Phase locking between atmospheric convectively coupled equatorial Kelvin waves and the diurnal cycle of precipitation over the Maritime Continent

Geophysical Research Letters Phase locking between atmospheric convectively coupled equatorial Kelvin waves and the diurnal cycle of precipitation over the Maritime Continent Dariusz B. Baranowski 1,2, Maria K. Flatau 3, Piotr J. Flatau 4, and Adrian J. Matthews 5 1JointInstituteforRegionalEarthSystemScienceandEngineering,UniversityofCalifornia,LosAngeles,California,USA, 2InstituteofGeophysics,FacultyofPhysics,UniversityofWarsaw,Warsaw,Poland, 3NavalResearchLaboratory,Marine MeteorologyDivision,Monterey,California,USA, 4ScrippsInstitutionofOceanography,UniversityofCalifornia,SanDiego, LaJolla,California,USA, 5CentreforOceanandAtmosphericSciences,SchoolofEnvironmentalSciencesandSchoolof Mathematics,UniversityofEastAnglia,Norwich,UK Abstract Convectively coupled Kelvin waves (CCKWs) are a major component of the tropical atmospheric circulation, propagating eastward around the equatorial belt. Here we show that there are scale interactions between CCKWs and the diurnal cycle over the Maritime Continent. In particular, CCKW packets that pass a base point in the eastern Indian Ocean at 90∘E between 0600 and 0900 UTC subsequentlyarriveoverSumatrainphasewiththediurnalcycleofconvection.Asthedistancebetween SumatraandBorneoisequaltothedistancetraveledbyaCCKWin1day,thesewavesarethenalsoinphase withthediurnalcycleoverBorneo.Consequently,thissubsetofCCKWshasaprecipitationsignaluptoa factorof3largerthanCCKWsthatarriveatothertimesofthedayanda40%greaterchanceofsuccessfully traversingtheMaritimeContinent. 1. Introduction The Maritime Continent is the archipelago between the Indian and Pacific Oceans, including Indonesia, Malaysia, Philippines, and New Guinea. It lies at the heart of the Indo-Pacific warm pool, the largest area of veryhigh seasurface temperaturesonEarth. Asa result,it isa majorcenter oftropicalatmospheric convec- tion and precipitation. The Maritime Continent is the central pivot of the Asian-Australian monsoon system [Chang et al., 2005]. Above the Maritime Continent rises the ascending branch of the Walker circulation, fluctuations of which are major ingredients of El Niño events and the Indian Ocean Dipole [Vinayachandran etal., 2007]. Atmospheric convection over the Maritime Continent provides the heat source for teleconnec- tionsthatinfluenceglobalclimate[ NealeandSlingo ,2003]. However, the difficulties of representing convective processes in atmospheric models, and their multiscale interactions in the tropics, combined with the land-sea contrasts and steep, high topography over the MaritimeContinent,makesimulationandpredictionofprecipitationoverthisregionparticularlyproblematic [Loveetal. ,2011; Birchetal. ,2015]. PrecipitationovertheMaritimeContinentisdominatedbythediurnalcycle[ YangandSlingo,2001],withdeep convectiondevelopingovertheislandsduringtheafternoonandevening,associatedwithseabreezecircu- lations. During the early morning, this precipitation moves offshore [NesbittandZipser , 2003]. Other modes of tropical variability such as the Madden-Julian Oscillation (MJO) have high amplitude over the Maritime Continentandinteractstronglywiththediurnalcycleofprecipitationthere[ Peatmanetal. ,2014]. Convectivelycoupledequatorialwaves[ Kiladisetal. ,2009]arealsoprominentmodesoftropicalatmospheric variability. In particular, convectively coupled equatorial Kelvin waves (CCKWs) propagate eastward along the equator with phase speeds in the range 10–15 m s −1 and have a high amplitude over the Maritime Continent[Roundy,2008].TheseCCKWsareanimportantpartofthetropicalcirculation.Aswellastheir“own” precipitation, they form part of the convective envelope of MJO events [Nakazawa, 1988;Gottschalcketal. , 2013]. There are also significant ocean-atmosphere interactions with surface fluxes from