1 JUNE 2008
DU ET AL.
2451
Interannual Variability of Sea Surface Temperature off Java and Sumatra in a Global GCM* YAN DU⫹
AND
TANGDONG QU
International Pacific Research Center, SOEST, University of Hawaii at Manoa, Honolulu, Hawaii
GARY MEYERS Integrated Marine Observing System, University of Tasmania, Hobart, Australia (Manuscript received 2 November 2006, in final form 23 October 2007) ABSTRACT Using results from the Simple Ocean Data Assimilation (SODA), this study assesses the mixed layer heat budget to identify the mechanisms that control the interannual variation of sea surface temperature (SST) off Java and Sumatra. The analysis indicates that during the positive Indian Ocean Dipole (IOD) years, cold SST anomalies are phase locked with the season cycle. They may exceed ⫺3°C near the coast of Sumatra and extend as far westward as 80°E along the equator. The depth of the thermocline has a prominent influence on the generation and maintenance of SST anomalies. In the normal years, cooling by upwelling– entrainment is largely counterbalanced by warming due to horizontal advection. In the cooling episode of IOD events, coastal upwelling–entrainment is enhanced, and as a result of mixed layer shoaling, the barrier layer no longer exists, so that the effect of upwelling–entrainment can easily reach the surface mixed layer. Horizontal advection spreads the cold anomaly to the interior tropical Indian Ocean. Near the coast of Java, the northern branch of an anomalous anticyclonic circulation spreads the cold anomaly to the west near the equator. Both the anomalous advection and the enhanced, wind-driven upwelling generate the cold SST anomaly of the positive IOD. At the end of the cooling episode, the enhanced surface thermal forcing overbalances the cooling effect by upwelling/entrainment, and leads to a warming in SST off Java and Sumatra.
1. Introduction The Indian Ocean Dipole (IOD) is an important factor influencing global climate (Saji et al. 1999; Webster et al. 1999; Allan et al. 2001; Yamagata et al. 2002; Saji and Yamagata 2003). As such, it has been intensively studied recently, but a quantitative understanding of the mechanisms that generate and maintain its sea sur-
* International Pacific Research Center Contribution Number 487 and School of Ocean and Earth Science and Technology Contribution Number 7218. ⫹ Current affiliation: LED, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou, China.
Corresponding author address: Dr. Yan Du, IPRC/SOEST, University of Hawaii at Manoa, 1680 East-West Road, Honolulu, HI 96822. E-mail:
[email protected] DOI: 10.1175/2007JCLI1753.1 © 2008 American Meteorological Society
face temperature (SST) anomalies is still lacking. The positive phase of IOD is characterized by a cold SST anomaly in the eastern basin and a warm SST anomaly in the western basin. The largest SST anomaly occurs in the eastern basin near the coast of Java and Sumatra and may fall below its long-term average by as much as ⫺3°C. The genesis of this anomaly is the subject of this study. Satellite-observed SST (Fig. 1) and chlorophyll concentration (Fig. 2) suggest that ocean dynamics plays a role. During the 1994 and 1997 positive IOD events, cold SST anomaly (⬍⫺1°C) extended to the west of 90°E along the equator (Fig. 1, center and right panels), while chlorophyll concentration suggests that this surface cooling signature originates from the subsurface in the coastal region off Java and Sumatra (Fig. 2b). The widespread surface cooling signature, however, does not occur during non-IOD years (Figs. 1a,d,g ; Fig. 2a). This seems to suggest that the southeastern tropical Indian Ocean (STIO) experiences a transformation from weak SST depressions off Java during normal
2452
JOURNAL OF CLIMATE
VOLUME 21
FIG. 1. Advanced Very High Resolution Radiometer (AVHRR) Oceans Pathfinder SST anomaly (°C) and Special Sensor Microwave Imager (SSM/I) 10-m wind velocity anomaly (m s⫺1). AVHRR SST is from 1985–99 with a 1/6° resolution, and SSM/I wind is from 1987–99 with a 1° resolution. Note that the SSM/I speeds are observed by the satellite instruments, whose directions are assigned by an optimal combination of these speeds, an atmospheric model’s output, and in situ data (Atlas et al. 1996).
years to strong SST depressions off Java and Sumatra during the positive IOD years, and upwelling off Java– Sumatra is associated with the anomalous cooling. The SST anomaly results in an abnormal condition both in the ocean and the atmosphere and consequently a large variation in regional and global climate (Saji et al. 1999; Yamagata et al. 2002; Saji and Yamagata 2003). Ocean dynamics during an IOD event is a key issue for climate research because it is the physical basis for predictability. The IOD anomaly persists for several months, like the El Niño–Southern Oscillation (ENSO) in the Pacific, and coupled numerical models can predict the IOD patterns as well as they predict ENSO. In a recent study, Luo et al. (2007) successfully predicted the 2006 IOD event using retrospective ensemble forecasts. But they failed to hindcast the strong 1997 IOD event. The predictions seemed to be dependent on the specification of subsurface conditions off Java and Sumatra. An understanding of the surface layer heat budget and in particular its relationship to thermocline
dynamics is required as a foundation for improving coupled model performance. During most years, the SST off Java and Sumatra decreases less than one degree in austral winter, despite the upwelling favorable wind (Fig. 1) and low net heat flux (NHF) into the surface ocean (Qu et al. 1994). The mean seasonal cycle in the mixed layer (ML) heat budget shows a delicate balance maintained between horizontal advection, upwelling/entrainment, and barrier layer (BL) formation (Du et al. 2005). The earlier study identified two distinct thermodynamic regimes in the STIO. South of Java, the warm advection by the Indonesian Throughflow (ITF) neutralizes the cold upwelling water, and consequently SST depression is suppressed. West of Sumatra, a BL is formed under the mixed layer all year-round (Qu and Meyers 2005b), impeding the cold water from entraining into the ML. Warm advection and BL formation combined with upwelling are the primary processes that control the seasonal variation of SST near Sumatra.
