Contrasting the Indian and East Asian monsoons ... · Contrasting the Indian and East Asian monsoons: implications on geologic timescales Bin Wanga ;b, Steven C. Clemensc, Ping Liu

Contrasting the Indian and East Asian monsoons:implications on geologic timescales

Bin Wang a;b;�, Steven C. Clemens c, Ping Liu b;1

a Department of Meteorology, University of Hawaii, Honolulu, HI, USAb International Paci¢c Research Center, School of Ocean Earth Science and Technology, University of Hawaii, Honolulu, HI, USA

c Department of Geological Sciences, Brown University, Providence, RI, USA

Accepted 19 June 2003

Abstract

The surface winds over the Arabian Sea and South China Sea are meaningful indicators for the strength of theIndian monsoon and East Asian monsoon, respectively. Paleo-monsoon variability has been studied through analysisof sediment records from these two monsoon regions. To facilitate interpretation of these records, we focus on theimpacts of ‘internal’ and ‘external’ forcing of the monsoon system by contrasting the annual cycle and interannualvariability of two subsystems: the monsoon over the Indian sector (40^105‡E) and over the East Asian sector (105^160‡E). Differences in the annual cycle within these subsystems arise primarily from the different landôceanconfigurations that determines atmospheric response to the solar forcing. Thus factors that drive intensities of themonsoonal annual cycle share common features with the external (geographic and orbital) forcing that controls paleo-monsoon variability. We show that the differences in interannual variations between the two monsoon subsystems areprimarily due to internal factors of the coupled atmosphereôcean^land system, such as remote impacts of El Nin‹o/LaNin‹a and local monsoonôcean interactions. The mechanisms that operate on interannual to interdecadal timescalesmay differ fundamentally from that on geologic/orbital timescales. The low-level flows over the East Asia andAustralia are essentially established by geographic forcing. The amplification of the Australia summer monsoonduring increased solar precession is likely caused by an enhanced East Asian winter monsoon, rather than followingan enhanced Indian summer monsoon as on the interannual timescale. It is also found that El Nin‹o influences thelow-level flow moderately over the Arabian Sea but to a greater extent over the South China Sea. As such, largechanges in the Pacific thermal conditions may significantly alter the intensity of the East Asian monsoon but not theIndian monsoon.9 2003 Elsevier B.V. All rights reserved.

Keywords: Indian monsoon; East AsianÂustralian monsoon; geographic forcing; orbital forcing; monsoonôcean interaction;ENSO forcing

0025-3227 / 03 / $ ^ see front matter 9 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0025-3227(03)00196-8

1 Also a⁄liated with LASG/Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China, 100029. The Inter-national Paci¢c Research Center is sponsored in part by the Frontier Research System for Global Change.* Corresponding author. Present address: International Paci¢c Research Center, University of Hawaii, Honolulu HI, 96822.

Fax: 1-808-956-9425.E-mail address: [email protected] (B. Wang).

MARGO 3364 9-9-03 Cyaan Magenta Geel Zwart

Marine Geology 201 (2003) 5^21

R

Available online at www.sciencedirect.com

www.elsevier.com/locate/margeo

mailto:[email protected]

http://www.elsevier.com/locate/margeo

1. Introduction

Paleo-monsoon studies have advanced throughthe study of sediment records on the ocean £oorin the Arabian Sea and South China Sea (Clem-ens and Prell, 1990; Clemens et al., 1996; Clem-ens et al., 1991; Overpeck et al., 1996; Prell et al.,1992; Schulz et al., 1998; Shimmield et al., 1990;Sirocko, 1991; Wang et al., 1999; Wang, 1999).Changes in sediment deposits in these regions re-£ect the in£uence of the summer- and winter-monsoon winds and precipitation on the physical,chemical, biotic, and isotopic characteristics ofocean surface and deep waters. On the annualtimescale the strongest seasonal reversal of windsbetween January/February and July/August oc-curs over the western Arabian Sea with a second-ary maximum over the South China Sea (Figs. 1and 4). Since the wind and precipitation varia-tions in these two regions are good indicators ofthe intensity of the Indian and East Asian mon-soons, the sediment changes to large extent re£ectvariations in these monsoons.

The classical view of monsoons contemplatesthe land/sea thermal contrast and annual cycleof solar radiation as fundamental factors that es-tablish monsoons. On geologic timescales the tec-tonic forcing shapes the landôcean distributionand controls the formation of the AsianÂustra-lian monsoon (AÂM hereafter) system. On theMilankovitch timescale changes in solar orbit pa-rameters (the obliquity or tilt of the Earth’s rota-tional axis, the date of perihelion and the eccen-tricity of the Earth’s orbit) induce large-amplitudemonsoon variability (e.g. Kutzbach and Guetter,1984). Interannual and decadal monsoon varia-tions are mainly caused by physical processes in-ternal to the earth’s climate system such as ElNin‹o-Southern Oscillation (ENSO) and local at-mosphereôcean^land interaction. Understandingmechanisms driving monsoon variability requiresbringing together information gained from inves-tigation of both instrumental and proxy data.

The most pronounced annual reversal in atmo-spheric circulation and precipitation on the Earthis located in the tropical region from 40‡E to160‡E and from 30‡S to 45‡N (Ramage, 1971;Krishnamurti et al., 1985; Wang, 1994; Webster

et al., 1998; Trenberth et al., 2000), which is re-ferred to as AÂM domain (Fig. 1). From themonsoon climate point of view the line of 105‡Elongitude, which runs along the eastern £ank ofthe Tibetan Plateau and through the Indo-China‘land bridge’, delineates the Indian monsoon andEast Asia^western North Paci¢c (WNP) monsoonregions (Wang and Lin, 2002). In the SouthernHemisphere the 105‡E longitude separatesroughly the maritime continent and the Austra-lian landmass from the Indian Ocean. The con-trasting land/ocean con¢guration to its east andwest is believed to have a rudimentary impact onthe monsoon climate over India and East Asia.On the interannual timescale the variations ofthe monsoon over the areas around India andthat over the East Asia^WNP sector (for brevityit will be referred to as East Asian sector here-after) exhibit strikingly di¡erent spatial and tem-poral structures (Wang et al., 2001). It is impor-tant to understand how the annual cycle in thesetwo regional subsystems di¡ers and, in particular,what these di¡erences imply for understanding ofthe paleo-monsoon variability. Careful analyses ofsuch di¡erences may shed light on the interpreta-tion of the paleo-monsoon records collected in theArabian Sea and South China Sea.

The South China Sea is intimately connectedwith the WNP warm pool and the local atmo-spheric general circulation is signi¢cantly in£u-enced by ENSO. Changes of the monsoon inten-sity induced by the factors internal to the coupledatmosphereôcean^land system such as ENSOwould distort an accurate estimation of thosecaused by the external geographic or orbital forc-ing. Thus, in order to more precisely interpret thepaleo-monsoon variability based upon the sedi-ment records under the South China Sea, it iscritical to know how much the East Asian mon-soon is a¡ected by changes in Paci¢c thermal con-ditions. This study is intended to investigate thenature of internal forcing and its impact on thewind indicators over the Arabian Sea and SouthChina Sea.

