1. Introduction
The sea surface temperature (SST) in the southwestern Indian Ocean is considered to be particularly important for several reasons. The northward migration of the intertropical convergence zone (ITCZ) toward the Indian subcontinent is delayed when the SST anomaly (SSTA) is positive there (Annamalai et al. 2005; Izumo et al. 2008). This means that the Indian summer monsoon is significantly influenced by the SSTA in this region. Also, the number of tropical cyclones in the western Indian Ocean is closely related to this SSTA (Xie et al. 2002). The SSTA in the southwestern Indian Ocean may also influence the global climate. In particular, its link with the Pacific–North American teleconnection pattern was discussed by Hurrell et al. (2004) and Annamalai et al. (2007).
In contrast to many other parts of the tropical Indian Ocean, the SSTA in the southwestern Indian Ocean is complex and cannot be simply explained by a surface heat flux anomaly (Klein et al. 1999). Although El Niño influences the basin-wide SST in the Indian Ocean, SSTA in the region is also one of the major components of the Indian Ocean dipole (IOD) structure (Saji et al. 1999; Yamagata et al. 2004). This is why several studies discussed variations of the subsurface thermocline (Reverdin and Fieux 1987; Woodberry et al. 1989; McCreary et al. 1993; Schott et al. 2009). Yokoi et al. (2008) called the unique thermocline feature in this particular region, the Seychelles Dome (SD). They have shown that its seasonal variation is dominated by a semiannual signal due to the local Ekman upwelling. On an interannual scale, past studies pointed out the importance of the westward propagation of Rossby waves (Masumoto and Meyers 1998; Xie et al. 2002; Rao et al. 2002; Rao and Behera 2005). However, Tozuka et al. (2010) have shown that regional Ekman pumping anomalies cannot be neglected even on the interannual time scale.
All of these studies improved our understanding of the subsurface variability of the SD, but no quantitative explanation of how the seasonal and interannual variations of the SD affect variations in the mixed layer temperature (MLT), and thus the SST, is presented. In this regard, Foltz et al. (2010) have analyzed the seasonal mixed layer heat balance using reanalysis, satellite, mooring, and in situ data. They showed that the net surface heat flux warms the mixed layer throughout the year with the strongest warming in austral spring and summer owing to stronger shortwave radiation and weaker latent heat loss. In addition, the horizontal advection contributes to the warming of the mixed layer in austral winter and spring, while the vertical turbulent mixing, which is estimated as a residual, results in the strongest cooling in austral winter owing to the strong surface wind and buoyancy forcing. However, their analysis with the mooring data was limited to only 2 yr, and they studied the heat balance in the mixed layer instead of the MLT balance. Also, Santoso et al. (2010) have analyzed the seasonal variation of the MLT balance of the Indian Ocean including the SD region using a coupled model. Their conclusion is mostly in accord with the results of Foltz et al. (2010), but the subsurface SD simulated in their model was shallower by as much as 20 m compared with the observations, and the cooling by the entrainment may be overestimated. The seasonal variation in the contribution from the entrainment to the MLT in Zhou et al. (2008) and Halkides and Lee (2011) is very strong, but their simulated MLDs vary seasonally by about 50 and 40 m, respectively, in contrast to about 30 m in the observations of Foltz et al. (2010). In this paper, we study the mechanism of seasonal and interannual variations in the MLT in the SD region using outputs from an ocean general circulation model (OGCM), which can simulate variations of the mixed layer depth (MLD) and thermocline depth realistically.
This paper is organized as follows. A brief description of our OGCM, together with validation of its performance, and details of the method used in this study are given in section 2. Seasonal and interannual variations of the MLT in the SD region are examined by calculating the MLT balance in sections 3 and 4, respectively. Summary and discussion are given in the final section.