1 JUNE 2008
DU ET AL.
2453
FIG. 2. Natural logarithm of the Sea-Viewing Wide Field-of-View Sensor (SeaWiFs) chlorophyll concentration (mg m⫺3) in October–December for (a) 1998–2003 and (b) 1997 with a horizontal 0.25° resolution. Two boxes indicate the region west of Sumatra (2°N–5°S, 95°–105°E) and the region south of Java (12°–7°S, 105°–120°E). The line between Sumatra and west Australia indicates the IX1 XBT line. The open circle presents the northern end of IX1 XBT line. For further information about SeaWiFs, please refer to the Web site (http://oceancolor.gsfc. nasa.gov/SeaWiFS/).
Can this delicate balance be maintained on the interannual time scale? Several earlier studies have addressed this issue. Among others, Murtugudde et al. (2000) investigated the dynamic and thermodynamic processes in the Indian Ocean during the 1997/98 IOD event. They found that the cooling in the eastern basin resulted from unusually strong upwelling along the equator and Sumatra. The strong Sumatra upwelling was forced both locally by the strong alongshore winds and remotely by equatorial and coastal Kelvin waves. Yu and Rienecker (1999) attributed the local southeast wind anomaly to the impact of ENSO. The enhanced southeast monsoon then forces the cold surface water away from the coastal area. Susanto et al. (2001) related the SST cooling off Java and Sumatra to decreased heat transport by the ITF. During El Niño years, the ITF carries colder than normal surface water into the region, helping induce a stronger than normal upwelling near the coast of Java and Sumatra. The cooling mechanisms identified in the above studies are potentially all active at different times. However, the studies were based on observations or model results for a relatively short time. A careful examination of the ocean’s role in the interannual SST variability for a reasonably long period is still lacking. For example, considering the fact that IOD and El Niño concurred in 1997, one might expect that the relationship between the ITF and the SST depression off Sumatra only occurs in this special case, as identified by Susanto et al. (2001), and not when IOD develops without El Niño. How upwelling strengthens and cold SST extends westward to the west of Sumatra during an IOD event is still an open question. Long-time in situ observations and modeling simulations are needed for further investigation.
Given the limitation of existing observations, this study analyzes the reanalysis products from a global general circulation model, the Simple Ocean Data Assimilation reanalysis version 1.4.2 (SODA; Carton and Giese 2008). Recently available satellite data and in situ observations are also used to give more detailed information about the ocean surface and atmosphere forcing. The results of the analysis are presented in the following sections: in section 2, we give model description and validation. In section 3, we describe the SST depression off Java and Sumatra and its variation on the interannual time scales, and introduce an IOD index. In section 4, we investigate the interannual variations of the upwelling processes in the region. The interannual variations of mixed layer, barrier layer, wind forcing, and upper-layer circulation are also presented in this section. In section 5, we examine the role of ocean processes in the mixed layer heat budget. Results are summarized in the last section.
2. SODA products and validation Results from SODA are used in this study to provide estimates of the heat budget in the STIO. The SODA is forced by daily wind stress and heat flux from bulk formulas from the 40-yr European Centre for MediumRange Weather Forecasts Re-Analysis (ERA-40) and spans a 44-yr period from January 1958 to December 2001. Surface freshwater flux is provided by the Global Precipitation Climatology Project (GPCP) monthly satellite-gauge merged product combined with evaporation from bulk formula, for the period from 1979 to 2001 (Adler et al. 2003). The assimilated observational dataset includes the historical archive of hydrographic profiles by ship, moored hydrographic observations,
2454
JOURNAL OF CLIMATE
and remotely sensed SST (Stephens et al. 2002; Boyer et al. 2002). The assimilation is carried out every 10 days, using the Incremental Analysis Update methodology (Bloom et al. 1996). Averages of model output variables (temperature, salinity, and velocity) are saved at 5-day intervals. The K-Profile Parameterization (KPP) mixing scheme is employed for the vertical diffusion of momentum, heat, and salt (Large et al. 1994). The lateral subgrid-scale processes are treated with a biharmonic mixing scheme (Smith et al. 2000). The computational domain extends to the North and South Poles of the earth, with a horizontal resolution of 0.4°⫻0.25°. Vertical resolution varies from 10 m in the upper levels to 500 m near the bottom, with a total of 40 vertical levels and a maximum depth of 5624 m. Bottom topography is reconstructed from the 1/30° analysis with modification for certain passages. For more details about the SODA, please see Carton and Giese (2008). The model outputs used for the present analysis are those available online. Model output variables are remapped onto a uniform 0.5° ⫻ 0.5° ⫻ 40-level grid with a temporal resolution in monthly averaged form. Water properties in the model remain conserved in the remapping. The study region is between 85°–130°E and 25°S–5°N. We calculate the depth of ML (MLD) by specifying a difference in potential density from the surface salinity and a specified temperature difference, ⌬T, from the SST. Similarly, the isothermal layer (ITL) is defined by specifying ⌬T only (Kara et al. 2000). Salinity stratification can generate an ML within the ITL leaving an intermediate layer known as the barrier layer (Lukas and Lindstrom 1991; Sprintall and Tomczak 1992). In the present study, we select ⌬T ⫽ 0.8°C. Though different values can also be used, our results are insensitive to the selection of ⌬T. Linear vertical interpolation is used to estimate the MLD and ITL depth. The MLD and barrier layer thickness (BLT) from SODA were validated with in situ observations (Qu and Meyers 2005b; Meyers et al. 1995) and with the tenth-degree resolution OGCM for Earth Simulator (OFES) driven by observed atmospheric forcing without ocean data assimilation (Masumoto et al. 2004). The validation focuses on two dynamically distinct regions (Du et al. 2005): west of Sumatra (2°N–5°S, 95°– 105°E) and south of Java (12°–7°S, 105°–120°E), as shown in Fig. 2. West of Sumatra, the BLT from SODA is consistent with observations in phase, but the amplitude of the semiannual variation relative to annual variation is somewhat larger (Fig. 3a). On the other hand, the amplitude of MLD variation is somewhat smaller, compared to both observations and OFES (Fig. 3c). South of Java, SODA’s simulation of MLD is
VOLUME 21
very good, because it is a data-rich region (Qu and Meyers 2005a, their Fig. 1). SODA simulates the barrier layer thickness near Java moderately well but with a phase lag relative to OFES and observations. On the interannual time scale, the simulated ITL depth is compared with in situ observations on the northern end of the IX1 XBT line (Meyers et al. 1995; Pigot and Meyers 1999). In the selected 1° ⫻ 2° bin (8.5°–6.5°S, 105°–106°E), four profiles are averaged each month to better present the general characteristics south of Java. Defined as the depth where water temperature is 0.8°C colder than its surface value, the simulated ITL depth is highly correlated (r ⫽ 0.90) with in situ observations, with the former leading by about one month (Fig. 3e). West of Sumatra, the simulated BLT also looks reasonable on the seasonal time scale (Fig. 3a). Despite some quantitative differences, results from SODA have a generally good representation of the mixed layer dynamics and thermodynamics in the region. This merits a further analysis of mixed layer heat budget as shown in the following sections.