Clemens and Kershaw (pers. commun.) ¢ndthat the phase of the summer-monsoons pollenrecords (three marine cores o¡ northern Austra-lia) matches the summer-monsoon phase mea-


B. Wang et al. /Marine Geology 201 (2003) 5^216

sured in sediment cores from the northern Ara-bian Sea (the Indian monsoon). In other words,on orbital timescale, enhanced Indian summermonsoons are ‘in phase’ with enhanced Australiansummer monsoons. On the ENSO timescale, astrong Indian summer monsoon is also followedby a strong Australian summer monsoon (Meehl,1987). Are these similar in-phase relationshipscaused by similar mechanisms? Is the orbital-scalevariability in the Australian summer monsoon re-lated to the changes in the Indian summer mon-soon? Clarifying these complex relationships re-quires realistic simulations of the interactionsbetween the AÂM and ENSO and among theAÂM subsystems with coupled atmosphereôcean models. Unfortunately the state-of-the-artcoupled atmosphereôcean general circulationmodels (CGCMs) cannot faithfully reproduce var-iations associated with such interactions. Most of23 coupled models participating the coupled mod-el intercomparison project (Meehl et al., 2000)have di⁄culty in reproducing the observed annualmean equatorial sea surface temperature (SST)and wind stress as well as the Indian^Paci¢c lagcorrelation (Davey et al., 2002). Few models haverealistically simulated the ENSO with respect toequatorial SST anomalies (Latif et al., 2001). Alsobecause large ensemble simulations are necessaryto reasonably simulate the monsoonÊNSO inter-action, it is extremely di⁄cult to address this openissue based upon the present-day CGCMs.

Asian subsystems are driven by di¡erentboundary thermal conditions associated withlandôcean con¢guration and topography, inves-tigation of their di¡erent characteristics over an-nual cycle may enhance our understanding howtectonic and solar orbital forcing a¡ects paleo-monsoon variation. Likewise study of di¡erentresponses and feedbacks of these monsoon sub-systems to ENSO and warm pool conditionsmay shed light on how internal dynamics of thecoupled atmosphereôcean system in£uences themonsoon interannual variability.

There is a rich source of literature describingannual variations of rainfall in these two mon-soon regions. The focus of this study is the longi-tudinal distinction in both annual cycles and in-terannual variations with particular attention paid

to the di¡erences between the Indian (40^105‡E)and East Asian (105^160‡E) sectors.

We ¢rst document the contrasting features intheir annual cycles (Section 3) and interannualvariations associated with ENSO (Section 4).We then discuss mechanisms accounting forsuch di¡erences (Section 5). Section 6 presents asummary of major results and further discussesthe implications for interpreting paleo-monsoonvariability.

2. Data and analysis methods

The National Centers for Environmental Pre-diction^National Center for Atmospheric Re-search (NCEP^NCAR) Reanalysis (Kalnay etal., 1996) provides the atmospheric data for theperiod of 1951^2000. Our analysis will focus onboundary layer winds at 925 hPa and precipita-tion. These data are on regular grids with a hor-izontal resolution of 2.5‡U2.5‡. The SST data areobtained from the Reynolds SST (Reynolds andSmith, 1994) in the NCEP^NCAR Reanalysis.Monthly mean precipitation data derived fromthe Climate Prediction Center Merged Analysisof Precipitation (CMAP; Xie and Arkin, 1997)data (January 1979^December 2000) were usedfor describing the annual cycle of precipitationand to check the quality of the Reanalysis precip-itation anomalies. It was found that the large-scale patterns of monthly mean precipitationanomalies derived from Reanalysis rainfall andCMAP are in a reasonably good agreement,which enables us to make a meaningful analysisof AÂM anomalies for the extended period from1951 to 2000.

Previous studies have been largely concernedwith regional monsoon variability, rather thanthe entire AsianÂustralian monsoon system.There are enormous regional di¡erences amongthe two subsystems. The Indian summer monsoonrainfall was found to be de¢cient in an El Nin‹odeveloping year (see review by Webster et al.,1998), although this relation broke down in thelast two decades (Kumar et al., 1999). De¢cientrainfall over Australian monsoon region occursduring northern fall and winter, often following


B. Wang et al. /Marine Geology 201 (2003) 5^21 7

a de¢cient Indian monsoon (Meehl, 1987). In con-trast to the Indian monsoon, the summer mon-soon in East Asia and WNP tends to be morea¡ected during the year after the El Nin‹o maturephase. Due to the strong season-dependence andENSO phase-dependence of the AÂM anoma-lies, we introduced a composite technique todocument the common features of evolvinganomalies of the AÂM system. This approachis distinguished from the previous analyses inthat it focuses on ¢nding a sequence of variabilitypatterns that are associated with the evolution ofENSO. The composite scenario presented in thispaper will better reveal the ENSO phase-depen-dent seasonal mean anomalies of the entire AÂM anomalies, thus providing a new perspectiveof the AÂM anomalies associated with a switchfrom developing to decaying phases of ENSO.This facilitates understanding of the AÂM vari-ability due to forcing internal to the coupled cli-mate system.

3. The contrasting annual cycles in the Indian andEast Asia sectors

The summer monsoon in the Indian and EastAsia sectors exhibits di¡erent spatial structures(Fig. 1). The Indian monsoon system is dominat-ed by a distinct clockwise monsoon gyre centeredin the equatorial Indian Ocean, which ties the In-dian monsoon trough, Somalia jet and southeastwinds associated with Mascarene High (Fig. 1;Krishnamurti and Bhalme, 1976). In contrast,the East Asia sector monsoon consists of threecirculation components: (a) the WNP monsoontrough (also known as the intertropical conver-gence zone) that is the con£uence zone betweenthe southwesterly monsoon and southeast windswhich originate from the Paci¢c trades, (b) themassive WNP subtropical ridge which is locatedto the north of the monsoon trough with its edgeextending to the coastal area of East Asia; and (c)the East Asian subtropical front (Meiyu front)

Fig. 1. Climatological JulyÂugust mean precipitation rates (shading in mm/day) and 925 hPa wind vectors (arrows) in the AÂM region. The precipitation and wind climatology are derived from CMAP (Xie and Arkin, 1997) (1979^2000) and NCEP/NCAR reanalysis (1951^2000), respectively. The three boxes de¢ne major summer precipitation areas of the Indian tropical mon-soon (5^27.5‡N, 65^105‡E), WNP tropical monsoon (5^22.5‡N, 105^150‡E), and the East Asian subtropical monsoon (22.5^45‡N, 105^140‡E).



