Page 1
INTRASEASONAL VARIABILITY OFTHE SOUTH
ASIAN MONSOON
by
DAVID MICHAEL LAWRENCE
B.S.,Universityof CaliforniaatSanDiego,1993
M.S.,Universityof Colorado,1996
A thesissubmittedto the
Facultyof theGraduateSchoolof the
Universityof Coloradoin partialfulfillment
of therequirementsfor thedegreeof
Doctorof Philosophy
Programin AtmosphericandOceanicSciences
1999
Page 2
This thesisentitled:
IntraseasonalVariability of theSouthAsian
Monsoon
writtenby David MichaelLawrence
hasbeenapprovedfor theProgramin Atmospheric
andOceanicSciencesby
PeterJ.Webster
Andrew M. Moore
Date
Thefinal copy of this thesishasbeenexaminedby thesignatories,andwefind that
boththecontentandtheform meetacceptablepresentationstandardsof scholarly
work in theabovementioneddiscipline.
Page 3
Lawrence,David Michael(Ph.D., Astrophysical,Planetary, and
AtmosphericSciences)
IntraseasonalVariability of theSouthAsianMonsoon
Thesisdirectedby ProfessorPeterJ.Webster
The SouthAsian monsoonexhibits pronouncedintraseasonalvariability
on timescalesrangingfrom a few daysto morethana month. A principalpurpose
of this studyis to provide a comprehensive analysisof low-frequency monsoonin-
traseasonalvariability andto determineits structurein spaceandtime. Large-scale
activeandbreakperiodsof rainfall areassociatedwith theslowly evolving Intrasea-
sonalOscillation(ISO)thatis characterizedduringnorthernsummerby anapparent
northward movementof convectionemanatingfrom the centralequatorialIndian
Ocean. The evolution of ISO convectioncanbe thoughof in termsof propagat-
ing equatorialmodes.Surfacefrictional convergenceinto a Rossbycell that is ex-
citedby equatorialISO convectiongeneratesa bandof convectionthat is oriented
southeastto northwestandstretchesfrom theequatorto about20�N. Viewedalong
any meridianthemodeappearsto propagatenorthwardwhile equatorialconvection
propagatesto theeast.
Interannualvariationsof summertimeISO activity are investigatedand
are found to be relatedto year-to-yearchangesin the numberof discreteevents.
Seasonsof high ISO activity exhibit significantlymorelow precipitationdaysand
consequentlydeficientseasonalrainfall thanseasonscharacterizedby little or no
ISO activity. ISO activity is found to be uncorrelatedto the El Nino-Southern
Oscillation(ENSO)or any othercontemporaneousseasurfacetemperature(SST)
variability. ThesummertimeISO activity doesexhibit a reasonablystronginverse
Page 4
iv
relationshipwith SouthAsianmonsoonstrength.Theyear-to-yearvariationsof ISO
activity alsoexhibit adominantbiennialtimescale.
A seconddominant mode of intraseasonalvariability is made up of
synoptic-scalewestward propagatingconvective disturbanceswith timescalesbe-
tween5 and10days.DuringtheactiveISOphaseoverIndia,westwardpropagating
synoptic-scalewave activity is above normal. Thedevelopmentof high-frequency
convectivedisturbancesover theBayof BengalandpeninsularIndia is attributedto
instability that is favoredin regionsof strongeasterlyvertical wind shearin con-
junctionwith equatorialheating.As ISOconvectionmovesinto thewesternPacific
Oceanregion westward propagatingsynoptic-scalewavesareexcited from which
point they propagateto thewestacrosssoutheastAsia into theBayof Bengalbring-
ing episodesof significantrainfall during the suppressed,and normally dry, ISO
phaseover India.
Page 5
ACKNOWLEDGMENTS
Firstandforemost,I wouldliketo thankmy advisor, ProfessorPeterWeb-
ster, who providedthemotivationanddirectionfor this work. Thefreedomto pur-
suea numberof my own ideas,oftendown thewrongpath,wasinvaluablein my
professionaldevelopment.Additionally, Peterprovidedmetheopportunityto take
partin theJointAir-SeaMonsoonINteractionExperiment(JASMINE) onboardthe
NOAA ShiptheRonaldH. Brown in theBayof Bengal,whichwasanenlightening
highlightof my graduateschoolcareer.
Thanksalsoto my thesiscommitteemembers:PeterWebster, GeorgeKi-
ladis,Andy Moore,JerryMeehl,Bob Grossman,andJeff Forbesfor their time and
expertise.
Specialthanksto George Kiladis and Harry Hendonwhosework and
knowledgeinspireda goodportion of the work of this thesis. Gil Compo,Matt
Wheeler, and Chris Torrencehelpedout immenselywith discussionson statisti-
cal techniques.Thanksalsoto Bob Tomas,KamranSahami,JohannesLoschnigg,
BrianMapes,PaquitaZuidema,JohnFasullo,Jeff Hicke,CristinaPerez,andChristy
Oelfke Clark for many rewarding discussions.I would also like to thankMurry
Salbyfor his excellentteachingof dynamicsandspectralanalysis,PeterWebster
andJudyCurry for their excellentandthoughtprovoking introductionsto thefield,
andAndy Moorefor his role in bringingtheForecastingWeatherandClimateclass
to the CDC Friday afternoonweatherandclimatediscussionsessions(and,on a
completelytangentialnote,Jim Corbridgefor his thought-provokingcourseonwa-
ter law).
Page 6
vi
I wouldliketo thankthePAOSofficestaff, especiallyBonnieGrebe,Kelly
Duong,RobertLambeth,SherryYearsley, andLois Macaria.
I wouldalsoliketo thankHai-RuChangandKamranSahamiwhocontin-
uallysolvedproblems,particularlymine,with thecomputersystemwhile theparade
of sys-opscameandwent.Thecomputingwasdoneat thePAOSComputingFacil-
ity primarilyusingIDL andFortran.ThedatawasobtainedfromtheNOAA Climate
DiagnosticCenterandtheNationalCenterfor AtmosphericResearcharchives.
Specialthanksgo to the office matesboth pastandpresentfor keeping
thingsfun, includingPaquita,Johannes,Kamran,Chris,Gil, Cristina,andChristy.