3. SST depressions off Java and Sumatra The eastern Indian Ocean is viewed as part of the warm pool extended from the Pacific, with SST higher than 28°C all year-round. In July–September, the prevailing southeast monsoon favors upwelling along the coast off Java and Sumatra. But, during normal years, the SST depression there is less than 1°. When positive IOD occurs, the cooling is much larger, and comparable with that in the eastern Pacific upwelling zones, with the low SST signature extending farther westward into the interior (Webster et al. 1999; Saji et al. 1999). Based on the criterion put forward by Saji et al. (1999), the Dipole Mode Index (DMI) (SSTwest ⫺ SSTeast) ⫺ mean(SSTwest ⫺ SSTeast) is calculated using SODA data (Fig. 4), where SSTwest is the SST averaged in (10°S–10°N, 50°–70°E), and SSTeast is the SST averaged in (10°S–0°, 90°–100°E). The 13-month mean filter was applied twice to remove the mean seasonal cycle. The DMI calculated from SODA shows a good agreement with Saji et al. (1999), as expected (figure not presented). Distinct peaks in DMI occur in 1961, 1967, 1972, 1994, and 1997, or so-called significant IOD years, when the averaged SST anomalies exceed 0.4°C in any three successive months during August–December (Saji et al. 1999). Large SST depressions off Sumatra (Fig. 4) contribute significantly to the positive DMI during these years. We notice that moderate IOD events also occur in 1963, 1976, and 1982. But the SST depressions are weaker, compared with those five significant IOD events. Another interesting feature is the phase difference
1 JUNE 2008
DU ET AL.
FIG. 3. Validation of (top) BLT, (middle) MLD, and (bottom) ITL depth from SODA with output from high-resolution model OFES (Masumoto et al. 2004), and observations (Qu and Meyers 2005b; Meyers et al. 1995). The ITL depth is averaged at the northern end of the IX1 XBT line (8.5°–6.5°S, 105°–106°E) and its position is shown in Fig. 2b. (e) The common period spans from May 1983 to December 2001. A 13-month running average has been applied twice to remove seasonal cycle. The mean ITL depth is 43 m from SODA and 50 m from the observations.
FIG. 4. SODA SST anomalies averaged in the region west of Sumatra and south of Java, superposed with the DMI. Here, the 13-month running average has been applied twice and the linear trend has been removed.
2455
2456
JOURNAL OF CLIMATE
VOLUME 21
FIG. 5. Volume transport anomalies of ITF along IX1 XBT line (Sv), superposed with SOI. The climatological annual mean, 15.0 Sv, has been removed before plotting. The direction toward the Indian Ocean is defined as positive. The position of the IX1 XBT line is shown in Fig. 2b. The five significant IOD events are shaded.
between the DMI and SST depression off Sumatra, with the latter leading by several months in some particular IOD events, such as 1967 and 1976. The persistent warming in the western Indian Ocean could be responsible for this phase difference. For example, as the cooling signature off Java and Sumatra turns weaker in late fall or early winter of 1976, the temperature difference between the two sides of the basin still remains large, presumably due to the westward propagation of downwelling Rossby waves that suppress the thermocline dome in the western Indian Ocean (Xie et al. 2002; Yamagata et al. 2004). Figure 4 shows the interannual variation of SST off Java and Sumatra. On the seasonal time scales, earlier studies have shown that these two regions are dominated by different dynamics and thermodynamics (Du et al. 2005). On the interannual time scale, we see that SST has a larger anomaly in the region west of Sumatra than in the region south of Java. During the three strongest IOD events, 1961, 1994, and 1997, negative SST anomaly exceeds ⫺0.5°C in the region west of Sumatra, while it reaches only ⫺0.25°C in the region south of Java. In 1963, 1967, 1972, and 1976, SST anomalies in these two regions are of equal importance. In 1982, negative SST anomalies can be seen only in the region south of Java.