(Tao and Chen, 1987). The term ‘Meiyu’ means‘plum rain’ which normally occurs from mid-Juneto mid-July when the East Asian subtropicalmonsoon front stagnates in a narrow zone extend-ing from the Yangtze River valley to southernJapan. Japanese called it Baiu and Koreans calledit Changma. The term ‘Meiyu’, ‘Baiu’, andChangma’ all mean the same East Asian subtrop-ical rain band that brings a major summer rainyperiod to these countries. In Fig. 1, the subtrop-ical front is weak, because the major subtropicalfront rainfall occurs before early July, thus does

not show up well in the JulyÂugust mean ¢eld.A clearer picture of the subtropical front and as-sociated precipitation can be identi¢ed from thenorth^south cross-section shown in Fig. 2b. TheEast Asia subtropical rain band occurs from mid-May to early July and migrates northward from20‡N to 35‡N. In between the subtropical andtropical monsoon rain is a relatively dry zone,which signi¢es the location of subtropical ridge(Figs. 1 and 2b).

The central monsoon system in the East Asiasector is the WNP monsoon trough and associ-ated heavy rainfall (Fig. 1). It is important torealize that the WNP rainfall and associated la-tent heat release is stronger than that over theIndian monsoon region. As shown in Fig. 1,two areas are chosen to represent the Indian mon-soon (5^27.5‡N, 65‡E^105‡E) and WNP monsoon(5^22.5‡N, 105^150‡E), respectively. The sizes ofthe two rectangular areas are almost even, yet themean precipitation rate in MayÔctober is 8.61mm/day over the WNP monsoon region whileonly about 6 mm/day over the Indian monsoonregion (Fig. 3). The strong WNP latent heatsource is dynamically consistent with the upper-level divergent circulation center near the Philip-pines as ¢rst revealed by Krishnamurti (1971) andrecently con¢rmed by Trenberth et al. (2000).

Note that the immense WNP heat source isconnected not only with the Indian monsoonbut also directly linked with Australian wintermonsoon trough and the pronounced low-levelnorthward cross-equatorial £ows over the SouthChina Sea, Celebes, and the northeastern part ofNew Guinea (Fig. 1; also see Tao and Chen,1987; Wang and Lin, 2002). The cross-equatorialclockwise gyre in the East Asia sector is not asstrong as that over the Indian sector, yet it doessuggest that the cold Australian land surface en-hances the hemispheric thermal contrast and con-tributes to the moisture convergence over theWNP summer monsoon. To the north, theWNP heat source is coupled with the WNP sub-tropical ridge and the East Asia subtropical mon-soon. The coupling between the tropical WNPmonsoon and the subtropical monsoon is partic-ularly evidenced by their in-phase sub-seasonalvariations in June and July (Fig. 3). At pentad

Fig. 2. Climatological pentad (5-day) mean precipitation rate(mm/day) averaged over (a) the Indian sector (65^105‡E),and (b) the western Paci¢c sector (105^145‡E). The dataused are derived from CMAP (Xie and Arkin, 1997) for theperiod of 1979^2000.



35 (June 20^24), the wet spell in the WNP (onsetin the southwest Philippine Sea) concurs with themaximum East Asia rainfall (peak Meiyu). Thedry singularity in the WNP at pentad 49 (end ofAugust) is directly followed by the retreat of EastAsia rainy season (the sudden decrease of themonsoon rainfall at pentad 50). The pronouncedsub-seasonal variations of the WNP summermonsoon were termed the climatological intrasea-sonal oscillation (CISO) that arises from phaselocking of transient intraseasonal oscillations bythe annual cycle (Wang and Xu, 1997).

The temporal evolutions of the monsoon sub-systems over the Indian and East Asian sectorsare also di¡erent. The Indian rainy season reachespeaks in early June and mid-July (Fig. 2a) withmaximum rainfall occurring in early June and thecorresponding rain band reaches the northern-most latitude (V25‡N) in mid-July. On the otherhand, over the East Asia sector, both the maxi-mum intensity and the northernmost location(V25‡N) of the tropical monsoon rain are foundin August (Fig. 2b), which matches the peaksrainfall in mid-August shown in Fig. 3. This

phase di¡erence in the annual cycle implies aneastward shift of convection centers from India(in June^July) to the WNP (in August) duringboreal summer. Accordingly, the Indian rainy sea-son ends in late-September, while the WP rainyseason retreats in late October (Fig. 3).

Another prominent di¡erence is that over theIndian sector, the summer rainfall is characterizedby double rain bands, one is located about 15‡Nand the other at 5‡S, whereas over the East Asiansector, only a single tropical monsoon rain bandis located north of the equator from June to Oc-tober. However, from late May to mid-July, anorthward migration of the Meiyu from 20‡N to35‡N can be clearly seen. This subtropical mon-soon rain belt is produced downstream of the Ti-betan Plateau, which prohibits such a rain bandformation over the Indian sector. Thus the Indiantropical monsoon rain has a close link with theequatorial convections throughout the entiresummer, whereas the monsoon rain in the tropicalWP monsoon has close linkage with the subtrop-ical monsoon rain in early summer.

While the Indian summer monsoon is stronger

Fig. 3. Climatological pentad mean precipitation rate (mm/day) averaged over three regions: Indian summer monsoon (5^27.5‡N,65^105‡E), WNP summer monsoon (5^22.5‡N, 105^150‡E), and East Asian summer monsoon (22.5^45‡N, 105^140‡E). The ab-scissa runs from pentad 1 (January 1^5) to pentad 73 (December 27^31). The data used are derived from CMAP (Xie and Arkin,1997) for the period of 1979^2000. The thick, thin, and long-dashed horizontal lines indicate, respectively, the annual mean rain-fall rates averaged for the WNP, East Asian, and Indian monsoon regions.



than the monsoon in East Asian sector (Fig. 1),the East Asian winter monsoon is much strongerthan Indian winter monsoon (Fig. 4). In Januaryand February, the most powerful winter monsoonoccurs over East Asia and adjacent marginal seas.This powerful winter monsoon is directly linkedto Australian summer monsoon westerly throughmarked cross-equatorial £ows over the SouthChina Sea and Celebes (Fig. 4), con¢rming thatthe annual cycle of the Australian monsoon has a¢rmer link to East Asian monsoon than to Indianmonsoon. Consistent with the monsoon circula-tion, the rainfall in the Australian sector is moreintense than that in the South Indian Ocean con-vergence zone (Fig. 4). Thus, the active convec-tion and rainfall region shifts from Indian sectorin boreal summer to the East Asian sector in aus-tral summer.

4. Di¡erences in interannual variability betweenIndian and East Asian monsoons

On the interannual timescale, the ENSO hasbeen recognized as a major cause of the AÂMvariability. In this section, we examine the di¡er-ence in the monsoon variability between the In-

dian and WP sectors. To this end, one has todocument the evolving AÂM anomalies associ-ated with ENSO turnabout.