CCKWs cooling the ocean,whichagainhasanimpactonMJOevents[ Baranowskietal. ,2016]. RESEARCH LETTER 10.1002/2016GL069602 Key Points: • Atmosphericequatorialconvectively coupled Kelvin waves (CCKWs) are phaselockedtothediurnalcycleover theMaritimeContinent • CCKWsinphasewiththediurnalcycle have a precipitation signal up to 3 timeslargerthanotherCCKWs • CCKWsinphasewiththediurnalcycle are 40% more likelyto successfully cross the Maritime Continent than otherCCKWs Supporting Information: • SupportingInformationS1 Correspondence to: D. B. Baranowski, dariusz.b.baranowski@jpl.nasa.gov Citation: Baranowski, D. B., M. K. Flatau, P. J. Flatau, and A. J. Matthews (2016), Phase locking between atmospheric convectively coupled equatorial Kelvin waves and the diurnal cycle of precipitation over the Maritime Continent, Geophys. Res. Lett., 43, 8269–8276, doi:10.1002/2016GL069602. Received25MAR2016 Accepted20JUN2016 Acceptedarticleonline24JUN2016 Publishedonline3AUG2016 ©2016. AmericanGeophysicalUnion. AllRightsReserved. BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8269

Geophysical Research Letters 10.1002/2016GL069602 However,thenatureofthedirectinteractionofCCKWswiththeMaritimeContinent,anditsdiurnalcycleof convectionisaknowledgegap.Becauseoftheirrelativelyhighphasespeedandspatiallylocalizedstructure, theconvectivelyactivepartofanindividualCCKWpassesthroughaparticularlongitudequickly,withatime scaleofapproximately1day[ Baranowskietal. ,2016].Hence,the timeofdayofthearrivalofaCCKW overthe MaritimeContinentmaybeimportantforitssubsequenti nteractionsanddevelopment.Ifitarrivesinphase withtheconvectivelyactivepartofthediurnalcycleovertheMaritimeContinent,theCCKWanddiurnalcycle may enhance each other. If the CCKW arrives during the convectively suppressed part of the diurnal cycle, theinteractionmaybeweak.Wealsonotethat thelongitudinalspacingbetweenthetwoequatorialislandsof SumatraandBorneoisapproximatelythedistancethataCCKWtravelsin1day .Hence,ifaCCKWarrivesinphase withthediurnalcycleoverSumatra,itwillalsobeinphasewiththediurnalcycleoverBorneowhenitarrives there 1 day later. This provides a potential mechanism for phase locking of CCKWs to the diurnal cycle over theMaritimeContinent.Inthispaper,weexplorethesehypothesesandinvestigatetheinteractionsbetween CCKWsandthediurnalcycleovertheMaritimeContinent. 2. Data and Methodology PrecipitationwasanalyzedusingtheTropicalRainfallMeasuringMission(TRMM)3B42precipitationproduct [Huffmanetal., 2007]. The data used were 3-hourly maps on a 0.25 ∘× 0.25∘grid from 1 January 1998 to 31 December2013. A Lagrangian database of CCKWs was constructed to analyze the interactions between the CCKWs and the diurnalcycleovertheMaritimeContinent.Thedetailsofthemethodologyaredescribedin Baranowskietal. [2016]andsummarizedhere.First,followingthemethodologyof WheelerandKiladis [1999],theTRMMprecip- itationdatawereaveragedoverthe2.5 ∘S–2.5∘Nequatorialband.Thenatwo-dimensionalFouriertransform in longitude and time was performed to transform to wave number-frequency space. Filtering was applied to isolate the CCKW signal, and the data reverse transformed back to longitude-time space. Finally, the Lagrangian CCKW database was then constructed by identifying coherent eastward propagating precipita- tionmaximainthisKelvinwave-filteredHovmöllerdataset.EachCCKWinthedatabaseconsistsofatrajectory: a set of time-longitude paired values, indicating the central position (longitude) of the CCKW at each time (with3hr esolution).Intotal,1840CCKWtrajectorieswereidentifiedinthe15yeardataset. 