Transcription
TAO,ProgressVol.5,No.3,349-363,Septemberin DendroclimaticROSANNED. D'ARRIGOl,StudisC.GORDON(Manuscriptin finalin IndonesiaJACOBylform1994and15 AprilPAUL.J. KRUSICl1994)ABSTRACTA 416-year(1514-1929)tree-ringwidth chronologyof teak (TectonagrandisL. F.) from Cepu, CentralJava,Indonesia(11135 宜 , 7 06'S),firstpublishedby Berlageover 60 yearsago, remainsone of the few high目。,lutionpaleoclimaticarchivesavailablefor the westernequatorialPacific.Herewepresentan updateof 曲is chronologyfrom1880-1991,whichextendsthe originalse 吋 es by 63 years.As wasfoundfor the originalBerlagerecord,由eupdatedchronologyis correlatedpositivelywithrainfalland inverselywithsea levelprl 田sureduringthe dry monsoonseason(aroundMayto October)justpriorto the periodof growthin Java.Thisteakrecordis also relatedto severalindicesof the EI Ni 晶 o-SouthernOscillation(ENSO),includingsealevel pressureat Darwin,Australia,Wright'srainfallindexfor the equatorialPacific,and the Quinnhistoricalrecordof ENSO.The years1737 and 1925,ratedas strongENSOeventsby Quinn,are rankedamongthe threelowestindexyearsin the Berlageteakrecord.1967,whichis not consideredto bean ENSOyear,has the lowestringwidthindexin the updatedseri 個 from1880-1991,and the low 回t dry seasonrainfallon recordfor J akar 妞, Indonesia(since1869).Thedry monsoondroughtduringthe extreme1982-83ENSOevent(althoughdevastatingfor muchof Indonesia)was not unusuallyseverebasedon eitherthe tree-ringor climate,Monsoon,Indonesia)INTRODUCTIONRainfallin the IndonesianArchipelagois directly linked to convectionassociatedwiththe IndonesianLow pressure cell, a key element of the El Ni 品。 -SouthernOscillation(ENSO)system(e.g.Hackert胡d Hastenrath1986, Diaz and K.iladis1992).Over muchofIndonesia,thereis a �at缸ound1Tree-Ringto take90%Laboratory,of thefor droughtseastwardintoplacein IndondroughtsLamont-Doherty巴to occur血e Pacific.sia由巴 warmENSOphase,whenConversely,abov 巳 normalrainfall血e coolduringin edPalisades,phase.withIt hasNewYorkbeenes位 matedEl N 的 0 events,regional10964,U.S.A.theand
350TAO,Vo1.5,No.3,September1994and about80% of such eventslinkedwith droughts(QuinnHalpert1987).The lirik betweenJavarainfalland ENSOappd 可 monsoons 巳ason,fromapproximatelyJuneto NovemberNicholls1981).Thisis also whenteak growthappearsto berainfalland ENSO(Murphyand Whetton1989,JacobyandOur understandingof ENSOhas improvedgreatlyasthe instrumentalclimaticrecord(e.g.Bradl 巳y et al. 1987,1990)and climatemodelingand prl 吋ictionof ENSOdynamicsBengtssonet al. 1993).Yi 前 , utionfromENSO-sensiresolvingthelong-term(annualor better)位 ve coreregionset ai. 1978,Ropelewski巳 arsto be strongestduring(Hackertand Hastenrathmostsensitiveto variationsD'Arrigo1990).a resultof recentanalysesAllan1988,Rasmusson(e.g.Zebiakand Caneandof 由e ENSOsystem.records你自 udiespurpose,cores,thisiceandthe1986,inofet al.1987,arenotlong,varves)(e.g.Indonesiaand elsewherein the low-latitudePacific)influencedby ENSOte1econnection(e.g.the southwesternUnitedStates,Loughneedto be furtherdevelopedand endevelopedfor the tropics,includingresultsin an ualIndonesia(Jacoby,1989).Thisis becausetemperatureseasonali旬, whichringformationat