We take boreal winter of the El Nin‹o year as areference season because the turnabout of ENSOcycle often occurs in boreal winter. As shown inFig. 5, twelve signi¢cant warm episodes (1951,1957, 1963, 1965, 1968/1969, 1972, 1976, 1982,1986/1987, 1991, 1994, 1997) occurring duringthe period of 1951^2000 all matured during Oc-tober to February except for the episode of 1986/1987. During the same period, nine signi¢cantcold episodes had occurred (1954/1955, 1964,1967/1968, 1970, 1973, 1975, 1984, 1988, 1998/1999); all reached their peaks in November toFebruary. Based on the strong phase-lock behav-ior of ENSO cycle, a composite scenario is madewith reference to the mature phases of ENSO,D(0)/JF(1), where year 0 represents the El Nin‹odeveloping year and year 1 means the followingyear. Fig. 6 shows the SST anomaly patterns com-posite for the 10 El Nin‹o events outlined by thesolid boxes in Fig. 5. The double-peak warmevents, 1968/1969 and 1986/1987 were excluded,because these events last two years, which is notcommon to others. The JJA(0), D(0)/JF(1), andJJA(1) SST anomalies represent those in the de-

Fig. 4. Same as in Fig. 1 except for January^February mean.



veloping, mature, and decay phases of the com-posite El Nin‹o event, respectively. In Fig. 7, wepresent monsoon anomalies in ¢ve consecutiveseasons, i.e. JJA(0), SON(0), D(0)/JF(1),MAM(1), and JJA(1). This composite scenariotakes into account the season-dependent andENSO phase-dependent characteristics of theAAM anomalies and is suitable for delineatingthe linkage and di¡erences between Indian andEast Asian sectors.

During the summer of El Nin‹o development,the anomalous low-level circulation is character-ized by an elongated anticyclonic ridge extendingfrom the maritime continent to the peninsula ofIndia (implying a moderate weak Indian summermonsoon) (Fig. 7a). On the other hand, to thenorth of the ridge, anomalous westerlies enhancethe WNP monsoon trough and convection (Fig.7a). Thus the observed frequency of tropicalstorm formation in the southeast quadrant ofthe tropical WNP (5^17‡N, 140^170‡E) drasticallyincreases (Chen et al., 1998; Wang and Chan,2002).

During the fall of an El Nin‹o developing year,the southeast Indian Ocean anticyclone, which

originated southwest of Sumatra in the previoussummer, grows rapidly, leading to a giant anticy-clonic ridge dominating the Indian Ocean with ananticyclone center located at (10‡S, 90‡E) and atilted ridge extending from western Australia allthe way to Arabian Sea (Fig. 7b). Intense easterlyanomalies develop along the equatorial IndianOcean; convection is suppressed in the easternIndian Ocean while enhanced in the western In-dian Ocean with the most suppressed convectionlocated slightly to the eastward and equatorwardof the anticyclone center. Comparison with clima-tology, Fig. 7b indicates that the monsoon rainfalland circulation of the entire AÂM system is se-verely reduced. Thus, the strongest ENSO impactson the Indian sector occur normally in the fall ofan ENSO developing year.

During the mature phase of El Nin‹o, D(0)/JF(1), the low-level circulation anomalies are con-trolled by two subtropical anticyclonic anomalieslocated in the southern Indian Ocean (SIO) andWNP, respectively (Fig. 7c). The former is a resultof the weakening and eastward retreat of the SIOanticyclone previously established in fall, whilethe latter results from the ampli¢cation and east-

Fig. 5. Time series of monthly mean Nin‹o 3 (black) and Nin‹o 3.4 (gray) sea surface temperature anomalies from January 1951to December 2000. The data used are from Reynold’s reconstructed SST (Reynolds and Smith, 1994). Anomalies are departuresfrom 1951^2000 climatology. Solid boxes outline the 10 El Nin‹o cases used in the composite. Two dashed boxes indicate two ElNin‹o events that are not selected.



ward migration of the Philippine anticyclone gen-erated in the previous fall. Compared with clima-tology (Fig. 4), both the East Asian winter mon-soon and the Australian summer monsoon areconsiderably weakened. However, Arabian Seawinter monsoon is enhanced.

During MAM(1) and JJA(1), the strongestmonsoon anomalies are found in the WNP andEast Asia (Fig. 7d,e). A salient feature is the per-sistence of the Philippine Sea anticyclone (Wanget al., 2000). The persistence of the Philippine Seaanticyclone a¡ects the subtropical high and pro-vides a prolonged impact on Meiyu/Baiu in JJ(1)when the El Nin‹o disappears. By JJA(1), subsi-dence controls the Philippine Sea and SoutheastAsia, leading to de¢cient summer monsoon rain-fall there. This agrees with results derived from

the previous studies, i.e. the WNP rainfall de-creases (Wang et al., 2000) while the Meiyu/Baiuis inclined to increase (Fu and Teng, 1988; Chenet al., 1992; Shen and Lau, 1995; Lau and Weng,2001).

How di¡erent are the impacts of El Nin‹o onthe monsoons over the Arabian Sea and SouthChina Sea? In the summer of an El Nin‹o devel-oping year, JJA(0), the southwesterly monsoonover the Arabian Sea weakens, while that overthe South China Sea strengthens (Fig. 7). How-ever, during JJA(1), the southwesterly in bothArabian and the South China Seas weakens(Fig. 7e). In boreal winter, the El Nin‹o-inducedanomalous winds strongly o¡set mean northeast-erly monsoons over the South China Sea, whilereinforcing the northeasterly monsoon over theArabian Sea (Fig. 7c).

In order to quantitatively measure the impactsof El Nin‹o, we selected two regions, (45^65‡E,0^15‡N) and (105^120‡E, 5^20‡N), representingwestern Arabian Sea and South China Sea, re-spectively. The El Nin‹o-induced composite anom-aly winds were then projected onto the localsummer (or winter) mean climatological wind vec-tors. The resulting di¡erence, normalized by theclimatological wind speed, denotes fractionalchanges of the annual cycle winds. Calculationswere performed for each grid and then area aver-aged over the western Arabian Sea and SouthChina Sea.

The results show that in JJA(0) and JJA(1), thecomposite El Nin‹o anomalies reduce Arabian Seamonsoon by about 1.9% and 0.9%, respectively,which is insigni¢cant. In the extreme events, thereduction of summer monsoon can reach about11% in JJA(0) (1972) and 7% in JJA(1) (1966).On the other hand, the composite El Nin‹oanomalies account for, on average, 10% of in-crease in JJA(0) and 9% of decrease in JJA(1) inthe South China Sea summer monsoon. In theextreme El Nin‹o events, the South China Seamonsoon was enhanced by 27% in JJA of 1972and reduced by 37% in JJA of 1998. As such, ElNin‹o may considerably alter the strength of theSouth China Sea summer monsoon.