2. Description of the model
a. Model description
Results from the OGCM of Tozuka et al. (2010) are used in this study. The OGCM is based on version 3.0 of the Modular Ocean Model (MOM3.0; Pacanowski and Griffies 1999), which was developed at the National Oceanic and Atmospheric Administration/Geophysical Fluid Dynamics Laboratory (NOAA/GFDL). This model covers most of the Indo-Pacific Oceans from 52°S to 30°N, with a horizontal resolution of 0.5°. The lateral eddy viscosity and diffusivity are based on the formulas given by Smagorinsky (1963), and the Gent and McWilliams (1990) parameterization scheme is adopted to incorporate the effects of mesoscale eddies. The present model does not treat the mixed layer physics explicitly; we adopt the convective adjustment scheme and the Pacanowski and Philander (1981) parameterization schemes for the vertical eddy viscosity and diffusivity. The OGCM is spun up for 20 yr by the monthly mean climatology of the wind stress from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis dataset (Kalnay et al. 1996) and by the surface heat flux, which is calculated by a bulk formula using the simulated SST and atmospheric variables obtained from the reanalysis data (Rosati and Miyakoda 1988). We note that this can act to relax the SST to the air temperature at 2-m height in the NCEP–NCAR reanalysis data, and the calculated heat flux may contain large errors if a model has poor skills. To rule out this possibility, we have compared the turbulent surface heat flux (sum of the latent and sensible heat fluxes) anomalies in the model and the reanalysis data. We find out that the correlation coefficient is 0.61, which is significant at the 99% confidence level. Thus, the surface heat fluxes used to force the OGCM are quite realistic. Also, our model results are not very sensitive to the atmospheric forcing fields, because the SD index (SDI) in our model is highly correlated (correlation coefficient = 0.73) with the SDI of the Simple Ocean Data Assimilation (SODA) product (Carton and Giese 2008), which uses European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis data for the wind forcing (Tozuka et al. 2010; Hermes and Reason 2009). The sea surface salinity (SSS) is restored to the monthly climatology with a relaxation time scale of 30 days, but the SST is not restored to the observed SST. The initial conditions are the annual mean climatology (Levitus and Boyer 1994; Levitus et al. 1994) with no motion. Then, the model is further integrated for 30 yr from 1978 to 2007 using the daily mean data from the NCEP–NCAR reanalysis dataset and the model SSS is restored to the monthly mean climatology. Outputs after 1980 are stored every 5 days and are analyzed here. More details about the model and its skill in simulating interannual variations are described by Tozuka et al. (2010).
b. Methodology
To obtain all terms except for the vertical terms in Eq. (1), each term in the model equation for temperature is accumulated at each model time step and saved at 5-day intervals for each grid cell, and these terms are integrated vertically over the diagnostically defined MLD at each horizontal grid point (Halkides and Lee 2009). The entrainment term is calculated offline based on the formulation given by Kim et al. (2006), but using 5-day snapshot data for temperature. Here, we choose to use the Kim et al. (2006) method to calculate entrainment, because it gives an exact estimate of the entrainment, whereas some conventional methods that use arbitrary temperature for entrained water do not permit closure of the mixed layer temperature balance. Then, the vertical diffusion term is calculated as a residual. To obtain the balance in the SD region, the volume-weighted average of Eq. (1) is taken. The monthly climatology of each term is then calculated from the monthly mean data from 1980 to 2006, and the monthly anomaly is computed by subtracting the monthly climatology. These anomaly data are used to construct composites in section 4.
3. Seasonal variation
Figure 1 shows the seasonal variations of the temperature profiles averaged in the SD region (10°–5°S, 50°–75°E) from both the World Ocean Atlas 1998 (WOA98; information online at http://www.nodc.noaa.gov) and the OGCM. In the present study, we tentatively define the MLD as a depth at which the temperature is 0.8°C lower than the SST, following Kara et al. (2000). We note that we have tried other definitions of the MLD (i.e., the depth at which the temperature is 0.5°C lower than the SST and the depth at which the density is 0.125 kg m−3 higher than the surface density), and obtained qualitatively the same results. This is because the contribution from salinity stratification is not important to the mixed layer in the SD region in both the observations (Mignot et al. 2007) and the current model. This is contrasted with the case in other areas of the tropical Indian Ocean, such as in the southeastern part (Halkides and Lee 2009). We find that the MLT averaged over the mixed layer shows its maximum (minimum) in April (August) (Fig. 2a) and corresponds well with the SST. Although the MLT from the OGCM is lower than that of the WOA98 from austral spring to fall, the phase of the seasonal variation is simulated well. The seasonal variation of the MLD in the same region from both WOA98 and the OGCM is shown in Fig. 2b. Again, the seasonal variation of the MLD is simulated well in the model; the mixed layer is thin from December to May at about 25 m or less, and reaches its maximum in September at about 40 m. These variations are also consistent with the mooring observations (Foltz et al. 2010). The reasonable success in resolving the mixed layer suggests that the Pacanowski–Philander parameterization for the vertical mixing works rather well in the present region. One interesting aspect of the seasonal variation is that the annual harmonic is dominant in the surface mixed layer. This is contrasted with the results that the semiannual harmonic is dominant in the subsurface dome (Fig. 2c; Yokoi et al. 2008; Hermes and Reason 2008). We need to explain this remarkable difference.
Vertical profile of the seasonal mean of temperatures in the SD region: (a) March–May (MAM), (b) June–August (JJA), (c) September–November (SON), and (d) December–February (DJF). Circles indicate the OGCM outputs and squares denote the WOA98.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
Seasonal variations of (a) MLT, (b) MLD, and (c) D20 from the OGCM outputs (thick solid line) and the WOA98 (thick dashed line) in the SD region. Also shown are those in the weak SD years (thin dashed line) and the strong SD years (dotted–dashed line).