3. Diurnal Cycle Over the Maritime Continent The dominance of the diurnal cycle in precipitation over the islands and seas of the Maritime Continent can be most clearly seen by comparing maps of mean precipitation at opposite phases of the diurnal cycle. At 0300 UTC, which corresponds to midmorning (0900–1100 local solar time, LST) across the Maritime Continent,theislandsofthewesternMaritimeContinen t(Java,Sumatra,Borneo,andSulawesi)andpeninsu- larMalaysiaarealmostrainfree,with3haverageprecipitationratesnear2.5mmd −1 (Figure1a).Bycontrast, thesurroundingcoastalseashavehighprecipitation,above10mmd −1. Twelve hours later, at 1500 UTC (approximately 2100–2300 LST), deep convection has developed over the islands(Figure1b),andthe3hmeanrainfallrateisabove15mmd −1.Simultaneously,precipitationoverthe surroundingcoastalseashasdecreasedtobelow5mmd −1. The diurnal cycle of precipitation over the open Indian Ocean to the west of the Maritime Content is much weaker,withmeanprecipitationremainingfairlyconstantthroughouttheday. The development of the diurnal cycle and land-sea contrast in precipitation over the Maritime Continent is shown using a “diurnal cycle Hovmöller diagram,” i.e., a longitude-time diagram of the mean diurnal cycle, averaged over the equatorial band (Figure 1c). In particular, precipitation develops over the mountains of western Sumatra (100∘E) during the afternoon then propagates westward over the adjoining sea during the night. Similarly, afternoon convection builds over Borneo (110–115∘E) and then moves offshore by a combinationofpropagationwestwardanda“jump”eastwardto118 ∘E. 4. Convectively Coupled Equatorial Kelvin Waves 4.1. Case Study During CINDY-DYNAMO A case study of the interaction between CCKWs and the diurnal cycle of precipitation over the Maritime Continent is shown for a 2 week period during November–December 2011 (Figure 2). This was during the BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8270

Geophysical Research Letters 10.1002/2016GL069602 Figure 1. Diurnal cycle of precipitation over the Maritime Continent. Mean precipitation rate (mm d−1; see legend for shading) at (a) 0300 UTC (approximately 0900–1100 LST) and (b) 1500 UTC (approximately 2100–2300 LST). The red lines at 2.5 ∘S and 2.5∘N indicate the equatorial band used in the further analysis. (c) Mean diurnal cycle of precipitation rate averaged over the equatorial band (2.5∘S–2.5∘N). The 12.5 mm d−1 contour is highlighted in red. The black solid (dashed) line indicates local midnight (noon). CINDY-DYNAMO field experiment [Yoneyama et al., 2013]. The two eastward propagating precipitation sig- nalsindicatedbytheblacklinesinFigure2havebeenpreviouslyidentifiedastwoCCKWs,whichthemselves formedpartoftheactivephaseofanMJOevent[ Gottschalcketal.,2013].Precipitationassociatedwiththefirst of the CCKWs stalls over Sumatra and this trajectory terminates within the Maritime Continent. The second trajectorypropagatesthroughtheregion. The mean diurnal cycle is shown by the 12.5 mm d−1 red contour, reproduced from Figure 1c and repeated everydayinFigure2.ThepassageoftheCCKWsovertheMaritimeContinentshowsstronginteractionwith the diurnal cycle. For example, on 23 and 24 November, before the arrival of the first CCKW, there is almost Figure 2. Hovmöller diagram of total precipitation rate (mm d−1; see legend for shading) averaged over the equatorial band (2.5∘S–2.5∘N) for the period 22 November to 5 December 2011. The tick marks on the vertical axis are at 0000 UTC each day. The two black lines indicate the propagation of the two CCKWs during this period. The red lines show the 12.5 mm d −1 contour from the mean diurnal cycle in Figure 1. BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8271