How is the Australian monsoon variability re-lated to the Asian monsoon variation on interan-

Fig. 6. Composite seasonal mean SST anomalies for selected10 El Nin‹o episodes: (a) JJA(0), (b) D(0)/JF(1), and (c)JJA(1), where 0 denotes the year in which El Nin‹o developsand 1 as the year El Nin‹o decays. The data used are fromReynold’s reconstructed SST (Reynolds and Smith, 1994).





nual timescales? To address this question, we de-¢ne, following the idea of McBride et al. (1995),an Australian monsoon index by using the 925hPa zonal winds averaged over the region (2.5^12.5‡S, 120^150‡E) where the annual reversal ofwinds is largest (Figs. 1 and 4). Fig. 8 shows acorrelation map of the anomalous meridionalwinds with reference to the Australian summermonsoon index for the period of 1951^2000 (en-semble size 50). The enhanced Australian summermonsoon is highly correlated with the strength ofthe northerly winds associated with the EastAsian winter monsoon. This connection is real-ized through the cross-equatorial £ows in threechannels located over the maritime continent:west of Sumatra, in between Sumatra and Bor-neo, and south of New Guinea. These cross-equa-torial £ows are essentially the same as seen on theannual timescale (Fig. 4). Fig. 8 also indicates noconnection exists between Indian winter north-easterly monsoon and Australian summer west-erly monsoon.

5. Discussion: causes of the di¡erences between thetwo AÂM subsystems

5.1. Annual cycle: geographic and orbital forcing

The e¡ects of the land^sea con¢guration andtopography are fundamental in understandingthe di¡erences in the monsoon annual cycle be-tween the Indian and the East Asian sectors.

Over Indian sector, the Asian landmass is lo-cated in the Northern Hemisphere (NH) whileocean occupies the Southern Hemisphere (SH).The landôcean distribution is opposite over theEast Asian sector: the landmass is in the SH whilethe ocean in the north. In addition, the TibetanPlateau is located to the north of the Indian sec-tor but to the west of the East Asian sector; theequatorial region of the Indian sector is occupiedby deep ocean, while the equatorial region of theEast Asian sector is occupied by the Indonesianarchipelago and shallow continental shelves; theIndian Ocean is bounded by land in both the west

Fig. 8. Correlation map of seasonal (DJF) mean meridional wind anomalies with reference to the Australian summer monsoon(DJF) index that is de¢ned by the zonal wind anomalies averaged over the region (2.5^12.5‡S, 120^150‡E). The correlation coe⁄-cients are computed for the period of 1951^2000.

Fig. 7. Seasonal mean precipitation rate (mm/day, contour) and 925 hPa wind anomalies composite for the 10 selected El Nin‹oepisodes shown in Fig. 6: (a) JJA (0), (b) SON(0), (c) D(0)/JF(1), (d) MAM(1), and (e) JJA(1), where year 0 denotes the year inwhich El Nin‹o develops and 1 as the year El Nin‹o decays. The data used are derived from NCEP/NCAR reanalysis (1951^2000)(Kalnay et al., 1996).



and east, while the ocean in the East Asian sectoris open to the east. These contrasts in geographicdistribution of land and ocean as well as the com-plex topography results in remarkable di¡erencesin the thermal forcing for the two monsoon sub-systems.

During boreal summer, the monsoon over theIndian sector is extremely powerful, because themeridionally di¡erential solar radiation heating isreinforced by the strong north^south thermal con-trast between the heated Asian land and cold SIO.In addition, the thermal e¡ects of the elevatedTibetan Plateau heat source further strengthenthis north^south thermal contrast. Thus, the ex-tremely large north^south thermal contrast duringboreal summer results in the strong Somalia jetand powerful Indian summer monsoon. Over theEast Asian sector, the summer monsoon is drivenby both the north^south thermal contrast betweencold Australian land and warmer WNP and theeast^west thermal contrast between the heatedEast Asian landmass and relatively colder Paci¢cOcean. This directionally ‘orthogonal’ forcing re-sults in a weaker summer monsoon with a tilt ofthe monsoon trough from the northern SouthChina Sea to (5‡N, 160‡E). However, because ofthe east^west thermal contrast, a giant North Pa-ci¢c subtropical ridge develops which separates(but also couples) the subtropical rain belt andtropical rain belt, forming a coupled tropicaland subtropical monsoon system with a large me-ridional extent. The landôcean distribution in theEast Asian sector renders the WNP monsoonmore constrained by the ocean thermal inertia.Thus, both the peak rainy season and fall transi-tion are about one^two months later than theircounterparts in the Indian sector.

During northern winter, the monsoon over theEast Asian sector is extremely powerful becauseof the contrast between the cold extratropicalAsian landmass and the relatively warm NorthPaci¢c Ocean. The heated Australian land in theSH and colder ocean in the NH generate low-levelcross-equatorial £ows through the southern SouthChina Sea. The monsoon in the Indian sector ismoderate because of absence of heated land in theSH and because of the blocking of cold Siberianair by the Tibetan Plateau. The Australian sum-

mer monsoon westerlies are intimately linked tothe East Asian winter monsoon but not the wintermonsoon from northern India Ocean (Fig. 4).

The above analysis suggests that the extremephase of the monsoon annual cycle is a responseto the solar radiation forcing and is primarily de-termined by landôcean con¢guration and topog-raphy. Therefore, the annual variability of thestrength of monsoon is primarily determined bychanges in the solar and geographic forcing.

5.2. Interannual variation: local monsoonôceaninteraction and El Nin‹o forcing

The way by which ENSO a¡ects the AÂM hasbeen considered primarily through changes in theWalker circulation (e.g. Ju and Slingo, 1995). Weargue that in addition to the direct remote forc-ing, local monsoonôcean interactions play a crit-ical role in accounting for the di¡erent interan-nual variation in the two subsystems: the Indianand East Asian sectors. Here we present evidenceto show the importance of local monsoonôceaninteraction.

In the Indian sector, the ENSO-related anoma-lies are stronger in boreal fall than in boreal win-ter, but the remote ENSO forcing is strongest inthe boreal winter when ENSO matures. This mis-match suggests that additional factors must be inplay besides the remote ENSO forcing. One ofthem is the local monsoonôcean interaction.During JJA(0), the remote ENSO forcing sup-presses Indonesian convection, in particular theconvection southwest of the Sumatra where clima-tological rainfall reaches maximum (Fig. 7a). Thesuppressed convection southwest of Sumatra ex-cites westward-propagating Rossby waves (Gill,1980). This causes descending motion and reinfor-ces the low-level anticyclonic anomaly in thesoutheast Indian Ocean and the southeasterlywind along the southern coast of Sumatra aswell as the equatorial easterly anomaly. This, inturn, induces anomalous coastal and equatorialupwelling and excessive evaporation and entrain-ment cooling in the equatorial eastern and south-eastern Indian Ocean (Fig. 6a). On the otherhand, in the western Indian Ocean, the low-levelcirculation anomalies tend to be against climato-



logical boreal summer monsoon, hence reducingthe total wind speed. As a result, the equatorialwestern Indian Ocean, western Arabian Sea, andthe southwestern subtropical Indian Ocean be-come warmer than normal (Fig. 6a). This is dueboth to suppression of the coastal and equatorialupwelling and evaporation cooling (over the Ara-bian Sea and west equatorial Indian Ocean), andto increased solar irradiance and suppressed evap-oration cooling (over the subtropical southwest-ern Indian Ocean). The resultant SST anomalypattern, referred to as the Indian Ocean Dipolemode or zonal mode (Saji et al., 1999; Webster etal., 1998), would further enhance the SIO anom-alous anticyclone and associated equatorial east-erly anomalies and Sumatra cross-equatorial£ows. The above-described positive feedback de-pends on boreal summer monsoon background£ow and is reversed quickly once the fall mon-soon transition occurs after November. Therefore,the positive feedback-induced Indian Oceananomaly pattern reaches maximum in borealfall. It is the local Indian monsoonôcean inter-action that remarkably ampli¢es Indian monsoonanomalies in boreal fall (Fig. 7b).