4. Ocean dynamics of the region a. Remote forcing conveyed by ITF The STIO is a unique tropical ocean, mostly because it has a direct impact from the Pacific through the Indonesian Throughflow (Gordon 1986; Wyrtki 1987). The ITF has been estimated to be about 10–12 Sv (1 Sv ⬅ 106 m3 s⫺1) (Godfrey 1996; Gordon 2001), and it transports about 0.4–1.2 PW of heat from the Pacific to the Indian Ocean (Vranes et al. 2002; England and Huang 2005). The warm upper layers in the STIO can be as-
cribed in a large part to the heat transport of the ITF and the deep thermocline transmitted from the western Pacific Ocean. For the mean seasonal cycle at least, it has been shown that the oceanic heat transport by the ITF is a key process balancing the cooling by upwelling in the region south of Java (Qu et al. 1994; Du et al. 2005). But, the ITF has only limited or no influence in the region west of Sumatra (Du et al. 2005). As discussed in the previous section, the large SST depression off Sumatra is a very important component of IOD. The simulated ITF by SODA is consistent with earlier observations (Meyers 1996; England and Huang 2005), with a long-term mean value of 15.0 Sv. As shown in Fig. 5, on the interannual time scale, the ITF variability is closely related to the Pacific ENSO. The correlation coefficient between the ITF transport and Southern Oscillation index (SOI) reaches 0.76 at the 95% confidence level. The ITF turns weaker in 1965, 1982–83, 1986, 1992–94, and 1997, when El Niño occurs. No significant correlation is found between the ITF transport and IOD. During three significant IOD events, 1961, 1994, and 1997, the ITF transport decreases only in 1997, when it occurs with a strong El Niño event. During three moderate IOD events, 1967, 1972, and 1976, the ITF transport is mostly normal or slightly increases. During the weaker IOD event of 1982, the ITF transport decreases. The somewhat independent behavior of ITF relative to IOD has been attributed to winds over the Indian Ocean (e.g., Wijffels and Meyers 2004). Those wind anomalies have a strong impact on oceanic conditions off Sumatra (e.g., Qu et al. 2008). SST anomaly has a strong signature in the region off Sumatra on the interannual time scales (Fig. 4). Can the ITF have a direct influence on the SST there? For a relatively short period of time from 1980 to 2000, the SST anomaly off Sumatra appears to have good correspondence with the ITF transport, consistent with what
1 JUNE 2008
DU ET AL.
2457
FIG. 6. Heat flux anomalies averaged in the region west of Sumatra and south of Java: (a) NHF and LHF and (b) shortwave flux (SWF) and DeltQ. The data are from NCEP spanning from 1958 to 2001, and from OAFlux from WHOI spanning from 1981 to 2002. For surface heat flux, positive values indicate downward fluxes.
has been found by Susanto et al. (2001). But, this does not seem to be true before the mid-1970s (Fig. 5). The cause of this change is not understood. It is likely that the SST anomaly off Sumatra is more related to the horizontal advection in the surface layer (Du et al. 2005) than to the total ITF transport (Fig. 5). The circulation in the STIO shows that the westward South Equatorial Current is strongest near 10°–11°S. Most of the surface ITF water flows westward around the western end of Java and has less influence farther north along the coast of Sumatra (Song et al. 2004; Qu and Meyers 2005a).
b. Local upwelling and horizontal advection Following the analysis by Qu (2003), we analyze the mixed layer heat budget based on the 44-yr SODA outputs. Since SODA provides no surface heat flux product, we cannot close the mixed layer heat budget. To gain a qualitative view of how surface heat flux works in generating SST anomalies, here we include the heat flux data from the National Centers for Environmental Prediction [NCEP; data provided by the National Oceanic and Atmospheric Administration/Earth System Research Laboratory (NOAA/ESRL) Physical Sciences Division, Boulder, Colorado, at http://www.cdc. noaa.gov/] and objectively analyzed air-sea heat flux data (OAflux) from the Woods Hole Oceanographic
Institute (WHOI; Yu and Weller 2007) in the region west of Sumatra (Fig. 6). The NCEP data span from 1958 to 2001 and the OAflux data span from 1981 to 2002. Qualitatively, the two datasets show similar results during their common period, except for a larger variability in OAflux. In both cases, latent heat flux (LHF) contributes most to the net heat flux variability. In 1961, 1994, and 1997, for example, when SST depression is enhanced (Fig. 4), the anomalies of the atmosphere and ocean specific humidity difference (DeltQ, specific humidity at 2 m above sea level minus that at the sea surface) are positive and approach its maximum strength (Fig. 6b). This significantly reduces the releasing of heat through evaporation and warms the ocean, despite the enhanced southeasterly wind. In the mean time, the incoming shortwave radiation increases as a result of decreased cloudiness (Fig. 6b). Thus, the mixed layer, rendered relatively cold by an upwelling–entrainment, also absorbs more radiation because of decreased cloudiness. The combined effect of the two processes results in positive net surface heat flux anomaly into the ocean (Fig. 6a). This surface heat flux anomaly counterbalances the cooling by ocean dynamics during the episode. The result is consistent with a previous study by Tokinaga and Tanimoto (2004). Figure 7 shows the horizontal advection and vertical entrainment in the two regions, west of Sumatra and
2458
JOURNAL OF CLIMATE
VOLUME 21
FIG. 7. SODA mixed layer heat budget (°C month⫺1) for the region (top) west of Sumatra and (bottom) south of Java: (left) Ent ⫹ Adv and (right) Ent and Adv. Ent stands for vertical entrainment and Adv for horizontal advection.
south of Java. The sum of advection and entrainment is on the lhs and the components are on the rhs, which are viewed as the role of ocean dynamics in the mixed layer heat budget. During the period from 1963 to 1993, horizontal advection neutralizes most of the cooling entrainment, and as a consequence, ocean dynamics appears to play no significant role in the mixed layer heat budget, which is supported by two previous studies on the seasonal time cycle (Qu et al. 1994; Du et al. 2005). But, this does not seem to be true when IOD occurs. In 1961 and 1997, for example, the horizontal advection becomes a significant cooling process, and its combination with vertical entrainment results in a large SST depression off Sumatra. South of Java, ocean dynamics helps to cool SST during all three significant IOD events, 1961, 1994, and 1997, as it does west of Sumatra (Fig. 7), though its amplitude is somewhat smaller. Using the NCEP heat flux and SODA MLD, we calculate the surface thermal forcing and compare it with model output (figure not shown). Although the mixed layer heat budget cannot be closed, the comparison does show that the surface thermal forcing plays a role in counterbalancing the cooling by horizontal advection and vertical entrainment during IOD years. The surface thermal forcing becomes dominant and changes the SST tendency after November/December in most cases.
