Role of SST over the Indian Ocean in Inﬂuencing the ...€¦ · mean monsoon ﬁelds (Krishnamurthy and Shu-kla 2000; Goswami and Ajayamohan 2001). This active-break cycle of the

Journal of the Meteorological Society of Japan, Vol. 85, No. 3, pp. 349--358, 2007 349

Role of SST over the Indian Ocean in Influencing the

Intraseasonal Variability of the Indian Summer Monsoon

Mathew ROXY

Graduate School of Environmental Earth Science, Hokkaido University, Sapporo, Japan

and

Youichi TANIMOTO

Faculty of Environmental Earth Science, Hokkaido University, Sapporo, Japan

(Manuscript received 28 September 2006, in final form 24 January 2007)

Abstract

Intraseasonal variability (10–60 days) of sea surface temperature (SST) over the north Indian Oceanand its influence on regional precipitation variability over the Indian subcontinent are examined. SST,cloud liquid water and precipitation over the Indian Ocean of the Tropical Rainfall Measuring Mission(TRMM), precipitation of Climate Prediction Center Merged Analysis of Precipitation (CMAP), and low-level atmospheric parameters of National Center for Environmental Prediction (NCEP) II reanalysis areutilized for this study. Western Ghats (WG) in the southwest and the Ganges-Mahanadi Basin (GB) inthe northeast of the Indian subcontinent are observed to be the regions of maximum precipitation withlarge standard deviations of the intraseasonal variability. Active (break) phases of precipitation occur inthese regions by the northward propagation of positive (negative) precipitation anomalies over the Ara-bian Sea and the Bay of Bengal. Latitude-time plots during the active phase of the WG region shows thatthe positive SST anomalies over the Arabian Sea formed by suppressed surface latent heat flux and in-creased downward shortwave radiation flux lead the positive precipitation anomalies. Surface air tem-perature anomalies follow the SST anomalies and then destabilize the lower atmosphere between1000 hPa and 700 hPa. These results indicate that, in the northward propagating dynamical surfaceconvergence, underlying SST anomalies tend to form a favorable condition for convective activity andmay sustain enhanced precipitation over the convergence region. This results in enhanced precipitationanomalies over the WG region that move further northeastward and merge with the northward propa-gating precipitation anomalies from the Bay of Bengal, enhancing the active phase of the GB region.

1. Introduction

Lying on both sides of the Indian subconti-nent, air sea interaction over the Arabian Sea

and the Bay of Bengal are found to influenceprecipitation variability associated with Indiansummer monsoon. The Arabian Sea is a princi-pal source of moisture fluxes across the westcoast of India, resulting in periods of heavyorographically forced rainfall over the West-ern Ghats (WG; Saha and Bavadekar 1977;Rakhecha and Pisharoty 1996). Rakhecha andPisharoty (1996) found that rainfall over the

Corresponding author: Mathew Roxy, Division ofOcean and Atmospheric Sciences, Graduate Schoolof Environmental Earth Science, Hokkaido Uni-versity, Sapporo, 060-0810, Japan.E-mail: [email protected]( 2007, Meteorological Society of Japan

Ganges-Mahanadi River Basins (GB) is associ-ated with tropical depressions formed over theBay of Bengal, with an effect principally limitedto the GB region. Vecchi and Harrison (2004)examined the role of SST anomalies in thenorthern Indian Ocean in the interannual vari-ability of precipitation in the WG and GB re-gions, where summer monsoon rainfall showslocal maximum in the seasonal mean and itsinterannual variability. For the WG region,they found a strong positive correlation withSST anomalies in the Arabian Sea at the onsetof summer monsoon. Meanwhile, they found nocorrelations between precipitation anomalies inthe GB region and SST anomalies in the Bay ofBengal. They suggest that interannual SSTvariations in the Bay of Bengal are poorly rep-resented in the SST product of National Cen-ters for Environmental Prediction (NCEP).

