1. Introduction
The sea surface temperature (SST) in the western Arabian Sea upwelling (WASU) region (Izumo et al. 2008), which corresponds to the whole area along the western boundary of the Arabian Sea, is known to undergo large seasonal variations, with cooler SST in boreal summer mainly owing to strong coastal upwelling (Fig. 1). SST in this area is important for various reasons. First, SST anomalies in this region are positively correlated with June–September precipitation anomalies over the Western Ghats region in India (Vecchi and Harrison 2004). This was confirmed by a sensitivity experiment with a coupled general circulation model (CGCM) (Izumo et al. 2008). The strength of monsoon rainfall in the Western Ghats increased by about 30% when the upwelling was suppressed along the Somalia–Oman coast, and the SST became warmer by up to 2°C. Also, the SST anomalies in the WASU region may influence precipitation over East Africa during the short rain season (Behera et al. 2005) and over the southern part of Iran during the first part of the rainy season (Pourasghar et al. 2012).
Seasonal march of SST in the OISST. Contour interval is 1°C, and temperatures lower than 27°C are shaded.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
Regarding the mechanism of interannual variation in SST anomalies over the WASU region, previous studies focused on variations in local winds, which alter the strength of coastal upwelling. Izumo et al. (2008) pointed out a link with SST anomalies over the Seychelles Dome (Yokoi et al. 2008, 2012; Hermes and Reason 2008; Tozuka et al. 2010), the doming of the thermocline in the South Indian Tropical Gyre, in the southwestern tropical Indian Ocean. When the SST over the dome region is warmer, the intertropical convergence zone tends to remain there longer (Annamalai et al. 2005), resulting in northeasterly wind stress anomalies in the WASU region. This induces coastal downwelling anomalies and thus positive SST anomalies.
Another possible mechanism for SST variations in the WASU region is the remote influence of incoming ocean waves. In the tropical Indian Ocean, dominant semiannual variability is one of the most remarkable features that is not seen in other basins. The semiannual variation in zonal wind stress (Wunsch 1977; Ogata and Xie 2011) along the equatorial Indian Ocean forces semiannual Kelvin waves and the associated eastward current is known as the Wyrtki jets (Yoshida 1959; Wyrtki 1973). The semiannual current variation is amplified by constructive interference between forced and reflected waves, which is often referred to as a basinwide resonance (Cane and Moore 1981; Han et al. 2011). Earlier studies examined the influence of waves on SST in the upwelling regions off Sumatra (e.g., Rao et al. 2002) and of the Seychelles Dome (e.g., Xie et al. 2002). However, no study to date has examined their impacts on the SST in the WASU region, despite its importance in the Indian Ocean climate. Here, we examine how the reflection of semiannual Kelvin waves at the eastern boundary of the Indian Ocean may influence the WASU region.
To understand the dynamics of the semiannual variations, many different kinds of models ranging from simple linear models to general circulation models (GCMs) have been employed. O’Brien and Hurlburt (1974) were the first to numerically reproduce the Wyrtki jets with a two-layer primitive equation model with a flat bottom and beta-plane approximation. Using a linearized primitive equations model forced by wind stress with only a semiannual component, Gent et al. (1983) had success in reproducing the semiannual current variation observed by Luyten and Roemmich (1982). Nagura and McPhaden (2010) also succeeded in reproducing the observed Wyrtki jets using a continuously stratified linear equatorial long-wave model forced by observed zonal wind stress. Those two studies confirmed that the jets are essentially dominated by linear long-wave dynamics, which are supported by observational estimates of the momentum balance (Nagura and McPhaden 2008). However, coupled models have difficulty in reproducing the above semiannual variation (Nagura et al. 2013). Regarding the basinwide resonance, Jensen (1993) used 3.5- and 1.5-layer Indian Ocean models and showed that the basinwide resonance of the second baroclinic mode forced by the semiannual wind stress plays an important role. More specifically, Han et al. (1999) revealed that direct wind forcing by the semiannual component of the wind accounts for 81% of the amplitude for the Wyrtki jets, while the remaining 19% is due to the reflected Rossby waves and the basinwide resonance. Based on a wave decomposition of their ocean general circulation model (OGCM) output, Yuan and Han (2006) examined the role of western boundary reflection. It was found that reflection of equatorial Rossby waves significantly contributes to the semiannual harmonic of the equatorial Kelvin waves.
