig. 2 in the paper of Matsumoto, 1997 ).To find out the onset and the variation of the rainy season over the region along 110-120°E，we performed the time-latitude cross-section of climatological pentad rainfallamount and the ratio of pentad rainfall to annual rainfall respectively (Fig. 2). In the .figure, the result in the domain north of 22°N was obtained by the cumulation of theclimatological daily rainfall of Chinese stations averaged from 1961 to 1995 to pentadamount，whereas in the domain south of 22°N was obtained by NCEP climatologicalpentad rainfall averaged from 1979 to 1995. For the domain near 22°N, a smooth has beenperformed since the two kinds of date were not completely joined. It showed that eitherthe rainfall amount (Fig. 2a) or the ratio (Fig. 2b), had the large value belt occurring overthe region between 25一27°N in early April (the nineteenth pentad), the pentad rainfallwas greater than 30 mm and the rainfall ratio greater than 2%. The belt maintained tillearly April, a center of above 40 mm and 2. 5% formed over 27 一28°N, then movedsouthward and reached 21一22°N of coastal region of South China in the fourth pentad ofApril，after maintained till mid May， it moved northward promptly and reachedChangjiang River Basin by the fourth pentad of June and North China by the fourth pentadof July. Therefore, during the period from mid April to May, the rainfall over regionbetween 23 - 28°N was bimodal, i. e. the first rainfall peak occurred in April, thendecreased, and increased again after mid May - early June. The former rain belt (April-May) propagated southward while the latter (May一July) propagated northward. Basedon this fact, we should distinguish the two kinds of rain belts. The former one shouldbelong to the rain belt of starting period of pre-flooding season over South China, while .42°N{39 (3)(b)P36。330300)24211811743Mar Apr. May Jun. Jul. Aug Scp. Oct.中国煤化工ul Aug. Sep.~ Oct.TYCHCNMHGog 10-12020 .Fig. 2. Time latitude cross section of climatological pentrau rauurass uuuil 1... alongand ratio of pentad rainfall (%): (a) pentad rainfall amount and (b) ratio of pentad rainfallto annual rainfall (%).No.4CHEN Longxun, LI Wei, ZHAO Ping and TAO Shiyan441the latter to the rain belt of peak period of pre- flooding season over South China. Theformer was formed by the convergence of cold air and the recurving southwesterly flowfrom western part of subtropical high and westerly flow from the so-called western troughof subtropical region occurring during winter to spring over South Asia, at this time thetropical summer monsoon over the regions of Asia did not outbreak yet, thus it belongedto the subtropical monsoon rain belt, the latter was formed by the seasonal northwardmovement of the subtropical high after the onset of the tropical summer monsoon. It alsobelonged to subtropical rain belt, but moved northward with the northward movemcnt ofsubtropical high.From above facts, based on the observational data of Chinese stations and NCEPprecipitation grid data, the onset date of the rainy season of subtropical monsoon duringthe period from April to May and the onset date of the rainy season of tropical andsubtropical monsoon after the onset of the tropical monsoon in May were calculated by themethod previously mentioned (Fig. 3). The result in Fig. 3 was mainly obtained based onour calculated onset date with reference of the reonset date of the rainy season over regionsin different part of China defined by different scholars being referenced (Data Center of theMeteorological Bureau of Guangdong Province 1987; Jiang 1991; Wang 1990; Xu 1991;Qian 1991; Bai 1988; Li 1991; Dai 1990; Zhang 1991).Seen from Fig. 3, it is indicated that by the end of March, rainy season has alreadyoccurred over the regions from the south of Changjiang River Basin, i. e. the south part ofZhejiang, Jiangxi, Hunan Provinces and Fujian Province to north of South China，then therain belt extended southward. The rainy scason started from north to south: it started inthe second pentad of April in middle part of Guangdong Province, the third-fourth pentadof April in coastal