Intraseasonal variability of the Indian sum-mer monsoon, delineated by active periods ofheavy rainfall interrupted by break periods ofdeficient, significantly modulates the seasonalmean monsoon fields (Krishnamurthy and Shu-kla 2000; Goswami and Ajayamohan 2001).This active-break cycle of the Indian summermonsoon is associated with the northwardpropagation of clouds and convection from theequatorial region to around 30�N over theSouth Asian monsoon region, at a phase speedof about 1� latitude per day (Yasunari 1979,1980; Sikka and Gadgil 1980; Gadgil and Asha1992; Ramesh Kumar et al. 2005). There havebeen many attempts to develop a theoreticalframework of atmospheric dynamics in thenorthward propagation of the cloud bands onthis time scale (Murakami 1984; Jiang et al.2004; Yokoi and Satomura 2006).

Kemball-Cook and Wang (2001) examinedthe influence of latent heat flux on the north-ward migration of outgoing longwave radiation(OLR) anomalies over the Indian Ocean. Theysuggest that the moisture flux from the oceanto the atmosphere indicated by negative latentheat flux anomalies favors convection by build-ing up moist static energy. However, since theirstudy is based on convective maximum over theequatorial Indian Ocean, the resultant compos-ite signals weaken towards higher latitudes,and hence may not be effective in explainingthe processes involving the regional precipita-tion over the Indian subcontinent. Hence a

study focusing on the regional precipitationmaximums over the Indian subcontinent andthe role of SST on it is sought here. Also, thereis a meridional variation of the mean windsfrom the equator to the north Indian Ocean.Contribution of the zonal wind anomalies tothe mean wind speed and the resultant latentheat fluxes should be carefully treated in theanalysis.

Vecchi and Harrison (2002) examined intra-seasonal variability of outgoing longwave radi-ation with respect to SST anomalies over theBay of Bengal. They pointed out that positive(negative) SST anomalies on the intraseasonalscale in the Bay of Bengal exhibit a strong sta-tistical relation with the active (break) periods.Much work on Indian monsoon precipitationhas focused on large-scale indicators of precipi-tation variability, often the total over India,known as All India Rainfall index (AIR, Vecchiand Harrison 2004). Because of the two oceanslying on both sides and orographic features ofthe Indian subcontinent, the intraseasonal pre-cipitation variability may have regional charac-teristics along the east and west coast. In thepresent study, we take into account the re-gional characteristics of intraseasonal precipi-tation variability, instead of using the AIRindex. This study examines the role of intrasea-sonal SST anomalies over the Arabian Sea andthe Bay of Bengal separately, with respect toregional precipitation variability on the intra-seasonal scale. Our analysis on intraseasonalscale is possible due to recent availability ofhigh-quality satellite data.

The rest of paper is organized as follows. InSection 2 we describe datasets and data pro-cessing methods used for this study. In Section3 we explore the regional structure of intrasea-sonal precipitation variability over the Indiansubcontinent. Our focus is the SST variabilityover the Arabian Sea and the Bay of Bengal,and its influence on the regional precipitationvariability over the Indian subcontinent. Sec-tion 4 gives concluding remarks and discussion.

2. Data and analysis methods

Until recently, comprehension of observedSST in the north Indian Ocean during thesouthwest monsoon season has been limiteddue to sparse ship and buoy sampling of SST(Harrison and Larkin 1996) and heavy cloudi-

350 Journal of the Meteorological Society of Japan Vol. 85, No. 3

ness associated with the monsoon seriouslyrestricting the number of satellite infraredobservations. However, since December 1997,Tropical Rain Measuring Mission (TRMM) Mi-crowave Imager (TMI) has provided a novelview of tropical SST variability, unaffected byclouds, aerosols, and atmospheric water vapor(Wentz et al. 2000). Because of its frequent-repeat, non-sun-synchronous orbit, the TMISST dataset provides uninterrupted pictures ofthe Indian Ocean SST under active atmosphericconvection, which is useful for analysis of intra-seasonal SST variability (Senan et al. 2001).

