The Indonesian Throughflow and the Global Climate System

676 JOURNAL OF CLIMATE VOLUME 11

The Indonesian Through¯ow and the Global Climate System

NIKLAS SCHNEIDER Climate Research Division, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California (Manuscript received 21 January 1997, in ®nal form 7 August 1997)

ABSTRACT The role of the Indonesian Through¯ow in the global climate system is investigated with a coupled ocean± atmosphere model by contrasting simulations with realistic through¯ow and closed Indonesian passages. The Indonesian Through¯ow affects the oceanic circulation and thermocline depth around Australia and in the Indian Ocean as described in previous studies and explained by Sverdrup transports. An open through¯ow thereby increases surface temperatures in the eastern Indian ocean, reduces temperatures in the equatorial Paci®c, and shifts the warm pool and centers of deep convection in the atmosphere to the west. This control on sea surface temperature and deep convection affects atmospheric pressure in the entire Tropics and, via atmospheric teleconnections, in the midlatitudes. As a result, surface wind stress in the entire Tropics changes and meridional and zonal gradients of the tropical thermocline and associated currents increase in the Paci®c and decrease in the Indian Ocean. The response includes an acceleration of the equatorial undercurrent in the Paci®c, and a deceleration in the Indian Ocean. Thus the Indonesian Through¯ow exerts signi®cant control over the global climate in general and the tropical climate in particular. Changes of surface ¯uxes in the Paci®c warm pool region are consistent with the notion that shading by clouds, rather than increases of evaporation, limit highest surface temperatures in the open ocean of the western Paci®c. In the marginal seas of the Paci®c and in the Indian Ocean no such relationship is found. The feedback of the through¯ow transport and its wind forcing is negative and suggests that this interplay cannot excite growing solution or lead to self-sustained oscillations of the ocean±atmosphere system.

1. Introduction transport of the through¯ow removes a signi®cant amount of the surface heat ¯ux the Paci®c received in The Indonesian Through¯ow is the only low-latitude the Tropics, one-third in a simulation of Hirst and God- oceanic connection between the major ocean basins of frey (1993) or 0.63 and 0.9 PW in the analysis of a today. Its transports of mass, heat, and freshwater from coupled model by Schneider and Barnett (1997). The the Paci®c into the Indian Ocean are important for the through¯ow forces the Leeuwin current, either due to oceanic circulation and sea surface temperatures, but its a particular form of vertical mixing (Kundu and role in the global climate system and its impact on sim- McCreary 1986) or by atmospheric cooling and the en- ulations of the climate by coupled ocean±atmosphere suing geostrophic adjustment of warm waters supplied models are unknown. In this study, the role of the In- by the through¯ow (Godfrey and Weaver 1991; Hughes donesian Through¯ow in the global ocean±atmosphere et al. 1992). Salt transports by the through¯ow represent system is examined by contrasting climates simulated a major salt sink for the Paci®c and a gain for the Indian by a coupled general circulation model with open and Ocean (Piola and Gordon 1984). Finally, water mass closed Indonesian passages. considerations (Gordon 1986; Gordon et al. 1992) and The role of the through¯ow in the ocean has been numerical experiments (Shriver and Hurlburt 1997) sug- recently reviewed by Godfrey (1996), so only a brief gest that the through¯ow takes part in the return branch summary is presented here. The mass transport of the of the thermohaline circulation. through¯ow affects the thermocline of the Indian Ocean These effects of the through¯ow on the oceanic cir- and is in part responsible for anomalous deep thermo- culation were con®rmed by Hirst and Godfrey (1993, cline and lack of cold upwelling off the western coast hereafter referred to as HG), who compared simulations of Australia (Godfrey and Golding 1981). The heat using ocean general circulation models with open and closed Indonesian passages. In addition, HG showed that the thermocline structures in the Indian Ocean and southern Atlantic were altered in a manner consistent Corresponding author address: Dr. Niklas Schneider, Climate Re- search Division 0224, Scripps Institution of Oceanography, Univer- with Sverdrup circulation (Godfrey 1989) and that sur- sity of CaliforniaÐSan Diego, La Jolla, CA 92093-0224. face temperatures changed in regions far removed from E-mail: [email protected] Indonesian waters where vertical mixing communicated