Paquita,in particular, hasbeena constantsourceof supportandenthusiasm.Jo-
hannes,hasbeen,well, Johannes.The snackcrew got me throughmany a slow
afternoon.ThematesatEdgewood: Diane,Matt, Brett,Courtney, Tim, Erin, Dave,
andAdammademy time away from schooljust that. Life wouldn’t have beenthe
samewithoutsomegreatsnow seasonsandthe10thMountainDivisionHut system
andmy soccerteams,theGunners,theBlues,theJazz,andScatter.
Finally, I want to thankmy family who meanmoreto meI canexpress.
My parents,GeorgeandKathy, whohavebeensosteadfastlysupportivethatI hardly
evennoticedit andthankfully convincedmeto apply to CU; my brotherSteve for
beingmy bestbuddy; andGramfor theScrabblegamesandbeingthebestgrand-
motherever. Last, but no wherenearleast, I would like to thank my long-time
friend andgirlfriend Dianewho hasanunmatchedspirit andwho is a prettygood
copy editorto boot.
Page 7
vii
THE MONSOONSOVEREIGN
Theatmospherewasstifling: theair wasstill asdeath,As theparchedjheelsemittedtheir foul andcharnelbreath.A mantleof redshadow envelopsall around—The trees, the grass,the hamlets,as the storm-cloudsforward
bound.Of asudden,comesawhirlwind, dancing,spinningrapidly;Thengustongustburstsquick, incessant,mad,rushingfuriously.A crash—andtheMonsoon’sonus,in torrentseverywhere,With thebellowing roarof thunder, andlightning,flareonflare.Thetempest’snow abated;ahushfalls o’er thescene;Thenmyriadbirdsstartchatt’ring andthegrassagainis green,Thefieldslikevast,still mirrors,in sheetsof waterlie,Thefrogs,in droningchorus,singhoarsetheir lullaby,Eachtankandpool is flooded,greatriversbursttheirbanks,King Summer’s reignis ended,theMonsoonsovereignranks.
— L. H. Niblett (1938),adapted
Page 8
CONTENTS
CHAPTER
1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 AnnualCycleandtheMeanMonsoon . . . . . . . . . . . . 9
1.2 Variability of theSouthAsianMonsoon . . . . . . . . . . . 16
1.2.1 TheIntraseasonalOscillation . . . . . . . . . . . . . 16
1.2.2 OtherIntraseasonalVariability . . . . . . . . . . . . 18
1.2.3 InterannualVariability . . . . . . . . . . . . . . . . 18
1.3 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2 DATA AND METHODS. . . . . . . . . . . . . . . . . . . . . . . 22
2.1 Datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.1.1 OLR . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.1.2 NCEP/NCARReanalysis. . . . . . . . . . . . . . . 23
2.1.3 SST . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1.4 Precipitation. . . . . . . . . . . . . . . . . . . . . . 24
2.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.1 StatisticalTechniques. . . . . . . . . . . . . . . . . 25
2.2.2 LaggedCross-CorrelationandLinearRegression . . 26
3 STRUCTUREAND EVOLUTION OF THE INTRASEASONAL
OSCILLATION . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.1 Theoriesfor SummertimeISOEvolution . . . . . . . . . . . 30
3.2 TimeseriesAnalysisandTemporalandSpatialFiltering . . . 33
3.2.1 FourierSpectralAnalysis . . . . . . . . . . . . . . . 33
Page 9
ix
3.2.2 Wavenumber-Frequency Analysis . . . . . . . . . . 36
3.2.3 Filtering . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3 SeasonalStructureof the PlanetaryScaleIntraseasonalOs-
cillation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3.1 Convection . . . . . . . . . . . . . . . . . . . . . . 50
3.3.2 Circulation . . . . . . . . . . . . . . . . . . . . . . 54
3.4 DetailedStructureof theSummerIntraseasonalOscillation . 61
3.4.1 Separationof Modes . . . . . . . . . . . . . . . . . 68
3.4.2 SeaSurfaceTemperature. . . . . . . . . . . . . . . 74
3.5 SummaryandDiscussion . . . . . . . . . . . . . . . . . . . 78
4 INTERANNUAL VARIATIONS OF THE INTRASEASONAL
OSCILLATION . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.1 InterannualandIntraseasonalVariability . . . . . . . . . . . 83
4.2 Measuresof BorealSummerISOActivity . . . . . . . . . . 88
4.3 MonsoonIndicesandMonsoon-ENSORelationships . . . . 98
4.3.1 MonsoonIndices . . . . . . . . . . . . . . . . . . . 98
4.3.2 Monsoon-ENSORelationships. . . . . . . . . . . . 102
4.4 InterannualVariationsof ISOActivity . . . . . . . . . . . . 104
4.4.1 Changesin ISOCharacteristics. . . . . . . . . . . . 104
4.4.2 Relationto InterannualSSTVariability . . . . . . . . 112
4.4.3 ENSO . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.4.4 OtherSSTVariability . . . . . . . . . . . . . . . . . 116
4.4.5 Relationto InterannualSouthAsianMonsoonVari-
ability . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.4.6 ISO,ENSO,andtheIndianMonsoon. . . . . . . . . 125
Page 10
x
4.5 SummaryandDiscussion . . . . . . . . . . . . . . . . . . . 125
5 MODULATION OFSYNOPTIC-SCALECONVECTION. . . . . 131
5.1 Synoptic-ScaleVariability . . . . . . . . . . . . . . . . . . . 131
5.2 Wavenumber-Frequency Spectra . . . . . . . . . . . . . . . 136
5.2.1 Method . . . . . . . . . . . . . . . . . . . . . . . . 136
5.2.2 OLR Wavenumber-Frequency VarianceSpectra . . . 137
5.2.3 Wavenumber-Frequency Filtering . . . . . . . . . . . 141
5.3 IntraseasonalOscillationandSynoptic-ScaleDisturbances. . 141
5.3.1 IntraseasonalOscillation . . . . . . . . . . . . . . . 141
5.3.2 Synoptic-ScaleWestwardPropagatingWaves . . . . 144
5.3.3 Synoptic-ScaleWavesandtheISOPhase . . . . . . 146
5.4 Synoptic-ScaleWaveActivity andISO . . . . . . . . . . . . 149
5.4.1 Measureof Synoptic-ScaleWaveActivity . . . . . . 149
5.4.2 ActiveISOPhase . . . . . . . . . . . . . . . . . . . 154
5.4.3 SuppressedISOPhase. . . . . . . . . . . . . . . . . 160
5.5 A CanonicalSequence. . . . . . . . . . . . . . . . . . . . . 166
6 CONCLUSIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.1 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.2 StructureandEvolutionof theISO . . . . . . . . . . . . . . 172
6.3 InterannualVariationsof theIntraseasonalOscillation . . . . 173
6.4 Modulationof Synoptic-ScaleWaves . . . . . . . . . . . . . 175