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
Monthly MLT balance in the SD region. Here, tendency, surface heat flux (Surf. flux), horizontal advection (Hor. adv.), entrainment (Entrain.), vertical diffusion (Ver. diff.), and horizontal diffusion (Hor. diff.) terms are shown.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
(a) Climatological temperature differences between the MLT and the temperature 10 m below the bottom of the mixed layer and (b) the vertical diffusion coefficient.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
(a) Seasonal variation in the contribution of the Ekman heat convergence and its zonal and meridional components to the MLT tendency: meridional heat convergence (meri), zonal heat convergence (zonal), and total Ekman heat convergence (total) in the SD region. (b) Wind stress and (c) MLT in austral winter. Contour interval is 1°C and temperatures less than 25°C are shaded in (c).
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
Figure 6a shows the monthly climatology of the contribution from four components of the surface heat flux term. It is clear that the seasonal variation in the surface heat flux in Eq. (1) is rather large for the contribution from the shortwave radiation as compared to other contributions. However, as seen in Fig. 6b, the actual amplitude of the seasonal variation in the shortwave radiation is not large. This apparent contradiction is understood by considering the seasonal variation of the MLD. The shortwave radiation is stronger (weaker) in austral summer (winter), when the mixed layer is thin (thick). Thus, the warming of the mixed layer by the shortwave radiation is more (less) efficient in austral summer (winter). Contributions from the latent heat flux and longwave radiation are relatively small compared with that from the shortwave radiation and always tend to decrease the MLT (Fig. 6a). Although the amplitude of the seasonal variation in the latent heat loss is as large as that of the shortwave radiation (Fig. 6b), its contribution to the MLT is almost constant throughout the year (Fig. 6a). This is because the MLD is thinner (thicker) when the latent heat loss is smaller (larger) in austral summer (winter). The contribution from the sensible heat flux term is negligible throughout the year. The above results are consistent with those from the observational study of Foltz et al. (2010), and we may conclude that the surface heat flux calculation based on the bulk formula in our OGCM is realistic.
Seasonal variations of (a) four components of the surface heat flux term in the MLT balance (10−7 °C s−1), and (b) the surface heat flux (W m−2) in the SD region. Sensible heat flux (SH), latent heat flux (LH), shortwave radiation (SW), and longwave radiation (LW) are shown.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
4. Interannual variation
To discuss the interannual variation, we use the SD index as introduced by Tozuka et al. (2010). The SDI is defined by use of the depth of 20°C isotherm (D20) anomalies averaged in the SD region. Since the interannual variation is locked seasonally to austral summer (December–February), anomalous years in which the SDI in austral summer is above (below) +0.9 (−0.9) standard deviations and the thermocline is deeper (shallower) are defined as the weak (strong) event years. Based on this definition, we have selected five weak SD years (i.e., 1982/83, 1994/95, 1997/98, 2002/03, and 2006/07) and six strong SD years (i.e., 1980/81, 1983/84, 1987/88, 1992/93, 1998/99, and 2005/06). Here, year (0) is the year in which the anomaly in the SD develops (e.g., 1982 in the 1982/83 event), and year (+1) represents the following year. As shown by Tozuka et al. (2010), the SD becomes anomalously weak (strong) owing to the downwelling (upwelling) anomaly caused by both the local Ekman pumping anomaly and the intrusion of Rossby waves. It is interesting to examine how the MLT variation is linked to the interannual variation of the SD.
a. Weak SD years
The MLT becomes anomalously warm during the weak SD years as shown in Fig. 7a; it reaches its maximum of about 0.6°C during January (+1) and February (+1). Since the SST in the SD region is close to 28°C, it may have a large impact on the atmospheric convective activity even though the amplitude of the anomaly is relatively small (Gadgil et al. 1984). In addition, the anomaly of the MLD in the weak SD years increases from austral winter to summer, reaching the maximum anomaly of 11 m in December (0) (Fig. 7b). Since this occurs during the season in which the MLD decreases, the annual variation of the MLD is much reduced (Fig. 2b). To examine what causes this MLD anomaly during a shoaling phase, we have calculated the Monin–Obukhov depth. Table 1 shows the Monin–Obukhov depth anomaly and the contribution from each term in Eq. (4) for December (0), when the MLD becomes the thinnest. It is demonstrated that Eq. (4) provides us with a reasonable value of the MLD anomaly, and that the MLD anomaly in December (0) is mainly explained by the second and third terms. We note that the contribution from the wind-stirring anomaly is negligible, showing a marked difference from that in the midlatitudes (Yasuda et al. 2000).