higherlati udes,is largelyabsentin the tropicalregions.Thustheseasonalnatureof monso nalrainfalor otherenvironmentalfactorsis requiredfor at lowlatitudes.Otherlimitationsare 血e relativescarcityof treespeciesWI 血 suitablelongevityandanatomicalstructu間 , loggingof ldi derablyfor manysubtropicalregionsof the globein rica,Morocco),augmentingthe moreextensivecoverageat higherlatitudes.Yet, coverageis still practicallynonexistentfor the equatorialbandof the tropicsfromlOoNto 1O0S.In 出is paperwe presentan updateof a tree-ringcombined500 datasectionsfromonlyconsideredthe finaltheTheMost(cross-date)the sectionsradii afteroriginalavailableprobablybe foundof lagewasWhettonchronologyof ringsampletheandmeasurements巳samplesfor reanalysis.correctbackuntilin zenumberwerenorInterpretingat leastMurphyin 1931 and has sincebeenJacobyand D' Arrigo(1990)is rom70of recordby Berlageamongthe samples,retainingconsistsof 10-22radiifromof actualtr, 己es usedis notnot includedw 巳re presentedthefirst published(1989),and(1514-1929)are onlyabout50 ye 盯s in ageabout400yearsin age.Thethe twooldestsections.Anthe ralmonsoon巳dagetree-ringwidthchronologyby DeBoer(1951),Murphyand2).S 巳 ctions.recordrecordrawmeasuremfinalby Berlage凹的 forhis1650original(Jacobyand Whettonin theandchronologyof radii from280 years,28 woodwith Some(Berlage(1931);to ourthe individualbasedonto matchamong叫d 28-29disks1931).knowledge,radiiarestudy,it wasinferredthatD'Arrigo1990).Additional(1989)and Jacobyand �go (1990).
D he wet monsoonof the priorandicaVinstrumental1966).DeBoerBerlage(r �前 ,establishedre rdtheandprimitivedespitedatingis a climaticsignalrelatedto monsoonrainfalland the EI Ni 品。 -Sou 血emin this chronology個erlage1931,DeBoer1951,Murphyand WhettonD'Arrigo1990).1966)founda correspondencebetweenhis chronologyand the lengthofseasonconcurrentwithgrowth,as well as with the intensityand durationsubsequentdry seasons.He also comparedhis teakvariationsto historrecordsrepresentativeof droughtand the 51al.血e edwithJavathenumberrainfalldataof drymonthsfor1864-1929,(r -0.43),foundannual血at thecloudinessth 巳 numberof rain days per y 巳訂 ( 但0.48).recently,Murphyand Whetton(1989)found a significantrelationshipbetweentheBerlageseriesand Wright's(1989)SouthernQscillationIndexor SOl (1852-1929).Jacobyand D' Arrigo(1990)determinedthat 出e Berlageseriescorrelatedwi 血 Java dry season(MayOctobe吋 precipitation(1892-1929,r 0.50),the lengthof the dry season(r -0.51),and th 巴SOL SOl-teakcorrelationswere oundto be significant(P .05) during the ationsJacobyandandENSO3.UPDATED3.1Tree-RingandIn Augustof 1992,teakBali, Indonesiafor tree-ringJAVATEAKDecember-Februaryare slightlyD' andHastenrathjustpriortO,andthe June-Augustis also the samplesanalysis.ing definedannualringsin responseWiththe assistanceof forestersfromtreesgrowingin a natural(unmanaged)wereTeakcollectedfromis deciduous,sites in centralandsheddingits leaves區可 .0'E15'1 D'NS'1.Mapof Indonesiaandvicinity.JK Jakarta,easternJavaand produc-d 可 monsoonseasgnin theseareas.( 也e state-runforest可 institute),teakin the hillsnearCepu(Figure1) wereto the urphybetweenCEP Ce 阱, SBN Surabaya.