6.5 Implicationsfor Prediction . . . . . . . . . . . . . . . . . . 177
6.6 FutureWork . . . . . . . . . . . . . . . . . . . . . . . . . . 179
6.6.1 JASMINE . . . . . . . . . . . . . . . . . . . . . . . 180
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Page 11
FIGURES
FIGURE
1.1 PrecipitationTimeseriesfor JJAS 1987and1988 . . . . . . . . . 5
1.2 SpectralAnalysisof OLR . . . . . . . . . . . . . . . . . . . . . 6
1.3 Climatologyof OLR and850-mbWind . . . . . . . . . . . . . . 10
1.4 Climatologyof PrecipiationEstimates . . . . . . . . . . . . . . 11
1.5 Climatologyof Zonal200-mbWind . . . . . . . . . . . . . . . 14
1.6 Climatologyof SST . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7 Time-SpaceSectionsof OLR Anomalies,Summer1996 . . . . . 19
2.1 Cross-CorrelationandLinearRegressionExample . . . . . . . . 28
3.1 RegionalSpectralAnalysisof OLR . . . . . . . . . . . . . . . . 35
3.2 Wavenumber-Frequency Spectraof OLR . . . . . . . . . . . . . 37
3.3 Climatologyof OLR��������� � . . . . . . . . . . . . . . . . . . . . 39
3.4 Wavenumber-Frequency Spectraof ZonalWind andDivergence. 41
3.5 AnnualCycleTime-LatitudeDiagramsof SSTandOLR��������� � . 42
3.6 Climatologyof OLR���������� � . . . . . . . . . . . . . . . . . . . . 44
3.7 OLR, 200-mbWind andStreamfunctionLaggedRegressions . . 47
3.8 OLR, 850-mbWind andDivergenceLaggedRegressions . . . . 48
3.9 OLR, 1000-mbWind andDivergenceLaggedRegressions. . . . 49
3.10 RegressedOLR VersusLatitude. . . . . . . . . . . . . . . . . . 52
3.11 Lag-LatitudeSectionof RegressedOLR and1000-mbDivergence 53
3.12 Simple Solutions for Heat-InducedTropical Circulation, Gill
(1980) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Page 12
xii
3.13 Lag-LongitudeSectionof RegressedOLR andv������ . . . . . . . 59
3.14 OLR and200-mbWind AnomaliesRegressedOntoOLR ���������� . 62
3.15 OLR and850-mbWind AnomaliesRegressedOntoOLR ���������� . 63
3.16 Lag-SpaceSectionsof RegressedOLR �������� . . . . . . . . . . . 65
3.17 Lag-SpaceSectionsof OLR �������� Segregatedby EventType . . . 71
3.18 Histogramof ISOEvents(SN,EN, E) PerFortnight . . . . . . . 73
3.19 MeanSSTBasicStateDifferenceMapof SN– EN Events . . . 75
3.20 LaggedSSTAnomaliesDuringSN Events . . . . . . . . . . . . 77
3.21 Lag-SpaceDiagramsof SSTandOLR �������� for SNEvents. . . . 79
4.1 PrecipitationTimeseriesfor JJAS 1987and1988 . . . . . . . . . 86
4.2 PrecipitationTime-LatitudeSectionsfor JJAS 1987and1988 . . 87
4.3 LeadingTwo EOFLoadingVectorsfor OLR �������� . . . . . . . . 91
4.4 LeadingThreeRotatedEOFLoadingVectorsfor OLR �������� . . . 93
4.5 WaveletAnalysisof OLR����� . . . . . . . . . . . . . . . . . . . 95
4.6 WaveletAnalysisof OLR ���������� . . . . . . . . . . . . . . . . . . 96
4.7 OLR �������� , OLR����� , andOLR ���������� VarianceMaps . . . . . . . 99
4.8 AIRI-OLR RegressionMap . . . . . . . . . . . . . . . . . . . . 101
4.9 SeasonalISOActivity Indices. . . . . . . . . . . . . . . . . . . 105
4.10 ISO-OLR �������� RegressionMap . . . . . . . . . . . . . . . . . 107
4.11 CompositePrecipitationSextile Histogram:India . . . . . . . . 109
4.12 CompositePrecipitationSextile Histogram:Bayof Bengal . . . 111
4.13 CompositeWaveletAnalysisfor WarmNino3SSTYears . . . . 113
4.14 CompositeWaveletAnalysisfor CoolNino3SSTYears . . . . . 114
4.15 Regressionof GlobalSSTAnomaliesOntoISOActivity Indices 117
4.16 ScatterDiagramof OLR ����� vs. OLR-India . . . . . . . . . . . 118
Page 13
xiii
4.17 [OLR ������] , AIRI, andOLR-IndiaComparison . . . . . . . . . 121
4.18 CompositeWaveletAnalysisfor WetAsianMonsoonYears . . . 123
4.19 CompositeWaveletAnalysisfor Dry AsianMonsoonYears . . . 124
4.20 Regressionof MeanOLR Onto[OLR ������] andJJAS Nino3SST 126
4.21 FourierSpectraof ISOActivity Indices,AIRI, andNino3SST . 129
5.1 PrecipitationTimeseriesandTime-SpaceSections:Summer1992 133
5.2 Wavenumber-Frequency Spectraof OLR at10�–20
�N . . . . . . 138
5.3 OLR and850-mbWind LaggedRegressionMapsof ISO . . . . 143
5.4 OLR, 850-mbWind, and850-mbRelative Vorticity LaggedRe-
gressionMapsof Synoptic-ScaleEvents . . . . . . . . . . . . . 145
5.5 OLR��� ��!#" LaggedRegressionSectionsfor ConvectiveandSup-
pressedPhasesof ISO . . . . . . . . . . . . . . . . . . . . . . . 148
5.6 OLR��� ��!#" , OLR��� ��!#"�, andOLR� ����$�!�% for Juneto July1992 . . 150
5.7 Climatologyof OLR��� ��!#" andZonalVerticalWind Shear . . . . 153
5.8 Modulationof OLR��� ��!#"�
DuringConvectivePhaseof ISO . . . 155
5.9 Lagged Composite Timeseries of ISO, OLR��� ��!#"�, and
OLR��� ��!�%�
at75�E, 15
�N . . . . . . . . . . . . . . . . . . . . . 156
5.10 PowerSpectraDuringConvectiveandSuppressedISOPhase . . 159
5.11 OLR��� ��!#"�
DuringSuppressedISOPhaseOver India . . . . . . 162
5.12 Modulationof OLR��� ��!#"�
DuringConvectivePhaseof ISOover
WesternPacificOcean. . . . . . . . . . . . . . . . . . . . . . . 163
5.13 Lagged Composite Timeseries of ISO, OLR��� ��!#"�, and
OLR��� ��!�%�
at125�E, 15
�N . . . . . . . . . . . . . . . . . . . . 165
5.14 CirculationAnomaliesDuringWestwardOLR��� ��!#"�
. . . . . . 167
5.15 CanonicalEvolutionof ISOandSynoptic-ScaleWaves . . . . . 168
Page 14
xiv
6.1 Time-LatitudeSectionof Meteosat-5OLR DataDuringJASMINE 182
Page 15
xv
TABLES
TABLE
3.1 Large-ScaleISOEventModeCharacteristics. . . . . . . . . . . 70
4.1 CorrelationsBetweenIndianMonsoonIndices . . . . . . . . . . 103
4.2 CorrelationsBetweenSeasonalISO Activity and Indian Mon-
soonIndices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Page 16
CHAPTER1
INTRODUCTION
”K eepa register of all changesof wind and weatherat all houres,by nightand by day, shewing the point the wind blowsfrom,whetherstrongor weak:TheRains,Hail, Snowandthe like ... especiallyHurricanes... but aboveallto take exactcare to observetheTrade-Wines,aboutwhatdegreesof LatitudeandLongitudethey firstbegin, whereandwhenthey cease, or change, or growstronger or weaker.”