Evolutions of (a) MLT, (b) MLD, and (c) D20 anomalies in the weak SD years. Anomalies exceeding 95% confidence level by a two-tailed t test are indicated by open circles.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
The Monin–Obukhov depth anomaly in the SD region in December (0) of the weak and strong SD years.
To understand the mechanism of the MLT anomaly (MLTA), we have used the MLT equation again. As shown in Fig. 8a, the magnitude of the tendency term is about 0.4–0.5 × 10−7 °C s−1 (0.10–0.13°C month−1) from July (0) to around January (+1). This value appears to be quite small, but the tendency of 0.4 × 10−7 °C s−1 for 6 months results in an MLTA of about 0.6°C, which reasonably explains the model MLTA. As shown in Fig. 8a, the vertical diffusion, in addition to the horizontal advection, plays an important role in this “anomalous warming.” This is because the temperature below the mixed layer becomes anomalously warm owing to the deeper thermocline (Fig. 7c) and the temperature difference (ΔT) between the MLT and the water temperature 10 m below the base of the mixed layer becomes smaller (Fig. 9b). Considering that the vertical resolution in our OGCM is 10 m in the upper 50 m, where the modeled mixed layer is found, we have decided to use 10 m in the above, as in Qu (2003). The D20 anomaly in December (0) is 25 m (Fig. 7c), which is much larger than the MLD anomaly of 11 m (Fig. 7b). The effects of the deeper thermocline are greater than those of the deeper mixed layer, and this is why ΔT becomes smaller (Fig. 9b). Furthermore, when the MLD is anomalously thick, the vertical turbulent diffusion across the bottom of the mixed layer becomes less efficient in cooling the water column of the mixed layer. In Fig. 8a, the horizontal advection anomaly warms the mixed layer during austral spring and summer in the weak SD years. This is consistent with the easterly wind stress anomaly associated with positive IOD events. The anomalous easterly causes the anomalous southward Ekman heat transport during austral spring (Fig. 10). Since the anomalous meridional temperature gradient (Fig. 10c) is much smaller than the mean temperature gradient, the contribution from the advection of the anomalous meridional temperature gradient is small. The remaining contribution must be due to the heat advection of the anomalous geostrophic flow associated with the intrusion of warm Rossby waves.
(a) Composites of the MLT balance anomaly of the SD region for the weak SD years. (b) Composites of the surface heat flux term anomaly owing to anomalous MLD and anomalous surface heat flux. Anomalies exceeding 95% and 90% confidence levels by a two-tailed t test are indicated by filled and open circles, respectively.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
Temperature differences between the MLT and the temperature 10 m below the bottom of the mixed layer in (a) weak and (b) strong years. Anomalies exceeding 95% confidence level by a two-tailed t test are indicated by open circles.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
(a) Composites of Ekman heat convergence anomaly and its zonal and meridional components in the weak SD years. Anomalies exceeding 95% confidence level by a two-tailed t test are indicated by open circles. (b) Wind stress anomaly and (c) MLTA in October (0). The contour interval is 0.2°C and negative anomalies are shaded for (c).
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
The contribution from the surface heat flux anomaly tends to cool the MLT, but it is smaller than the contribution from the sum of the warming terms such as the horizontal advection and vertical diffusion terms until around January (+1) (Fig. 8a). This is why the anomalous warming occurs. The negative anomaly of the surface heat flux contribution is mostly due to the contribution from the shortwave radiation (Fig. 11a). This perplexing aspect is due to the fact that the shortwave radiation becomes less efficient in warming the thicker mixed layer in the weak SD years (Fig. 6b). In addition, the cooling by the climatological latent heat flux is less efficient when the MLD is thick and is counteracted by the interannual negative (cooling) anomaly in the latent heat flux itself. The contribution of the longwave radiation to the surface heat flux anomaly is very small throughout a year (Fig. 11b). However, its contribution to the MLT shows a clear positive maximum during the austral summer (Fig. 11a). This is again related to the thicker MLD. As a whole, the anomalous contribution from the surface heat flux term is mainly due to the anomalous MLD [i.e., 
As in Fig. 6, but for composites of the weak SD years. Anomalies exceeding 95% confidence level by a two-tailed t test are indicated by open circles.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
b. Strong SD years
The case in the strong SD years is almost a mirror image of that in the weak SD years. As shown in Fig. 12a, the MLT becomes anomalously cool with the minimum of about −0.4°C in December (0). Also, the MLD becomes anomalously thin in austral summer, with its minimum of about −6 m in February (+1) (Fig. 12b). From the calculation of the Monin–Obukhov depth, we find that the MLD anomaly in December (0) is mainly due to the buoyancy anomaly represented by the third term in Eq. (4), rather than the wind-stirring anomaly represented by the first term (see Table 1). The residual term of Eq. (4) is larger than the case in the weak SD years, because higher-order terms are larger with larger δQ.