This paper is organized as follows. A brief description of our OGCM, data, theoretical model, and mixed layer heat balance calculation is given in section 2. Section 3 discusses the impact of the reflected Rossby waves on the SST in the WASU region. A discussion and summary is given in sections 4 and 5, respectively.
2. Model, data, and method
a. Model and data description
The OGCM is based on version 3.0 of the Modular Ocean Model (MOM3.0; Pacanowski and Griffies 1999) developed at the National Oceanic and Atmospheric Administration (NOAA)– Geophysical Fluid Dynamics Laboratory (GFDL). The model domain encompasses most of the Indo-Pacific Oceans from 52°S to 30°N with horizontal resolution of 0.5°. A Laplacian-type horizontal mixing parameterization is adopted with eddy viscosity and diffusivity of 2.0 × 103 and 1.0 × 103 m2 s−1, respectively. These values are exactly the same with those used by the 0.5° × 0.5° Indo-Pacific OGCM of Masumoto (2002), which were successful in reproducing oceanic variability in the domain. For vertical mixing parameterization, we adopt the scheme developed by Pacanowski and Philander (1981). The OGCM is first spun up for 20 yr from the annual-mean climatology (Levitus and Boyer 1994; Levitus et al. 1994) with no motion using the monthly-mean climatology of the wind stress from the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis data (Kalnay et al. 1996). The surface heat flux is calculated based on a bulk formula using the simulated SST and atmospheric variables obtained from the NCEP–NCAR reanalysis data (Rosati and Miyakoda 1988). The simulated sea surface salinity (SSS) is restored to the observed monthly climatology with a relaxation time scale of 30 days. Then, the model is integrated from 1978 to 2012 using the daily-mean reanalysis data, and the model SSS is again restored to the monthly-mean climatology. Outputs after 1982 are stored every 5 days and used to construct climatologies. We note that the OGCM used in this study is similar to that of Tozuka et al. (2010) and Yokoi et al. (2012), but the integration is extended by 5 yr to include recent years, and the horizontal mixing parameterization is changed.
In this study, we have performed two experiments. One is a control (CTRL) experiment, where the model is integrated as described above. On the other hand, to check how the reflection of equatorial Kelvin waves affects the WASU region, we have conducted another experiment, where the horizontal viscosity in the eastern equatorial Indian Ocean (7.5°S–7.5°N, 80°E–Indonesian coast) is increased by 10 times so that waves are damped or eliminated. This latter experiment is called the Damping near the Eastern Boundary (DAMPEB) run. Nagura and McPhaden (2010) pointed out that most of the observed variability in the equatorial Indian Ocean can be described by considering Kelvin and first meridional Rossby modes, which give rise to sea level and velocity variability in 5°S–5°N. The region of damping covers those latitudes of the dominant wave modes. Also, Qiu et al. (1997) showed that Laplacian-type horizontal mixing acts as a linear damping on long Rossby waves, which provides a theoretical basis for our experimental setting. The western limit of the damping region is set such that the region is outside the Arabian Sea and wide enough to damp the Rossby waves. The validity of our experiments is checked by comparing results from the OGCM with those from a theoretical model (described below), which analytically computes boundary-generated waves.
For model validation, we use the optimum interpolation sea surface temperature (OISST; Reynolds et al. 2002) for SST and the drifter-derived monthly climatology for near-surface zonal currents (Lumpkin and Johnson 2013).
b. Theoretical model
The theoretical model we adopt is a continuously stratified linear long-wave model on an equatorial β plane of Nagura and McPhaden (2010). In the model, the equations are solved analytically first and then integrated using observed winds. We use Quick Scatterometer (QuikSCAT) winds from 1999 to 2009 to force the theoretical model, rather than the NCEP–NCAR reanalysis in consistency with the OGCM, because the theoretical model is tuned and validated for the QuikSCAT run (Nagura and McPhaden 2010). The basin of the theoretical model has meridional walls at 40° and 100°E. Wave reflections at the western boundary are calculated using the condition of no normal mass flux, whereas the condition of no normal velocity is used for eastern boundary reflections. More detailed descriptions of the theoretical model are given by Nagura and McPhaden (2010), who also show that the model results compare well with satellite altimetry and in situ velocity observations east of 50°E. As their model aims to reproduce large-scale circulation such as the Wyrtki jets, short waves are filtered out by the use of the long-wave approximation, and thus the model shows less skill west of 50°E.