regions of South China, the fifth pentad of April in Leizhou Peninsula ofChina and the first pentad of May in Hainan Island，therefore the rain belt expanded fromnorth to south. Besides, it expanded southwestward to Guangxi Region. Although there40N{b(a)35| 05~30{.45434125|020{| 6215-↑~5%560 55 54一58°51/10 95100105110115120125130E7080>0100120130EFig. 3. Distribution of the onset datc of rainy season over recions in Asia from spring to summer:(a) distribution of the onset date of rainy scason中国煤化工lonsoon; and(b) distribution of the onset date of rainy season or.YHCNMHGhefirstdigitof numerals at contours denotes month and the sccond denotes pentad. The onset date ofthe summer monsoon over the region wesl of Indo-China Peninsula adopted from IndianScholar. )442ACTA METEOROLOGICA SINICAVol.15was no evidence which indicated that the rainy season could expand to the Indo-ChinaPeninsula from the mainland of China, the rainy season over the region started during theperiod from the sixth pentad of April to the first pentad of May and the date wascompletely same as that of the coastal region of South China. Besides, seen from thedistribution of the onset date of the rainy season, the onset of the rainy season over themainland of China and Indo-China Peninsula was conjoint, so was the rain belt. Figure 4was the climatological distribution of OLR during the period of 26- 30 April and 850 hPastreamline field (adopted from the result of Matsumoto ,1997). Seen from the distributionof OLR, there existed a belt of continuous low value of OLR from South China to Indo-China Peninsula which indicated that the cloud belt was conjoint. Only the cloud south ofit (south of 10°N) belonged to the cloud system of tropical monsoon. On the 850 hPastreamline field (Fig. 3b), three flow areas were divided, i. e. north subtropical monsoonarea, equatorial westerly area and trade wind area. From the Indo-China Peninsula toSouth China, the flow was formed by the recurved southwesterly flow from western partof subtropical high and westerly flow from the so-called western trough of subtropicalregion occurring during winter to spring over South Asia, both of them belonged to thesubtropical flow rather than the tropical one. In the figure of Matsumoto (1997), thepattern characteristics maintained till the second pentad of May. Therefore, the onset ofthe rain belt and rainy season over the region from Indo-China Peninsula to South Chinaduring the period from April to early May was subtropical , not tropical precipitation. Theearliest onset of this subtropical monsoon rain belt occurred over Hunan province andsouth of Jiangxi Province of mainland of China in early April, not over Indo-ChinaPeninsula，It was unreasonable to take Indo-China Peninsula as the region where theearliest onset of tropical monsoon rainfall or summer tropical monsoon rainfall took place.Seen from Fig. 4a，another two rainfall areas are located over the northeastern andsoutheastern Qinghai-Xizang (Tibetan) Plateau respectively. The rainy season over thenortheastern part belonged to spring rainfall while the one over the southeastern part wascaused by Indian westerly trough and should belong to subtropical rainfall (Li and Chen2001).Figure 3b is the distribution of the onset date of rainy season over the regions afterthe onset of tropical summer monsoon. Note that after the onset of tropical summermonsoon, the rain belt over the mainland of East China was still the rain belt of north ofsubtropical high ，therefore belonged to subtropical monsoon rain belt ，however it movednorthward with the northward movement of the subtropical high. The rain belt south ofthe subtropical high was exactly the tropical monsoon rain belt. The most dominantfeature of the onset of tropical summer monsoon over SCS was the retreat of thesubtropical high from Indo-China Peninsulaor even from SCS，the continuous rainfalloccurring over the SCS after the onset of the tropical summer monsoon over SCS should bedefined as tropical monsoon rain belt. We can中国煤化工et of the tropicalsummer monsoon over SCS in mid-May, the equMYHCN MH Gaorthward to 