As in Fig. 7, but for the strong SD years.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
To clarify the variation mechanism of the MLTA, we have again used the MLT equation (Fig. 13). The tendency term is negative from July (0) to November (0). The vertical diffusion term contributes most to this cooling. This is because ΔT becomes larger than the climatology (Fig. 9c), as a result of a stronger SD with a shallower thermocline (Fig. 12c). The horizontal advection anomaly is negative, but its amplitude is smaller than in the case of the weak SD years. This horizontal advection anomaly is consistent with the meridional Ekman heat transport anomaly related to the negative IOD (Fig. 14), which has a negative anomaly until March (+1). The contribution from the surface heat flux anomaly is positive, but it is weaker than the vertical diffusion term and the cooling of the MLT occurs during the austral spring. The contribution from the anomalous shortwave radiation term is dominant among the four components of the surface heat flux term and warms the thinner mixed layer actively in austral summer and fall (Fig. 15a). The anomalous surface heat flux itself is dominated by the positive latent heat flux anomaly associated with the cooler SST (Fig. 15b). However, the effective role of the latent heat flux anomaly is actually to cool the mixed layer. This intriguing aspect is understandable by considering that thinner MLD is more efficiently cooled by the climatological latent heat flux (Fig. 15a). In contrast to the weak SD years, both the anomalous MLD and net surface heat flux contribute to the surface heat flux term anomaly (Fig. 13b).
As in Fig. 8, but for the strong SD years.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
As in Fig. 10, but for the strong SD years.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
As in Fig. 11, but for the strong SD years.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
5. Summary and discussion
Using outputs from the OGCM, we have investigated the seasonal and interannual variations of the MLT above the SD. In the seasonal climatology, the MLT warms from August to April and cools from May to July. The surface heat flux plays the most important role in this seasonal variation, and its major contribution is from the shortwave radiation. This is why the annual cycle is dominant in the MLT and SST variations. The horizontal advection causes strong warming in austral winter. This is related to the southward Ekman heat transport induced by the strong meridional SST gradient and the easterly wind stress associated with the Indian summer monsoon. The cooling by the vertical turbulent diffusion becomes most effective in austral summer owing to the thinner mixed layer.
The seasonal variation of the mixed layer heat balance in the SD region was discussed by Foltz et al. (2010) using observation data. The importance of the surface heat flux and the seasonal variation of the horizontal advection are consistent with our model results. According to Fig. 12 of Foltz et al. (2010), the heat flux of the vertical processes is −75 W m−2 in June–August and −45 W m−2 in January–March. Thus, in terms of the contribution to the mixed layer heat content, the cooling by the vertical processes is twice as strong in June–August. However, the contribution of the vertical processes to the MLT is almost the same throughout the year, because the MLD is about 40 m in June–August and 20 m in January–March. Therefore, our model results are consistent with the observational study of Foltz et al. (2010). However, we note that we have focused on the MLT instead of the mixed layer heat content. Santoso et al. (2010) discussed the seasonal variation of the SD region using a coupled model. The contribution of the entrainment to the MLT undergoes significant seasonal variation with the strongest cooling in May. However, the strong cooling in May can be due to their model bias in the thermocline depth; the simulated thermocline depth in May is about 20 m shallower than that of the observation and the effects of simulated entrainment in the SD region around May may be overestimated. Zhou et al. (2008), by use of an OGCM, showed that the seasonal variation in the contribution from the entrainment to the MLT is very strong. This may be related to the model bias in the seasonal variation of the MLD; their model MLD varies seasonally by about 50 m in contrast to only about 30 m in the observation (Foltz et al. 2010). Therefore, the modeled entrainment rate must be much higher than the observation. Also, the model MLD reaches about 75 m in austral winter. This means that the base of the mixed layer becomes closer to the thermocline, where the vertical temperature gradient is stronger, and much colder water may be entrained into the mixed layer. In the seasonal variation of the MLT in the Jet Propulsion Laboratory’s (JPL) Estimating the Circulation and Climate of the Ocean (ECCO) product (Halkides and Lee 2011), the subsurface component plays an important role. The stronger contribution in their model may be due to a larger variation in the MLD. It varies by about 40 m, which is twice as large as that in our model and in WOA98.