1961, 1994, and 1997 (Fig. 8). When these strong IOD events occur, the MLD gets shallower by more than 5 m, and the BLT gets thinner by more than 3 m. The shoaling ML makes SST more susceptive to the entrainment from below, while the thinning BLT reduces the barrier between the mixed layer and the thermocline, allowing the subsurface water to have a direct influence on the SST. A similar phenomenon is seen in the MLD but not in the BLT south of Java, where the BLT remains almost unchanged throughout the events. SODA underestimated the amplitude of BLT seasonal variation south of Java (Fig. 3b); consequently, the BLT interannual variation presented in Fig. 8 (upper panel) might have been underestimated. The BLT also decreases in 1963, 1982, and 1991 west of Sumatra. Using the Climate Prediction Center
c. Mixed layer and barrier layer The MLD and BLT are important factors influencing the SST. West of Sumatra, there is a thick BLT impeding the cold thermocline water from entering the mixed layer and this has been interpreted as evidence why upwelling does not have a significant impact on the SST during normal years (Du et al. 2005). Both MLD and BLT shoal during IOD years, especially in the events of
FIG. 8. (top) MLD and (bottom) BLT anomalies (m) off Java and Sumatra from SODA.
1 JUNE 2008
DU ET AL.
2459
FIG. 9. BLTs off Sumatra from SODA and rain-rate anomalies (mm month⫺1) from the CPC CMAP superposed with the normalized SOI.
(CPC) Merged Analysis of Precipitation (CMAP; data provided by the NOAA/ESRL Physical Sciences Division), which spans from 1979 to 2004 (Fig. 9), we find that the variation of BLT is closely related to rainfall. The correlation coefficient between the two reaches 0.60 at the 95% confidence level. Meanwhile, no significant correlation is seen between the MLD and rainfall (figure not shown), indicating that other important processes are involved in the formation of MLD. Upwelling is an important process influencing the MLD and BLT. In the region south of Java and west of Sumatra, upwelling driven by coastal wind tends to uplift the thermocline and thus reduce the MLD and BLT during most IOD events (Fig. 8). To further investigate the spatial distribution of BLT and vertical entrainment, we conduct an empirical orthogonal function
(EOF) analysis (Fig. 10). The first EOF mode of vertical entrainment shows a strong signature near the coast of Java and Sumatra, with all peaks occurring during IOD years 1961, 1963, 1967, 1972, 1977, 1982, 1994, and 1997. Vertical entrainment is narrowly confined in the coastal region, reflecting the direct influence of coastal upwelling forced by the southeast monsoon. The first mode of BLT shows negative anomalies west of Sumatra in 1961, 1963, 1967, 1972, 1977, 1982, 1994, and 1997, corresponding well with the enhanced vertical entrainment. Their correlation reaches 0.67 at the 95% confidence level.
d. Wind stress and current Horizontal circulation is an important process redistributing cold upwelling water from the coastal region,
FIG. 10. The first EOF mode of vertical entrainment : (a) Ent and (b) BLT spatial mode, and (c) temporal coefficient.
2460
JOURNAL OF CLIMATE
VOLUME 21
FIG. 11. Mixed layer zonal current (U; cm s⫺1) anomalies and ERA-40 zonal wind stress (WSU; Pa) anomalies in the region (top) west of Sumatra and (bottom) south of Java.
and has large variations during IOD events (Fig. 11). The enhanced southeast monsoon is mostly responsible for the westward current anomaly. West of Sumatra, surface current anomaly is westward in the years 1961, 1976, 1982, 1994, and 1997 (Fig. 11a), when the enhanced southeast monsoon prevails. South of Java, westward current anomaly is most evident in the years 1972, 1994, and 1997, corresponding with a large zonal wind anomaly (Fig. 11b). The anomalous surface current advects cold water away from the coastal region and spreads it westward toward the interior ocean. During the strongest IOD events, 1961, 1994, and 1997, horizontal advection is of comparable strength with the vertical entrainment west of Sumatra (Fig. 7). Westward current anomalies also develop south of Java in 1964, 1967, 1969, 1975, and 1982, and west of Sumatra in 1967 and 1971, despite the eastward wind anomalies (Fig. 11). These westward current anomalies may be attributed to remote processes. Among others, the Wyrtki Jets (Wyrtki 1973) originating from the central equatorial Indian Ocean are of particular importance. The current near the Sumatra coast can also be influenced by the circulation in the Bay of Bengal. This detail needs to be investigated further by research.
5. Heat budget analysis The large SST depression observed in the STIO is always phase locked with the seasonal upwelling in the
region off Java and Sumatra. During an IOD event, with the northward movement of enhanced southeasterly wind anomalies (Yu and Rienecker 1999), the active upwelling center moves from the region south of Java to the region west of Sumatra. As a consequence, the region west of Sumatra encounters a strong upwelling, which further has a notable impact on the local mixed layer heat budget.
a. Area-averaged heat budget Five significant IOD events (1961/62, 1967/68, 1972/ 73, 1994/95, and 1997/98) are combined to produce a composite IOD event. Figures 12a,b show the seasonal cycle of the mixed layer heat budget of this composite event averaged in the two areas. South of Java, the SST decreases from April to August. Heat advection by the ITF has almost no influence on the SST before October and only mildly warms it after that season. A strong vertical entrainment occurs in July, and it continues until next January, with its maximum strength in November. The vertical entrainment neutralizes the weak horizontal warming advection and is primarily responsible for the negative temperature tendency from April to August. Surface heat flux intensifies from September through December, which overbalances the effect of vertical entrainment and leads to a warming in the ML. West of Sumatra, the temperature tendency reaches its minimum in August, about two months after it does south of Java. Cold entrainment spans from April to
1 JUNE 2008
DU ET AL.