15-20°N，later the effect on the Indo-China Peninsula " turnea coming from subtropicalmonsoon flow to coming from tropical monsoon flow, the rainfall intensified once andprogressed westward to the Bay of Bengal. From the north Calcutta of India toNo.4CHEN Longxun. LI Wei, ZHAO Ping and TAO Shiyan44310EQ20^N(6)--八Fig. 4. Distribution of the climatological OLR (a) and 850 hPa streamline (b) during the periodof 26- 30 April given by Matsumoto (1997).Xishuangbanna of Yunnan Province of China via Burma, the onset date of rainy seasondefined by the Meteorological Station of Xishuangbanna was the first pentad of June.Therefore, the onset of the tropical monsoon rainy season propagated from east to west.The subtropical rain belt previously located over the inshore South China movednorthward with the northward movement of subtropical high after the onset of the tropicalmonsoon over SCS. It progressed northward from south since late May and arrived overthe middle and low reaches of Changjiang River Basin by mid June. As a matter of fact，before the mid June, during April to May, the rainy season over the Changjiang River hasalready started although the rainfall was not strong. According to Fig. 2a, the pentadrainfall along 110- 120E was less than 40 mm，however the intensity of rain beltenbanced after the onset of tropical summer monsoon with its value greater than 40 mm.The rainy season over the Tibet also progressed westward and northward with the onset ofthe tropical summer monsoon in early June (Li and Chen 2001).From the above discussion, there was no any discussion of the onset of summermonsoon over SCS but only the discussion of the process of the tropical monsoon over thetropics after the onset of the summer monsoon over sCS and the northward movement ofthe subtropical monsoon over the mainland of China, thus the conclusion of them wasbasically correct.IV. ONSET OF RAINY SEASON OVER ASIA IN 1998 .Many Chinese scholars defined that the onset中国煤化工over SCS of1998 is in the period of 19- 23 May. The date isa li:THCN MH Ge normal. Forthe reason that the rainfall data of Indo-China Peninsula in 1998 are not available, it isimpossible for us to define the onset date of rainy season over that region, instead wecould define the onset of rainy season over the regions by TBB/GMS data. For the sake of444ACTA METEOROLOGICA SINICAVol.15it, the figure of distribution of mean TBB every three days during the period of 4- - 27 Maywas constructed. For the convective cloud belt only occurred near the equator during thetime of 4- 15 May and the distribution was similar, only the figure from the period of 10-27 May was given (see Fig. 5). During the period from 10一12 May (not shown), asubtropical cyclone developed rapidly over the south boundary of India. By 13- 15 May, aconvective cloud cluster appeared over west of 90°E and south of 13°N. During the sameperiod, a cloud belt of cold front occurred over the coastal region of South China. By 16-18 May, the center of cloud cluster over the Bay of Bengal moved northward to (15°N,90°E), while the cloud belt over the coastal region of South China extended westward tothe east and middle part of Indo-China Peninsula. By 19- 21 May, a little convective cloudcluster still stayed over the East Bay of Bengal while the cloud belt over the South Chinamoved southward and reached northof SCS (15- 20°N) and east of Indo-China Peninsula.This indicated that the rainfall over the SCS started over the north part in 1998，it wasapparently formed by the north coming subtropical cloud belt. We can see that theconvective cloud cluster over the Bay of Bengal only maintained 3 days (16- 18 May) overthe West of Indo- China Peninsula, afterwards, weakened and disappeared on 19- 21 May.Therefore it is unreasonable to define 16- 18，May as the onset date of tropical sunmermonsoon over Indo-China Peninsula. By 22- 24 May, there was nearly no convectivecloud over Indo China Peninsula while the north coming cloud belt over the sCs split intotwo parts, the cloud cluster over the sCS continuously moved southward to 10- 15°N,this was the period of onset of