On the interannual time scale, the MLT becomes anomalously warm (cool) in the weak (strong) SD years. The mechanism is summarized in the schematic diagram of Fig. 16. The thermocline becomes deeper (shallower) owing to local Ekman downwelling (upwelling) and the arrival of downwelling (upwelling) Rossby waves (Fig. 16a; Tozuka et al. 2010). As a result of a deeper (shallower) thermocline, the temperature below the mixed layer becomes warmer (colder), and leads to less (more) cooling by means of vertical diffusion (Fig. 16b). Also, the above local wind is associated with the easterly (westerly) wind stress anomaly that induces southward (northward) Ekman heat transport and warms (cools) the mixed layer (Fig. 16b). The warmer (colder) MLT results in enhanced (suppressed) latent heat loss as well as less (more) shortwave radiation reaching the ocean surface because of more cloudiness (Fig. 16c). As a result, the MLD becomes thicker (thinner), and this leads to less (more) efficient cooling by the vertical diffusion, longwave radiation, and latent heat flux. Thus, positive (negative) MLT grows further (Fig. 16d). This MLTA is partly damped by less (more) efficient warming by shortwave radiation (Fig. 16d). In contrast to the seasonal variation, we may conclude that the vertical diffusion term plays the most important role and causes the anomalous warming (cooling).
Schematic diagram of the interannual variation of the SST above the SD during (a),(b) austral spring and (c),(d) summer of the weak SD years.
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
It is interesting to discuss how the MLTA in the SD region links with the large-scale climate modes. The ordinary correlation between the dipole mode index (DMI), which is defined by the difference in the SSTA between the tropical western Indian Ocean (10°S–10°N, 50°–70°E) and the tropical eastern Indian Ocean (10°S–equator, 90°–110°E) following Saji et al. (1999), and the MLTA in the SD is 0.59 when the DMI leads by 1 month. This is reasonable because the positive (negative) IOD may induce the weak (strong) SD, whose D20 is anomalously deep (shallow) (Tozuka et al. 2010). Then, the vertical displacement of the thermocline may influence the interannual variation of the MLTA through vertical mixing between the mixed layer and the seawater under the bottom of the mixed layer. We must note that the ordinary correlation between the Niño-3 index, which is defined as the area-averaged SSTA over the tropical eastern Pacific (5°S–5°N, 90°–150°W), and the MLTA in the SD is 0.55 when the Niño-3 index leads by 2 months. As discussed by Xie et al. (2002), and clarified in more detail by Rao and Behera (2005), the ENSO may generate oceanic processes by influencing the wind field over the southern Indian Ocean through its teleconnection.
To examine how the existence of SD contributes to the variations of the MLT, we have analyzed the temperature difference between the MLT and the temperature 10 m below the base of the mixed layer (Fig. 17a). The difference exceeds 2°C and is larger than that in the surrounding ocean. This is because the thermocline is shallow in the SD region and the cold subsurface water is closer to the surface mixed layer. Therefore, the vertical mixing may play a more important role compared with the surrounding region. In addition, the MLD is thinner in the SD region (Fig. 17b). This is why the MLT in this region is very sensitive to subsurface oceanic processes and may explain large interannual SST variations there (Izumo et al. 2008).
Horizontal distribution of the climatological (a) temperature difference between the mixed layer and below the MLD at a depth of 10 m below the base of the mixed layer and (b) MLD in austral summer. The contour interval is 0.25°C and temperature differences larger than 1.5°C are shaded in (a), while the contour interval is 10 m and MLDs shallower than 30 m are shaded in (b).
Citation: Journal of Climate 25, 2; 10.1175/JCLI-D-10-05001.1
The present study provides a detailed mechanism that generates the MLTA and thus the SSTA (equivalent to the MLT here) in the SD region. The SSTA is known to influence the amount of rainfall and the onset of the Indian summer monsoon by affecting the meridional migration of the ITCZ (Annamalai et al. 2005; Izumo et al. 2008). Therefore, the present study may contribute to improvements in our understanding and predictability of the Indian summer monsoon.
This study is based mostly on the OGCM simulations. To understand the details of the ocean–atmosphere interaction over the SD region, we need to use a coupled GCM. However, this task is not easy at the present stage, because the seasonal variation of the SD is not well simulated in all coupled GCMs, as shown by Yokoi et al. (2009). As we have shown here, variations in the MLT are related to variations in the thermocline depth. The variations of the SD are related to a remote (as well as local) wind field, and this wind field is related to the MLT (i.e., SST). Therefore, we need to improve the skill of coupled GCMs in simulating the SD and its seasonal variability. Also, the relative role of vertical processes in the seasonal variation of the MLT differs among models. Since it depends on the vertical mixing parameterization, more effort is required to improve the parameterization. Ongoing field programs in this region (Vialard et al. 2009; McPhaden et al. 2009) are expected to improve our understanding of ocean–atmosphere coupled processes and mixed layer processes in the SD region.