352TAD,\ 句11.5,No.3,September1994sampledin orderto updatethe originalBerlagechronology.Accordingto theseforesters,these teak trees are amongthe oldest(about100-150yr) remainingon the islandof Java for由is commercial yvaluablespecies.Figure2 showsthe originalBerlagechronology(1514-1929)and Figure3 a comparisonwi 也 the recently-updatedseries(1880-1991).For the periodof overlap(1880-1929),thecorrelationbetweenthe Berlagechronologyand the updatl 吋 seriesis 0.57(Fi 伊re 3). 四Ieye 訂s ectvaluesin his ori 斟naldata set, appe 缸 to showagreementbetweenthe two series:Berlageindexin 1928 1.61,higher曲anp 血viousye 缸 1929 .65;upda記 indexin1928 1.189,higherthanpreviousyear,"1929 ogy20-year(1931)teakobtainedsmoothingbeginsThe updatedchronologyin 1870,was truncated四Ie19coresarea subsetchronologyfromspline.is comprisat 1880whenofthosesampledfromBerlage巴d ofthereat1514-1929.(1931)hasbeen四Ie originalstandardizedringwidthusing19 coresfrom10 虹'ees.Theseries,is a sample刮目 of 4' coresfromfour由e th each other.Someattritionof s 缸npleswas also mentionedby Berlage(1931).甘Ie meancorrelationamongall the coresin COFECHA,a gram(Holmes1983),is 0.5. By convention,eachannualring is consideredto havebeenformed扭曲eye 前 in whichgrowthbegins.In theseteak treesthe growingseason,whichcoincideswiththe periodof mostabundantrainfall,is believedto beginaroundNovember(ye 缸。 and extendsthroughto approximatelyMay of the followingye 叮 (ye 前 t 1) (Coster1927,1928).We decidedto focusprimarilyon 由e higher-frequencyvariationsassociatedwith the ENSObandwidthof around3-7 ye 缸s (Rasmussonet aI. 1990,Barnett1991).Forthis pu 中 O 峙, the tre -ringdatawerede 甘endedusing 20-yearsmoo 曲ing splineswhichremove50% or mo 時 of 血e varianceat frequenciesover 20 years(Cookand Peters1981,CooketaI. 1990).The detrendingwas performedas p 位t of the 985).Thestandarddeviationof 血e updatedchronologyis .18,the meansensitivity( 組 indicationof high-frequencyvariance- Fritts1976)is .20 and re.02and-.19.
D' ArrigoThe originaland updatedsco 血s (obtainedby subtractingindicesof eachseriesfor thetwo se 個 of scoresfor this interval.Berlageseriesand the later (1930end of 由e 1880-1929interval.353et ai.chronologieswere mergedby first creatingdimensionlessthe meanand dividingby the standarddevi 甜on) from出epriodof overlapfrom1880-1929,and thenaveragingtheScoresor growthdeparturesfor the earlier(1514-1879)1991)updateds 巳nes were 由en splicedtogetheron eitherThe resultingmergedchronologyis shownin Figure4.BERlRGERND UPDRTEDCEPU.dRUR,JERKCHRONOLOGIESI 8819281948196819882888YEARS3.Plotshowingogy gseriesfortwoAUERAGEOREAL四冒 6 自由AGEchronologyandperiodof overlapfromthis p 巳riod is ationCHRONOLOGTEAKI ES432闡︼固︼自個國(-1-2-3-415冒 7882 國B 國19881888"YERRSFig.4.Combinedfor updatedmergeddividingteakchronologiesbythe' two seriesfor the 1880-1929inthe earlierperiod(Berlage,1514-1879)1991)werethen splicedto bothends且byforrst calculatingthe standard記 raging.and laterp 咄咄 (update,of the overlappingdata.forJavascoreseachofScoresfor1930-