the tropical monsoon over SCS. During the period theprevious convective cloud cluster over the Bay of Bengal disappeared, an equatorial cloudcluster maintained over the south of 5°N，no cloud cluster occurred over 10- 15°N. So itcan be seen from it that, in terms of tropical cloud cluster during the onset period oftropical summer monsoon in 1998, it occurred earliest over the sCS, rather than over theIndo- China Peninsula and the Bay of Bengal. It is not until 25一27 May that the convectivecloud cluster over the mid SCS intensified, the cloud cluster over the region from Indo-China Peninsula to east of the Bay of Bengal appeared, and the two started to connect.Therefore，a tropical convective cloud belt occurred over the region of 10 - 15°N.95- 125E.From the above discussion, it can be found that the convective cloud belt during theperiod of the onset of the tropical summer monsoon over SCS in 1998 was formed by thecloud belt of continental cold front moving southward to the mid-SCS and developing oncemore after being affected by the onset of the tropical summer monsoon over SCS. Chen etal. (1999) analyzed the source of the convective cloud belt during the period of the onset ofthe tropical summer monsoon over SCS and indicated that during the 19 years from 1975 to1993，there were 5 years with the cloud coming from the north, and 13 years with thecloud coming from the region near equator (including cross equatorial flow and flow camefrom West Pacific). Besides, during May of 1998中国煤化τpf development ofthe convective cloud over Indo China Peninsula ，:DHCNMHGre 21 May whichwas caused by the storm over the Bay of Bengalion of the cloudbelt during the pre-lood period of the South China. The characteristic of the former wasmainly subtropical while the latter apparently belonged to the tropical one. The secondNo.4CHEN Longxun, LI Wei, ZHAO Ping and TAO Shiyan44530N(@)(b).25-20-15105EQ(J)20(8BQ10'SJ890100 11120130 140E 801001i0120 130 140EFig. 5. Distibution of lhe three day -mean TBB (in K) during the period of 10- 27 May of 1998.development was during 25一27 May, the cloud belt evidently belonged to the period oftropical monsoon onset and was the tropical cloud belt. In addition , it developed first overSCs during 22- 24 May and then over the west of Indo China Peninsula during 25 - 27May.V. CHARACTERISTICS OF THE FLOW PATTERN DURING THE .ONSET OF THE RAINYSEASON IN 1998Figure 6 depicts three day- mean 850 hPa streamlines during the period of 4-27 Mayof 1998，the data obtained from SCSMEX Data Analysis Center. The streamline fromNCEP data was also performed to discuss the streamline over the region west of 90E.During the period of 4- 9 May of the year，subtropical anticyclone controlled the wholeAsian region of 7- 25N, an anticyclone center was located over the Arabian Sea , anotheranticyclone controlled the East India. The equatorial westerly only occurred over thetropical region south of 5°N and west of 90E. Since中国煤化工:veloped overSouth India, till 15 May, the center was located overYHCNMHGidwetcenterof Oze was located over southeast of the cyclone，the cold tongue extended into the center ofthe cyclone from the north with 0。less than 340 K. So till 13- 15 May, the cyclone wassubtropical one. During 16 - 18 May, the cyclone moved northward and developed446ACTA METEOROLOGICA SINICAVol.15 .30N25卡2015弋0fEQ t10S .25153Q卡10s25 t20 t1500-10S15+Q十10S N7090100 110 120 130E7010020 130E中国煤化工Fig.6. Three day- mean 850 hPa streamline and 0.. dun.MYHCNM H Gyof 198 (a)4- 6May; (b)7-9; (c) 10-12; (d)13-15; (e)16-18; (f) 19-21; (g) 22 -24; and(h) 25-27 May.No.4CHEN Longxun, L1 Wei, ZHAO Ping and TAO Shiyan447strongly, the center transformed to warm center (θ地less than 345K)，the previous coldcenter retreated to northwest of the cyclone. The southwest flow of the east part ofcyclone was located in north of the anticyclone over Indian Ocean and the anticyclone overSCS-West Pacific at all times. The flow was constituted by two branches, the northbranch was formed by the recurved northwest flow of the subtropical anticyclone over the .Arabian Sea which belonged to the subtropical