We have conducted two experiments. In the control run, we force the model assuming a reflection efficiency of 85% in terms of amplitude on the basis of the observational study of Le Blanc and Boulanger (2001). For comparison with the DAMPEB experiment of the OGCM, we conducted a sensitivity experiment, in which we use 0% for the reflection efficiency at the eastern boundary. The difference between the sensitivity and control experiments is presented below.
c. Mixed layer temperature balance calculation
3. Results
Prior to investigating the impacts of the reflected Rossby waves, it is necessary to check the validity of the current OGCM. Figures 1 and 2 show the seasonal march of observed and simulated SST, respectively. In the observations, the SST becomes the warmest in May, exceeding 27°C everywhere in the western tropical Indian Ocean with a maximum temperature above 30°C. However, after the onset of the Indian summer monsoon, the SST along the western boundary cools due to the strong southwesterly winds that favor coastal upwelling, vertical mixing, and evaporative cooling. This cooling decays when the Indian summer monsoon retreats. The northern Arabian Sea starts to cool in December, and the SST lower than 27°C covers most of the area north of 10°N. The above observed seasonal variation is well simulated by our OGCM, except that the SST is systematically lower by about 1°C.
As in Fig. 1, but for the CTRL run.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
Because our focus is on the equatorial surface zonal current, its monthly climatology in the observation and the CTRL run is compared in Figs. 3a and 3b. As in the observation, the simulated equatorial zonal current is dominated by a semiannual cycle; the maximum eastward (westward) currents are seen in May and November (February and July) in the central Indian Ocean in the CTRL run. Thus, the phase of the semiannual cycle is also well reproduced. There are some discrepancies in the amplitude of surface currents. One such example is the weaker fall Wyrtki jet in the CTRL run. This may be related to salinity stratification (Han et al. 1999; Masson et al. 2003). Because the current model uses a restoring boundary condition for SSS and the model vertical resolution is 10 m in the upper 50 m, the barrier layer that develops in boreal fall is not well simulated. Nevertheless, because of the good agreement between the model and observation data, we expect that this model can provide useful insight into the role of reflected Rossby waves on the SST in the WASU region.
Monthly climatology of the surface zonal current along the equator in (a) observation, (b) CTRL experiment, and (c) DAMPEB experiment. Positive values signify eastward currents and negative values are shaded. Contour interval is 0.1 m s−1.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
Figure 4 shows the difference in simulated SST between CTRL and DAMPEB experiments. The difference is relatively small throughout the year, except from August to November in the WASU region; the SST in the DAMPEB run is colder by as much as 0.4°C in October. Although this is not very large, this may have a significant impact on the overlying atmosphere, considering that the SST in this region during this time of the year is around 27°C (Figs. 1, 2) (Gadgil et al. 1984; Graham and Barnett 1987). As will be discussed in more detail, one reason for this relatively small difference is that we use a bulk formula for the heat flux calculation in our OGCM; this can act to relax the SST to the air temperature at the 2-m height of the NCEP–NCAR reanalysis data.
Difference in the monthly climatology of SST (°C) and surface current (m s−1) between CTRL and DAMPEB experiments (DAMPEB minus CTRL).
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
To examine the causes of the SST difference along the East African coast, we have made an artificial box (2°S–5°N, African coast–50°E) that covers the area with the SST difference greater than 0.3°C in October (Fig. 5). As is clear from Fig. 4, the SST difference is smaller than −0.05°C from January to June (Fig. 6). Then, the difference grows rapidly and reaches −0.26°C in October, but decays rapidly in November and December.