᭧ 1998 American Meteorological Society

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC APRIL 1998 SCHNEIDER 677 changes of the thermocline to the surface. The signature scribes the coupled model, and sections 3 and 4 intro- of the Indonesian Through¯ow on surface temperatures duce the experiment and discuss the signi®cance of re- in the southeastern Indian Ocean and in the Great Aus- sults in light of internal variability of the coupled sys- tralian Bight were con®rmed by simulations of the Ant- tem. Changes of the mean climate and seasonal cycle arctic Ocean (Ribbe and Tomczak 1997). are presented in section 5 and feedbacks in section 6, In HG's experiment, surface ¯uxes of heat and fresh- followed by discussion and conclusions. water were approximated by Newtonian relaxation to observed values, and surface wind stresses were pre- 2. Coupled model scribed from observations. Thus changes of the surface ¯uxes of heat and freshwater were those obtained from The model (ECHO, Latif et al. 1994) was developed this Newtonian formulation, and ¯uxes of momentum at the Max-Planck-Institut fuÈr Meteorologie, Hamburg, were held constant. The sensitivity of the coupled Germany, and consists of an atmospheric general cir- ocean±atmosphere system to the Indonesian Through- culation model (ECHAM3; Roeckner et al. 1992; DKRZ ¯ow could therefore not be determined in HG's study. 1992) and a primitive equation, ocean general circula- Results of HG, together with the observed atmo- tion model. ECHAM3 has 19 levels in the vertical and spheric response to interannual anomalies of surface is run at T42 (2.8Њϫ2.8Њ) resolution. The horizontal temperature, yield a hypothesis for the role of the resolution of the ocean model is variable with latitudinal through¯ow in the global climate system. Observations spacing of 0.5Њ within 10Њ of the equator and expanding of interannual anomalies show that an increase of sur- to 2.8Њ resolution poleward of 20Њ latitude. The longi- face temperature in the tropical central Paci®c, a re- tudinal resolution is 2.8Њ and there are 20 levels in the sponse seen by HG, causes an eastward shift of the vertical (see Latif et al. 1994 for additional details). centers of deep convection and indicate that the through- ECHO is a global model and couples atmosphere and ¯ow controls positions of the western Paci®c warm pool ocean globally and between 60ЊN and 60ЊS without ¯ux and centers of deep convection. Atmospheric telecon- correction. At higher latitudes surface temperatures and nections between interannual anomalies of surface tem- salinity are additionally relaxed to climatology (cf. Lev- perature in the equatorial Paci®c and atmospheric pres- itus 1982). sure in midlatitudes (Horel and Wallace 1981; Karoly A simulation of ECHO spanning 125 yr was inves- 1989) suggest that the through¯ow affects midlatitudes tigated in a number of studies (Latif et al. 1994; Latif by its control of tropical surface temperatures. Similarly, and Barnett 1994; Schneider et al. 1996; Xu et al. 1998). observed links between surface temperature in the trop- While the model suffers from a tendency, typical for ical Paci®c and the summer monsoon (Palmer et al. coupled models, to split the intertropical convergence 1992), and between summer monsoon rainfall and tem- zone in the Paci®c and to produce waters off South peratures in the eastern Indian Ocean north of Australia America that are too warm, it generates all major fea- (Nicholls 1995), suggest that the monsoons are affected tures of the atmosphere and ocean with approximately by the through¯ow. Rainfall anomalies over Australia the right strength and spatial structure. The Indonesian have been linked to surface temperature anomalies in Through¯ow of the last 105 yr of this simulations with the Indonesian waters, Indian Ocean, and central Paci®c ECHO has been studied in detail by Schneider and Bar- (Nicholls 1989). Thus, control of sea surface tempera- nett (1997). It transports on average 13.8 Sv (Sv ϵ 106 ture by the through¯ow implies that rainfall over Aus- m3 sϪ1) and exports 0.9 PW of heat from the Paci®c to tralia is affected. the Indian Ocean, and compares favorably with obser- A test of this hypothesis requires experiments with a vation and other simulations. Overall, the simulation of coupled ocean±atmosphere model, preferably without the world's climate and through¯ow by ECHO are ¯ux correction, as the latter might compromise results. amazingly realistic and give con®dence that results re- Schneider and Barnett (1997) investigated the through- ported here have some bearing for the earth's climate. ¯ow in a simulation with such a coupled ocean±atmosphere model and found it realistic in both its transport 3. Experiment of mass and heat. This simulation serves as a reference state for an experiment that determines changes of the In ECHO, passages between New Guinea and the climate system induced by a closure of Indonesian wa- model's Asia and between New Guinea and Australia ters. At the outset it has to be pointed out that the ther- connect the western Paci®c and the Indian Ocean. In mohaline circulation in the coupled model is constrained March of year 90 of the reference run, at a time when by observed surface temperature and salinity in high the through¯ow was near its average, these straits were latitudes and can therefore not change. However, the closed with walls and the through¯ow was forced to model is well suited to study the sensitivity to the vanish. The model with closed through¯ow was inte- through¯ow of atmosphere and upper ocean equator- grated for 10 yr, and the average of the last 5 yr is ward of 60Њ lat, a task that also limits simulation time investigated. to a realizable effort. In the Tropics and high latitudes this integration time The outline of the paper is a follows: section 2 de- allowed the coupled system to reach approximately a

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC 678 JOURNAL OF CLIMATE VOLUME 11

thermocline in the Paci®c are enhanced. In the midlatitude Paci®c and Northern Atlantic, the trace of baroclinic waves is seen whose changes of the heat content are not signi®cant, however. Thus, the duration is of suf®cient length to determine the response of upper ocean and atmosphere to changes of the through¯ow. Changes expected from a longer integration include eastward migrations of the Paci®c warm changes, further relaxation of the zonal gradient of the thermocline in the Paci®c, and an enhancement of changes in the midlatitude gyres.

4. Signi®cance A few weeks after closure of Indonesian waters, differences of simulated SST and of surface ¯uxes between reference integration (run REF hereafter) and integration of ECHO with closed Indonesian seas (referred to as NOTF) occur throughout the world. This is because closure of Indonesian waters not only changes the phys- ical characteristics but perturbs the coupled ocean±atmosphere system, which then departs from the reference run due to its nonlinear nature and sensitivity to initial conditions. To distinguish effects of closure of the through¯ow from internal variability of the coupled system, we compare the last 5 yr of NOTF with statistics of 105 yr of REF integration. Signi®cance of changes of a scalar quantity, for ex- ample SST, is determined by the rank of the average of the last 5 yr of the NOTF run in the distribution of 5- yr averages from REF. The latter are determined with 4 yr overlap, such that the 105-yr time series yield a sample size of 101. Alternatively, the Mann±Whitney test (Conover 1971) is employed, a nonparametric test that determines if the expected value of two random samples differ signi®cantly. The test is based on the sum of ranks of the ®rst random sample in the concat- enation of both samples, and it assumes that individual FIG. 1. (a) Leading empirical orthogonal functions (EOF 1: solid, 2: dashed, and 3: dotted) of SST change (60ЊS±60ЊN) after closure samples are independent from each other. Strictly speak- of Indonesian waters with percent variance explained indicated. EOFs ing, this is not true for results of the simulations since were determined from the 5-yr running mean SST from years 86±99 interannual and decadal signals are present. However, that had its average removed. (b) Reconstruction of 5-yr running mean both tests yield very similar conclusions, and the ®rst SST changes due to EOF 1 centered on year 97.4. (c) Reconstruction of 5-yr running mean SST changes due to EOFs 2±10 centered on method, which takes into account autocorrelations, will year 97.4. Positive values are shaded. be used in the following. The test is applied at every position, and results are mapped only if signi®cance exceeds 95% con®dence limits. new equilibrium. The leading empirical orthogonal Vectors, such as the wind stress, are rotated along function of the 5-yr running mean of the changes of sea axis of maximum covariance of their components before surface temperature (SST) after closure of the through- signi®cance tests are applied to each component. Sig- ¯ow has reached steady value (Fig. 1a), explains 75% ni®cance of changes of the vector results from the prod- of the change, and captures nearly the entire change of uct of the probabilities of each component. SST (Fig. 1b). Higher modes, which have not reached Changes of annual cycles are determined by com- equilibrium (Fig. 1a), have largest action in midlatitudes parison of the annual amplitude and phases based on 5- (Fig. 1c) but alter SST in the Tropics and high latitudes yr segments of the NOTF run, and 101 5-yr segments only in a minor way. At the end of the experiment, 5- of the REF run. Annual amplitudes and phases of REF yr averages of oceanic heat content of the upper 250 m are interpreted as polar coordinates of a vector, and sig- are still evolving such that the cooling in the Indian ni®cance of changes is determined using the procedure Ocean and the relaxation of the zonal gradient of the for vectors outlined above.