flow, the south branch was equatorialwesterly, the common boundary of the two branches was near 5°N. Till 19-21 May, thecenter of the cyclone transformed to cold center once more, meanwhile a new cycloneformed in the north of the anticyclone over the West Pacific of east of Taiwan island ofChina. By 22一24 May it moved westward to SCS，there was cold air infall into thenorthwest of the cyclone， the NW flow of the cyclone and the recurved west flowcontrolled north of the SCS，meanwhile a convective cloud belt reached the middle andnorth SCS from north to south (Fig. 5). The cyclone was still located north of theanticyclone, therefore the SW flow over the north sCS was not the tropical one. By 22 -24 May, the cross equatorial SW flow over sCS coming from Australia intensifiedsuddenly, strong SW flow appeared over south sCS, thus there apparently occurred twowest flows over SCS, the middle and north parts were controlled by the subtropical NWflow caused by the development of the cyclone over SCS while the south part wascontrolled by the recurved cross equatorial tropical SW flow over SCS. The two convergedover the mid SCS, tropical convective cloud cluster appeared. By 25 - 27 May, the twobranches combined completely. Therefore the onset of the summer monsoon over SCS wasmainly formed by the combination of the subtropical west flow of the south of the cyclonedeveloped over the north SCS and recurved cross-equatorial tropical SW monsoon overSCS, but had no relationship with the Indian monsoon.The problem needs to be discussed is whether the west flow coming from the Indianregion is tropical or subtropical? To solve the problem we need to see the Arabian flow.According to NCEP streamline field, till 24 May, the Arabian Sea was still controlled byanticyclone，the northwest flow in north of the anticyclone reached 10°N, then was locatedover east of the Bay of Bengal. So the west flow over north of 10°N of South Asia wassubtropical monsoon rather than tropical one.Combined with TBB data, it can be seen that in 1998， no matter from convectivecloud belt or flow, the onset of the summer monsoon over SCS was the earliest in Asianregion, the onset of the monsoon over SCS was mainly caused by the seasonal variation ofthe members of the East Asian monsoon circulation system.VI. CONCLUSIONSFrom the above discussion, we obtained the following conclusions :(1) In climatology, during the period from April to June, Asian region experiencedthe impact of two rain belts. Since April，subtropic中国煤化工:urred over thesoutheast part (25一27°N) of mainland of China ,|Y片C N M H Gouthward afterits formation causing the pre-flooding rainy season over South China, meanwhile ilpropagated southwestward to Guangxi Region. The rainy season occurring over Indo-China Peninsula in April was conjoint with it. This was caused by the convergence of cold448ACTA METEOROLOGICA SINICAVol. 15air and the recurved southwesterly flow from subtropical high and westerly flow from theso-called western trough of subtropical region of South Asia occurring during winter tospring，and belonged to the typical subtropical rain belt. After the mid May, with theonset of the tropical summer monsoon over SCS, tropical summer monsoon propagatedwestward, 80 did the corresponding rainy season, The rainfall over Indo-China Peninsulashowed a feature of bimodal distribution, i. e. after the rain peak of the subtropicalmonsoon appeared in April, the rainfall decreased, the second rain peak occurred when therainfall was activated after the onset of the tropical summer monsoon with its westwardpropagation. With the onset of the summer monsoon over SCS, the subtropical highmoved northward seasonally , the previous subtropical monsoon rain belt propagated fromnorth to the coastal region of South China progressed northward in turn, and formed thesecond time of rain peak over south of 27°N. The northward moving rain belt was still thesubtropical monsoon rain belt. This was identified as Meiyu rain belt over the middle andlow reaches of Changjiang River Basin.