In the East Asian sector, the El Nin‹o weakensnorthern winter monsoon and the Australiansummer monsoon (Fig. 7c). Pronounced anticy-clonic monsoon anomalies occur over the Philip-pine Sea in MAM(1) and JJA(1) after the maturephase of ENSO when the remote forcing in theeastern Paci¢c decays rapidly (Fig. 6c). This factimplies that the Philippine Sea anticyclonicanomalies cannot be attributed simply to the re-mote forcing from the El Nin‹o; rather, other pro-cesses must play a critical role in maintaining thePhilippine Sea anticyclonic anomalies. Given thechaotic nature of atmospheric motion and the de-caying remote forcing by El Nin‹o, what mecha-nisms sustain the anomalous WNP subtropicalanticyclonic anomaly from mature phase to ensu-ing summer? Wang et al. (2000) have attributedthe development and persistence of the WNP anti-cyclonic anomaly to a positive thermodynamicfeedback between atmospheric Rossby wavesand the underlying warm-pool ocean. The in-creased total wind speed east of the anomalousanticyclone center cools the ocean surface due to

excessive evaporation and entrainment. The cool-ing, in turn, suppresses convection and reduceslatent heat release, which excites descending at-mospheric Rossby waves that propagate westwardand reinforce the WNP anomalous anticyclone.

In summary, the Indian monsoon weakeningreaches peak phase in the fall of El Nin‹o devel-oping year and primarily over the SIO due torapid intensi¢cation of the anomalous subtropicalanticyclone over the SIO. On the other hand, theweakening monsoon over the WNP is prominentin the spring and summer during the decay of ElNin‹o due to rapid intensi¢cation and maintenanceof the anomalous subtropical anticyclone over theWNP. This di¡erence in the Indian and EastAsian sectors is due to monsoonôcean interac-tion regulated by the local mean background sea-sonal cycle and their respective geographic loca-tions relative to El Nin‹o forcing. Thus, themonsoonôcean interaction is critical to accountfor the di¡erences in the interannual variations ofthe Indian and East Asian summer monsoons.

6. Conclusion and discussion

The low-level monsoon £ows over the ArabianSea and South China Sea are meaningful indica-tors of the strength for the Indian and East Asiansummer monsoon, respectively. The sediments re-corded in these two monsoon regions provide val-uable information about the monsoon variabilityon geologic timescales. To facilitate interpretationof the results revealed by paleoclimatic records,we attempted to explore the di¡erences betweenthe ‘internal’ and ‘external’ forcing mechanisms ofthe monsoon system. This objective is achieved bycontrasting the behavior of the two monsoon re-gions: the monsoon over the Indian sector (40^105‡E) and the monsoon over the East Asian sec-tor (105^160‡E). These two monsoon regions havedi¡erent landôcean con¢gurations. Their annualcycle and interannual variability are expected tobe quite dissimilar. Understanding the dissimilarbehavior in the annual cycle and interannual var-iability may help to understand the e¡ects of theexternal vs. internal forcing of the monsoon sys-tem at longer time scales.



The seasonal cycle di¡ers in the two sectors ofthe AÂM system. The peak of the summer mon-soon occurs in June^July in Indian sector but inJulyÂugust in the East Asian sector. The Indiansummer monsoon consists of a single tropicalmonsoon system (Indian trough, equatorial gyrewith Somalia jet, and Mascarene High), while thesummer monsoon in the East Asian sector con-sists of a coupled tropical monsoon system (WNPtrough, cross-equatorial £ows and AustralianHigh) and a subtropical monsoon system (WNPsubtropical high and the East Asian subtropicalmonsoon front). While the summer monsoon isstronger in the Indian sector, the boreal wintermonsoon is stronger in the WNP sector. The Aus-tralian summer monsoon is more closely coupledwith powerful East Asian winter monsoon; andthe Australian winter monsoon tends to becoupled with the WNP summer monsoon. Thesedi¡erences result arguably from the atmosphericresponse to di¡ering landôcean distribution andtopography (geographic forcing) and the annualvariation of the solar radiation as elaborated inSection 5.1.

On the interannual timescale, the variability inthe Indian and East Asian sectors strongly de-pends on phases of ENSO. In the summer of anEl Nin‹o developing year, an anomalous anticy-clone extends from the maritime continent tothe southern tip of India, which increases theWNP monsoon westerly but weakens the Indiansummer monsoon. In the mature phase of an ElNin‹o, two anomalous anticyclones located overthe SIO and WNP, respectively, dominate theAAM anomalies. These suppress the East Asianwinter monsoon and Australian summer monsoonsimultaneously. During the boreal spring andearly summer after the El Nin‹o mature phase,the anomalies are primarily found in the WNPand East Asia. The associated anomalous anticy-clone is over the Philippine Sea, suppressing localconvection while enhancing rainfall over the EastAsian subtropical front. Thus, from the develop-ment to decay phase of El Nin‹o, the variabilitycenter of the AÂM anomalies shifts from theIndian to East Asian sector. Note that, Meehl(1987) and others pointed out that a strong(weak) Indian summer monsoon is followed by

a strong (weak) Australian summer monsoon.This is due to the eastward shift of the AÂManomalies from the developing to mature phaseof ENSO, which has to do with the phase lockingof ENSO to the annual cycle. The phase lockingof El Nin‹o/La Nin‹a peaks to boreal winter is oneof the most robust features of ENSO cycle. It isprimarily due to the seasonally varying basic statein the Paci¢c Ocean (Tziperman et al., 1998; Anand Wang, 2001).

The internal factors that a¡ect the strength ofAÂM include not only remote forcing fromENSO but also the monsoon and warm oceaninteraction. The strongest Indian monsoonanomalies occur during the fall of an El Nin‹odeveloping year, while the strongest East Asianstrongest anomalies occur in the spring after theEl Nin‹o year. These di¡erences are attributed to:(1) the manner in which the regions are a¡ectedby remote forcing of the AÂM circulation, and(2) the ENSO-induced local monsoonôceaninteraction that substantially modi¢es directENSO impacts as discussed in Section 5.2.