The 3-day running mean SST and cloud liq-uid water based on TMI on 0.25� grid is usedin this study. We used the pentad precipitationbased on the Climate Prediction Center MergedAnalysis of Precipitation (CMAP), available on2.5� grid (Xie and Arkin 1996, 1997). To havehigher resolution data of precipitation, 3-dayrunning mean TMI precipitation data on 0.25�

grid is also utilized over the ocean. The daily2.5� gridded surface winds, surface latent heatflux (SLHF), downward shortwave radiationflux (DSWRF), and surface air temperature(SAT) based on NCEP Reanalysis II (NCEP II)are also used. Equivalent potential tempera-ture ðyeÞ is calculated based on air temperatureand specific humidity of NCEP II. The unstablecondition of the lower atmosphere is estimatedas a difference in ye between 1000 hPa and700 hPa,

Dye ¼ ye1000 � ye700

where ye1000 and ye700 are ye at 1000 hPa and700 hPa, respectively.

Data of SST, cloud liquid water, and precipi-tation between June 1 and Sept 30 (summermonsoon) from 1998 to 2002 are interpolatedto daily for compatibility among variables.Anomalies are obtained by removing the sea-sonal means for each of individual years andthen those anomalies are band pass filteredusing a boxcar filter to retain intraseasonalvariability between 10–60 days. These intrasea-sonally filtered anomalies are used in our study.

3. Results

3.1 Intraseasonal precipitation variabilityover the Indian subcontinent

To analyze the regional characteristics ofprecipitation over the Indian subcontinent on

intraseasonal scales, we plot the summertimestandard deviations (June to September) of in-traseasonal variability in Fig. 1. Large stan-dard deviations of the intraseasonal variabilityemerge over two regions in the Indian subconti-nent, over the WG (72.5–77.5�E, 7.5–20�N) andGB (77.5–87.5�E, 20–27.5�N) regions. The timeseries of the intraseasonal variability averagedover these two regions (Fig. 2) show larger am-plitudes in the WG region than in the GB re-gion and are not simultaneously correlated toeach other (correlation of 0.03). However, laterin this study, it is shown that precipitationanomalies at these two regions are having alag-lead correlation. Since there is no simul-taneous correlation but a lag-lead correlationbetween the two regions of precipitation, andsince the two oceans lying on either side couldhave different influence on them, examiningthe intraseasonal variability of precipitation ineach region is essential instead of a use of theAIR index.

3.2 WG regionTo examine temporal evolution of the intra-

seasonal precipitation variability over the WGregion, we define the WG active phase as thedays when the intraseasonal precipitationanomalies in that region (Fig. 2, black curve)

Fig. 1. Standard deviation of intraseaso-nal variability of precipitation (unit:mm day�1) for JJAS (1998–2002).Shading convention is represented onthe side of the panel. Contour lines (in-terval: 2 mm day�1) of precipitation aresuperimposed. The southwest (north-east) inset rectangle represents the WG(GB) region.

June 2007 M. ROXY and Y. TANIMOTO 351

exceeds a threshold value of 0.6 standard devi-ation (4 mm day�1), representing strong anom-alous precipitation. Most of the WG activephases have 5–7 days duration. Thirteen (13)events of the WG active phases are identifiedduring 1998–2002. We also define the WG pre-active (post-active) phase as 10 days before (af-ter) each day of the active phase because thisperiod (20 days) roughly represents half of theactive-break cycle. Figure 3 shows compositemaps of the intraseasonal precipitation, surfacewind, and SST anomalies during these threephases. During the WG pre-active phase (Fig.3a), weakly positive anomalies (@1 mm day�1)of precipitation are found in the southeasternpart of the Arabian Sea and in the southwest-ern part of the Bay of Bengal. These precipita-tion anomalies propagate northward and beginto intensify (5@7 mm day�1) during the WGactive phase, eventually triggering enhancedprecipitation near west coast of the WG region.Another local precipitation maximum with lessintensity (@5 mm day�1) is also located overthe Bay of Bengal during the active phase.These anomalies propagate further northwardduring the WG post-active phase, and thenweakened anomalies (@1 mm day�1) locateover the GB region (Fig. 3c).