2461
FIG. 12. Mixed layer heat budget (°C month⫺1) anomalies in (top) a composite IOD event and (bottom) the 1997/98 IOD event for the region (left) south of Java and (right) west of Sumatra. Ent stands for vertical entrainment, Adv for horizontal advection, Tt for temperature tendency, and NHF for surface net heat flux.
next February. Differing from what was discussed for the seasonal cycle (Du et al. 2005), the entrainment contributes significantly to the temperature tendency on the interannual time scale. Horizontal advection is negative, and it explains more than half of the cooling from August to November. The cooling horizontal advection cannot originate remotely from the ITF; instead, it results from the intensified southeast monsoon. Heated by the incoming surface heat flux, the temperature tendency changes its sign in mid-September against the cooling of ocean dynamics. This result is consistent with earlier studies on the role of surface heat flux in the turnabout of IOD events (Murtugudde et al. 2000; Tokinaga and Tanimoto 2004). As a special case, we examine the 1997 IOD event and compare it with the composite event. South of Java, SST decreases from mid-May to September, mostly due to the cooling by vertical entrainment. Horizontal advection and surface heat flux are the primary heating processes. Different from what discussed for the composite event, surface heat flux keeps warming the surface ocean nearly all the time in 1997. Despite the cooling by vertical entrainment, the combined effect of surface heat flux and horizontal advection generates a warming in SST from October to December, which is similar to the composite event. West of Sumatra, the SST anomaly reaches its minimum in mid-October, with a phase delay of about one month relative to that
in the composite event (Figs. 12c,d). Cold entrainment starts to appear in mid-May, and keeps cooling the region until next February. Its maximum occurs in December. Horizontal advection cools the region mostly from July to next January, with its maximum in November. Surface net heat flux reaches its maximum in November, turning the SST tendency from negative to positive (Tokinaga and Tanimoto 2004). The shoaling of MLD provides a favorable condition for this conversion by trapping the incoming surface net heat flux in a thinner surface layer (Fig. 8).
b. Spatial distribution The 1997 IOD event is one of the strongest since 1960. Despite some quantitative differences, it represents most common features of the IOD events (Fig. 12). We select this event as an example to discuss the spatial distribution of the mixed layer heat budget. In July–October, the negative temperature tendency first appears near the coast of Java and Sumatra and then extends farther westward to the interior ocean near the equator (Fig. 13). It turns positive in November–December. Consonant with this SST change, we see cold horizontal advection in July–August south of Java and in November–December west of Sumatra (Fig. 13). The cold entrainment occurs only in the coastal area (Fig. 14a), and remains almost unchanged from September to December. During this period, the
2462
JOURNAL OF CLIMATE
VOLUME 21
FIG. 13. (top) Temperature tendency (Tt) and (bottom) horizontal advection (Adv) anomalies (°C month⫺1) averaged over two months in 1997: (left) July–August, (middle) September–October, and (right) November–December. For horizontal advection, positive values indicate warming the mixed layer.
MLD drops by more than 10 m near the coast of Java and Sumatra (Fig. 15a). With a significant decrease of rainfall (Fig. 14b), the barrier layer disappears west of Sumatra, with a negative anomaly exceeding ⫺10 m (Fig. 15b). The surface current anomaly can be accounted for by the westward extension of the cold signature. Careful examination of surface current shows a strong anticyclonic circulation anomaly in the region between Australia and Indonesia, mostly to the east of 110°E. The formation mechanisms of this circulation are not understood. Both the mean wind stress curl (Fig. 16b) and intraseasonal variations (Feng and Wijffels 2002; Yu and Potemra 2006) likely play a role in its formation. Farther northwest, we see a widespread westward sur-
face current anomaly, a unique feature of the 1997/98 event. In July–August, as the MLD and BLT shoal (Fig. 15), cold upwelling water reaches the sea surface along the coast of Java and Sumatra. The surface current then advects the cold signature westward. As a consequence, horizontal advection shows a cooling effect in a large area between 100° and 110°E (Fig. 13d). To the east, near the outlet of Lombok Strait, the warming by the ITF counterbalances the cooling by the upwelling (Fig. 13d), and thus the temperature tendency turns positive (Fig. 13a). In September–October, the MLD and BLT continue to shoal and the vertical entrainment gets stronger. Surface current advects cold signature as far westward as
FIG. 14. Anomalies of (a) vertical entrainment (isoclines ⫺4°, ⫺2°, ⫺1°, and ⫺0.5°C month⫺1 are given and area cooling effect stronger than ⫺0.5°C month⫺1 is shaded), and (b) CMAP rainfall (isoclines over 50 mm month⫺1 are given and shaded, with contour interval 50 mm month⫺1) and net heat flux (contour interval 20 W m⫺2) from OAflux, averaged from September to December 1997. In (b), positive values indicate downward fluxes.
1 JUNE 2008
DU ET AL.
2463
FIG. 15. Anomalies of (a) MLD and (b) BLT (m) in September–December 1997.
85°E (Fig. 13e). East of about 110°E, the ITF keeps warming the region (Fig. 13e), and counterbalances most of the cooling by the upwelling (Fig. 14). South of Java, the westward surface current is apparently related to the anticyclonic circulation anomaly shown in Fig. 16a. West of Sumatra, the westward surface current seems to be a direct response to the wind near the equator (Fig. 16). During this period, the horizontal advection has more influence on the SST than the vertical entrainment does. In November–December, the ML has dropped by more than 20 m and the BL completely disappears. This makes the SST more susceptive to the atmospheric forcing and oceanic processes. West of Sumatra, the vertical entrainment and horizontal advection continue to cool the SST (Figs. 13f and 14a). However, because of the heat gained from the atmosphere (Fig. 14b), mostly by enhanced latent heat flux and shortwave radiation (Fig. 6), the temperature tendency turns positive at this episode, with a maximum of about 1°C month⫺1 near the coast of Sumatra (Fig. 13c). The southeastern Indian Ocean begins its warming episode.