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC APRIL 1998 SCHNEIDER 679

FIG. 2. Difference of sea level in cm of the integration with closed through¯ow and the reference run with open Indonesian waters. Solid FIG. 3. Changes in surface temperature in K due to closure of the lines mark positive values, dashed represent negative numbers. There through¯ow. Warming is depicted by solid contours and cooling by is no zero contour. Shading indicates differences that are signi®cant dashed lines. Shading highlights changes that are signi®cant at the at the 95% level. 95% level.

Comparison of NOTF run with the entire 105 yr of responsible for weak upwelling and vertical shear of REF leads to rather stringent signi®cance requirements, currents in the eastern equatorial Indian Ocean. since changes due to through¯ow closure starting from year 90 of REF are required to be signi®cantly different from internal variability of the model, including changes 2) SEA LEVEL on decadal timescales (Latif and Barnett 1994) and the In response to closure of the through¯ow, sea level time-varying datum they supply. This ensures that the increases in the Paci®c, decreases in the Indian Ocean, comparison is conservative. and yields a sea level difference in excess of 30 cm between the western Paci®c and the Indonesian Sea (Fig. 5. Climate sensitivity 2). Concurrently, sea level gradients south of Australia are reduced. In the Indian Ocean sea level is decreased a. Mean state by more than 25 cm centered at 20ЊS. The wedge shape 1) SUMMARY of the sea level difference is consistent with changes in surface velocity of HG and results from lack of circu- Closure of the Indonesian Through¯ow causes an lation around Australia. South of Madagascar sea level eastward shift of the western Paci®c and eastern Indian decreases in the Aghulhas retro¯ection, again consistent Ocean warm pool, with accompanying changes of cen- with HG. ters of deep convection and their expression in surface ¯uxes of heat and freshwater. The eastward shift of the centers of deep convection move tropical atmospheric 3) SURFACE TEMPERATURE pressure patterns associated with the Walker circulation eastward and so affect the entire Tropics. Teleconnec- The surface equatorial Paci®c warms by up to 1 K tions to midlatitudes communicate these tropical signals with maximum heating east of the dateline, while In- over the North Paci®c and over the southern oceans. As donesian seas and eastern Indian Ocean cool by up to a result, surface wind stress changes globally. The Pa- 0.6 K (Fig. 3). These changes represent an eastward ci®c trades relax, while trades in the Atlantic and Indian shift of the warm pool, a relaxation of the SST gradient increase. along the equator in the west, and an increase in the In the southern Indian Ocean and along the eastern eastern Paci®c. Qualitatively, this response is expected coast of Australia, oceanic heat content and currents from a shutdown of the through¯ow, since it transports re¯ect the changes of the Sverdrup circulation due to heat from the Paci®c into the Indian Ocean. In the southeastern boundary in¯ow into the Indian Ocean and cir- ern Indian and Atlantic Oceans, cooler surface waters culation around Australia. The sensitivity of the surface stretch from Australia to South America, while in the wind stress to the Indonesian Through¯ow causes a re- south Paci®c, surface temperatures are increased slight- laxation of spatial gradients of the heat content in the ly. Even though barely signi®cant, the North Paci®c off Paci®c Ocean and a steepening in the Indian Ocean and, North America is warmed by up to 0.6 K, while the to a lesser extent, in the Atlantic. Thus, surface currents central North Paci®c is cooled. and the equatorial undercurrent weaken in the Paci®c, Over land, changes are largest over western Australia, while in the tropical Indian Ocean these are enhanced with warming of up to 2 K. Over northeastern Asia slightly. These results indicate that the control of the surface temperature cooled, and there is some indication Indonesian through¯ow on the surface winds is partially for warming over the Himalayas and Brazil.

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC 680 JOURNAL OF CLIMATE VOLUME 11

FIG. 4. Changes in precipitation due to closure of the through¯ow FIG. 5. Changes of net heat ¯ux in W mϪ2 in response to closure in mm dayϪ1. Increase of rainfall is shown by solid lines, decreases of the Indonesian seas. Oceanic heating is shown by solid contours are indicated by a dashed line, and the zero contour is omitted for and cooling by dashed lines. There is no zero contour, and shading clarity. Shading highlights changes that are signi®cant at the 95% marks areas that are signi®cant at the 95% level. level.

over the western Paci®c and in the Indian Ocean off 4) PRECIPITATION western Australia by less than 20 W mϪ2. Latent heat The cooling of the eastern Indian Ocean and warming losses are slightly increased over the eastern Indian of the western Paci®c are associated with an eastward Ocean at 10ЊS and over the western Tasman Sea. Over shift of the centers of deep convection, as shown by western Australia, latent heat losses are reduced in an changes in precipitation (Fig. 4). The eastern Indian area where rainfall is decreased and surface tempera- Ocean and western Australia experience precipitation tures are increased, and they indicate a drying of this de®cits of up to 3 mm dayϪ1, while precipitation in the area. Over western Australia, changes of the sensible western Paci®c increases by more than this amount. In heat ¯ux cool by an additional 15 W mϪ2, and, together the equatorial Paci®c, precipitation increases, whereas with increased longwave heat losses, balance heat gains the intertropical convergence zones are weakened. Con- due to shortwave radiation and reduce latent heat ¯uxes. vective precipitation dominates the signal in the Tropics and is augmented by changes in large-scale precipitation 7) NET HEAT FLUX (due to large-scale convergence) of the same sign, but of half the magnitude. Over western Australia, in the The response of the net heat ¯ux (Fig. 5) over the Tasman Sea, and in the Paci®c south of Japan large- ocean is correlated with and counteracts changes in sur- scale precipitation changes signi®cantly by 0.3 mm face temperature; over land changes are small. This re- dayϪ1 (not shown). sponse is reminiscent of the damping in¯uence of surface heat ¯ux seen in the studies of El NinÄo (Barnett et al. 1991; Schneider and Barnett 1995) and corresponds 5) RADIATION to a heating of 10±20 W mϪ2 in the eastern Indian Ocean The convective rainfall into the western and central and a cooling of 10 W mϪ2 in the equatorial Paci®c. equatorial Paci®c is accompanied by changes of short- Interesting enough, a large warming in the Tasman Sea and longwave radiation expected from the change in (Fig. 3) is associated with vigorous cooling of more cloudiness (not shown). Patterns of these ®elds are very than 60 W mϪ2 and indicates large heat transfers from similar over the ocean. In the western Paci®c shortwave the ocean. Changes in the net ¯ux are much smaller than radiation is reduced by up to 30 W mϪ2, whereas eastern those obtained by a Newtonian formulation in HG and Indian Ocean shortwave radiation is increased by almost indicate that feedbacks within the atmosphere moderate the same amount. Longwave radiation over the ocean the surface ¯ux response. partially offsets shortwave changes, with increased heat Changes of the net heat ¯ux (Fig. 5) and precipitation gains of 6 W mϪ2 in the western Paci®c and decreases (Fig. 4) have similar spatial patterns in the tropical east- of4WmϪ2 in the eastern Indian Ocean. Over land, in ern Indian Ocean and western Paci®c. Precipitation can- particular Australia and Siberia, increases in surface cels approximately two-thirds of the surface buoyancy temperature are accompanied by increased longwave ¯ux due to change in net heat ¯ux and indicates that heat losses and increased shortwave heat gains. changes of surface heat and freshwater ¯uxes associated with the transfer of centers of deep convection from the Indian Ocean to the Paci®c alter the surface buoyancy 6) TURBULENT HEAT FLUX ¯ux little. This is in accordance to changes of the surface Changes in the turbulent heat ¯uxes are dominated buoyancy ¯ux associated with interannual migrations of by the latent heat ¯ux over the ocean and by sensible the centers of deep convection (Schneider and Barnett heat ¯ux over land. Latent heat ¯ux losses are decreased 1995).