— Instructionsfrom the Royal Societyof London’s – Directions forSea-Men,Boundfor Far Voyages(1666)– indicatinganearlyappreciationoftheimportanceof understandingthevariability of themonsoon.
TheAsiansummermonsoonis themostvigorousweathersystemin the
world,profoundlyaffectingmostof thenationsof SouthandSoutheastAsia. Across
theregion,over75%of thetotalannualrainfall occursduringthesummermonsoon.
Nearly60%of theplanet’s populationrelieson thesoakingmonsoonrainsto sup-
port agriculturalproduction,to provide adequatedrinking water for humansand
Page 17
1. INTRODUCTION 2
livestock,and to generatehydroelectricpower that drivesagriculturaland indus-
trial production.Thesignificanceof themonsoonin people’s daily livescannotbe
overestimatedasemphasizedby KhushwantSingh (FeinandStephens1987):
Themonsoonis themostmemorableexperiencein thelivesof Indians... Whatthe four seasonsof themonsoonmeanto theEuropean,theoneseasonof themonsoonmeansto theIndian.Thesummermonsoonisprecededbydesolation;it brings with it the hopesof spring; it hasthe fullnessof summerand thefulfillment of autumnall in one. (p.38, Singh1987)
The monsoon’s influence,however, is not restrictedto the Asian con-
tinent. The intenseheatingassociatedwith the condensationof water vapor in
the upperatmosphereimpactsthe circulationpatternsthroughoutthe tropicsand
the extra-tropics (e.g., Yasunariand Yuji 1992). Additionally, observationalev-
idencepoints to a complex interrelationshipbetweenthe Asian monsoonandthe
El Nino-SouthernOscillation(ENSO),a relationshipthatwasfirst observedby Sir
Gilbert Walker (1924)andthathasbeenthesubjectof intensestudyover the last
decade (for review seeWebsteret al. 1998). The relationshipbetweenIndian
rainfall and the phaseof ENSO is not entirely clear althoughthere is somein-
dication that the monsoon’s role is active ratherthan passive (Normand1953;
Yasunari1990). The global impactsof ENSOfrom both a climatological (e.g.,
amongcountlessothersRopelewski andHalpert1987; Kiladis andDiaz1989)and
asociologicalperspective (e.g.,Glantz1996)arewell-documented,althoughby no
meanscomprehensive. If themonsoonindeedinfluencesENSO,thentheimpactsof
themonsoonon globalclimatewould bewidespread.Regardless,theAsianmon-
soonis an importantcomponentof the tropical coupledatmosphere/oceansystem
andneedsto beaccuratelyrepresentedin modelsin orderto increasepredictability
on theglobalscale.
Page 18
1. INTRODUCTION 3
The Asian monsoonexhibits high amplitudevariability on virtually all
timescalesfrom synopticto interdecadal.Theobservedfluctuationsof themonsoon
havebeenstudiedfor over200years(for historicalaccountof SouthAsianmonsoon
studies,seeKutzbach1987)with moreintensive studiesin the last100 yearsmo-
tivatedby a desireto understandandpredictlarge-scaledroughtssuchasthegreat
Indiandroughtandtheresultingfamineof 1877.On theinterannualtimescale,the
standarddeviationof totalIndiansummerrainfall is only 10%of thelongtermmean
of 853mm (Mooley andParthasarathy1984).However, evensuchsmallvariations
in seasonalprecipitationcansignificantly impactcrop production (Parthasarathy
et al. 1988; Gadgil 1996). For example, Websteret al. (1998)show that total
Indianrainfall is positively correlatedwith Indianrice productionat 0.61over the
last four decadeswith a 10%reductionin total rainfall leadingto anaverage15%
reductionin riceproduction.
While the stressis often on how the total seasonalrainfall impactscrop
productionandsocietyin general,it is eventsthatoccurwithin theseasonthatoften
have the greatestoverall impact. For example,the following is an excerptfrom a
FoodandAgricultureOrganizationreporton the1996Indianmonsoonseason:
Torrentialmonsoonrains in July over centraland north easternpartsof thecountrycausedseriousflooding.Latestestimatesindicatethatoverall 2.4mil-lion peoplewereaffected,with some900peopleand15,000livestockkilled.Around3.3 million hectaresof cropareawereaffectedand600,000hectaresseverelydamaged.Overall, the1996monsoonbeganon time ... providing fa-vorableplantingconditionsfor 1996/97paddyandcoarsegrains.However, ineastandwestMadhyaPradesh,low rainfall in JuneandearlyJuly resultedindelayedplantingof themaincropwhichmayaffectoverallproduction.
Apparently, both excessive rainfall anddeficientor delayedrainfall canadversely
affect cropproduction,which highlightstheneedfor betterunderstandingandpre-
dictionof boththeshortandlong termfluctuationsin precipitation.