Difference in the monthly climatology of SST between CTRL and DAMPEB experiments (DAMPEB minus CTRL) in October. Contour interval is 0.05°C, and negative values are shaded. The box shows the region where the mixed layer temperature balance is calculated.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
SST difference (°C) between CTRL and DAMPEB experiments (DAMPEB minus CTRL) in the box region.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
For quantitative understanding, we have first calculated a mixed layer temperature balance in this box using outputs from the CTRL run (Fig. 7), because it is necessary to understand how the SST is determined in this region prior to examining the difference between CTRL and DAMPEB experiments. The tendency term is dominated by a semiannual harmonic with a warming (cooling) from February to April (May to July) and August to October (November to January). This semiannual variation is almost in phase with both surface heat flux and horizontal advection terms. As shown in Fig. 3, the equatorial zonal current has a dominant semiannual harmonic, and the horizontal advection has a warming effect when the zonal current is westward and transports warmer water from offshore. The positive maximum in August is stronger than that in March because of the larger zonal SST gradient in August associated with cooler SST along the coast. Also, western boundary currents such as the East African Coastal Current and Somali Current transport cooler waters from the South Equatorial Current northward and contribute to cooling of the box region, especially in early boreal summer. On the other hand, the semiannual variation of the surface heat flux term is partly associated with that of the MLD with its seasonal minima in April and November (Fig. 8a). When the mixed layer is thinner, it becomes more sensitive to the surface heat flux. Also, latent heat flux causes less cooling during the monsoon breaks around April and October, likely due to the weak winds (Fig. 9). The maximum in surface heat flux term is stronger in April than in October because of stronger shortwave radiation (Fig. 9). The vertical term cools the box region throughout the year, and its seasonal variation is due to interplay between the MLD (Fig. 8a) and the temperature difference between the mixed layer and the layer below (Fig. 8b). The cooling by the vertical term is strongest in May because the mixed layer is relatively thin and the temperature difference is close to its seasonal maximum. The horizontal diffusion term also contributes to the cooling, but its contribution is very minor.
Monthly climatology of the mixed layer temperature balance (10−7 °C s−1) in the box region derived from the CTRL run. Here, tendency, surface heat flux, horizontal advection, horizontal diffusion, and vertical process terms are indicated by thick solid, thin dashed, thick dashed, dotted–dashed, and thin solid lines, respectively.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
Seasonal march of (a) MLD (m) and (b) the temperature difference (°C) between the mixed layer temperature and temperature at 10 m below the base of the mixed layer in the box region in the CTRL run.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
Monthly climatology of the net surface heat flux and its four components in the box region (W m−2) used to force the OGCM in the CTRL run.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
Next, we check the difference in the mixed layer temperature balance in the same box region for the two experiments (Fig. 10). It is clear that the box region becomes colder in the DAMPEB experiment in boreal fall because the warming by horizontal advection is smaller. This is somewhat a surprising result because the damping of equatorial Rossby waves would result in a change in the thermocline depth, and thus we had expected that the vertical processes are the main cause of the SST difference. However, this may be explained by the fact that the mixed layer is shoaling rapidly from September to November (Fig. 8a), and there is basically no entrainment of water from below. Therefore, the effect of the shallower thermocline, and associated colder water below the mixed layer, cannot be felt by the mixed layer. On the other hand, the surface heat flux acts as a negative feedback.
As in Fig. 7, but for the difference in each term of the mixed layer temperature balance equation between CTRL and DAMPEB experiments (DAMPEB minus CTRL).
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
The large difference in the horizontal advection term is intimately linked with that of the zonal current. Figure 11a shows the difference in the surface zonal current (i.e., the difference between Figs. 3b and 3c). Around June, a large difference of about −0.2 m s−1 is found to originate from 80°E, which corresponds to the western boundary of the strong damping region. This signal propagates westward and reaches the western boundary of the equatorial Indian Ocean from August to November, when the difference in the horizontal advection term is also the largest. The amplitude is slightly less than 0.10 m s−1 near the western boundary. In addition, a positive velocity difference can be found in boreal winter, resulting in semiannual variability in the difference to the west of 80°E.