Geophysical Research Letters 10.1002/2016GL069602 Figure 3. Lagged composite maps of daily mean total precipitation rate anomalies (mm d−1; see legend for shading) for convectively coupled equatorial Kelvin waves on day (a)−1, (b) 0, (c) 1, (d) 2, (e) 3, and (f) 4. Day 0 corresponds to the time when the Kelvin wave passes through 90∘E. It means that precipitation is averaged for the entire day (eight data points) at which CCKW is at 90∘E for lag 0 (see supporting information for details of compositing and anomalies calculation). noprecipitation(i.e.,noblueshading)andnodiurnalcycleatSumatranlongitudes(96–102 ∘E)tomatchthe meandiurnalcyclethere(redcontour). On25November,theenvelopeofprecipitationassociatedwiththefirstCCKWarrivesoverSumatra.However, the response over the Maritime Continent is not one of continuous precipitation throughout the day. Instead, the diurnal cycle itself is strongly enhanced. On 25 November, precipitation over Sumatra develops during the day to a 3 h average rate of over 60 mm d−1 and then propagates westward overnight over the sea,inphasewiththemeandiurnalcycleshownbytheredcontouronthatday(whichisaprecipitationrate ofonly12.5mmd −1).Hence,thediurnalcycleoverSumatraon25Novemberisenhancedfromthemeanby afactorofover5.OvernighttheprecipitationoverSumatradropsbacktonearzero,andthenthenext2days (26 and 27 November) also experience a very strong diurnal cycle. A similar response is seen over Bornean longitudes(110–115 ∘E).PriortothearrivaloftheCCKWenvelope(uptoandincluding26November),there is almost zero precipitation and no diurnal cycle. On 27 and 28 November, with the arrival of the CCKW, the diurnalcycleisstronglyenhanced. After the first CCKW has passed (28 November for Sumatra and 29 and 30 November for Borneo), there is a suppressionofprecipitationandthediurnalcycle.WiththearrivalofthesecondCCKW,themeanprecipitation anddiurnalcycleareenhancedagain,especiallyoverBorneo. Hence, there are strong interactions between the passage of CCKWs over the Maritime Continent and the diurnal cycle of precipitation over the land and sea. The main interaction appears to be one way, with the previously existing CCKW propagating eastward from the Indian Ocean and modulating the diurnal cycle overtheMaritimeContinent.However,theremayalsobescopeforaninteractionintheoppositesense.The precipitationthatdevelopsoverSumatraon26November,withintheenvelopeofthefirstCCKW,propagates westwardasusualovertheneighboringseas.Butinsteadofpeteringout,itcontinueswestwardto83 ∘Eover thenext2daysandinfluencesprecipitationaspartofthesecondCCKW. 4.2. Composite Mean CCKW The case study in section 4.1 showed intriguing interactions between the passage of CCKWs and the diurnal cycle of precipitation over the Maritime Continent. In this section, a statistical analysis of the CCKWs ispresented,toallowforasystematicevaluationoftheseinteractions. CCKWtrajectoriesfromtheLagrangiandatabasewereselectediftheypassedthroughtheregionimmediately “upstream”(inthesenseofaneastwardpropagatingwave)oftheMaritimeContinent,beginningwestof85 ∘E and ending east of 100∘E. In total, 447 trajectories satisfied this criterion and were included in the analysis. Withinthisdomain,abasepointwaschosentobeat90 ∘asthisisclosetotheMaritimeContinentwhilebeing outsideofthedirectinfluenceofitsdiurnalcycle.LaggedcompositedailymeansoftheseCCKWeventswere BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8272

Geophysical Research Letters 10.1002/2016GL069602 Figure 4. Mean anomalous total precipitation rate (mm d−1) for CCKWs averaged over the passage of the wave (days−1 to 4), as a function of longitude. The thick black line is calculated from all CCKWs. The green dotted (red dashed) line is calculated only from CCKWs arriving at 90 ∘E between 0600 and 0900 UTC (0000–0300 UTC) (see supporting information for details of compositing and anomalies calculation). constructed,suchthatday0correspondstothetimewhentheCCKWpassesthroughtheupstreambasepoint overtheIndianOceanat90 ∘E. On day−1 (Figure 3a), the active precipitation center of the CCKW is over the central Indian Ocean (west of 85∘E), and the precipitation over the Maritime Continent is near normal. On day 0 (Figure 3b), the CCKW haspropagatedeastward,anditsprecipitationanomalyisamaximumat90 ∘E,bydesignoftheanalysis.On subsequent days, the CCKW continues to propagate eastward, until by day 4 (Figure 3f), it has crossed the