Page 19
1. INTRODUCTION 4
The strong intraseasonalvariationsin precipitationin the SouthAsian
monsoonareobviousfrom inspectionof daily precipitationtimeseriessuchasthose
shown for the1987and1988seasonsin Fig. 1.1.During1987,therearethreeclear
”active” periodsof rainfall, whicharedefinedin thisstudysimplyasprolongedpe-
riodsof heavy precipitationover centralIndia. Theactive periodsareseparatedby
”break” periodswhich aredefinedin this studyasprolongedperiodsof low rain-
fall. The three1987active periodsareseparatedby approximately40 dayseach.
Theactiveperiodsof precipitationhavebeenlinkedto northwardmoving envelopes
of convectionfrom the equatorialIndianOceanwhich in turn have beenlinked to
large-scaleeastwardpropagatingconvectionenvelopesalongtheequator(e.g.,Ya-
sunari1979; JulianandMadden1981)(seeSection1.2.1andChapter3 for further
detailsanddiscussion).The entiresystemis termedthe IntraseasonalOscillation
(ISO).
While thenear40-dayvariability isadominantmodeof intraseasonalvari-
ability of rainfall in the SouthAsianmonsoonregion, othertimescalesareimpor-
tantaswell. For example,evidenceof shortertimescaleoscillationsis apparentin
Fig. 1.1 asheavy precipitationeventsseparatedby about6–9 days. Both the near
40-dayandthe6–9-daytimescalesof variability show up asstatisticallysignificant
spectralpeaksin outgoinglongwaveradiationrecords(Fig. 1.2).
Evidenceof interannualvariability is alsoapparentin Fig. 1.1. For this
region in centralIndia, the total precipitationwas 896 mm in 1988, which is in
excessof 50%morerain thanthe554mm thatfell in 1987. Thetwo yearsexhibit
notablydifferentintraseasonalfluctuations.While 1987containsat leastthreewell-
definedactive periodsseparatedby well-definedbreakperiods,1988appearsto be
essentiallydevoid of any obviousactive or breakperiodsexceptperhapsanactive
Page 20
1. INTRODUCTION 5
1987
Jun&01' Jun
&11
Jun&21
Jul&01' Jul
&11
Jul&21
Jul&31
Aug(10
Aug(20
Aug(30
Sep09' Sep
19Sep29
0'
5)
10
15
20*
mm
/day
1988
Jun&01' Jun
&11
Jun&21
Jul&01' Jul
&11
Jul&21
Jul&31
Aug(10
Aug(20
Aug(30
Sep09' Sep
19Sep29
0'
5)
10
15
20*
mm
/day~40 days ~40 days
6-9 days
6-9 days
Figure1.1. Timeseriesof daily precipitationrateestimatesaveragedover centralIndia (75+ –80+ E, 10+ –15+ N) for JunethroughSeptember, 1987and1988. Precipi-tationestimatesfrom stationdata(Section2.1.4).
Page 21
1. INTRODUCTION 6
0.30,
0.25,
0.20,
0.15,
0.10,
0.05,
Frequency (cycles/day)
0,
20-
40
60.
80
100
Spe
ctra
l Pow
er (
W m-2
)
3.3/
4 50
6.7.
10 20 Period (days)
5 days1
6-7 days
9-10 days2
35-50 days
Figure 1.2. Ensembleaveragedvariancespectrumof June-Septemberoutgoinglongwave radiationfor the years1974-1997in the Indian peninsularegion 703 –803 E, 103 –203 N. Dashedline is the 95% significancelevel theoreticalred-noisespectrum.Datafrom NOAA OutgoingLongwaveRadiationdataset(Section2.1.1)
Page 22
1. INTRODUCTION 7
periodin the middle of June. The starkdifferencesbetweenthesetwo seasonsin
bothtotal precipitationandtheintraseasonalvariability have stimulatedsubstantial
research(Krishnamurtiet al. 1989; Krishnamurtiet al. 1990)andareoftenused
asmeasuringsticksto assessthesimulationskill of GCMs.
Swaminathan(1987)suggeststhat the effectsof abnormalmonsoonson
crop productionand economicoutput could be mitigatedwith skillful forecasts
of seasonalmonsoonprecipitation.Unfortunately, accurateseasonalpredictionof
monsoonrainfall usingdynamicforecastmodelshasprovenelusive (Websteretal.
1998), althoughempirical forecastshave faredsomewhat better (Shukla1987;
Das1987). Thepotentialreasonsfor poor forecastskill in the Asianmonsoonre-
gion arevaried. For example, Fennessyet al. (1994)find that GCM simulations
of theAsianmonsoonarehighly sensitive to theformulationof orographyandsoil
moisture.TheAtmosphericModel IntercomparisonProject(AMIP) comparedthe
variousatmosphericGeneralCirculationModels(GCMs)andfoundthat thesimu-
latedrainfall distributionsovertheBayof BengalandtheIndiansubcontinentvaried
widely betweenmodels (Lau et al. 1996).Thelack of skill in theAsianmonsoon
region may be partly dueto the fact that the modelsdo not simulatewell the ob-
served northward propagationof monsoonconvective zones. Even singlemodels
thatcapturethestrongdry monsoon(1987)to wet monsoon(1988)transitionshow
little or no predictabilityin otheryears (SperberandPalmer1996). Websteret al.
(1998)arguethatthehigh degreeof spreadseenin monsoonforecastsis likely due
eitherto difficulty in modelingmonsoonregionsor to nonlinearerror growth due
to regionalhydrodynamicinstabilities.To answerthis question,continuedprogress
towardsa morecompleteunderstandingof thecomplex observedvariability of the
Page 23
1. INTRODUCTION 8
monsoonis requiredsuchthat the importantprocessesthatgovernmonsoonalpre-
cipitationcanbecorrectlyincorporatedinto forecastmodels.
The purposeof this dissertationis to examinein detail the intraseasonal
variability of theSouthAsianmonsoon.A numberof aspectsof the intraseasonal
variability will beaddressed.Theprincipalaimsareto:
4 describethespatialandtemporalstructureof thesummertimeISO,
4 illustratetheprimarysimilaritiesanddifferencesbetweenthesummertime
andwintertimeISOs,
4 interpretthenorthwardmovementof summertimeISOconvectionin terms
of equatorialwaves,
4 examinetherelationshipbetweenyear-to-yearfluctuationsof summertime
ISO activity andinterannualvariationsof SouthAsianmonsoonstrength,
and
4 investigatethemodulationof synoptic-scaleconvectionby theISO.