Acknowledgments
We thank two anonymous reviewers for helpful comments to our original manuscript. This study benefited from discussions with Dr. Yukio Masumoto. The OGCM was run on the HITACHI SR11000/J1 at the Information Technology Center, University of Tokyo, under a cooperative research agreement with the Atmosphere and Ocean Research Institute, University of Tokyo. The present research is supported by the Japan Society for Promotion of Science through Grant-in-Aid for Scientific Research (B) 20340125.
REFERENCES
Annamalai, H., P. Liu, and S.-P. Xie, 2005: Southwest Indian Ocean SST variability: Its local effect and remote influence on Asian monsoons. J. Climate, 18, 4150–4167.
Annamalai, H., H. Okajima, and M. Watanabe, 2007: Possible impact of the Indian Ocean SST on the Northern Hemisphere circulation during El Niño. J. Climate, 20, 3164–3189.
Carton, J. A., and B. S. Giese, 2008: A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA). Mon. Wea. Rev., 136, 2999–3017.
Foltz, G. R., J. Vialard, B. P. Kumar, and M. J. McPhaden, 2010: Seasonal mixed layer heat balance of the southwestern tropical Indian Ocean. J. Climate, 23, 947–965.
Gadgil, S., P. V. Joseph, and N. V. Joshi, 1984: Ocean–atmosphere coupling over monsoon regions. Nature, 312, 141–143.
Gent, P. R., and J. C. McWilliams, 1990: Isopycnal mixing in ocean circulation models. J. Phys. Oceanogr., 20, 150–155.
Halkides, D. J., and T. Lee, 2009: Mechanisms controlling seasonal-to-interannual mixed layer temperature variability in the southeastern tropical Indian Ocean. J. Geophys. Res., 114, C02012, doi:10.1029/2008JC004949.
Halkides, D. J., and T. Lee, 2011: Mechanisms controlling seasonal mixed layer temperature and salinity in the southwestern tropical Indian Ocean. Dyn. Atmos. Oceans, 51, 77–93.
Hayes, S. P., P. Chang, and M. J. McPhaden, 1991: Variability of the sea surface temperature in the eastern equatorial Pacific during 1986–88. J. Geophys. Res., 96, 10 553–10 566.
Hermes, J., and C. J. C. Reason, 2008: Annual cycle of the south Indian Ocean (Seychelles-Chagos) thermocline ridge in a regional ocean model. J. Geophys. Res., 107, C04035, doi:10.1029/2007JC004363.
Hermes, J., and C. J. C. Reason, 2009: The sensitivity of the Seychelles-Chagos thermocline ridge to large-scale wind anomalies. ICES J. Mar. Sci., 66, 1455–1466.
Hurrell, J. H., M. P. Hoerling, A. S. Phillips, and T. Xu, 2004: Twentieth century North Atlantic climate change. Part 1: Assessing determinism. Climate Dyn., 23, 371–389.
Izumo, T., C. de Boyer Montégut, J.-J. Luo, S. K. Behera, S. Masson, and T. Yamagata, 2008: The role of the western Arabian Sea upwelling in Indian monsoon variability. J. Climate, 21, 5603–5623.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.
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.
Kim, S.-B., I. Fukumori, and T. Lee, 2006: The closure of the ocean mixed layer temperature budget using level-coordinate model fields. J. Atmos. Oceanic Technol., 23, 840–853.
Klein, S. A., B. J. Soden, and N.-C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917–932.
Kraus, E. B., and J. S. Turner, 1967: A one-dimensional model of the seasonal thermocline: II. The general theory and its consequences. Tellus, 19, 98–106.
Kunze, E., E. Firing, J. M. Hummon, T. K. Chereskin, and A. M. Thurnherr, 2006: Global abyssal mixing inferred from lowered ADCP shear and CTD strain profiles. J. Phys. Oceanogr., 36, 1553–1576.
Levitus, S., and T. P. Boyer, 1994: Temperature. Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS 4, 117 pp.
Levitus, S., R. Burgett, and T. P. Boyer, 1994: Salinity. Vol. 3, World Ocean Atlas 1994, NOAA Atlas NESDIS 3, 99 pp.
Masumoto, Y., and G. Meyers, 1998: Forced Rossby waves in the southern tropical Indian Ocean. J. Geophys. Res., 103, 27 589–27 602.
McCreary, J. P., P. K. Kundu, and R. L. Molinari, 1993: A numerical investigation of dynamics, thermodynamics and mixed-layer processes in the Indian Ocean. Prog. Oceanogr., 31, 181–244.
McPhaden, M. J., and Coauthors, 2009: RAMA: The Research Moored Array for African–Asian–Australian Monsoon Analysis and Prediction. Bull. Amer. Meteor. Soc., 90, 459–480.