6. Summary In the present study, we explore the role of ocean dynamics in determining the SST interannual variabil-
ity off Java and Sumatra. The SODA products, multisources satellite measurements, and in situ observations have been used for this analysis. During the IOD years, with the intensified southeast monsoon, the SST anomaly off Java and Sumatra can reach as large as ⫺3°C, comparable with that in the eastern Pacific during ENSO years. The delicate balance between advection, upwelling–entrainment, and barrier layer formation, which dominates the seasonal heat budget off Java and Sumatra, cannot be maintained at the interannual time scales. The SST interannual variability is phase locked with the seasonal cycle. The mature phase of the positive IOD always occurs in the southeast monsoon season. As the intertropical convergence zone moves out of the region, rainfall is significantly reduced, which goes against the formation of the barrier layer. Surface southeasterly wind enhances the coastal upwelling– entrainment. The ML shoals and surface water are more responsive to the thermocline entrainment, wind forcing, and surface heat flux. Among these processes, the effect of horizontal advection is remarkable, and this is different from what has been discussed for the seasonal cycle. The SODA output reveals widespread westward surface current anomalies to the west of Java and Sumatra, which advect cold upwelling water from the coastal region to the ocean interior and cool down
FIG. 16. Anomalies of (a) mixed layer surface current (cm s⫺1), and (b) wind stress (Pa) and its curl (107 Pa m⫺1) from ERA-40 in September–December 1997.
2464
JOURNAL OF CLIMATE
the SST over a large part of the STIO. The enhanced, prolonged southeast monsoon is mostly responsible for the current anomalies both to the region south of Java and to the region west of Sumatra. The finding of the westward surface current anomalies will provide an important insight into the IOD dynamics, but its full implications for the ocean–atmosphere interaction in the region require further investigation. At the end of cooling episode, the surface net heat flux counterbalances the cooling effect by horizontal advection and vertical entrainment before the southeast monsoon reverses, then terminates the cooling episode of an IOD event, and begins a warming episode in the eastern Indian Ocean. Acknowledgments. The AVHRR Oceans Pathfinder SST data were obtained from the Physical Oceanography Distributed Active Archive Center (PO.DAAC) at NASA’s Jet Propulsion Laboratory, Pasadena, California (http://podaac.jpl.nasa.gov). The SODA, SeaWiFs, and OAflux data were obtained from APDRC in IPRC-SOEST, University of Hawaii (http://apdrc.soest. hawaii.edu). This work was supported by the National Aeronautics and Space Administration through Grant NAG5-12756, by the Japan Marine Science and Technology Center through its sponsorship of the International Pacific Research Center (IPRC), by the Chinese Academy of Sciences, and by the Integrated Marine Observing System, University of Tasmania. REFERENCES Adler, R. F., and Coauthors, 2003: The Version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 1147–1167. Allan, R., and Coauthors, 2001: Is there an Indian Ocean dipole, and is it independent of the El Niño–Southern Oscillation? CLIVAR Exchanges, 6, 18–22. Atlas, R., R. Hoffman, S. Bloom, J. Jusem, and J. Ardizzone, 1996: A multiyear global surface wind velocity data set using SSM/I wind observations. Bull. Amer. Meteor. Soc., 77, 869– 882. Bloom, S. C., L. L. Takacs, A. M. da Silva, and D. Ledvina, 1996: Data assimilation using incremental analysis updates. Mon. Wea. Rev., 124, 1256–1271. Boyer, T. P., C. Stephens, J. I. Antonov, M. E. Conkright, L. A. Locarnini, T. D. O’Brien, and H. E. Garcia, 2002: Salinity. Vol. 2, World Ocean Atlas 2001, NOAA Atlas NESDIS 50, 165 pp. Carton, J. A., and B. S. Giese, 2008: A reanalysis of ocean climate using SODA. Mon. Wea. Rev., in press. Du, Y., T. Qu, G. Meyers, Y. Masumoto, and H. Sasaki, 2005: Seasonal heat budget in the mixed layer of the southeastern tropical Indian Ocean in a high-resolution ocean general circulation model. J. Geophys. Res., 110, C04012, doi:10.1029/ 2004JC002845. England, M. H., and F. Huang, 2005: On the interannual variabil-
VOLUME 21
ity of the Indonesian Throughflow and its linkage with ENSO. J. Climate, 18, 1435–1444. Feng, M., and S. Wijffels, 2002: Intraseasonal variability in the south equatorial current of the east Indian Ocean. J. Phys. Oceanogr., 32, 265–277. Godfrey, J. S., 1996: The effect of the Indonesian throughflow on ocean circulation and heat exchange with the atmosphere: A review. J. Geophys. Res., 101, 12 217–12 238. Gordon, A. L., 1986: Interocean exchange of thermocline water. J. Geophys. Res., 91, 5037–5046. ——, 2001: Interocean exchange. Ocean Circulation and Climate: Observing and Modelling the Global Ocean, G. Siedler, J. Church, and J. Gould, Eds., Elsevier, 303–314. Kara, A. B., P. A. Rochford, and H. E. Hurlburt, 2000: An optimal definition for ocean mixed layer depth. J. Geophys. Res., 105, 16 803–16 821. Large, W. G., J. C. McWilliams, and S. C. Doney, 1994: Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization. Rev. Geophys., 32, 363–403. Lukas, R., and E. Lindstrom, 1991: The mixed layer of the western equatorial Pacific Ocean. J. Geophys. Res., 96 (Suppl.), 3343– 3358. Luo, J.