In the region of the enhanced precipitationanomalies, an anomalous convergence of sur-face winds is observed (Fig. 3a). The intrasea-

sonal surface wind anomalies over the southArabian Sea during the pre-active phase arenortheasterlies (@1 ms�1), which results in re-duced wind speed because the seasonal meanwinds are southwesterly. This weak wind speedcauses the positive SST anomalies (@0.3�C)over the same region by the reduced heatrelease from the ocean. This is followed byan abrupt shift to anomalous southwesterlies(@1 ms�1) in the southern part of the ArabianSea during the active phase, forming a meri-dional gradient with negative (positive) SSTanomalies in the south (north) of the ArabianSea (Fig. 3b). The local SST maximum leadsrelative to the northward-migrating enhancedprecipitation, indicating that a warm oceansurface induced by weak wind speed may con-tribute to a favorable condition of the convec-tive activity as in the tropics.

To examine the phase relationship amongthe intraseasonal precipitation and SSTanomalies over Arabian Sea, we made hov-moller plots of daily composite variables aver-aged over 60–70�E (left inset rectangle in Fig.3c), from the WG pre-active to the WG post-active phase. Coherent northward migrations(0.9� per day) of the anomalies over the ArabianSea are observed (Fig. 4a) where positive SSTanomalies (0.2–0.4�C) lead by 1/4 of thecycle the positive precipitation anomalies (4–6 mm day�1). Though less intense, the negative

Fig. 2. Intraseasonal time series of precipitation anomalies (unit: mm day�1) averaged over the WG(black curve) and the GB (gray curve) regions. An active phase for each of the regions is defined aswhen the intraseasonal precipitation is above 0.6 standard deviation.


SST and precipitation anomalies follow the pos-itive anomalies, making an intraseasonal cycleof 40@50 days. For simplicity, we will focus onthe active phase with the positive precipitationand SST anomalies.

For a further evaluation of the role of SSTanomalies in influencing the intraseasonal pre-cipitation variability, the same hovmoller plotsare made using zonal surface winds, SLHF,DSWRF, SAT, Dye, and cloud liquid wateranomalies (Fig. 5). All panels in Fig. 5 depict acoherent northward propagation of the intra-seasonal anomalies from 5�N to 20�N. Coherentnorthward propagation was not observed in themeridional wind anomalies (not shown). Fig-ures 5a and 5b show that anomalous easterlysurface winds that oppose the climatologicalsouthwesterlies induce reduced evaporativecooling as shown by positive SLHF anoma-lies (30 Wm�2, downward positive). PositiveDSWRF anomalies (Fig. 5c, 30 Wm�2) appearat the same time, with intensity comparable tothe SLHF anomalies, warming the ocean sur-face. The SLHF anomalies are prominent northof 15�N, while the DSWRF anomalies areprominent around 15�N. Positive sensible heatnet flux and net longwave radiation fluxanomalies were also observed to be in phasewith the other surface heat fluxes, but they aretoo small in comparison and hence not shownhere. This suggests that SLHF and DSWRFanomalies contribute to the evolution of posi-tive SST anomalies (@0.3�C, contours in Fig. 5)with the SLHF anomalies playing significantrole over the northern Arabian Sea. The extentto which the surface heat flux anomalies canaccount for the observed intraseasonal varia-tions of SST were examined using the SST ten-dency equation

qTs

qt¼ Ftot

rcph

where Ts is the SST, Ftot is the total heat flux, %is density of water, cp is the specific heat ofwater at constant pressure, and h is the depthof the mixed layer averaged as 40 m for the In-dian Ocean during June to September (Karaet al. 2003). The observed SST tendency variedcoherently with those derived from the surfaceheat fluxes, both in amplitude and phase (Figs.5d and 5e), supporting the aforementioned rela-tionship. For instance, with an Ftot of 50 Wm�2,