Theremainderof thischapteris devotedto adescriptionof theannualcy-
cle andthemeansummertimemonsoonanda review of thecurrentunderstanding
of the summertimeISO. Chapter2 describesthe datasetsusedandintroducesthe
statisticalmethodsusedin thisstudy. Chapter3 presentsthetemporalandstructural
evolutionof thesummertimeISOandcomparesit to thewintertimeISO.Theinter-
annualfluctuationsof summertimeISO activity areexaminedin Chapter4, which
also includesa moreextensive review of interannualSouthAsian monsoonvari-
ability. The synoptic-scalevariability in the SouthAsian monsoonregion andits
modulationby the ISO is investigatedin Chapter5. The final chapterprovidesa
synthesisof theobservationsandconclusionsaswell asa brief discussionof future
work.
Page 24
1. INTRODUCTION 9
1.1 Annual Cycle and the Mean Monsoon
The term ”monsoon” appearsto have originatedfrom the Arabic word
mausimwhichmeansseason.TheAsianmonsoonsystemcontainstwo distinctsea-
sons: the ”wet” andthe ”dry.” The wet occursduring borealsummerwhenwarm
andmoistwindsblow acrossSouthAsia from thesouthwest.During thedry win-
ter season,the winds reverse,blowing cool anddry air from the winter continent
acrossthe Indiansubcontinentfrom the northeast (for discussionof annualcycle
andfundamentalphysicalprinciplesrequiredto explain monsooncirculations,see
Webster1987). The focusof this studywill be on the wet summerseason,but it
is importantto keepin mind that thesummerAsianmonsoonis partof thegreater
Asian-Australianmonsoonsystem(Websteretal. 1998).
The climatologicalsummer, Juneto September(JJAS), andwinter, De-
cemberto March(DJFM),meandistributionsof 850-mbwinds,outgoinglongwave
radiation(OLR), andprecipitationareshown in Figs.1.3and1.4. During north-
ern summer, the low level circulationover the Indian Oceanand the Indian sub-
continentis dominatedby strongcross-equatorialflow and southwesterlywinds
acrossthe ArabianSea,the Indian subcontinent,andthe Bay of Bengal. Halley
(1686)first hypothesizedthat the cross-equatorialflow is causedby the tempera-
ture contrastbetweenthe cool southernoceansandthe hot continentallandmass,
a land-seatemperaturedifferencethatgeneratesa pressuregradientthatdrivesthe
winds. Hadley (1735)advancedthetheoryof themonsoonby includingtheeffects
of the rotationof the earthwhich explainsthe characteristicsouthwesterly, rather
thanpuresoutherlyflow. Themonsoonalcirculationactsasa moisture”conveyer
belt,” transportingmoisturefromthesouthIndianOceanandtheArabianSeatoward
the SouthAsian landmassandthe Bay of Bengal (e.g.,CadetandGreco1987;
Page 25
1. INTRODUCTION 10
JJAS: OLR and 850-mb Wind5
06 o 50
6 oE 1006 oE 150
6 oE 1606 oW
40oS
20oS
EQU
20oN
40oN
180
2007
2207
2407
W m8 -210. m/s
DJFM: OLR and 850-mb Wind5
06 o 50
6 oE 1006 oE 150
6 oE 1606 oW
40oS
20oS
EQU
20oN
40oN
180
200
2207
2407
W m8 -210. m/s
Figure1.3. Climatologyof OLR and850-mbwind for 23 borealsummers(JJAS),1975–1997(excluding 1978for OLR) and23 borealwinters(DJFM), 1974–1997(excluding1977and1978for OLR). Datafrom NOAA OutgoingLongwave Radi-ationdataset(Section2.1.1)andNCEP/NCARreanalysis(Section2.1.2).
Page 26
1. INTRODUCTION 11
JJAS: Precipitation9
0: o 50
: oE 100: oE 150
: oE 160: oW
40oS
20oS
EQU
20oN
40oN
0:2468101214mm/day;
DJFM: Precipitation9
0: o 50
: oE 100: oE 150
: oE 160: oW
40oS
20oS
EQU
20oN
40oN
0:2468101214mm/day
Figure1.4. Climatologyof precipitationestimatesfor 1979–1995JJAS andDJFM.Precipitationestimatesfrom stationdataover land and MSU satellitedataoveroceans(Section2.1.4).
Page 27
1. INTRODUCTION 12
FasulloandWebster1999a). The relatively steadyflux of moist, unstableair ad-
vectedover the SouthAsian landmasssupportsdeepconvectionandprecipitation
acrosstheSouthAsianmonsoonregion. MeansummertimeOLR lessthan220W
m <�= (OLR valueslessthan220W m <�= aregenerallyconsideredto indicaterainfall,
e.g. Arkin andArdanuy(1989))is locatednorthof theequatorovera largeportion
of theAsianmonsoonregion. Theclimatologicalsummerrainfall (Fig. 1.4)is char-
acterizedby averagedaily precipitationratesexceeding5 mmday< > throughoutthe
southandsoutheastAsian regionsaswell asthroughmuchof the northernIndian
OceanincludingtheeasternArabianSeaandespeciallytheBay of Bengal,where
thehighestmeanprecipitationratesin theentiremonsoonarefound. Theareasof
maximumprecipitationtendto be locatedupstreamof low, coastalmountainsthat
inducelow-level lifting of conditionallyunstableair (GrossmanandDurran1984;
GrossmanandGarcia1990).An intriguingmaximumin precipitationis alsofound
in thewinterhemispherejustsouthof theequatorat80? –90? E.
Themeanconvectionandprecipitation,duringnorthernwinter, shiftsfrom
southernAsia to alongandsouthof the equatorover the maritimecontinentand
northernAustraliaandextendinginto theSouthPacificConvergenceZone(SPCZ).
The850-mbcirculationis now markedby north-easterlywindsover theIndiansub-
continentandcross-equatorialflow from northto southover themaritimecontinent
whichprovidesmoisturefor theAustralianmonsoonrainfall.