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC APRIL 1998 SCHNEIDER 681

FIG. 6. Changes in height of the 850-mb isobar in meters due to the closure of the through¯ow. Increases are marked by solid lines and decreases by dashed lines. The data has been smoothed with a three-point boxcar ®lter, and the zero contour is omitted for clarity. Shading highlights changes that are signi®cant at the 95% level.

8) ATMOSPHERIC PRESSURE The eastward shift of the centers of deep convection and the warming in the central Paci®c have a dramatic effect on atmospheric pressure around the globe (Fig. 6). Consistent with the approximately 90Њ long between the cold and warm changes on the equator (Fig. 3), atmospheric pressure response has a global wavenumber FIG. 7. Deviations from zonal mean of geopotential at 200 (top), 2, with upward displacements of the 850-mb isobaric 500 (center), and 850 mb (bottom), averaged between 15ЊS and 15ЊN, surface west of the dateline and over Brazil and much for the reference run (solid line) and closed through¯ow run (dashed). NCEP/NCAR reanalysis (Kalnay et al. 1996) averaged from 1974 to of the tropical Atlantic, and negative perturbations over 1995 is shown as a dashed±dot line. the eastern Paci®c, Africa, and the western Indian Ocean. The eastward shift of deep convection excites teleconnections with the midlatitudes, reminiscent of the response to interannual anomalies in the tropical Paci®c over Brazil are accompanied by stronger southward (Horel and Wallace 1981; Karoly 1989). The North Pa- winds from the Atlantic. ci®c low is intensi®ed, as is the Siberian high, and there In midlatitudes, changes of surface pressure increase is an indication of changes over North America. In the continental out¯ows over the North Paci®c and north- Southern Hemisphere, lows over the circumpolar cur- ward stresses over the Gulf of Alaska. The circulation rent are intensi®ed. over the Atlantic is altered by a cyclonic circulation Further aloft, 500- and 200-mb surface show similar with southward stress over the eastern United States, changes, with increasing amplitudes over the western and northward stress over the central North Atlantic. In Paci®c and the Atlantic, and decreasing amplitudes over the Southern Hemisphere, the enhancement of the lows the eastern Paci®c and Indian Ocean (Fig. 7). The fact at 50ЊS leads to a northward shift of the westerlies, with that displacements of isobaric surface over the warm westward changes at 40ЊS and eastward changes at 60ЊS. pool are of the same sign at 850, 500, and 200 mb Equatorward winds at the western coast of Australia are suggests that this is a response to atmospheric heating enhanced, as the eastward winds in the Tasman Sea. due to convection. These changes of the wind stress suggest a possible feedback of the through¯ow on its wind forcing that will be further investigated in section 6. 9) WIND STRESS

The global changes in the atmospheric pressure due 10) HEAT CONTENT AND CURRENTS to the closure of the through¯ow elicit global changes of the surface wind stress (Fig. 8). In the tropical western Consistent with changes of the surface temperature, and central Paci®c, winds converge onto the equator closure of through¯ow cools the upper-water column of and are more eastward, as expected from the appearance the Indian Ocean and warms the Paci®c (Fig. 9). The of deep convection there, and offset the divergence of increase of heat in the Paci®c is largest in the eastern winds over the equatorial cold tongue found in the ref- half, along the equator, and under the intertropical con- erence run (Schneider et al. 1996). Over Indonesia and vergence zones and indicates that the zonal slope of the in the eastern Indian Ocean, northward winds strengthen thermocline in the equatorial Paci®c and the ridges at as seen in the southeast trades and the southwest winds the poleward edges of the countercurrents of both hemi- in the Bay of Bengal. The deepening of low pressures spheres are reduced. Consistent with these changes of

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC 682 JOURNAL OF CLIMATE VOLUME 11

FIG. 8. Changes of surface wind stress in 10Ϫ2 NmϪ2 in response to closure of the Indonesian Through¯ow. Colors indicate magnitude of changes. Areas of signi®cant changes at the 95% level are shaded. heat content, transports of the south equatorial currents closure of the Indonesian through¯ow (Fig. 11). The and the countercurrents are reduced (Fig. 10). relaxation of zonal gradient of the thermocline along In the Atlantic, warming to the north of the equator the equator in the Paci®c is accompanied by a weak- indicates that the ridge associated with the North Equa- ening of the Equatorial Undercurrent. While changes torial Countercurrent is reduced. In addition, the South are small compared to observed values of the speed of Atlantic shows a band of cooling that extends beyond the Equatorial Undercurrent, compared to the weak ¯ow the southern tip of Africa into the Indian Ocean (Fig. in the coupled model, they are of the order of 10%. In 9). the Indian Ocean, and to a lesser extent in the Atlantic, The heat content and circulation around Australia due the increase of east±west slope of the equatorial ther- to the Indonesian Through¯ow and its continuation in mocline results in an increase of the vertical shear of the Indian Ocean is shown by the changes of the heat velocity. In fact, in the eastern Indian Ocean a subsur- content and currents on the eastern side of Australia and face maximum of eastward velocity is nonexistent in across the Indian Ocean to the coast of Madagascar. the integration with open through¯ow, in accordance North of 10ЊS changes of the heat content and transports with weak average winds and observations (Knox 1976; suggest a strengthening of the equatorial currents and a Knox and Anderson 1985). Closure of the through¯ow steepening of the east±west slope of the equatorial ther- results in increases of upwelling and vertical shear in mocline (Fig. 9). the eastern equatorial Indian Ocean and partially re- In addition to the changes of the transport in the upper lieves the anomalous dynamical conditions of the east- ocean, vertical shear along the equator is sensitive to ern Indian Ocean compared to the eastern Paci®c and Atlantic.