(a) Time–lon diagram of the difference in the surface zonal current along the equator between CTRL and DAMPEB experiments (DAMPEB minus CTRL). Note that strong damping is imposed from 80°E to the eastern boundary in the DAMPEB experiment. (b) Time–lon diagram of the difference in the zonal current at 10-m depth between the sensitivity and control experiments (former minus latter) of the theoretical model for the first two gravest baroclinic modes. Contour intervals are 0.05 m s−1, and negative values are shaded.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
To examine whether the above difference in the zonal current can be explained in terms of linear wave dynamics, we have calculated the contribution from the reflected Rossby waves using the theoretical model of Nagura and McPhaden (2010). Figure 11b shows the zonal current velocity at the 10-m depth associated with the reflected Rossby waves for the first two gravest baroclinic modes. The theoretical model shows the dominant semiannual signals that radiate from the eastern boundary and propagate westward, which is in agreement with the OGCM results to the west of 80°E. Regarding the signal that reaches the western boundary in boreal fall, it starts from the eastern boundary with a magnitude of 0.55 m s−1 and gradually decays as it propagates westward. When it reaches the western boundary, it retains a magnitude of 0.15 m s−1. Although the difference in the zonal velocity in the two OGCM experiments is about half of this value, the theoretical model qualitatively supports our hypothesis that the reflected Rossby waves originating from the semiannual Kelvin waves play an important role in the zonal current near the western boundary and the SST in the WASU region.
The results of the two models are noticeably different to the east of 80°E. This is because the methods to eliminate waves are quite different in the two models. In the DAMPEB experiment of the OGCM, damping is uniformly increased to the east of 80°E. Incoming Kelvin waves and reflected Rossby waves are gradually damped as they propagate, and the difference from CTRL is largest at 80°E, where Rossby waves originate from the damping region. Also, downwelling (upwelling) Kelvin waves accompanied by eastward (westward) currents on the equator tend to superimpose on downwelling (upwelling) Rossby waves accompanied by westward (eastward) currents near the eastern boundary. Therefore, damping of both Kelvin and Rossby waves does not result in large differences there. On the other hand, the sensitivity experiment of the linear model completely eliminates reflected Rossby waves at the eastern boundary by manipulating the eastern boundary reflectivity. This is why signals in the linear model are largest at the eastern boundary. To the west of 80°E, both models simulate free wave propagation, and the results agree well.
It is interesting to note that large SST differences between the two experiments can be seen only in boreal fall, even though the downwelling Kelvin waves are excited twice a year. This asymmetry may be because the zonal SST gradient is large in boreal fall (Figs. 1, 2), owing to strong coastal upwelling caused by the Findlater jet (Findlater 1969), whereas the SST is more uniform in the zonal direction in boreal spring. Also, the reflected downwelling Rossby waves originating from the downwelling Kelvin waves generated in boreal spring and reaching the western boundary about 6 months later are stronger than those originating from the downwelling Kelvin waves generated in boreal fall (Nagura and McPhaden 2010).
Also, we note that the horizontal resolution of 0.5° in the OGCM is sufficient to resolve equatorial waves and their impacts, but it is not sufficient to fully resolve the western boundary currents such as East African Coastal Current and Somali Current. These currents transport cooler waters from the South Equatorial Current northward and play an important role in setting up the large zonal SST gradient around the box region in boreal summer to fall. Therefore, similar sensitivity experiments with a high-resolution model may result in a larger zonal SST gradient and thus a larger difference in the zonal advection between the two experiments.
4. Discussion
a. Sensitivity to surface heat flux forcing
The largest difference in SST between DAMPEB and CTRL experiments is about −0.4°C. This may be an underestimation of the true impacts of the reflected Rossby waves on the SST in the WASU region because we use a bulk formula for the heat flux calculation in our OGCM, and the simulated SST in both experiments are effectively restored to the same atmospheric temperature from the reanalysis data.
When SST becomes higher (lower), surface heat loss due to the latent heat flux, sensible heat flux, and longwave radiation is expected to increase (decrease), but these processes cannot be represented by an OGCM if it is forced by the heat flux from observation–reanalysis data. This is why we use bulk formula for heat flux calculation in our OGCM. However, because we are prescribing model air temperature and specific humidity using the NCEP–NCAR reanalysis data, the atmosphere does not change in response to surface heat flux changes.
As expected, the SST difference between these two experiments becomes larger, with a maximum difference of −0.6°C (Fig. 12). While areas with an SST difference larger than −0.4°C are confined to the African coast in the DAMPEB minus CTRL cases (Fig. 4), they extend to 60°E in the DAMPEB_NOR minus CTRL_NOR case (Fig. 12). Therefore, the reflected downwelling Rossby waves may have stronger impacts on the SST in the WASU region. We note that there are several ways in which we can conduct the above additional sensitivity experiments, including experiments with longer effective damping time scale, but the experiments we have conducted are more drastic and may give an upper bound of the SST change in the context of OGCM.