The factors that control monsoon intensity (theamplitude of the annual reversal of the surfacewinds and/or the contrasting precipitation ratebetween wet summer and dry winter) may be clas-si¢ed in two groups: the forcing external to thecoupled atmosphereôcean^land system (geo-graphic and solar orbital forcing) and the forcinginternal to the coupled climate system, such as(remote) ENSO forcing, local monsoonôcean in-teraction, landâtmosphere interaction, and extra-tropical in£uences (ice or snow cover). The mech-anisms operating on the annual and interannualtimescales are dominated, respectively, by the ex-ternal and internal forcing. The di¡erences be-tween the Indian and East Asian monsoons isessentially determined by the relative strengthsof the external vs. internal forcing. The factorsthat determine the intensity of the monsoonal an-nual cycle (the ocean^land con¢guration and to-pography and annual cycle of solar radiation)share common features with the paleo-monsoonvariability caused by geographic/orbital forcing.The understanding gained from the aforemen-tioned investigation may shed light on the mech-anisms of monsoon variability that operate di¡er-



ently on the short-term (interannual to interdeca-dal) and geological timescales.

The variability associated with ENSO (or thethermal conditions in the Paci¢c) has a moderateimpact on the strength of the Arabian Seasummer monsoon (on average, it weakens byabout 1^2%), whereas it can substantially modifythe strength of the South China Sea summer mon-soon: by about 10% on average and about 40% inindividual strong events. The latter suggests that alarge change in the Paci¢c thermal conditions(disappearance of El Nin‹o, for instance) couldcause considerable changes of the amplitude ofthe East Asian monsoon. This di¡erence mayhave signi¢cant implications for interpreting thepaleo-monsoon records, which are used to deducethe geologic evolution of the Indian and EastAsian monsoon.

The phase of the summer-monsoon-in£uencedpollen records o¡ northern Australia match ex-actly the summer-monsoon phase measured incores from the Indian Ocean (Clemens and Ker-shaw, pers. commun.). How to interpret this link-age on orbital timescale? Does the in-phase rela-tionship between the Indian and the Australianmonsoons on the orbital timescale share commonmechanism(s) with that on the interannual time-scale? Our results suggest that although the in-phase relationship occurs both on orbital and in-terannual timescales, the mechanisms that estab-lish these phase relationships are not necessarilythe same. The quasi-in-phase relationship on theinterannual timescale results primarily from theimpacts of the remote ENSO. A typical warmevent develops in boreal summer and matures to-ward the end of the year (as shown in Fig. 7).From boreal summer to winter, the remote Nin‹oforcing continuously suppresses the maritime con-tinent convection through changing Walker circu-lation. During Nin‹o development, the suppressedmaritime continent convection extends to thesouthern tip of Indian (Fig. 7a), causing weakIndian summer monsoon. During the ensuing aus-tral summer, the same suppressed maritime con-tinent convection weakens the Australian summermonsoon. Hence, on the interannual timescale thelinkage between the Indian and Australiansummer monsoons arises from the Nin‹o forcing.

This mechanism can be responsible for the orbit-al-scale phase coupling between the Indian andAustralian summer monsoons only if long-termENSO variability dominates the sediment signalas suggested in Clement et al. (1999). In contrast,the results derived from the present study suggestthat the coupling of the East Asian winter andAustralian summer monsoon is a robust featureon the annual timescale. This linkage is estab-lished by geographic forcing. It is likely that thiscoupling would operate on orbital timescales.Understanding the extent to which ENSO versusorbital/geographic forcing is responsible for thephase coupling of Indian and Australian summermonsoons will require establishing the phase ofthe East Asian winter monsoon at orbital time-scales as well as further modeling e¡orts usingfully coupled GCM’s.

Acknowledgements

B. Wang is in part supported by the NOAAPaci¢c program and the NSF Climate Dynamicsprogram.

References

An, S.-I., Wang, B., 2001. Mechanisms of locking the El Nin‹o/La Nin‹a mature phases to boreal winter. J. Clim. 14, 2164^2176.

Chen, T.C., Weng, S.P., Yamazaki, N., Kiehne, S., 1998. In-terannual variation in the tropical cyclone formation overthe western North Paci¢c. Mon. Wea. Rev. 126, 1080^1090.

Chen, L.-X., Dong, M., Shao, Y.-N., 1992. The characteristicsof interannual variations on the East Asian monsoon.J. Meteor. Soc. Jpn. 70, 397^421.

Clemens, S.C., Prell, W.L., 1990. Late Pleistocene variabilityof Arabian Sea summer monsoon winds and continentalaridity: Eolian records from the lithogenic componenet ofdeep-sea sediments. Paleoceanography 5, 109^145.

Clemens, S.C., Prell, W.L., Murray, D.W., Shimmield, G.,Weedon, G., 1991. Forcing mechanisms of the Indian Oceanmonsoon. Nature 353, 720^725.

Clemens, S.C., Murray, D.W., Prell, W.L., 1996. Nonstation-ary phase of the PlioPleistocene Asian monsoon. Science274, 943^948.

Clement, A.C., SeUger, R., Cane, M.A., 1999. Orbital controlson the El Nin‹o/southern oscillation and the tropical climate.Palaeoceanography 14, 441^456.



Davey, M.K., Huddleston, M., Sperber, K.R., Braconnot, P.,Bryan, F., Chen, D., Colman, R.A., Cooper, C., Cubasch,U., Delecluse, P., DeWitt, D., Fairhead, L., Flato, G., Gor-don, C., Hogan, T., Ji, M., Kimoto, M., Kitoh, A., Knut-son, T.R., Latif, M., LeTreut, H., Li, T., Manabe, S., Me-choso, C.R., Meehl, G.A., Power, S.B., Roeckner, E.,Terray, L., Vintzileos, A., Voss, R., Wang, B., Washington,W.M., Yoshikawa, I., Yu, J.Y., Yukimoto, S., Zebiak, S.E.,2002. STOIC: A study of coupled model climatology andvariability in tropical ocean region. Clim. Dyn. 18, 403^420.

Fu, C., Teng X., 1988. The relationship between ENSO andclimate anomaly in China during the summer time (in Chi-nese). Sci. Atmos. Sinica (Special issue), pp. 133^141.

Gill, A.E., 1980. Some simple solutions for heat-induced trop-ical circulation. Quart. J. R. Meteorol. Soc. 106, 447^462.

Ju, J., Slingo, J.M., 1995. The Asian summer monsoon andENSO. Quart. J. R. Meteorol. Soc. 121, 113^168.

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven,D., Gandin, L., Iredell, M., Saha, S., White, G., Woollen, J.,Zhu, Y., Leetmaa, A., Reynolds, B., Chelliah, M., Ebisuza-ki, W., Higgins, W., Janowiak, J., Mo, K.C., Ropelewski,C., Wang, J., Jenne, R., Joseph, D., 1996. The NCEP/NCAR 40-year reanalysis project. Bull. Am. Meteorol.Soc. 77, 437^471.

Krishnamurti, T.N., 1971. Tropical east^west circulations dur-ing the northern summer. J. Atmos. Sci. 28, 1342^1347.

Krishnamurti, T.N., Bhalme, H.N., 1976. Oscillations of amonsoon system. Part I: Observational aspects. J. Atmos.Sci. 33, 1937^1954.

Krishnamurti, T.N., Surge, N., Manobianco, J., 1985. Annualcycle of the monsoon over the global tropics. WMO worldclimate research programme publications Ser. 4, WMOTD65, Part IV-I-IV-21.