Mignot, J., C. de Boyer Montégut, A. Lazar, and S. Cravatte, 2007: Control of salinity on the mixed layer depth in the world ocean: 2. Tropical areas. J. Geophys. Res., 112, C10010, doi:10.1029/2006JC003954.
Moisan, J. R., and P. P. Niiler, 1998: The seasonal heat budget of the North Pacific: Net heat flux and heat storage rates (1950–90). J. Phys. Oceanogr., 28, 401–421.
Morioka, Y., T. Tozuka, and T. Yamagata, 2011: On the growth and decay of the subtropical dipole mode in the South Atlantic. J. Climate, 24, 5538–5554.
Pacanowski, R. C., and S. G. H. Philander, 1981: Parameterization of vertical mixing in numerical models of tropical oceans. J. Phys. Oceanogr., 11, 1443–1451.
Pacanowski, R. C., and S. M. Griffies, 1999: MOM 3.0 manual. GFDL Ocean Group Tech. Rep. 4, NOAA/Geophysical Fluid Dynamics Laboratory, Princeton, NJ, 680 pp. [Available from GFDL, Princeton University, Princeton, NJ 08542.]
Paulson, C. A., and J. J. Simpson, 1977: Irradiance measurements in the upper ocean. J. Phys. Oceanogr., 7, 952–956.
Qiu, B., and K. A. Kelly, 1993: Upper-ocean heat balance in the Kuroshio extension region. J. Phys. Oceanogr., 23, 2027–2041.
Qu, T., 2003: Mixed layer heat balance in the western North Pacific. J. Geophys. Res., 108, 3242, doi:10.1029/2002JC001536.
Rao, S. A., and S. K. Behera, 2005: Subsurface influence on SST in the tropical Indian Ocean: Structure and interannual variability. Dyn. Atmos. Oceans, 39, 103–135.
Rao, S. A., S. K. Behera, Y. Masumoto, and T. Yamagata, 2002: Interannual subsurface variability in the tropical Indian Ocean with a special emphasis on the Indian Ocean Dipole. Deep-Sea Res. II, 49, 1549–1572.
Reverdin, G., and M. Fieux, 1987: Sections in the western Indian Ocean—Variability in the temperature structure. Deep-Sea Res., 34, 601–626.
Rosati, A., and K. Miyakoda, 1988: A general circulation model for upper ocean simulation. J. Phys. Oceanogr., 18, 1601–1626.
Saji, N. H., B. N. Goswami, P. N. Vinayachandran, and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Nature, 401, 360–363.
Santoso, A., A. S. Gupta, and M. H. England, 2010: Genesis of Indian Ocean mixed layer temperature anomalies: A heat budget analysis. J. Climate, 23, 5375–5403.
Schott, F. A., S.-P. Xie, and J. P. McCreary, 2009: Indian Ocean circulation and climate variability. Rev. Geophys., 47, RG1002, doi:10.1029/2007RG000245.
Smagorinsky, J., 1963: General circulation experiments with the primitive equations. 1. The basic experiments. Mon. Wea. Rev., 91, 99–164.
Tozuka, T., T. Yokoi, and T. Yamagata, 2010: A modeling study of interannual variations of the Seychelles Dome. J. Geophys. Res., 115, C04005, doi:10.1029/2009JC005547.
Vialard, J., and Coauthors, 2009: Air–sea interactions in the Seychelles–Chagos thermocline ridge region. Bull. Amer. Meteor. Soc., 90, 45–61.
Woodberry, K. E., M. Luther, and J.-J. O’Brien, 1989: The wind-driven seasonal circulation in the southern tropical Indian Ocean. J. Geophys. Res., 94, 17 985–18 002.
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, J.-J. Luo, S. Masson, M. R. Jury, and S. A. Rao, 2004: Coupled ocean–atmosphere variability in the tropical Indian Ocean. Earth Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., Vol. 147, Amer. Geophys. Union, 189–212.
Yasuda, I., T. Tozuka, M. Noto, and S. Kouketsu, 2000: Heat balance and regime shifts of the mixed layer in the Kuroshio extension. Prog. Oceanogr., 47, 257–278.
Yokoi, T., T. Tozuka, and T. Yamagata, 2008: Seasonal variation of the Seychelles Dome. J. Climate, 21, 3740–3754.
Yokoi, T., T. Tozuka, and T. Yamagata, 2009: Seasonal variation of the Seychelles Dome simulated in the CMIP3 models. J. Phys. Oceanogr., 39, 449–457.
Zhou, L., R. Murtugudde, and M. Jochum, 2008: Seasonal influence of Indonesian Throughflow in the southwestern Indian Ocean. J. Phys. Oceanogr., 38, 1529–1541.