Fig. 3. Composites of intraseasonal SST(colors; unit: �C), surface winds (arrows)and precipitation (contour lines; unit:mm day�1) anomalies during the WGactive phase. The inset rectangle inFig. 3c indicates the domains for thehovmoller plots of the Arabian Sea(60–70�E, Fig. 4) and the Bay of Bengal(85–95�E, Fig. 7). Contour lines of 1.0,3.0, 5.0, and 7.0 mm day�1 are plotted.Scaling of arrows (unit: ms�1) is givenon top of the figure. Coloring conventionis represented at the bottom of the fig-ure.


h of 40 m, and standard values of % and cp, theheat flux forcing estimated is in the order of0.025�C day�1. This is comparable to the SSTvariability of 0.02�C day�1, estimated for anSST change of 0.8�C in 40 days.

The positive SST anomalies are followed bycomparable magnitude of positive SAT anoma-lies (Fig. 5f ), indicating ocean-to-atmosphereeffect near the sea surface. These positive SATanomalies induce the positive anomalies in Dye(0.6�C) over the same region (Fig. 5g), a condi-tion favoring convective activity. Even thoughpositive anomalous Dye maximum appears tobe collocated with SLHF anomalies, the phasechange of positive Dye anomalies leads thephase change of negative SLHF anomalies.This indicates that the negative SLHF anoma-lies cannot account for the observed phasechange in the Dye anomalies, suggesting theeffect of SAT anomalies on the Dye anomalies.Indeed, ye1000 anomalies were observed to benearly equal to SAT anomalies. The positiveDye anomalies render adequate unstable condi-tion before the active phase. The destabiliza-tion of the lower atmosphere may provide theuplift of the moisture content from the lower at-mosphere and the consequent condensation re-sulting in positive cloud liquid water anomalies(@0.04 mm, Fig. 5h), giving rise to enhancedprecipitation. Indeed, positive precipitation

anomalies coincide with unstable condition inthe lower atmosphere (Fig. 4b) and abundantcloud liquid water (Fig. 4c). These findings sug-gest an SST effect on the convective cloudsthrough changes in the vertical structures ofthe lower-tropospheric temperature and mois-ture content. The suggested SST effect appearsto work effectively over the off-equatorial re-gions in the Arabian Sea, in the 10 to 25�Nband, where the SST anomalies are significant.

3.3 GB regionIn Fig. 3, the composite precipitation anoma-

lies after the WG active phase further propa-gate northward and are eventually displacedinto over the GB region. With the same mannerof the WG active phase, we define the GB activephase based on the gray curve in Fig. 2, whenthe threshold value exceeds 0.6 standard devia-tion (1.4 mm day�1). Sixteen (16) events of theGB active phases are identified. As for the WGactive phase, we also define the GB pre-activeand post-active phases. Figure 6 shows compos-ite maps of the intraseasonal precipitation, sur-face wind, and SST anomalies during thesethree phases. During the GB pre-active phase(Fig. 6a), two local maxima of precipitationanomalies locate along WG and in the Bay ofBengal. In a broad sense, these two maximaare also present during the WG active phase

Fig. 4. Hovmoller plots of intraseasonal anomalies of (a) SST (colors; unit: �C), (b) Dye (colors; unit:K) and (c) cloud liquid water (colors; unit: mm) over the Arabian Sea during the WG active phase.Contour lines of TMI precipitation anomalies (interval: 1 mm day�1) are superimposed, with nega-tive values dashed. Coloring convention is represented at the side of each panel.


(Fig. 3b). These anomalies are observed tomove northward and eventually merged ataround 22�N, 80�E during the GB active phase(Fig. 6b). Though less intense, the merged pre-cipitation anomalies are present during theWG post-active phase (Fig. 3c). These resultsshow that the GB active phase tends to occur5–10 days after the WG active phase. Indeed,the highest lag correlation coefficient is 0.44when the WG time series leads the GB time se-ries by 10 days (Fig. 2).