TAO,354Vol. 5, No.3,September1994Palmerand Murphy(1993)have also updatedthe Berlageobtainedfromten 仕'ees at a teakplantationnearBojonegoro,developeda chronologybasedon theses 缸npleswhichextendsthe originalseriesby ten years.3.2Climaticseries.Theireastof Cepu.from1920-1989sampleswereTheyhaveand overlapsResponseTheupdated出at foundCeputime扭曲e originalseries(Figurechronology.3) demonstratesa similarIt is Javarainfall(r .39,P 0.01)averagedoverthe dry monsoonmonthsof Juneto September(Figure5). Jakartarainfallwas usedsinceit is a mo 間 completerainfallrecord也祖 the stationsinEast Java,allowinga comparisonfrom1880-1988.Wi 由 this longerdataset it appearsthatrainfallovermuchof the dry 扭扭on, rather曲an only the transitionalmonths,are influentialto growth(Jacobyand D' Arrigo1990).As was foundpreviously,teak se 巳ms to be relativelyinsensitiveto the amountof wet seasonrainfall,whenmoistu間 is generallyadequateforgrowth(Figure5) (Berlage1931,1966,Jacobyand D' Arrigo1990).Thesecorrelationsandthoseshownbelowdid not differsignificantlyfollowingprewhiteningof 血e instrumentaldata (Boxand Je 暐ins 1970).Th 巳 updal 記 d teak recordis also significantlycorrelatedwith sea level pressure(SLP)inJakar 徊 , Surabaya(eastof Cepuin EastJava)andDarwin,northernAustralia(Figures1 and6). HigherSLPvaluesat theselocationsare as ociatedwiththe thof teak.Darwinis one of two PURNO1988τERKD.5目.3國.串串.1 "ZE E4- D.l︼,-EE--司-D.3-D.5JF同自 阿dJRSONBdF阿n阿阿 ONTHSFig.5. ionand teak updatedseriesfor the commonperiodfrom1880-1988.Comp 缸1son is fromJanuaryof 曲eye 前由at growthbegins(yeart) untilMayof 血e following ye 缸 (t l).Growingseasonfor teakis fromaroundNovember(yeart) untilMay(yeart l).Significantcorrelationsat or abovethe .05 (*)and .01 (* 叮 levels.are hrainfallduringthe dry (eastmonsoon)seasonimmediatelyprecedinggrow 曲 , fromJuneto September.
D' Arrigo355et al.5國8.38.I-圓圓,-8.3-8.5開國.3︻﹄區 ��申『回-8.5JFMRMJJRSONDJFMHM例。 weenfor threestationsforSurabaya,East JavaLevelsof significancestrongest(negative)levelpressu阻 duringcorrelationsthe drymonthlysea levelpressureand teakupdatedavailablecommonperiods:Jakar恤 , WestJava(middle)and Darwin,northernAustralia(botas in Figure5. As for rainfall (Figure 5), 也eareseasonfoundbetweenteakjust priorto p 巳riodgrowthof growth.andseagenerallyusedto estimatethe stateof ENSOin the SOl(Wright1989).Wright's(1989)rainfallindexfor the centralequatorialPaci 且c, anotherindicatorof large-scaleENSO,showsan inversecorrespondencewithJavarainfallindicesand withteakgrowth(Figures7 and 8).Thisinvers巳 relationshipreflec扭曲e tendencyfor the IndonesianLowto migrateeastwardduringthe warmENSOphase,resultingin droughtover Indonesiabut abundantrainfalloverthe centraland 巳asternPacific(e.g.Diazand Kiladis1992).In mostof our comparisonsbetweenteakgrowthand the aboveindices,we havefounda declinein the strengthsofthe correlationsfromaround1930to the 1950s.Alsonotedby otherinvestigators,thisdeclinemay reflecta w 巳akening扭曲e ampli扭曲 of ENSOduring由is time(MichaelsenandThompson1992, Cole et al. 1993).
NrRLLINDEH冒 &JJFM&JRS7.ZerolagJ onsbetweentogrowthWRI GHTperiodandatRR I f growthonsetI NDEH冒國Wright's由e January-March(JAS)and October-Decemberas in Figure5. l討 I-June.seasons.Levelsandof sig-in d 可 seasonoccurperiod.RNDindex(.曲的, July-CEPU.由smonTERKdRUR894-1982n‘z- ::I1I. - 8.1 ::I自.-8.3串串-8.5dFh rrelationsbetweenWright's(1989)rainfallindexof 由e EINiiw-SouthernOscillationfor the centralequatorialPacificand the teakupdatedseries.Levelsof significanceas in Figure5. In this case therelationshipwithrainfallis inverse,becausethe migrationof 由e Indonesian Lowduringthe warmENSOphaseis associatedwithdroughtanddecreasedtree growthin Java.