-J., S. Masson, S. Behera, and T. Yamagata, 2007: Experimental forecasts of Indian Ocean Dipole using a coupled OAGCM. J. Climate, 20, 2178–2190. Masumoto, Y., and Coauthors, 2004: A fifty-year eddy-resolving simulation of the world ocean: Preliminary outcomes of OFES (OGCM for the Earth Simulator). J. Earth Simul., 1, 35–56. Meyers, G., 1996: Variation of Indonesian throughflow and the El Niño–Southern Oscillation. J. Geophys. Res., 101, 12 255– 12 263. ——, R. J. Bailey, and A. P. Worby, 1995: Geostrophic transport of Indonesian throughflow. Deep-Sea Res. I, 42, 1163–1174. Murtugudde, R., J. P. McCreary Jr., and A. J. Busalacchi, 2000: Oceanic processes associated with anomalous events in the Indian Ocean with relevance to 1997–1998. J. Geophys. Res., 105, 3295–3306. Pigot, L., and G. Meyers, 1999: Analysis of frequently repeated XBT lines in the Indian Ocean. CSIRO Marine Laboratories Rep. 238, 43 pp. Qu, T., 2003: Mixed layer heat balance in the western North Pacific. J. Geophys. Res., 108, 3242, doi:10.1029/2002JC001536. ——, and G. Meyers, 2005a: Seasonal characteristics of circulation in the southeastern tropical Indian Ocean. J. Phys. Oceanogr., 35, 255–267. ——, and ——, 2005b: Seasonal variation of barrier layer in the southeastern tropical Indian Ocean. J. Geophys. Res., 110, C11003, doi:10.1029/2004JC002816. ——, ——, J. S. Godfrey, and D. X. Hu, 1994: Ocean dynamics in the region between Australia and Indonesia and its influence on the variation of sea surface temperature in a global general circulation model. J. Geophys. Res., 99, 18 433–18 445. ——, Y. Du, J. P. McCreary Jr., G. Meyers, and T. Yamagata, 2008: Buffering effect and its related ocean dynamics in the Indonesian Throughflow region. J. Phys. Oceanogr., 38, 503– 516. Saji, N. H., and T. Yamagata, 2003: Possible impacts of Indian Ocean dipole mode events on global climate. Climate Res., 25, 151–169. ——, B. N. Goswami, P. N. Vinayachandran, and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Nature, 401, 360–363.
1 JUNE 2008
DU ET AL.
Smith, R. D., M. E. Maltrud, F. O. Bryan, and M. W. Hecht, 2000: Numerical simulation of the North Atlantic Ocean at 1/10°. J. Phys. Oceanogr., 30, 1532–1561. Song, Q., A. L. Gordon, and M. Visbeck, 2004: Spreading of the Indonesian Throughflow in the Indian Ocean. J. Phys. Oceanogr., 34, 772–792. Sprintall, J., and M. Tomczak, 1992: Evidence of the barrier layer in the surface layer of the tropics. J. Geophys. Res., 97, 7305– 7316. Stephens, C., J. I. Antonov, T. P. Boyer, M. E. Conkright, R. A. Locarnini, T. D. O’Brien, and H. E. Garcia, 2002: Temperature. Vol. 1, World Ocean Atlas 2001, NOAA Atlas NESDIS 49, 167 pp. Susanto, R. D., A. L. Gordon, and Q. Zheng, 2001: Upwelling along the coasts of Java and Sumatra and its relation to ENSO. Geophys. Res. Lett., 28, 1599–1602. Tokinaga, H., and Y. Tanimoto, 2004: Seasonal transition of SST anomalies in the tropical Indian Ocean during El Niño and Indian Ocean dipole years. J. Meteor. Soc. Japan, 82, 1007– 1018. Vranes, K., A. L. Gordon, and A. Ffield, 2002: The heat transport of the Indonesian Throughflow and implications for the Indian Ocean heat budget. Deep-Sea Res. II, 49, 1391–1410. Webster, P. J., A. M. Moore, J. P. Loschnigg, and R. R. Leben, 1999: Coupled ocean–atmosphere dynamics in the Indian Ocean during 1997–98. Nature, 401, 356–360. Wijffels, S., and G. Meyers, 2004: An intersection of oceanic
2465
waveguides: Variability in the Indonesian Throughflow region. J. Phys. Oceanogr., 34, 1232–1253. Wyrtki, K., 1973: An equatorial jet in the Indian Ocean. Science, 181, 262–264. ——, 1987: Indonesian through flow and the associated pressure gradient. J. Geophys. Res., 92, 12 941–12 946. Xie, S.-P., H. Annamalai, F. A. Schott, and J. P. McCreary, 2002: Structure and mechanisms of South Indian Ocean climate variability. J. Climate, 15, 864–878. Yamagata, T., S. K. Behera, S. A. Rao, Z. Guan, K. Ashok, and H. N. Saji, 2002: The Indian Ocean dipole: A physical entity. CLIVAR Exchanges, No. 7, International CLIVAR Project Office, Southampton, United Kingdom, 15–18. ——, ——, J. J. Luo, S. Masson, M. R. Jury, and S. A. Rao, 2004: Coupled ocean–atmosphere variability in the tropical Indian Ocean. Earth’s Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., Vol. 147, Amer. Geophys. Union, 189– 211. Yu, L. S., and M. M. Rienecker, 1999: Mechanisms for the Indian Ocean warming during the 1997–98 El Niño. Geophys. Res. Lett., 26, 735–738. ——, and R. A. Weller, 2007: Objectively analyzed air–sea heat fluxes for the global ice-free oceans (1981–2005). Bull. Amer. Meteor. Soc., 88, 527–539. Yu, Z., and J. Potemra, 2006: Generation mechanism for the intraseasonal variability in the Indo-Australian basin. J. Geophys. Res., 111, C01013, doi:10.1029/2005JC003023.