TheborealsummerSouthAsianmonsoontendsto bestrongerthanthebo-
realwinter Australianmonsoonbothin termsof total precipitationandthestrength
of the monsoonalcirculation. The differencesare due in large part to the pres-
enceor absenceof anelevatedheatsource.TheSouthAsianmonsoonis strongly
influencedby the elevatedTibetanplateauheatsource. The onsetof the boreal
Page 28
1. INTRODUCTION 13
summermonsoonis coincidentwith the reversalof the meridional temperature
gradient in the upper tropospheresouth of the Tibetan plateau (Flohn 1957;
Li andYanai1996). The temperaturegradientsin the southernhemispherenever
reversedueto theabsenceof anelevatedheatsourcein Australia. However, there
arestill strongcross-equatorialpressuregradientsthatdrive theborealwintermon-
soon. Thegradients,not aslargeasthoseoccurringin theborealsummer, arethe
result of the intenseradiationalcooling over north Asia during winter (Webster
etal. 1998).
Theclimatological200-mbwindsareshown in Fig. 1.5for boththesum-
merandwinterseasons.Thepredominantfeatureis thepersistentupperlevel west-
erly jetspolewardof 20@ in bothhemispheres.OverSouthAsiaandAfrica, thereis
anupperlevel easterlyjet duringsummerthatis replacedby anupperlevel westerly
jet duringwinter. Overall,theupperlevel flow in bothseasonsis largelyoppositein
directionto thelow level flow yielding aneasterlyverticalwind shearduringsum-
merandawesterlyverticalwind shearduringwinter. Cross-equatorialreturnflow is
alsoevidentduringsummer, asignificantportionof which is divergent(notshown),
andthuscontributesto thelocalHadley cell (Krishnamurti1971).
TheclimatologicalSSTdistributionsareshown in Fig. 1.6.MeanSSTsin
excessof 28@ C areconfinedto equatoriallatitudes.In the IndianOceanbasinand
the westernPacific Ocean,a clearlatitudinalshift of warmSSTsinto the summer
hemisphereisevident.TheentireBayof BengalandtheeasternArabianSeapossess
warmSSTsduringsummer, extendingfrom 5@ Sto theBangladeshcoastat20@ N. In
winter, theareaof warmSSTis moreevenlydistributedabouttheequator, extending
from 10@ N to 15@ S. Also of noteis the expansionof warm SSTseastward in the
centralequatorialPacific Oceanandwestwardin theIndianOceanduringnorthern
Page 29
1. INTRODUCTION 14
200-mb WindAJJASB
06 o 50
6 oE 1006 oE 150
6 oE 1606 oW
40oS
20oS
EQU
20oN
40oN
-30-20-10061020730C4050D60Em sF -130. m/s
DJFMG
06 o 50
6 oE 1006 oE 150
6 oE 1606 oW
40oS
20oS
EQU
20oN
40oN
-30-20-1006102030C4050D60Em sF -130. m/s
Figure1.5. Climatologyof 200-mbwind (vectors)andzonal200-mbwind (con-tours) for 23 borealsummers(JJAS), 1975–1997and23 borealwinters (DJFM),1974–1996.Windsfrom NCEP/NCARreanalysis(Section2.1.2).
Page 30
1. INTRODUCTION 15
SST JJAS
100H oE 160H oW 60oW
40oS
20oS
EQU
20oN
40oN
0H161820I22I24I26I2830JoC
SST DJFM
100H oE 160H oW 60oW
40oS
20oS
EQU
20oN
40oN
0H161820I22I24I262830JoC
Figure1.6.Climatologyof SSTfor 16summers(JJAS) andwinters(DJFM),1982–1997.SSTfrom Reynoldsseasurfacetemperature(Section2.1.3).
Page 31
1. INTRODUCTION 16
winter. While the climatologicalSST distributions are presentedhere, thereare
alsosignificantintraseasonalchangesin SSTthatareimportantandwhich will be
discussedin Section3.2.2.
1.2 Variability of the South Asian Monsoon
1.2.1 The Intraseasonal Oscillation Thesouthwestor SouthAsian
monsoonis markedby episodesof prolongedabundantprecipitation(activeperiods)
separatedby periodsof prolongedreducedrainfall (breakperiods).Thetransitions
from active to breakperiodsandvice versaevolve slowly suchthat therearetyp-
ically 3 to 4 active periodsover the courseof a singlemonsoonseason,May to
September(Websteret al. 1998). Prolongeddry spellsat critical life stagesare
foundto adverselyaffectcropdevelopmentandgrowth, andhenceyields (seee.g.,
Lal et al. 1999). Consequently, understandingthe transitionsaswell as the tim-
ing of rainy anddry spellshassociologicalimportanceandis particularlyrelevant
for farmersandwatermanagerssinceadvanceinformationof forthcomingactive
andbreakspellscould be usedto implementagriculturalandwatermanagement
strategies.
Thelow-frequency active/breakcyclesoccurontimescalesof about30–40
days (Raghavanet al. 1975; Yasunari1979; Yasunari1980; Yasunari1981; Kr-
ishnamurtiandSubrahmanyam1982; LauandChan1986; GadgilandAsha1992)
with transitionsbetweenthetwo statestakingabout15–20days.Spectralpeaksin
monsoonalparametershavebeenfoundin the30–40-dayperiodbandin anumberof
studies(e.g., Yasunari(1979),cloudiness;Cadet(1986),precipitablewater; Knut-
sonetal. (1986),OLR anduK�L�M ; HartmannandMichelsen(1989),precipitation).It
will turn out not to bea coincidencethat theperiodof oscillationis approximately
Page 32
1. INTRODUCTION 17
thesameasthatof theMadden-JulianOscillation (MJO,MaddenandJulian1971;
MaddenandJulian1972),alsotermedtheIntraseasonalOscillation(ISO).
Climatologically, theISO is strongestduringtheborealwinterandspring
seasonswhenit appearsasa strictly eastwardly propagatinglarge-scalesystemof
convectionalongtheequator, extendingfrom theIndianOceaneastto thedateline
(HendonandSalby1994).During thesouthwestAsianmonsoonseason,theISOis
typically weaker andof morecomplex character(Madden1986). A fundamental
anduniquecharacteristicof the summerISO is a northward propagationof con-
vection,beginningin thecentralequatorialIndianOceanandendingat thefoot of
the Himalayanmountainsin northernIndia. The northward propagationof con-
vectionhasbeenobservedanddescribedby many authors (e.g.,Murakami1976;
Yasunari1979; Yasunari1980; Yasunari1981; SikkaandGadgil1980; Krishna-
murti andSubrahmanyam1982; SinghandKripalani 1985; Lau andChan1986;
WangandRui 1990). Thesimilarity betweenthetimescaleof the ISO andthecy-
cling timefrom active to breakto activeperiodover India led Yasunari(1979),and
subsequentlyJulianandMadden(1981)and Lau andChan(1986),to suggestthat
thenorthwardpropagationof convectionis associatedwith theeastwardpropagat-
ing cloudsalongtheequator (for review seeMaddenandJulian1994). Recently,
Websteretal. (1998)observedthatthenorthwardpropagationof precipitationfrom
theequatoris accompaniedby a concomitantsouthwardpropagationof convection
which lastsfor ashorterdurationandextendsonly to 15N S,whichmayexplainwhy
thereis aclimatologicalprecipitationsummermaximumsouthof theequatorin the
easternIndianOcean.