D'Arrigo3.3Rankingof ExtremeTeakYears357et aI.vs.the回storicalQuinnENSORecordWe haverankedthe mostex 加me low and high growthyears 扭曲e teakseriesand缸'ed themto Quinn's(1992)listingof. large-scaleENSOevents.Basedon historicalfromsitesacrossthe nIS mo 肥 likelyto relateto the teakrecord血an his regionallistingof EI Ni 品 os for westernSouthAmerica(Quinnet al.1987).Previously,Jacobyand D'Arrigo(1990)comparedthe originalBerlagerecordto the regionalEI Ni 晶。 listing(Quinnet al. 1987),Figure9showsa graphicalcomparisonbetweenthe upda 旭 d Java teak chronologyand Quinn's(1992,Table6.1) largescaleENSOsfor 1880-1991.Mostbelow-averageteak departureswerecomprecordsfoundwi 出 ENSOto coincidecorrespondenceconsidering(Quinnet α 1. tin Figure9). �QU I NNalsosome】 EN SO(1992droughtdiscrepancies,asYERRSHE圖.UAE. EdRNDareis a ve 可 l ENSO,E2ND55H'" 自回冒﹂暉晶1可一19281948目 onologyscoresdesignated1虹ge-scaleENSO years (starredvalues)listed in Quinn (1992, Table 6.1). In mostcases, below-averageteak growth occurs during ENSO years. Excep 世onsincludeyearsof drought(D) and below-averageteakgrowth(but notENSO)and ENSOyearsof aboveaverageteak growthbut no drought(ND)(see 紀xt for details).For theindicesoccurENSOye 訂s(Quinnet aI.only a lBerlagechronologyfrom1514-1929,the first to thirdlowestgrowthin 1737,1634 and 1925.Both1737 and 1925 are listedas strong(S) large-scalein Quinn(1992),although1737wasnot listedin the regionalEI Ninorecord1987).1634is an S eventin the regionallisting.of Quinnet aI. (1987)but(M) event(in 1635)in the large-scalelisting(Quinn1992).The 1876-1878as ve 可 strongor VS ratelyin 1642,decreased1928andteak1641growth.( 血e latterTheis actuallythreehighestratedteakvalues,as an S ENSOin descendingyearby Quinn
TAD,358thirdY 己 arevents,the1972No.3,泊d.5,September1994For the more recentintervalcoincidingwith the updatedseries(1880-1991),the first tolowestindexyearsare 1967,1976 and 1929.1967 is not consideredto be an ENSO(Allanand Pariwono1990,Quinn1992),while1976and 1929are ratedas M and M respectively(Quinn1992).TheENSOyears1972(S )and1925( 的 rate amongten lowestteakvaluesfrom1880-1991.In the Palmerarid Murphy(1993)update,1967,and ntrings."eThegrowthve 可 strong(VS)of ,effectsin ,of the globe,Yet the teakwerenotextremewhen takeninto long-termcontext(Figures3-4).This is in agreementwith the resultsofThe 1940-41VS eventwas also only slightlybelowav 己rage inPalmerand Mu 中hy (1993).our teak record;indiceswerenear-averagein the Palmerand Murphy(1993)seriesin theseye 訂s. Thethre 巳 highestteakvaluesin the recentrecord,in descendingorder,are 1978,1904 and 1990.1978 and 1904 are yearsof above-averageJakartadry seasonr 血 nfall;1990datawerenot available.For the combinedreco 吋 from1514-1991,whichwereevaluatedas normalizeddepartures,1737and1634remainthe lowestgrowthye 缸 s. TheENSOyears1967,1976and1925訂'e thesixth,ninthand 13th lowestgrowthdeparturesin the combinedrecord.