MaritimeContinent. Thestrongestprecipitationsignalsareobservedonday0.However,thisisanartifactofthecompositemean procedure.TheCCKWsareallinphaseonday0,bydesign.Athigherlags,CCKWswithslightlydifferentphase speedswillbegintodestructivelyinterferewhenaveraged,reducingthestrengthofthemeansignal. ThenetanomalousprecipitationovertheMaritimeContinentassociatedwithCCKWswasthencalculatedby averagingoverthelifetimeofeachCCKWevent(fromday −1today4)andthenaveragingoverlatitude(the 2.5∘S–2.5∘Nequatorialband).ThereisaclearlongitudinaldependencetothisCCKW-associatedprecipitation (thick black line in Figure 4). The maximum precipitation associated with the CCKW is between 90 and 97∘E (the sea west of Sumatra, consistent with the anomaly maps in Figure 3). There is then a relative minimum in precipitation at 104 ∘E over eastern Sumatra, before net CCKW precipitation increases again over Borneo (110–115∘E). 5. Phase Locking Between Diurnal Cycle and Kelvin Waves Given the relatively narrow time window of approximately 1 day that a CCKW is active for at any particular longitude(Figure3),itispossiblethatanyinteractionbetweentheCCKWandthediurnalcycleisdependent on thetimeofthedayatwhichtheCCKWarrivesovertheMaritimeContinent . If the peak of the CCKW arrives during the convectively active phase of the diurnal cycle, it might be expected to constructively interact withthediurnalcycle.Conversely,iftheCCKWpeakarrivesduringtheconvectivelysuppressedphaseofthe diurnal cycle, its interaction with the diurnal cycle may be weak. Also, given that the longitudinal distance fromSumatratoBorneoisapproximatelyequaltothedistanceaCCKWpropagatesin1day,thenifaCCKW arrivesatafavorabletimeforconvectionoverSumatra, itwillalsoarriveatafavorabletimeoverBorneo,and aphaselockingmayemerge. BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8273

Geophysical Research Letters 10.1002/2016GL069602 Figure 5. Mean trajectory of CCKWs (green dotted (red dashed) line for CCKWs arriving at 90∘E between 0600 and 0900 UTC (0000–0300 UTC)) derived from observations, superimposed on mean diurnal cycle of precipitation over the Maritime Continent (same as in Figure 1c). Toinvestigatethis,the447selectedCCKWeventsfromsection4.2werefurthersubsetted,dependingonthe time of day of their arrival at the base point at 90∘E, to the west of the Maritime Continent. Based on their arrival time, each CCKW event was allocated to one of four 6 h bins: 0000 or 0300 UTC; 0600 or 0900 UTC; 1200 or 1500 UTC; and 1800 or 2100 UTC. The distribution of events was even, with 102, 117, 120, and 108 trajectories in each bin, respectively. Hence, there was no preferred time of day for the arrival of the CCKW eventsatthewesternedgeoftheMaritimeContinent. The net anomalous precipitation contribution from CCKWs arriving at 90 ∘E in the 0600–0900 UTC window is shown by the green dotted line in Figure 4. It is up to 50% higher than the average contribution from all CCKWs(blackline)overtheislandsoftheMaritimeContinent.ThisisconsistentwithCCKWdynamics.ACCKW passing90 ∘Eat0730UTC(middleofthe0600–0900UTCbin)andpropagatingeastwardwithatypicalphase speedof10–15ms −1 willarriveat97 ∘E,justoffshoreofwesternSumatra,at2200–0500UTC(greendotted line in Figure 5). This is at the time of maximum precipitation in the diurnal cycle here (shading in Figure 5). Hence, CCKWs arriving in this window are at the right time to constructively enhance the diurnal cycle and thereforethemeanprecipitationoffshoreofSumatra.TheseCCKWsthengoontoenhancethediurnalcycle andprecipitationoverBorneo(110–115 ∘E;greendottedlinesinFigures4and5),consistentwiththephase lockinghypothesisdescribedabove. Conversely, the net precipitation contribution from CCKWs arriving 6 h earlier (the 0000–0300 window; red dashedlineinFigure4)isverydifferent.OverSumatra,theseCCKWsarenotinphasewiththediurnalcycleand haveaprecipitationsignaltheresimilartothatofthemeanCCKW(blackline).However,overBorneo,although