b. Seasonal cycle The seasonal cycle of heat ¯ux of ECHO's through- ¯ow is smaller by a factor of 4 than the seasonal cycle of the surface heat ¯ux in the Indian Ocean (Schneider and Barnett 1997). Oceanic changes of annual cycles due to closure of the through¯ow result therefore largely from an altered mean state described above. In general, amplitudes of the annual cycle are affected, while the phase remains largely unaltered since it is tied to the solar cycle. FIG. 9. Changes in average temperature in K over the top 250 m due to closure of the Indonesian Sea. Shading indicates changes that Changes of the annual cycle of 850-mb height are are signi®cant at the 95% level. marked by a sequence of areas with alternatively larger

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC APRIL 1998 SCHNEIDER 683

FIG. 10. Changes in transport in mϪ2 sϪ1 over the top 250 m due to closure of the through¯ow. Arrows denote direction of transport change; their length is scaled with an exponential function for better clarity. Changes of the transport magnitude are indicated by color code given on the right. Changes in areas without color are not signi®cant at the 95% level. and smaller annual cycles stretching from the North changes of the net heat ¯ux. They are dominated by Paci®c over North America into the North Atlantic (Fig. changes in shortwave radiation in the low latitudes as- 12). These changes indicate strengthening and shifting sociated with the seasonal cycle of deep convection, and of atmospheric pressure patterns and are reminiscent of by change in the latent heat ¯ux in midlatitude oceans. atmospheric teleconnections of interannual changes in SST in the tropical Paci®c and winter time pressure in 6. Feedbacks the midlatitudesÐthat is, of the excitation of the PNA pattern (Horel and Wallace 1981) and, similarly, con- a. Ocean±atmosphere coupling nections with the southern ocean (Karoly 1989). Closure of the through¯ow alters the oceanic topog- Over the Indian Ocean sector, annual cycles of 850- raphy and thereby affects the oceanic Sverdrup circu- mb height are enhanced over the eastern Indian Ocean, lation. This sensitivity to the Indonesian Through¯ow Indonesia and Australia, and over Asia and suggest a has been documented in experiments with an ocean strengthened monsoon, consistent with the warming in model by HG. In the coupled model resulting changes the central Paci®c (Palmer et al. 1992). Indeed, surface of the surface temperature will alter the state of the wind stress over the Bay of Bengal and Indonesian wa- atmosphere and of the surface ¯uxes and feed back on ters are signi®cantly increased during both monsoons the ocean. Thus a comparison of the results of HG with in the NOTF run (Fig. 13). the sensitivity to the closure of the through¯ow in the Changes of the seasonal cycle of SST correspond to coupled model identi®es feedbacks of the coupled system. Closure of the Indonesian Through¯ow in HG alters the oceanic Sverdrup circulation in the Indian Ocean, southern Atlantic, and on the eastern coast of Australia. The response of the coupled model in these areas is very similar to results of HG; even the wedge shape of anomalous transports in the Indian Ocean that are due to the arresting of westward propagation of Rossby waves by the mean circulation (HG) are reproduced. This indicates that this response of the coupled model to closure of the through¯ow results mainly from changes of the Sverdrup transport due to changes of the topography. FIG. 11. Changes in zonal velocity in 10Ϫ2 msϪ1 along the equator. Positive numbers indicate eastward ¯ow. Shading implies changes In the equatorial Indian Ocean, HG report no signif- that are signi®cant at the 95% level. icant changes and in the Paci®c away from Australia

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC 684 JOURNAL OF CLIMATE VOLUME 11

FIG. 12. Change of the annual cycle of 850-mb height due to closure of the through¯ow. Changes are represented as vectors, whose angle from north indicates changes in phase. An eastward-pointing arrow corresponds to a 3-month increase of the annual phase. Color is present in areas where results are signi®cant at the 95% level and represents changes in amplitude in m. changes of steric height and currents are restricted to through¯ow transport by the winds is in large part de- the equator and re¯ect changes of the temperature of scribed by Godfrey's Island Rule (Godfrey 1989; the upwelled water. In the coupled model the sensitivity Schneider and Barnett 1997), changes of the island rule to the through¯ow of SST, heat content, and currents is transport in response to the closure are a measure for much greater and indicates a sensitivity through feed- the sign of this feedback. The average wind-induced backs with atmospheric forcing, most notably wind island rule transport of the reference run is 18 Sv stress. Speci®cally, along the equator, warm SST signals (Schneider and Barnett 1997) and is slightly, but sig- centered over 170ЊW are associated with eastward wind ni®cantly, smaller than the wind forcing of 19.5 Sv of stress to its east (Fig. 14) and a convergence of the the NOTF experiment (Fig. 15). This suggests that a meridional wind stress on the equator. Upwelling in the decrease of the through¯ow yields an increase of its western Paci®c weakens, and the east±west thermocline wind forcing and implies that the feedback of through- slope along the equator relaxes (Fig. 14). This together ¯ow and island rule winds is negative and stabilizing. with the changes of off-equatorial curl in the intertrop- This in turn suggests that the interplay of the through- ical convergence zones leads to the changes of the trans- ¯ow transport and winds cannot lead to self-sustained port and vertical structure of the currents on the equator oscillations. This conclusion is not applicable to the and further enhances the surface temperature changes. baroclinic component of the through¯ow since the latter In the Indian Ocean, the westward wind stress is in- is governed by a different set of dynamics (Schneider creased and leads to a shallowing of the thermocline in and Barnett 1997). the west and a relative deepening in the east (Fig. 14). Thus, closure of the through¯ow results in the devel- c. Warm pool temperature opment of a weak cold tongue in the eastern Indian Ocean. This feedback of the coupled system in the equa- Closure of the through¯ow in a coupled general cir- torial Paci®c and Indian Ocean to the perturbation of culation model perturbs the oceanic heat budget in the the through¯ow is reminiscent of the hypotheses that area of highest SSTÐthe Indo-Paci®c warm poolÐand the delineation of warm pool and cold tongue in the results in a tendency to warm the western Paci®c and equatorial Paci®c results from coupled interactions cool the eastern Indian Ocean. The response of the sur- (Dijkstra and Neelin 1995). face ¯uxes to this perturbation reveals feedbacks of the coupled ocean±atmosphere system of ECHO that control, and potentially limit, highest SST. Results pre- b. Through¯ow forcing sented are for 5-yr averages and therefore should not The response of the coupled model gives guidance show the generation of highest SST by clear skies (Wal- on the feedback between the through¯ow and the at- iser and Graham 1993) that have lifetimes of at most a mospheric circulation that forces it. Since forcing of the few months (Waliser 1996).