(M ),withincludeandJakartabelowthe(19 2)Years1939below-averageseveralthe 05個 ),and1968-9(M-)large-scale(Figure9). TheseENSO且,whichwerenotas ociatedwithdroughtbasedof eastmonso ndroughtin Table6.3 of Quinseasonrainfalexce dedtwostandarddeviationsin 1905(althoughlistedas a droughtin TableENSOscoincidedwithdroughtyearsin Indon6.3 or the Jakartarainfalrecord.of ENSO(M i 凹 , as weltook placein 1967,a non-ENSO1990).Droughtalso occurred由e fifth in the updatedteak(Figure9). NoENSO1963 (Quinn1992,are not explainedor theJakartarainfallIn somecases,decouplingbetweenin Indonesia(Rasmussonof ENSO,monsoonare also uncertaintiesrecords(JacobyandnoasENSO:the mostThelowestsevereye 前 exceptforin 1961(the fourthseries)and in 1963,teakdrought1907ENSOnone(M ),of(M ),did1932coincideevents,recordexam 阱 ,slightly1939norQuin 'snotstrongrainfal(19 2)(Figu間的 . Forin 1904,andwasonly6.3).Neitherthe巳E 泊 , ac ordingto eitheronthetre lueJakartaforrainfalthe18 0record,the regionof Australasia(Allanand Pariwonomostsevere扭曲e Jakartarainfallrecordanda yearof slightlybelow-averageteakgrowthwas recordedin 1961,and onlyTable6.1).1893-94and 1938 wereby ENSO'sand/ordry seasondrought,a weakENSOeventwas recordedinyearsof decreasedteak growthwhichbasedon he teak and Quinnrecordsmay be explainedbylarge-scaleENSOand its regionalmanifestationas east monsoondroughtand Wallace1983,Quinn1992).Diffl 巳rences扭曲e seasonal,phasingrainfallor teak growthmay also accountfor somediscrepancies.Therewhichhavebeenacknowledgedin boththe Quinn(1992)and teakD'Arrigo1990).
D'Ardgo3.4IntegratingHigh-ResolutionTakenva 可 et aLProxyRecordsof ENSOrecordsPacificof ENScoreregionapaleoclimaticthe equatorialcanrevealandhowoverindividualareasENSOof ENSOeventst 巴leconnection(Baumgartneret at. 1989,Diazet at. 1992).As one examplewe lopedfor sitesin th 巳 equatorial(Coleet at. 1992,1993).SeveralEl Ni 晶 olENSOeventsabsentin the teak seriesthe-alsonot represented(Coleet al.1993).Quinn et at.1992).in a otopecoralrecord扭曲eye 訂s 1907wererecordedarefromTarawa,at ION,172Eand1932(bo 出 M ye 訂smin shortcoralrecordsfromtheGalapagos,Tarawaand Bali (just east of Java)in 1963,1965,1969,1972,1976,1982-83叩d1986-87(Coleet α l. 1992).All but th 巳 relativelyweak1969eventare alsorepresentedinthe teak record.A weakENSOyear in 1963is recordedin the Tarawaand Galapagoscorals(Coleet al. 1992)and in the teak record.1965 (S) is a strongereventthan 1963 in both theteak and coralrecords.The 1972 and 1976 eventsare well representedin both the teak andcoralrecords,with1976beingthe strongestof the two in bothcases(althoughonlylistedas an M eventin Quinn1992).Yet in th 巳 Bali coral,the 1972eventis strongerthan1976.The extreme1982 品3 event,onlyslightlybelownormalin the teakrecordand in Jakartadryseasonrainfall,wasnear-normalin theal. 1992).As in the teak,the 1982-831972.The 1986-87(M) eventis indicatedweaklyrecordedat .Asspectralmentioned,igo1990)討dueThusIn SSA,toanalysis(SSA)a 20-yearsplinein Tarawa.on the teakanywereis appliedspectral(Vautardpropertiesand Ghilwithin(averagetheinvestigatedto and D' ArrigodatausingSSA.functionof ENSO.of Mexicoet ai.of ectralpeakswereidentifiedin the combinedncy) bandwidthof ENSO(Rasmussonfor over 20%of the variancein theteak巳Barnett1991).Together,出ey accountand are similarto a p 巳 ak of 5.8 ye 訂s detectedThe peakin the teak recordis amplitude-modulatedye 缸s of thiscentu可 (Cookand D' Arrigo1993).recordsthe Gulf(Cleavelandthe 5.68),et al. 1990,chronology,et α t. ionwhichreflectsstrongspectralpeakat 4 yearsseriesin standardizing巴巴 