theseCCKWsdostillhaveanetpositiveprecipitationanomaly,itismuchweakerthanthe0600–0900CCKWs. Themeantrajectoryofthe0000–0300CCKWs(reddashedlineinFigure5)isonlyplottedupto102 ∘E.Eastof thispoint,theanomaliesareveryweak,andtheassociatedspreadofthetrajectoriesinthissubsetbecomes solargethattheirmeanpositionwithinthediurnalcyclebecomesmeaningless. TheeffectofthephasingoftheCCKWswiththediurnalcycleonthelongevityoftheCCKWswasthenexam- ined,bytestingwhatfractionofCCKWssuccessfullytraversedtheMaritimeContinent(i.e.,terminatedeastof 125 ∘E).The47%ofthe0600–0900CCKWs(inphasewiththediurnalcycle)successfullytraversedtheMaritime Continent,whileonly33%ofthe0000–0300CCKWs(outofphasewiththediurnalcycle)managedto. The two other CCKW subsets (1200–1500 and 1800–2100 arrivals at 90∘E; not shown) are intermediate in behavior.Totesttherobustnessoftheanalysis,thedatasetwassplitintotwo(1998–2005and2006–2012). Eachindependentdatasetwasanalyzedseparately.Theresultswerequalitativelysimilar. BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8274

Geophysical Research Letters 10.1002/2016GL069602 Thecasestudyexaminedearlierisconsistentwiththiscompositebehavior.ThefirstCCKWinthecasestudy (Figure 2) arrives at 90∘E outside the 0600–0900 UTC window. It is not exactly in phase with the diurnal cycle and peters out between Sumatra and Borneo. However, the second CCKW arrives at 90∘E inside the 0600–0900 UTC window. This wave interacts strongly with the diurnal cycle and successfully crosses the MaritimeContinent. 6. Discussion WehaveshownasystematicimpactonCCKWsovertheMaritimeContinentthatdependsonthetimeofday oftheirarrival.CCKWsthatpassabasepointovertheIndianOceanat90 ∘Eareinphasewiththediurnalcycle over Sumatra and Borneo. Compared with CCKWs that arrive 6 h earlier, these favored waves have a precip- itation signal that is up to 3 times larger, and they are 40% more likely to successfully traverse the Maritime Continent. These scale interactions between the diurnal cycle and CCKWs are an example of rectification in theclimatesystem. The magnitude of the phase locking effect varies throughout the year. The difference between the two groups is largest during boreal fall (September–November) and winter (December–February) seasons and weaker during boreal spring (March–May) and summer (June–August) seasons; nevertheless, this effect never reverses—the 0600–0900 UTC group has always larger precipitation anomaly than 0000–0300 UTC group.TheseasonaldifferencesinthephaselockingbetweenCCKWsandthediurnalcyclearelikelydueto seasonalchangesinthebackgroundcirculationandCCKWscharacteristics.Thatwillbeasubjectofourfurther studies. Our analysis focuses on CCKW events which propagate along the equator. However, there are other events whichpropagatenorthwardorsouthwardalongthecoastofSumatra.Forexample, RidoutandFlatau [2011] described a CCKW event, which propagated northward along a dynamic equator (a 0 planetary vorticity line). In boreal winter season some of the events propagate southward along the coast of Sumatra due to interaction with topography [Guo et al., 2014]. Therefore, it is possible that we underestimate number of CCKWs, which propagated across the Maritime Continent, because they detoured part or the entire region. However,sensitivitytestswithrespecttothemeridionalextendoftheequatorialbandusedintrajectorycal- culations(seesupportinginformation)indicatethatoverI ndianOceanandMaritimeContinentthispotential underestimationissmall.Therefore,itmayhaveonlylimitedeffectontheresults. ThedynamicalnatureoftheseinteractionsbetweenCCKWsandthediurnalcycleisunknown.However,the forthcominginternationalYearsoftheMaritimeContinentfieldexperimentshouldprovideacomprehensive observational data set that can be used to tackle this problem. The ability of atmospheric models to cap- turethese interactions, andparticularly theirdependence on horizontalresolution andwhether convection isparameterizedorexplicitlyresolved,willalsobeofgreatinterest. References Baranowski,D.B.,M.K.Flatau,P.J.Flatau,andA.J.Matthews(2016),Impactofatmosphericconvectively-coupledKelvinwavesonupper oceanvariability, J.Geophys.Res.Atmos. ,121,2045–2059,doi:10.1002/2015jd024150. Birch, C. E., M. J. Roberts, L. Garcia-Carreras, D. Ackerley, M. J. Reeder, A. P. Lock, and R. Schiemann (2015), Sea breeze dynamics and convectioninitiation:Theinfluenceofconvectiveparameterizationinweatherandclimatemodelbiases, J.Clim.,28,8093–8108. Chang, C.