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC APRIL 1998 SCHNEIDER 685

This dynamic of ECHO supports the analysis and hypotheses of Zhang and Grossman (1996) that evaporation under the centers of deep convection is reduced, and is consistent with suggestions that highest SST are limited by reduction of incident radiation due to clouds (Graham and Barnett 1987; Ramanathan and Collins 1991). There is, however, a large scatter around the median values. West of 150ЊE this scatter is even larger, and for the highest SST median tendencies are nil or reversed: shortwave radiation is slightly enhanced and partially balanced by decreases of longwave radiation (Fig. 16b). The areas that correspond to the high SST are found in the enclosed Indonesian waters and South China Sea, rather than the open ocean east of 150ЊE and suggest that the feedback of solar radiation and SST is not a universal phenomenon but may be restricted to the open ocean.

7. Discussion and conclusions The role of the Indonesian through¯ow in the global climate is investigated by comparison of coupled ocean± atmosphere model simulations with open and closed In- donesian seas. While the experiments were cast and described in terms of changes due to the closure of the through¯ow, the reverse interpretation holds equally, and the discussion will be presented in terms of effects of inclusion of the through¯ow in a coupled model. The Indonesian Through¯ow and its associated circulation around Australia and New Guinea deepen the thermocline in the Indian Ocean to the west of Australia; as expected from Sverdrup relation, and mass compen- sation shallow the thermocline everywhere else. These changes of the thermocline affect the surface in regions of upwelling or convection and result in increases of SST in the Indian Ocean and decreases of SST in the tropical Paci®c (HG). This oceanic response was found by an ocean-only investigation of Hirst and Godfrey (1993) and sets the stage for the response of the coupled model. The warming in the eastern Indian Ocean shifts the warm pool to the west and drags with it centers of deep convection with their signatures of increased precipitation, reduced shortwave radiation, and convergent FIG. 13. Bimonthly averages of wind stress in 10Ϫ2 NmϪ2 in the winds. Thus, the position of the warm pool is in part Bay of Bengal (a) and over Indonesian waters (b) for the reference controlled by the Indonesian Through¯ow. The shift of run, indicated by variance ellipses, and for the NOTF run (arrows). the atmospheric centers of deep convection and diabatic Bimonthly periods are indicated by symbols listed in legend. Variance ellipses are based on all 5-yr averages of the reference run, and the heating cause a global readjustment of the atmospheric major axes have lengths of one standard deviation. pressure of wavenumber 2 in accordance to the 90Њ long distance between cooling in the Paci®c and warming in the Indian Ocean. This global readjustment is associated East of 150ЊE changes of heat ¯ux normalized by with change in the surface wind stress in the Tropics. changes in SST (Fig. 16a) indicate that shortwave ra- Midlatitudes are affected by atmospheric teleconnec- diation has the largest cooling tendency in the high SST tions excited by the shift of deep convection and show range. It is partially balanced by heating tendencies of signature of the PNA and South Paci®c patterns known longwave radiation, due to changes in opacity of the from interannual anomalies. The sensitivity to the atmosphere, and, more importantly, by a reduction of through¯ow of the tropical wind stress increases hori- latent heat loss under the centers of deep convection. zontal gradients of heat content in the Paci®c and de-

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC 686 JOURNAL OF CLIMATE VOLUME 11

FIG. 14. Changes in the Indo-Paci®c of SST in K (solid line), mean temperature of the upper 250 m in K (dashed line), and zonal wind stress in 10Ϫ2 NmϪ2 (dotted line) due to closure of the through¯ow and averaged from 3ЊNto3ЊS. creases these gradients in the northern Indian Ocean and (1989) ®nds that rainfall over western Australia has a to a lesser extent in the Atlantic. An open through¯ow positive correlation to the SST difference between In- leads thereby to an acceleration of the tropical currents donesian Seas and the central Indian Ocean. Opening including the Equatorial Undercurrent in the Paci®c, of the through¯ow results in cooling of the Paci®c, while these currents are reduced in magnitude in the warming of the Indian Ocean, and an increase of rainfall Indian Ocean. In fact, the eastward winds, warm conover western Australia. This contradiction is either due ditions, deep thermocline, and weak vertical shear in to differences in SST patterns or might be an artifact the eastern equatorial Indian Ocean result in part from of the model. Closure of the through¯ow also affects control of the winds by the Indonesian through¯ow. the amplitude of the annual cycles, most notably mid- Over Australia, simulated temperature and precipi- latitude quasi stationary pressure cells and, as expected tation are sensitive to the Indonesian through¯ow. Qual- from the warming of the Paci®c, the strength of the itatively, this is expected from interannual anomalies of Asian monsoon. Australian rainfall (Nicholls 1989). However, Nicholls In summary, the Indonesian Through¯ow directly affects the position of the warm pool and the centers of deep convection due to its in¯uence on the Sverdrup transport and thermocline depth. By its control on the position of the centers of deep convection, the through- ¯ow exerts control on the tropical and midlatitude atmospheric pressure, and thereby on the surface wind stress. This latter forcing feeds back on the ocean and alters currents and the thermal structure in the entire Tropics. Thus the Indonesian Through¯ow affects the climate of the entire Tropics and parts of the midlatitudes. In addition to the determination of the sensitivity of the climate to the through¯ow, closure of the through- ¯ow perturbs the oceanic heat budget in the warm pool. The adjustment of surface heat ¯uxes reveals the sensitivity of the coupled system in this high SST range. In the Paci®c shortwave radiation is reduced in response to an increase of SST and thus can limit growth in SST. FIG. 15. Five-year averages of the island rule transport in Sv due Longwave radiation and latent heat ¯uxes display a pos- to wind stress (Godfrey 1989) for the reference run (solid thin) and itive feedback with surface temperature and can there- for the experiment with closed Indonesian Sea (short thick line, ex- fore not limit sea surface temperatures in the warm pool tended by dashed line). Five-year averages of the reference run are represented by their center point, and NOTF results are shown as a of the western Paci®c. However, in the very warm wa- solid line indicating years 95±99. ters of the South China Sea and Indonesian waters no