ffectsof mponentsfromwas performedto removemultiplierprocessin orderto identifyand phasemodulationsover(p .05),BaliANALYSESSingular1993).D'Aeffectsin thea slightlydatawas a less strongeventthanbyring 扭曲e teak recordand ismoderately叫Duringthe ve 叮 low growthand belowaveragedry seasonin Javain 1967,therewerecool-phas巴 anomaliesin Tarawaand the Galapagos,but drierconditionsindicativeof the warmphasewererecordedin theBali coraland instrumentalrecords,suggestingdecouplingof the westernmostPacificfrom血e large-scaleENSOsystemin 1967(Allanand Pariwono1990,Coleet 叫. 1992).1961also app 巳ars to be an exampleof a moreregionalized,westernPacificevent.4.ENSO扭曲e TarawawithTheteakcoralrecordteak(Colea strongsignal扭曲espec 恤 diffl 缸, however,For example,a tree-ringbasedregionof erSOlshowsaindicate
360TAO,possibleinvestigated.shiftsin theVo1.5,of 由e antinvestigations1989,(95%(1993)ye 缸sWhettonforcing,level)periodicitiesin the ENSObandwidthof 血e teak chronology(DeBoer1951,Palmerand Murphy1993).For example,Palmerlevel)peaksat 5.7 and 5.3 ye 缸趴呵呵 closetocoralse 吋 es.Lower-frequencyfluctuationswithweredetectedby DeBoer(1951),and morerecently(1989)and Palmerand Murphy(1993).Theaccordingto De Boer(1951)and Murphyandspectralpeaksat 50 yearsand at around30 yearsWhetton(1989)dominantdevelopment50-year5. FUTUREandPalmerandpeak couldof the rebeingfurther(1993)be an artifactn 。但也at血eof chronologyPROSPECTSFuturedendroclimaticresearchin Indonesiawill requireremotesiteswithseasonalrainfall,whereold growthtreesin additionto teakwhichare capableof producingwell-definedbe identi 晶 ed.Besidessamplingat Cepuin 1992,we collectedrem 。他 forestreserveswithpronounceddry seasons,includingeasternmostJava,and in westernBali.All of the teak s 阻npledtrees,driesttime(3-7 years)werealso foundin 0 由 erBerlage1966,Murphyand Whettonand Murphy(1993)foundsignificantour findingsand to 由e Colee.t al.cyclesof 缸ound50 yearsand 89describedin detailby Murphyand89-ye訂 cyclemay be linkedto sol 訂Whetton(1989).Significant(95%are describedby bothMurphyandPalmerbe climaticin originBerlageseries.1994onlya fewp 缸t of Java,decadeswasanin age,exception,from.plantationsites.由e identificationof additionalmaystill exist.Treespeciesannualringsalso needtoteaksamplesfromseveralBaluranand AlasPurwoinat theseBaluran,loationswere扭曲e no 吋1easternyoung個dwh 巳re samplesfroma few isolatedteaktreesgrowingin morenaturalwoodedsettingsw 巴re obtained.Still,theses 缸npIesonlydatebackto 血e1930s.Sulawesi;an Indonesianislandless developedthanJava,has pronouncedrainfallseasonalityand may yieldremainingspecimensof naturalteak or othersp 巳 cies.It mayalsobe possibleto derivechronologiesby obtainingsamples仕om historicaland arch;ieologicalruins or structures.Severaltree speciesendemicto Irian laya, Indonesiaand Papua NewGuinea(Ogden1982), includingsome which grow in alpine environments,may be sensitiveto variationsin temperatureor precipitationrelatedto changesin the monsoonand ENSO.AsadditionalnetworkSoutheast6. smchronologiesof chronologiesAsia(s 臼 otherareof teak,papersdevelopedpine andthis volume).forothIndonesia,r era growingsitesin曲直ARYAn update of a teak chronologyfrom Java has been presented.The combinedseriesnearly 500 years of climatic