Difference in the monthly climatology of SST between CTRL_NOR and DAMPEB_NOR experiments (DAMPEB_NOR minus CTRL_NOR) in October. Contour interval is 0.15°C, and negative values are shaded.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
Also, when the SST change due to the difference in the horizontal advection is estimated by integrating the difference from July to November based on Fig. 10, we obtain −0.74°C. Although this is a rough estimate, this indicates that if there is no restoring by the surface heat flux, the SST difference between CTRL and DAMPEB experiments could be about twice as large.
b. Damping in the eastern basin
We have checked the sensitivity of our results to the strength of damping in the eastern Indian Ocean. When we repeated the DAMPEB experiment by increasing the horizontal viscosity by a factor of 20 instead of a factor of 10, the maximum SST difference with the CTRL run increases only by 0.02°C.
Also, there is a possibility that the sudden increase in viscosity at 80°E in the DAMPEB experiment will generate spurious signals because this jump may generate changes in large-scale horizontal divergence and circulation in the ocean. To check this possibility, we have compared Figs. 3b and 11a. If the spurious signals are dominant, differences in the surface zonal current around 80°E in Fig. 11a should resemble the surface zonal current around 80°E in Fig. 3b because the spurious signals should be roughly proportional to zonal current speed (Fig. 3b) at 80°E. However, this is not the case, and the signals around 80°E in Fig. 11a are quite different from that of Fig. 3b.
c. Impacts on flow along the western boundary of the Arabian Sea
Based on observational data analyses, Beal et al. (2013) recently pointed out that northward flow along the western boundary of the Arabian Sea appears in April, despite the fact that the southwest monsoon has not started yet, and affects the SST along the coast. They suggested that the appearance of the northward flow is associated with the arrival of the annual downwelling Rossby waves propagating along 8°N (Brandt et al. 2002; Rao et al. 2010). Because the damping in the eastern Indian Ocean also damps the Kelvin waves that propagate along the northern Indian Ocean waveguide, the annual downwelling Rossby waves may also be damped in the DAMPEB experiment.
To check this interesting possibility, we have prepared Fig. 13, which shows differences in surface meridional velocity between CTRL and DAMPEB experiments in the western Arabian Sea in April (differences in surface current in other months and in the wider domain are shown in Fig. 4). Figure 13 clearly shows southward velocity anomalies along the western boundary around 8°N. This implies that the northward flow along the western boundary of the Arabian Sea is weakened when the damping is introduced in the eastern equatorial Indian Ocean in the DAMPEB experiment. Therefore, our model experiments support the conjecture by Beal et al. (2013).
Difference in the monthly climatology of meridional current between CTRL and DAMPEB experiments (DAMPEB minus CTRL) in April. Contour interval is 0.01 m s−1, and negative values are shaded.
Citation: Journal of Physical Oceanography 44, 5; 10.1175/JPO-D-13-0127.1
5. Summary
Based on two OGCM experiments with and without strong damping in the eastern equatorial Indian Ocean, we have examined how the reflection of semiannual Kelvin waves at the eastern boundary impacts the SST in the WASU region. It is found that the SST in this region becomes colder, especially in boreal fall when the damping is introduced. Our mixed layer temperature balance calculation indicates that this SST difference is mainly due to a difference in horizontal advection. The westward zonal current in boreal fall becomes weaker in the western equatorial Indian Ocean without the reflected Rossby waves. As a result, warmer water is not transported toward the WASU region, and the SST becomes colder by about 0.4°C in an experiment with the damping (i.e., DAMPEB experiment). The impact on the SST is limited to boreal fall because the zonal SST gradient is weaker in other seasons and/or the amplitude of the reflected Rossby waves is smaller. The importance of horizontal advection is somewhat a surprising result because the damping of equatorial Rossby waves also leads to changes in the thermocline depth and thus changes in the contribution from the vertical processes. However, the mixed layer is shoaling in boreal fall, and there is basically no entrainment of water from below. Therefore, the shallower thermocline and colder water below the mixed layer do not significantly affect the mixed layer temperature.