Page 33
1. INTRODUCTION 18
Many of thesefeaturescanreadilybeseenin observations.Figure1.7ais
a time-latitudesectionof OLR anomaliesalongIndianpeninsularlongitudes,75O –
85O E, during JuneandJuly, 1996. Both northward andsouthward propagationof
convectionis evidentin theraw data.Figure1.7bis a time-longitudesectionalong
a latitudinal swath betweenthe equatorand5O N. Two eastward propagatingcon-
vectiveeventsoccurduringthe2-monthperiod.During bothevents,thenorthward
propagationof convectionbegins subsequentto the passingof a large-scaleequa-
torial convective system. Figure1.7c, a time-longitudesectionalong 10O –15O N,
suggeststhatthis interpretationof theISOevolutionmaybeincorrector incomplete
becausea westto eastpropagationof convection,that lagsthe eastward propaga-
tion of convectionalongthe equatorby about5-10 days,is alsoapparentat these
latitudes.Oneof the goalsof this studyis to interpretthe off-equatorialeastward
propagationin thecontext of known ISOdynamics.
1.2.2 Other Intraseasonal Variability The second intraseasonal
mode that is prevalent during the Asian summer monsoon is composedof
high-frequency, 5–10-day, wavelike phenomena,including westward propagat-
ing synoptic-scalevorticity waves (Lau and Lau 1990) as well as monsoonde-
pressionsand lows (for review seeMak 1987). Thesestormstypically form
in the Bay of Bengalor in the westernPacific Oceanand propagatewestward
and north-westward into the GangesRiver valley (Krishnamurtiet al. 1977;
Sahaet al. 1981). Largeprecipitationamountsaccompany thesedisturbancesand
canleadto flooding,especiallyin thelowlandsof northeastIndiaandBangladesh.
1.2.3 Interannual Variability The interannual variability of the
SouthAsian monsoonhasbeenthe subjectof extensive research(for review see
Page 34
1. INTRODUCTION 19
OLR Anomaly 199675o-85oE
JunP01Q Jun
P08Q Jun
P15
JunP22
JunP29
JulP06Q Jul
P13
JulP20
JulP27
20R oS
10oS
EQUS10oNT
20oNT
30oNT(a)
Equator-5oN
60U oE 80oE 100oE 120oE 140oE
Jun-01PJun-08PJun-15PJun-22PJun-29PJul-06PJul-13PJul-20PJul-27P
(b)
10o-15oN
60U oES
80oES
100oES
120oES
140oES
Jun-01PJun-08PJun-15PJun-22PJun-29PJul-06PJul-13PJul-20PJul-27P
(c)
Figure 1.7. Time-spacediagramsof raw OLR anomaliesfor June–July, 1996.Anomaliesarecalculatedby removing the meanandthe first threeharmonicsoftheannualcycle (365.25,182.625,and121.75days).Contourintervalsareevery20W m V�W with darkshadesindicatingnegativeOLR anomaliesandlight shadesindi-catingpositive OLR anomalies.OLR anomaliesareaveragedalong(a) 75X –85X E,(b) Equator– 5X N, (c) 10X –15X N.
Page 35
1. INTRODUCTION 20
Websteret al. 1998). A numberof studieshave investigatedthecomplex relation-
shipbetweeneasternPacificSSTsandIndianmonsoonstrength(e.g.,Websterand
Yang1992; JuandSlingo1995; WainerandWebster1996).Althoughthesepara-
tion of causeandeffect is difficult, awarmENSOeventtendsto suppressmonsoon
convection (Webster1995)while, conversely, a strongmonsoontendsto inhibit
warm ENSOevents (Yasunari1990). Furthercomplicatingthe monsoon-ENSO
relationshipis that it appearsto possessinterdecadalvariability (Elliot andAngell
1988; TorrenceandWebster1999).
Websteret al. (1998)point out thatnearlyall El Nino yearsaredrought
yearsin India but not all droughtyearscorrespondto El Nino years. Therefore,
althoughthe ENSO-monsoonrelationshipappearsstrong it doesnot explain all
the interannualvarianceof monsoonstrength.Consequently, othersourcesof in-
terannualmonsoonvariability have beensoughtincluding links betweenmonsoon
strengthandotherSSTanomalies.A numberof studiesshow strongpositive lead
correlationsbetweenIndian OceanSSTandIndian rainfall (e.g.,averageMarch-
May ArabianSeaSST, RaoandGoswami (1988);precedingfall andwinter Indian
OceanSST, HarzallahandSadourny (1997)and Clarketal. (1999)).
Apart from the ENSOtimescale,the monsoonexhibits distinct biennial
fluctuations (Mooley andParthasarathy1984). The biennialcomponentmay be
relatedto the troposphericbiennialoscillation(TBO) that is found in many atmo-
sphericvariablesincludingprecipitation,surfacepressure,troposphericwinds,and
SST (Meehl1987; Meehl1997)or to biennialvariationsin Eurasiansnow cover
(Vernekaretal. 1995; Yang1996).
Page 36
1. INTRODUCTION 21
1.3 Summary
Themeanmonsoonandtheannualcycle,with emphasisonthedifferences
betweentheborealsummerandwinter, in rainfall distribution,OLR,circulation,and
SSThave beendescribed(Figs.1.3–1.6).Evidenceof thewide-rangingvariability
of theSouthAsianmonsoonfrom synopticto interannualvariability waspresented
(Figs.1.1–1.7).Thecurrentunderstandingof thesummertimeISO, synoptic-scale
variability, andinterannualvariability of theSouthAsianmonsoonwasreviewed.
The next chapterexaminesthe large-scalestructureandevolution of the
summerISO, the dominantmodeof intraseasonalvariability in the SouthAsian
monsoon.