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC APRIL 1998 SCHNEIDER 687

FIG. 16. Change of surface ¯uxes in W mϪ2 in the tropical Indo-Paci®c normalized by changes of surface temperature from differencing 5-yr averages of the run without through¯ow and of the reference run. Points with signi®cant changes of SST entered the calculation, and results have been bin averaged with respect to average SST of the NOTF and REF simulation. Shown are results for sensible, latent, longwave, and shortwave heat ¯uxes, bin medians are shown by thick lines, and 0.05 and 0.95 levels of the probability density function are depicted by thin lines. Positive values correspond to a transfer of heat into the ocean. Panel (a) shows results for the tropical Paci®c and (b) for the marginal seas of the western Paci®c and for the Indian Ocean. such relationship is found. This suggests that the control ci®c is rapid such that integration time is suf®cient for of the highest SST is not universal and is, within the equilibrium to be established. However, because of the limits of the coupled model, restricted to the open ocean slow interplay of SST, thermocline depth and the wind away from boundaries. stress, the trend of 5-yr averages indicate that the west- The experiment with the coupled model also deter- ern Paci®c heat content is decreasing at the end of the mines the sign of the feedback between the through¯ow coupled integration, and a longer integration of the mod- and its forcing of by wind, as described by the Island el will display in the Paci®c a further relaxation of the Rule (Godfrey 1989). Closure of the through¯ow in- thermocline and the concomitant weakening of the trade creases wind forcing and suggests that the feedback of winds. In midlatitudes, most notable in the Agulhas and the through¯ow transport and winds is negative. Results Kuroshio region, the oceanic adjustment time is of the also only apply to the barotropic transport of the order of 20 yr (Hirst and Godfrey 1994) even though through¯ow, which are governed by the island rule. The all patterns of change in the ocean are established after role of feedbacks of the baroclinic component of the 5 yr. The stability analysis of the through¯ow and its ¯ow and the wind cannot be addressed in this experi- wind forcing is based on the reference run and on the ment. run with closed through¯owÐthat is, two points only. Results have to be viewed within the limitations of It is possible that smaller variations of the through¯ow the model and the experiment. The sensitivity of the transport might lead to a different sign of the feedback. coupled system to the Indonesian through¯ow might be The coupled model has a few idiosyncrasies that have exaggerated by the strong through¯ow in the reference an unknown effect on coupled processes. It has strong run: its transport of 13.5 Sv is larger than most obser- numerical diffusion in the Tropics, simulates a cold vational estimates of 5±10 Sv. The integration time of tongue extending too far to the west, and has too few 10 yr allows only partial adjustment of midlatitude stratus clouds. These issues have been addressed in a oceans and of the tropical thermocline. Hirst and God- new version of this model (Frey et al. 1997) and suggest frey (1994) show that the adjustment of surface tem- that rather than extending the current simulation, it perature in the eastern Indian Ocean and equatorial Pa- might be fruitful to repeat this experiment with the new-

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC 688 JOURNAL OF CLIMATE VOLUME 11 er version of the coupled model. A remaining problem (NERSC), and support by the Environmental Science with all global coupled models is their coarse resolution Division of U.S. Department of Energy (DOE DE- that does not resolve the complicated topography and FG03-ER61198) as part of the Atmospheric Radiation pathways in the Indonesian Seas. This issue will need Measurement Program, and by the National Science to be addressed by comparison of ocean simulations Foundation (NSF ATM-93-14495) are gratefully ac- with different resolutions. knowledged. Closing the Indonesian Through¯ow in a coupled model is a thought experiment to determine its role in the climate system. Use of results presented here in the REFERENCES interpretation of variations of the coupled system has to consider that the through¯ow was forced to vanish Barnett, T. P., M. Latif, E. Kirk, and E. Roeckner, 1991: On ENSO physics. J. Climate, 4, 487±515. by blocking Indonesian waters. The oceanic and at- Conover, W. J., 1971: Practical Nonparametric Statistics. John Wiley mospheric response to this perturbation includes a sea and Sons, 462 pp. level difference between the western Paci®c and Indian Dijkstra, H. A., and J. D. Neelin, 1995: Ocean±atmosphere interaction Oceans, of 30 cm and is not the same as climate vari- and the tropical climatology. Part II: Why the Paci®c cold tongue ations that reduce or bring to zero the through¯ow trans- is in the east. J. Climate, 8, 1343±1359. DKRZ, 1992: The ECHAM3 atmospheric general circulation model. port. Such variations are associated with small or zero Deutsches Klimarechenzentrum Rep. 6, 188 pp. [Available from sea level difference across the Indonesian seas. Results DKRZ, Bundesstrasse 55, D-20146 Hamburg, Germany.] presented might suggest that the through¯ow can be Frey, H., M. Latif, and T. Stockdale, 1997: The coupled model ECHO- used to control the westward extent of the equatorial 2. Part I: The tropical Paci®c. Mon. Wea. Rev., 125, 703±720. cold tongue in a coupled model. Even though the ten- Godfrey, J. S., 1989: A Sverdrup model of the depth-integrated ¯ow for the World Ocean allowing for island circulations. Geophys. dency of the model to produce a split intertropical con- Astrophys. Fluid Dyn., 45, 89±112. vergence zone is reduced, atmospheric pressure is de- , 1996: The effect of the Indonesian through¯ow on ocean cir- graded (Fig. 7) and argues against usage of the through- culation and heat exchange with the atmosphere: A review. J. ¯ow to tune a coupled model. Geophys. Res., 101, 12 217±12 237. The role of the Indonesian Through¯ow in interan- , and T. J. Golding, 1981: The Sverdrup relationship in the Indian Ocean, and the effect of Paci®c±Indian Ocean through¯ow on nual variability cannot be addressed with this experi- Indian Ocean circulation and on the east Australian current. J. ment, due to its short duration and since interannual Phys. Oceanogr., 11, 771±779. anomalies are affected both by changes of the mean state , and A. J. Weaver, 1991: Is the Leeuwin current driven by Paci®c and by the lack of an oceanic connection through the heating and winds? Progress in Oceanography, Vol. 27, Per- gamon Press, 225±272. Indonesian Archipelago. An investigation of the latter Gordon, A. L., 1986: Interocean exchange of thermocline water. J. process requires experiments where the plumbing in the Geophys. Res., 91, 5037±5046. Indonesian seas allows a mean through¯ow but prohibits , R. F. Weiss, W. M. Smethie, and M. J. Warner, 1992: Ther- passage on interannual signals. Model studies that se- mocline and intermediate water communication between the lectively turn off coupling in the atmosphere in the In- South Atlantic and Indian Oceans. J. Geophys. Res., 97, 7223± 7240. dian Ocean (Nagai et al. 1995; Latif and Barnett 1995) Graham, N. E., and T. P. Barnett, 1987: Sea surface temperature, suggest that interannual anomalies in the Indian Ocean surface wind divergence, and convection over tropical oceans. do not play an active role in the development of El Science, 238, 657±659. NinÄo, and thus interannual anomalies of the through¯ow Hirst, A. C., and J. S. Godfrey, 1993: The role of the Indonesian should be of little consequence. Experiments with a re- Through¯ow in a global GCM. J. Phys. Oceanogr., 23, 1057± 1086. duced gravity model (Verschell et al. 1995) showed in- , and , 1994: The response to a sudden change in Indonesian terannual anomalies of upper-layer thickness in the Pa- Through¯ow in a global ocean GCM. J. Phys. Oceanogr., 24, ci®c are insensitive to closure of the through¯ow, while 1895±1910. anomalies in the Indian Ocean depend critically on open Horel, J. D., and J. M. Wallace, 1981: Planetary scale atmospheric Indonesian passages. This argues also against an active phenomena associated with the Southern Oscillation. Mon. Wea. Rev., 109, 813±829. role of the through¯ow in the development of interan- Hughes, T. M. C., A. J. Weaver, and J. S. Godfrey, 1992: Thermo- nual anomalies in the Paci®c. However, these conclu- haline forcing of the Indian Ocean by the Paci®c Ocean. Deep- sions need to be veri®ed by further and exciting studies Sea Res., 39, 965±995. with fully coupled ocean±atmosphere models. Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc., 77, 437±471. Karoly, D. J., 1989: Southern Hemisphere circulation features as- Acknowledgments. This work has bene®ted from dis- sociated with El NinÄo±Southern Oscillation events. J. Climate, cussions with Drs. T. Barnett, N. Barth, A. Gershunov, 2, 1239±1252. J. Sprintall, and T. Stockdale, and would not have been Knox, R. A., 1976: On a long series of measurements of Indian Ocean possible without the enthusiasm and computational ex- equatorial currents near Addu Atoll. Deep-Sea Res., 23, 211± pertise of J. Ritchie. The coupled model has been gen- 221. , and D. L. T. Anderson, 1985: Recent advances in the study of erously made available by the Max-Planck-Institut fuÈr the low-latitude ocean circulation. Progress in Oceanography, Meteorologie, Hamburg. Computations were performed Vol. 14, Pergamon Press, 259±317. at the National Energy Research Super computer Center Kundu, P. K., and J. P. McCreary Jr., 1986: On the dynamics of the