(2) In 1998, the onset of summer monsoon over SCS was caused by the combinationof the subtropical westerly flow which was in the south of the cyclone developed over theWest Pacific and moved into the north SCS, and recurved cross -equatorial tropical SWmonsoon over south sCS. The corresponding cloud belt also entered into sCs from northto south and intensified with the onset of the SW monsoon over south SCS, the appearanceof the convective cloud belt over Indo-China Peninsula was 3- 5 days lag behind that overSCS. Therefore the onset of the summer monsoon over sCS in 1998 was mainly controlledby the East Asian monsoon system. During the period of the onset of monsoon over SCS,the Indian region was controlled by subtropical flow and had little impact on the onset ofthe tropical monsoon over SCS. In fact, till 23- 24 May (i. e. the onset of summermonsoon over SCS) the Arabian Sea was controlled by the subtropical anticyclone. Theregion from the Arabian Sea to the Bay of Bengal north of 10°N was still controlled by thesubtropical flow. Therefore, for the whole region of South Asia, the onset of the tropicalsummer monsoon started earliest over SCS, then propagated westward after the onset ofmonsoon over SCS.REFERENCESBai Zhaoye (1988), Weather of Northwestern China, China Meteor. Press, Beiing, pp. 1- 442 (inChinese).Chen Longxun et al. (1999) ,Preliminary study of the characteristics and mechanism of the onset of thesummcr monsoon over South China Sea and its adjacent area, Acta Meteor. Sinica, 57: 16-29 (inChen Longxun et al. (1991), East Asia Monsoon, China Meteor. Press, Beijing， pp. 1- 362 (inDai Jjiaxi (1990), Climate of Tibet, China Meteor. Press中国煤化工hinese).Data Center of Meteorological Bureau of Guangdong FYHCNMHG;uangdong Province,China Guangdong Science and Technology Press, pp IJiang Delong (1991),Climate of the Middle and Lowwer Reaches of Yangtze Rirver, China Meteor. Press,Beiing, pp. 1 - 426 (in Chinese).Lau，K.-M. and Yang Song (1997)，Climatology and interannual variability of the southeast AsianNo.4.CHEN Longxun, LI Wei, ZHAO Ping and TAO Shiyan449Summer Monsoon, Adv. Atmos. Sci., 14: 141- 162.Liang Jianyin and Wu Shangsen (1999), Analysis of characteristics of the onset of the southwesternmonsoon over South China Sea in 1998, Onset of the Monsoon over South China Sea and Its Interactionwith the Ocean.，China Meteor. Press , Beijing. pp. 52- 56 (in Chinese).Li Jiangien (1991), Climate of Xinjiang，China Meteor. Press, pp. 1- 302 (in Chinese)Li Wei and Chen Longxun (2001), Characteristics of the seasonel variation of rainy season over theTibetan Plateau in 1998 with its impact on East Asian weather , Acta Meteor. Sinica, 15:293- 309.Lin Zhiguang (1987), Further study of the process of onset and withdrawal of the summer monsoon rainbelt over East China, Collected Papers of Meteorological Technology of Chinese Academy ofMeteorological Sciences, No. 10, 24- 31 (in Chinese).Matsumoto Jun(1997), Seasonal transition of summer rainy season over Indochina and adjacent monsoonregion, Adv. Atmos. Sci., 14: 231一 245.Qian Linqing (1991), Climate of Loess Plateau, China Meteor. Press, pp.1- 426 (in Chinese).PRC -US Scientifie Working Group (1995), Scientific Plan of the South China Sea Monsoon Experiment 1一65Tao Shiyan and Chen Longxun (1987) ,A review of recent research on the East Asian summer monsoon inChina , Monsoon Meteorology, Oxford University Press, pp. 60- 92.Wang Wenhui (1990)， Meteorology of Iner Mongolia, China Meteor. Press, Beiing, pp. 1-427 (inChinese).Webster. P.J. et al. (1998), Monsoon: processes, predictability, and the prospects for prediction, J.Geo. Res. ，103: 14,451- 510.Xie An, Liu Xia and Ye Qian (1996), Summer Monsoon over South China Sea, New Prugress in the AsianMonsoon research ，China Meteor. Press, Bejing, pp. 132 - 141 (in Chinese).Xu Yuhua (1991), Climate of Southwest China , China Meteor. Press, Beijing, pp. 1- 297 (in Chinese).Yan Junyue (1997), Climate features of the onset of southwestern monsoon over South China Sea, ActaMeteor. Sinica, 55:174- 186 (in Chinese).Zhang Jiacheng (1991) ，Introduction to Climate in China, China Meteor. Press, Bejing, pp. 1-476 (in中国煤化工MYHCNMHG