Factors other than the further northeastward

propagation of precipitation from the WG re-gion may also contribute to the GB activephase, such as SST anomalies over the Bay ofBengal. Local maximum of SST anomalies inthe Bay of Bengal is northward displaced rela-tive to the enhanced precipitation in Figs. 6aand 6b. A hovmoller plot of composite SST andprecipitation anomalies averaged over the Bayof Bengal (85–95�E; right inset rectangle inFig. 3c) during the GB active phase (Fig. 7)shows a similar ocean-to-atmosphere effectover the off-equatorial regions of the Bay of

Fig. 5. Hovmoller plots of intraseasonal anomalies of (a) surface zonal wind (colors; unit: ms�1), (b)surface latent heat flux (colors; unit: Wm�2), (c) downward shortwave radiation flux (colors; unit:Wm�2), (d) surface heat flux forcing (colors; unit: �C day�1), (e) SST tendency (colors; unit:�C day�1), (f ) surface air temperature (colors; unit: �C), (g) Dye (colors; unit: K) and (h) cloud liquidwater (colors; unit: mm) over the Arabian Sea during the WG active phase. Contour lines of SSTanomalies (interval: 0.1�C) shown in Fig. 4a are superimposed, with negative values dashed. Color-ing convention is represented at the side of each panel.


Bengal, in a similar manner as over the Ara-bian Sea (Fig. 4). Though the SST anomaliesare weak (@0.2�C in 15–25�N), compared tothose observed over the Arabian Sea, they arefollowed by enhanced precipitation anomalies(@5 mm day�1). This anomalous precipitationmoves further northward, merges with the pre-cipitation anomalies from the WG region (Fig.6b), and provides in-land anomalous precipita-tion at 25�N during the GB active phase. The

result implies that ocean-to-atmosphere effectover the Arabian Sea and the Bay of Bengal in-fluences the GB active phase.

4. Concluding remarks and discussion

Using observed SST, precipitation, surfacewind, SLHF, DSWRF, SAT, equivalent poten-tial temperature, and cloud liquid water, wehave analyzed the relationship between theSST over the Indian Ocean and the Indiansummer monsoon on intraseasonal timescales.The intraseasonal precipitation anomalies overthe WG and GB regions are simultaneouslyuncorrelated to each other, with the former in-fluenced by the ocean-to-atmosphere effectover the Arabian Sea and the latter by both theArabian Sea and Bay of Bengal. This suggeststhat intraseasonal precipitation variability overeach region should be examined separately forstudies involving the variability over the In-dian subcontinent as a whole. Over the ArabianSea, positive SLHF induced by reduced windand positive DSWRF anomalies are observedto contribute to the evolution of positive SSTanomalies. These northward propagating (at0.9�/day) positive SST anomalies are found tolead the positive SAT anomalies, indicatingocean-to-atmosphere effect near the sea sur-face. The positive SST anomalies are followedby the positive Dye anomalies, suggesting theactive role of SST anomalies in inducing unsta-ble conditions over the lower atmosphere, which

Fig. 6. Same as in Fig. 3, but for the GBactive phase.

Fig. 7. Same as in Fig. 4a, but for overthe Bay of Bengal during the GB activephase.


results in enhanced precipitation over theArabian Sea and the WG region. The positiveprecipitation anomalies during the WG activephase propagate further northward, contribu-ting to the GB active phase, 10 days after theWG active phase. Also, a similar ocean-to-atmosphere effect over the Bay of Bengal maycontribute to the GB active phase. A quantita-tive estimation of how much the Arabian Seaand the Bay of Bengal contribute to the GBactive phase is left for future research.