A comparison with results from a theoretical model suggests that we underestimate the magnitude of velocity associated with reflected Rossby waves in the OGCM experiments. Also, an additional model experiment shows that the formulation of surface heat flux in the DAMPEB experiment weakens the SST response by restoring it to the atmospheric temperature of the reanalysis data. Therefore, we suggest that the impact of Rossby waves on SST in the WASU region can be larger than we have estimated here.
In a coupled model or in the real world, a change in SST in the WASU region is known to influence the Indian monsoon (Vecchi and Harrison 2004; Izumo et al. 2008), and this may influence the atmospheric circulation over the Indian Ocean. As a result, alongshore winds in the WASU region are also modified. Then, the SST in the WASU region may further be modified through changes in coastal upwelling, horizontal advection by western boundary currents, and surface heat flux (latent and sensible heat flux). Thus, it will be interesting to examine the impacts in a CGCM. However, most models that participated in phase 5 of the Coupled Model Intercomparison Project (CMIP5; Taylor et al. 2012) could not simulate the dominant semiannual cycle in the zonal wind stress along the equatorial Indian Ocean (Nagura et al. 2013). This means that the effect discussed in this paper is not well simulated by state-of-the-art CGCMs, and this may be one of the reasons why it is difficult to simulate as well as to predict interannual variation of the rainfall associated with the Indian summer monsoon.
Acknowledgments
We thank two anonymous reviewers for providing useful comments. The OGCM was run on SR16000 system of Information Technology Center, The University of Tokyo, under the cooperative research with Atmosphere and Ocean Research Institute, The University of Tokyo.
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, doi:10.1175/JCLI3533.1.
Beal, L. M., V. Hormann, R. Lumpkin, and G. R. Foltz, 2013: The response of the surface circulation of the Arabian Sea to monsoonal forcing. J. Phys. Oceanogr., 43, 2008–2022, doi:10.1175/JPO-D-13-033.1.
Behera, S. K., J.-J. Luo, S. Masson, P. Delecluse, S. Gualdi, A. Navarra, and T. Yamagata, 2005: Paramount impact of the Indian Ocean dipole on the East African short rains: A CGCM study. J. Climate, 18, 4514–4530, doi:10.1175/JCLI3541.1.
Brandt, P., L. Stramma, F. Schott, J. Fischer, M. Dengler, and D. Quadfasel, 2002: Annual Rossby waves in the Arabian Sea from TOPEX/POSEIDON altimeter and in situ data. Deep-Sea Res. II, 49, 1197–1210, doi:10.1016/S0967-0645(01)00166-7.
Cane, M. A., and D. W. Moore, 1981: A note on low-frequency equatorial basin modes. J. Phys. Oceanogr., 11, 1578–1584, doi:10.1175/1520-0485(1981)011<1578:ANOLFE>2.0.CO;2.
Findlater, J., 1969: A major low-level air current near the Indian Ocean during the northern summer. Quart. J. Roy. Meteor. Soc., 95, 362–380, doi:10.1002/qj.49709540409.
Gadgil, S., P. V. Joseph, and N. V. Joshi, 1984: Ocean–atmosphere coupling over monsoon regions. Nature, 312, 141–143, doi:10.1038/312141a0.
Gent, P. R., K. O’Neill, and M. A. Cane, 1983: A model of the semiannual oscillation in the equatorial Indian Ocean. J. Phys. Oceanogr., 13, 2148–2160, doi:10.1175/1520-0485(1983)013<2148:AMOTSO>2.0.CO;2.
Graham, N. E., and T. P. Barnett, 1987: Sea surface temperature, surface wind divergence, and convection over tropical oceans. Science, 238, 657–659, doi:10.1126/science.238.4827.657.
Han, W., J. P. McCreary, D. L. T. Anderson, and A. J. Mariano, 1999: Dynamics of the eastern surface jets in the equatorial Indian Ocean. J. Phys. Oceanogr., 29, 2191–2209, doi:10.1175/1520-0485(1999)029<2191:DOTESJ>2.0.CO;2.
Han, W., J. P. McCreary, Y. Masumoto, J. Vialard, and B. Duncan, 2011: Basin resonance in the equatorial Indian Ocean. J. Phys. Oceanogr., 41, 1252–1270, doi:10.1175/2011JPO4591.1.
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., 113, C04035, doi:10.1029/2007JC004363.
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, doi:10.1175/2008JCLI2158.1.