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC APRIL 1998 SCHNEIDER 689

through¯ow from the Paci®c into the Indian Ocean. J. Phys. Through¯ow in the ®ne resolution antarctic model. J. Phys. Oceanogr., 16, 2191±2198. Oceanogr., 27, 445±455. Latif, M., and T. Barnett, 1994: Causes of decadal variability over Roeckner, E., and Coauthors, 1992: Simulation of the present-day the North Paci®c and North America. Science, 266, 634±637. climate with the ECHAM model: Impact of model physics and , and , 1995: Interactions of the tropical oceans. J. Climate, resolution. Max-Planck-Institut fuÈr Meteorologie, Rep. 93, 171 8, 952±964. pp. [Available from MPI, Bundesstrasse 55, D-20146 Hamburg, , T. Stockdale, J. Wolff, G. Burgers, E. Maier-Reimer, M. M. Germany.] Junge, K. Arpe, and L. Bengtsson, 1994: Climatology and vari- Schneider, N., and T. Barnett, 1995: The competition of freshwater and radiation in forcing the ocean during El NinÄo. J. Climate, ability in the ECHO coupled GCM. Tellus, 46A, 351±366. 8, 980±992. Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA , and , 1997: Indonesian Through¯ow in a coupled general Professional Paper 13, 173 pp. [Available from U.S. Government circulation model. J. Geophys. Res., 102, 12 341±12 358. Printing Of®ce, Washington, DC 20402.] , , M. Latif, and T. Stockdale, 1996: Warm pool physics in Nagai, T., Y. Kitamura, M. Endoh, and T. Tokioka, 1995: Coupled a coupled GCM. J. Climate, 9, 219±239. ocean atmosphere model simulations of El NinÄo/Southern Os- Shriver, J. F., and H. E. Hurlburt, 1997: The contribution of the global cillation with and without an active Indian Ocean. J. Climate, thermohaline circulation to the Paci®c to Indian Ocean Through- 8, 3±14. ¯ow via Indonesia. J. Geophys. Res., 102, 5491±5511. Nicholls, N., 1989: Sea surface temperature and Australian winter Verschell, M. A., J. C. Kindle, and J. J. O'Brien, 1995: Effects of rainfall. J. Climate, 2, 965±973. Indo-Paci®c through¯ow on the upper tropical Paci®c and Indian , 1995: All-India summer monsoon rainfall and sea surface tem- Ocean. J. Geophys. Res., 100, 18 409±18 420. peratures around northern Australia and Indonesia. J. Climate, Waliser, D. E., 1996: Formation and limiting mechanism for very 8, 1463±1467. high sea surface temperature: Linking the dynamics and ther- Palmer, T. N., C. Brankovic, P. Viterbo, and M. J. Miller, 1992: Mod- modynamics. J. Climate, 9, 161±188. , and N. E. Graham, 1993: Convective cloud systems and warm- eling interannual variations of summer monsoons. J. Climate, 5, pool SST: Coupled interactions and self-regulation. J. Geophys. 399±417. Res., 98, 12 881±12 893. Piola, A., and A. Gordon, 1984: Paci®c and Indian Ocean upper layer Xu, W., T. P. Barnett, and M. Latif, 1998: Decadal variability in the salinity budget. J. Phys. Oceanogr., 14, 747±753. North Paci®c as simulated by a hybrid coupled model. J. Climate, Ramanathan, V., and W. Collins, 1991: Thermodynamic regulation 11, 297±312. of ocean warming by cirrus clouds deduced from observations Zhang, G. J., and R. L. Grossman, 1996: Surface evaporation during of the 1987 El NinÄo. Nature, 351, 27±32. the Central Equatorial Paci®c Experiment: A climate scale per- Ribbe, J., and M. Tomczak, 1997: Effect of the missing Indonesian spective. J. Climate, 10, 2522±2537.

Unauthenticated | Downloaded 09/24/21 09:29 PM UTC