Prior to the WG active phase, an anomaloussurface convergence associated with northeast-erlies (southwesterlies) of the surface winds inthe northern (southern) part of the ArabianSea enhances convective precipitation. Accord-ing to Jiang et al. (2004), the easterly verticalshear (the upward increase of easterlies) overthe anomalous surface convergence in the loweratmosphere tends to generate anomalous mois-ture convergence north of the center of precipi-tation anomalies, leading to the northwardshift of the anomalous surface convergence andprecipitation. Their study shows that the north-ward migration of precipitation anomalies iscaused by this vertical shear mechanism. Hsuet al. (2004, 2005) suggest that the eastwardextension of the northward propagating precip-itation anomalies from the Arabian Sea couldbe driven by the anomalous surface conver-gence located to the east of the deep convectionregion, which is in turn induced by the liftingand frictional effects of the geography over thewest coast of the Indian subcontinent. Rossbywave response to heating sources over theequatorial Indian Ocean have been linked in re-motely enhancing the northward propagatingprecipitation anomalies (Lawrence and Web-ster 2002; Annamalai and Sperber 2005). Inthe present study, it is shown that the underly-ing positive SST anomalies enhance the north-ward propagating precipitation anomalies overthe Arabian Sea in a thermodynamical way.

Kemball-Cook and Wang (2001) had earlierpointed out that negative SLHF anomalies con-tribute to enhancing the convective activitythrough an increase in the moist static energy.In their study, it is given that anomalous east-erlies coincide with the development of nega-tive SLHF anomalies, which results in an in-crease in the moist static energy. Since theclimatological zonal winds are westerly at

the surface, the anomalous surface easterliesagainst the mean flow result in positive SLHFanomalies. In that case, the wind induced posi-tive SLHF anomalies cannot directly accountfor the positive moist static energy anomaliesobserved in their study. As also shown in thepresent study, the observed anomalous easter-lies induce the positive SLHF anomalies, re-sulting in positive SST anomalies (Figs. 5a and5b). These positive SST anomalies play a role inenhancing the positive Dye anomalies of thelower atmosphere (Fig. 5g), resulting in a con-dition favorable for enhanced precipitation(Figs. 4a and 5h). It is also to be noted that anincrease in SST enhances the moisture holdingcapacity of the lower atmosphere and henceis significant while considering moist staticenergy.

Thus, the results presented herein suggestthat SST anomalies over the north IndianOcean are significant in locally enhancing thenorthward propagating precipitation anomalies.The significant lead of positive SST anomaliesto the WG and GB active phases provides auseful precursor of the active spells in theIndian summer monsoon.

Acknowledgements

The TRMM data is obtained from RemoteSensing Systems and the NCEP/NCAR Reanal-ysis data from the NOAA Climate DiagnosticsCenter. The CMAP data is supplied by theNOAA Climate Diagnostics Center. Insightfulcomments by the two anonymous reviewers ledto substantial improvements in the manuscript.This work was partly supported by Grand-In-Aid for Scientific Research defrayed by the Min-istry of Education, Culture, Sports, Science andTechnology of Japan (17340137) and the Sumi-tomo foundation. We gratefully acknowledge fi-nancial support from the Center of Excellence,Graduate School of Environmental EarthScience, Hokkaido University.

References

Annamalai, H. and K.R. Sperber, 2005: Regionalheat sources and the active and break phasesof boreal intraseasonal (30–50 day) variability.J. Atmos. Sci., 62, 2726–2748.

Gadgil, S. and G. Asha, 1992: Intraseasonal Varia-tion of the Summer Monsoon I: ObservationalAspects. J. Meteor. Soc. Japan, 70, 387–397.


Goswami, B.N. and R.S. Ajayamohan, 2001: Intra-seasonal oscillations and interannual variabil-ity of the Indian summer monsoon. J. Climate,14, 1180–1198.

Harrison, D.E. and N.K. Larkin, 1996: The COADSsea level pressure signal: A near-global ElNino composite and time series view. 1946–1993. J. Climate, 9, 3025–3055.

Hsu H.-H., C.-H. Weng, and C.-H. Wu, 2004: Con-trasting characteristics between the northwardand eastward propagation of the intraseasonaloscillation during the boreal summer. J. Cli-mate, 17, 727–743.