-P., Z. Wang, J. Mcbride, and C.-H. Liu (2005), Annual cycle of southeast Asia–Maritime Continent rainfall and the asymmetric monsoontransition, J.Clim.,18,287–301. Gottschalck,J.,P.E.Roundy,C.J.SchreckIII,A.Vintzileos,andC.Zhang(2013),Large-scaleatmosphericandoceanicconditionsduringthe 2011–12dynamofieldcampaign, Mon.WeatherRev. ,141,4173–4196. Guo,Y.,X.Jiang,andD.E.Waliser(2014),ModulationoftheconvectivelycoupledKelvinwavesoverSouthAmericaandthetropicalAtlantic OceaninassociationwiththeMadden-JulianOscillation, J.Atmos.Sci. ,71,1371–1388. Huffman, G. J., R. F. Adler, D. T. Bolvin, G. J. Gu, E. J. Nelkin, K. P. Bowman, Y. Hoong, E. F. Stocker, and D. B. Wolff (2007), The TRMM multisatellite precipitation analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales, J.Hydrometeorol.,8,38–55. Kiladis,G. N.,M.C. Wheeler,P.T.Haertel,K.H.Straub, andP.E.Roundy(2009),Convectively coupledequatorialwaves, Rev.Geophys.,47, RG2003,doi:10.1029/2008RG000266. Love, B. S., A. J. Matthews, and G. M. S. Lister (2011), The diurnal cycle of precipitation over the Maritime Continent in a high-resolution atmosphericmodel, Q.J.R.Meteorol.Soc. ,137,934–947. Nakazawa,T.(1988),TropicalsuperclusterswithinintraseasonalvariationsoverthewesternPacific, J.Meteor.Soc.Japan ,66,823–839. Neale,R.B.,andJ.M.Slingo(2003),TheMaritimeContinentanditsroleintheglobalclimate:AGCMstudy, J.Clim.,16,834–848. Nesbitt,S.W.,andE.J.Zipser(2003),ThediurnalcycleofrainfallandconvectiveintensityaccordingtothreeyearsofTRMMmeasurements, J.Clim.,16,1456–1475. Peatman,S.C.,A.J.Matthews,andD.P.Stevens(2014),PropagationoftheMadden–JulianOscillationthroughtheMaritimeContinentand scaleinteractionwiththediurnalcycleofprecipitation, Q.J.R.Meteorol.Soc. ,140,814–825. Acknowledgments D.B.B. has been supported by Office of Naval Research Global NICOP program (award N62909-15-1-2021 to the University of Warsaw). M.K.F. has been supported by the Chief of Naval Research through the NRL Base Program, PE 0601153N. P.J.F. has been supported by ONR DRI (Unified Physical Parameterization for Extended-Range Prediction). Development of the Kelvin wave trajectories database has been supported by Poland’s National Science Centre (Narodowe Centrum Nauki; decision 2012/07/N/ST10/03303). Analysis of the DYNAMO field campaign data has been supported by the Office of Naval Research under Program Element 601153N. BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8275

Geophysical Research Letters 10.1002/2016GL069602 Ridout, J. A., and M. K. Flatau (2011), Convectively coupled Kelvin wave propagation past sumatra: A June case and corresponding compositeanalysis, J.Geophys.Res. ,116,D07106,doi:10.1029/2010JD014981. Roundy,P.E.(2008),AnalysisofconvectivelycoupledKelvinwavesintheindianoceanMJO, J.Atmos.Sci. ,65,1342–1359. Vinayachandran,P.N.,J.Kurian,andC. P.Neema(2007),IndianOcean responseto anomalousconditions in2006, Geophys. Res. Lett.,34, L15602,doi:10.1029/2007GL030194. Wheeler, M., and G. N. Kiladis (1999), Convectively coupled equatorial waves: Analysis of clouds and temperature in the wavenumber-frequencydomain, J.Atmos.Sci. ,56,374–399. Yang,G.Y.,andJ.M.Slingo(2001),Thediurnalcycleinthetropics, Mon.WeatherRev. ,129,784–801. Yoneyama,K.,C.Zhang,andC.N.Long(2013),TrackingpulsesoftheMadden-Julianoscillation, Bull.Am.Meteorol.Soc. ,94,1871–1891. BARANOWSKIETAL. PHASELOCKINGOFCCKWS 8276