Kumar, K.K., Rajagopalan, B., Cane, M.A., 1999. On theweakening relationship between Indian monsoon andENSO. Science 284, 2156^2159.

Kutzbach, J.E., Guetter, P.J., 1984. The sensitivity of monsoonclimates to orbital parameter changes for 9000 years B.P.:Experiments with the NCAR GCM. In: Berger, A., Imbrie,J., Hays, J., Kukla, G., Sultzman, B. (Eds.), Milankovitchand Climate: Understanding the Response to AstronomicalForcing. Reidel, Dordrecht, pp. 801^820.

Lau, K.-M., Weng, H., 2001. Coherent modes of global SSTand summer rainfall over China: An assessment of the re-gional impacts of the 1997^1998 El Nin‹o. J. Clim. 14, 1294^1308.

Latif, M., Sperber, K., Arblaster, J., Braconnot, P., Chen, D.,Colman, A., Cubasch, U., Cooper, C., Delecluse, P., Dewitt,D., Fairhead, L., Flato, G., Hogan, T., Ji, M., Kimoto, M.,Kitoh, A., Knutson, T., LeTreut, H., Li, T., Manabe, S.,Marti, O., Mechoso, C., Meehl, G., Power, S., Roeckner, E.,Sirven, J., Terray, L., Vintzileos, A., VoM, R., Wang, B.,Washington, W., Yoshikawa, I., Yu, J., Zebiak, S., 2001.ENSIP: the El Nin‹o simulation intercomparison project.Clim. Dyn. 18, 255^276.

McBride, J.L., Davidson, N.E., Puri, K., Tyrell, G.C., 1995.The £ow during TOGA COARE as diagnosed by the

BMRC tropical analysis and prediction system. Mon.Wea. Rev. 123, 717^736.

Meehl, G.A., 1987. The annual cycle and interannual variabil-ity in the tropical Paci¢c and Indian Ocean region. Mon.Wea. Rev. 115, 27^50.

Meehl, G.A., Boer, G.J., Covey, C., Latif, M., Stou¡er, R.J.,2000. The coupled model intercomparison project (CMIP).Bull. Am. Meteorol. Soc. 81, 313^318.

Overpeck, J.T., Trumbore, S., Prell, W.L., 1996. The south-west Indian monsoon over the last 18,000 years. Clim. Dyn.12, 213^225.

Prell, W.L., Murray, D.W., Clemens, S.C., Anderson, D.M.,1992. Evolution and variability of the Indian Ocean summermonsoon: Evidence from the western Arabian Sea drillingprogram. In: Duncan, R.A., et al. (Ed.), The Indian Ocean:A Synthesis of Results from the Ocean Drilling Program.Am. Geophys. Union, Washington, DC, pp. 447^469.

Ramage, C., 1971. Monsoon meteorology. International Geo-physics Series Vol. 15. Academic Press, San Diego.

Reynolds, R.W., Smith, T.M., 1994. Improved global sea sur-face temperature analyses using optimum interpolation.J. Clim. 7, 929^948.

Saji, N.H., Goswami, B.N., Vinayachandran, P.N., Yamagata,T., 1999. A dipole mode in the tropical Indian Ocean. Na-ture 401, 360^363.

Schulz, H., von Rad, U., Erienkeuser, H., 1998. Correlationbetween Arabian Sea and Greenland climate oscillations ofthe past 110,000 years. Nature 393, 54^57.

Shen, S., Lau, K.M., 1995. Biennial oscillation associated withthe East Asian summer monsoon and tropical sea surfacetemperatures. J. Meteorol. Soc. Jpn. 73, 105^124.

Shimmield, G.B., Mowbray, S.R., Weedon, G.P., 1990. A 350ka history of the Indian Southwest Monsoon ^ evidencefrom deep-sea cores, northwest Arabian Sea. Trans. R.Soc. Edinb. 81, 289^299.

Sirocko, F., 1991. Deep-sea sediments of the Arabian Sea: Apaleoclimatic record of the southwest-Asain summer mon-soon. Geol. Rundsch. 80, 557^566.

Tao, S., Chen, L., 1987. A review of recent research on theEast Asian summer monsoon in China. In: Chang, C.P.,Krisnamurti, T.N. (Eds.), Monsoon Meteorology. OxfordUniversity Press, Oxford, pp. 60^92.

Trenberth, K.E., Stepaniak, D.P., Caron, J.M., 2000. Theglobal monsoon as seen through the divergent atmosphericcirculation. J. Clim. 13, 3969^3993.

Tziperman, E.L., Cane, M.A., Zebiak, S.E., Xue, Y., Blumen-thal, B., 1998. Locking El Nino’s peak time to the end of thecalendar yearin the delayed oscillator picture of ENSO.J. Clim. 11, 2191^2199.

Wang, B., 1994. Climate regimes of the tropical convectionand rainfall. J. Clim. 7, 1109^1118.

Wang, B., Lin, H., 2002. Rainy season of the Asian^Paci¢csummer monsoon. J. Clim. 15, 386^398.

Wang, B., Chan, J.C.L., 2002. How Strong ENSO Eventsa¡ect tropical storm activity over the western North Paci¢c.J. Clim. 15, 1643^1658.

Wang, B., Xu, X., 1997. Northern Hemisphere summer mon-



soon singularities and climatological intraseasonal oscilla-tion. J. Climatol. 10, 1071^1085.

Wang, B., Wu, R., Fu, X., 2000. Paci¢cÊast Asian tele-connection: How does ENSO a¡ect East Asian climate?J. Clim. 13, 1517^1536.

Wang, B., Wu, R., Lau, K.M., 2001. Interannual variability ofAsian summer monsoon: Contrasts between the Indian andWestern North Paci¢cÊast Asian monsoons. J. Clim. 14,4073^4090.

Wang, L., Sarnthein, M., Erlenkeuser, H., Grimalt, J.,Grootes, P., Heilig, S., Ivanova, E., Kienast, M., Pelejero,C., P£aumann, U., 1999. East Asian monsoon climate dur-ing the Late Pleistocene: High-resolution sediment recordsfrom the South China Sea. Mar. Geol. 156, 245^284.

Wang, P., 1999. Response of Western Paci¢c marginal seas toglacial cycles: Paleoceanographic and sedimentological fea-tures. Mar. Geol. 156, 5^40.

Webster, P.J., Magana, V.O., Palmer, T.N., Shukla, J., Tomas,R.A., Yanai, M., Yasunari, T., 1998. Monsoons: Processes,predictability, and the prospects for prediction. J. Geophys.Res. 103, 14451^14510.

Xie, P., Arkin, P.A., 1997. Global precipitation: A 17-yearmonthly analysis based on gauge observations, satellite esti-mates and numerical model outputs. Bull. Am. Meteorol.Soc. 78, 2539^2558.



Documents

Contrasting the Indian and East Asian monsoons ... · Contrasting the Indian and East Asian monsoons: implications on geologic timescales Bin Wanga ;b, Steven C. Clemensc, Ping Liu