informationfor the data-sparseequatorialwestern Pacific時gion.Despitereservatio間 , particularlyconcerningtheearlypartof the teak血cord,由eclimaticsensitivityof 曲is recordhas nowbeenclearlydemonstratedthrougha numberofstudies.Therelationshipwithpriordry seasonrainfallis foundto holdthroughthe recentperiod.Becausedry seasonrainfallin Javais closelylinkedto the behaviorof the ENSOsystem(e.g.Hackertand Hastenrath1986);the Javateakchronologyis alsofoundto be 個indexof ENSO-relatedvariabilityin the westernequatorialPacific.Over20% of the varianceprovide
D-r n g o e 4Eu a361'.Aband-passfilteringprocedure,was retainedin the teak recordusingSSA as a data. adaptivewiththe aim of emphasizingthe IIENSO"仕equencyband.Comparisonwith coralrecordsfrom the equatorialPacificrevealsgoodagreementand severalexamplesof decouplingofwesternPacificanomaliesfromthe larger-scaleENSOsystem.A densernetworkof highresolutionpaleoclimaticdata fromboth the core and teleconnectionregionsof ENSOwillgreatlyimproveour understandingof long-tennENSOtemporal曲d spatialvariabilityin 由efuture.*Note:Datais availableon requestfromthe authors.國AcknowledgementsernmentWef Indonesiafor己thanktheirtheIndonesianInstituteof Scienceson 由is project.assistanceWealso(LIPI)gratefully曲e gov-andacknowledgethecooperationand aid of forestersMamanand othersat PerumPerhiJtaniin Ce 阱 , Indonesia.WethankE. R. Co kforadviceandas istanceandM. Moore and R. Allan for providingmeteorologicaldata and assistance.Fundingwas providedby the Lamont-DohertyEarObservatory'sClimateCenter,and the rantNo. nNo. 5182.曲REFERENCESAll 血, R.Phys.J.,198 :Geogr.,sia.Nino-Southern4-40.1991:The4, 269-285.P.,T. st.P.,G. E.P.,Holden-Day,Bradley,巳TheSouthern88,M. K.,em Oscillationr, M.Latif,scalesin the甘 opicalandAustrala-In: D.Americas,climatesystem.R. E. stems:H. ,OscillationBarne鈍,N. Gr, 咄缸凡 of tree-ringsof Djatifrom the Dutch).andweather.worldAt �s Analysis,ForecastingandControl,Francisco.R. 且 ,H. F. Diaz,G. N. Kiladis,temperatureand precipitationrecords.Cleaveland,Progr.in low-latitudeG. T. Shen,climatethe relationshipTectona,24,G. M. Jenkins,SantimeL. G. Thompson,E. Roeckn每 rhand.,"interactionof multipleto long-range1966:andOcean-atmosphereregion.個U. Schlese,Med.扭曲e Australasian145-178.interaction1931:Onon Java.rainfallinfluencesclimaticchangeby high-resolutionice layers,and marinevarves.扭曲e Pacificand the WesternMon.,話 , 1 14.approach阱 , H. P.,Box,1990:J. Michaelsen,recordingof interannualcoralbands,glacialof ClimateVariabilityWash.D. C. Geophys.BerlaOscil ationInt.仗 , T.J. Climate,Baumg12,J., 祖d J. I. P 訂iwono,J. ClimatoJ.,10,Al 曲,R.BarneElE. R.detectedCook,andin tree-ringandJ. K. Eischeid,Nature,327,D. W. S 岫Ie,data from1992:Mexico1987:ENSOsignalin continental497-501.Secularvariabilityand the southernofUnited出e Sou曲States.In:,
362TAD,H. E Diaz,andSouthernOscillation,theJ.Cole,且,G.T.J. E.,Oscillation:Cook,R.E.R.,Thesis,V. MarkgrafCambridgeNo.3,(Eds.),Univ.G. Fairbanks,isotopi
350 TAO, Vo1.5, No.3, September 1994 and about 80% of such events linked with droughts (Quinn et ai. 1978, Ropelewski and Halpert 1987). The lirik between Java rainfall and ENSO app 巳ars to be strongest during the d 可monsoon s巳ason, from approximately June