Hsu, H.-H. and M.-Y. Lee, 2005: Topographic effectson the eastward propagation and initiation ofthe Madden-Julian Oscillation. J. Climate, 18,795–809.

Jiang, X., T. Li, and B. Wang, 2004: Structures andmechanisms of the northward propagating bo-real summer intraseasonal oscillation. J. Cli-mate, 17, 1022–1039.

Kemball-Cook, S. and B. Wang, 2001: Equatorialwaves and air–sea interaction in the borealsummer intraseasonal oscillation. J. Climate,14, 2923–2942.

Krishnamurthy, V. and J. Shukla, 2000: Intraseaso-nal and interannual variability of rainfall overIndia. J. Climate, 13, 4366–4377.

Murakami, T., T. Nakazawa, and J. He, 1984: Onthe 40–50 day oscillations during the 1979Northern Hemisphere summer. I: Phase propa-gation. J. Meteor. Soc. Japan, 62, 440–468.

Rakhecha, P.R. and P.R. Pisharoty, 1996: Heavyrainfall during monsoon season: Point and spa-tial distribution. Curr. Sci., 71, 179–186.

Ramesh Kumar, M.R., S.S.C. Shenoi, and J. Schultz,2005: Impact of convection over the equatorialtrough on summer monsoon activity over India.Int. J. Remote Sens., 26, 4747–4762.

Saha, K.R. and S.N. Bavadekar, 1977: Moisture fluxacross the west coast of India and rainfall dur-ing the southwest monsoon. Quart. J. Roy.Meteor. Soc., 103, 370–374.

Senan, R., et al., 2001: Validation of SST and windspeed from TRMM using North Indian Oceanmoored buoy observations, Indian Institute ofScience, Bangalore, India. 29 pp.

Sikka, D.R. and S. Gadgil, 1980: On the maximumcloud zone and the ITCZ over Indian longi-tudes during the southwest monsoon. Mon.Wea. Rev., 108, 1840–1853.

Vecchi, G.A. and D.E. Harrison, 2002: Monsoonbreaks and subseasonal sea surface tempera-ture variability in the Bay of Bengal. J. Cli-mate, 15, 1485–1493.

Vecchi, G.A. and D.E. Harrison (Eds.), 2004: Interan-nual Indian rainfall variability and IndianOcean sea surface temperature anomalies,247–260, American Geophysical Union, Geo-physical Monograph, Washington D.C.

Webster, P.J., et al., 1998: Monsoons: Processes, pre-dictability and the prospects of prediction. J.Geophys. Res., 103, 14451–14510.

Wentz, F., et al., 2000: Satellite measurements of seasurface temperature through clouds. Science,288, 847–850.

Xie, P. and P.A. Arkin, 1996: Analyses of globalmonthly precipitation using gauge observa-tions, satellite estimates, and numerical modelpredictions. J. Climate, 9, 840–858.

Xie, P. and P.A. Arkin, 1997: Global precipitation: a17-year monthly analysis based on gauge ob-servations, satellite estimates, and numericalmodel outputs. Bull. Amer. Meteorol. Soc., 78,2539–2558.

Yasunari, T., 1979: Cloudiness fluctuation associatedwith the Northern Hemisphere summer mon-soon. J. Meteor. Soc. Japan, 57, 227–242.

Yasunari, T., 1980: A quasi-stationary appearance of30–40 day period in the cloudiness fluctuationsduring the summer monsoon over India. J.Meteor. Soc. Japan, 58, 225–229.

Yokoi, S. and Satomura, T., 2006: Mechanisms of thenorthward movement of submonthly scale vor-tices over the Bay of Bengal during the borealsummer. Mon. Wea. Rev., 134, 2251–2265.


Documents

Role of SST over the Indian Ocean in Inﬂuencing the ...€¦ · mean monsoon ﬁelds (Krishnamurthy and Shu-kla 2000; Goswami and Ajayamohan 2001). This active-break cycle of the