Jensen, T. G., 1993: Equatorial variability and resonance in a wind-driven Indian Ocean model. J. Geophys. Res., 98, 22 533–22 552, doi:10.1029/93JC02565.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
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, doi:10.1029/2000JC900072.
Le Blanc, J.-L., and J.-P. Boulanger, 2001: Propagation and reflection of long equatorial waves in the Indian Ocean from TOPEX/POSEIDON data during the 1993–1998 period. Climate Dyn., 17, 547–557, doi:10.1007/s003820000128.
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.
Lumpkin, R., and G. Johnson, 2013: Global ocean surface velocities from drifters: Mean, variance, ENSO response, and seasonal cycle. J. Geophys. Res. Oceans, 118, 2992–3006, doi:10.1002/jgrc.20210.
Luyten, J. R., and D. H. Roemmich, 1982: Equatorial currents at semi-annual period in the Indian Ocean. J. Phys. Oceanogr., 12, 406–413, doi:10.1175/1520-0485(1982)012<0406:ECASAP>2.0.CO;2.
Masson, S., C. Menkes, P. Delecluse, and J.-P. Boulanger, 2003: Impacts of salinity on the eastern Indian Ocean during the termination of the fall Wyrtki jet. J. Geophys. Res., 108, 3067, doi:10.1029/2001JC000833.
Masumoto, Y., 2002: Effects of interannual variability in the eastern Indian Ocean on the Indonesian Throughflow. J. Oceanogr., 58, 175–182, doi:10.1023/A:1015889004089.
Moisan, J. R., and P. P. Niiler, 1998: The seasonal heat budget of the North Pacific: Net heat flux and heat storage rates (1950–1990). J. Phys. Oceanogr., 28, 401–421, doi:10.1175/1520-0485(1998)028<0401:TSHBOT>2.0.CO;2.
Nagura, M., and M. J. McPhaden, 2008: The dynamics of zonal current variations in the central equatorial Indian Ocean. Geophys. Res. Lett., 35, L23603, doi:10.1029/2008GL035961.
Nagura, M., and M. J. McPhaden, 2010: Wyrtki jet dynamics: Seasonal variability. J. Geophys. Res., 115, C07009, doi:10.1029/2009JC005922.
Nagura, M., W. Sasaki, T. Tozuka, J.-J. Luo, S. Behera, and T. Yamagata, 2013: Longitudinal biases in the Seychelles Dome simulated by 35 ocean–atmosphere coupled general circulation models. J. Geophys. Res. Oceans, 118, 831–846, doi:10.1029/2012JC008352.
O’Brien, J. J., and H. E. Hurlburt, 1974: Equatorial jet in the Indian Ocean: Theory. Science, 184, 1075–1077, doi:10.1126/science.184.4141.1075.
Ogata, T., and S.-P. Xie, 2011: Semiannual cycle in zonal wind over the equatorial Indian Ocean. J. Climate, 24, 6471–6485, doi:10.1175/2011JCLI4243.1.
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, doi:10.1175/1520-0485(1981)011<1443:POVMIN>2.0.CO;2.
Pacanowski, R. C., and S. M. Griffies, 1999: MOM 3.0 manual. GFDL Ocean Group Tech. Rep. 4, GFDL, Princeton University, Princeton, NJ 08542, 680 pp.
Pourasghar, F., T. Tozuka, S. Jahanbakhsh, B. Sari Sarraf, H. Ghaemi, and T. Yamagata, 2012: The interannual precipitation variability in the southern part of Iran as linked to large-scale climate modes. Climate Dyn., 39, 2329–2341, doi:10.1007/s00382-012-1357-5.
Qiu, B., W. Miao, and P. Müller, 1997: Propagation and decay of forced and free baroclinic Rossby waves in off-equatorial oceans. J. Phys. Oceanogr., 27, 2405–2417, doi:10.1175/1520-0485(1997)027<2405:PADOFA>2.0.CO;2.
Rao, R. R., M. S. Girish Kumar, M. Ravichandran, A. R. Rao, V. V. Gopalakrishna, and P. Thadathil, 2010: Interannual variability of Kelvin wave propagation in the wave guides of the equatorial Indian Ocean, the coastal Bay of Bengal and the southeastern Arabian Sea during 1993–2006. Deep-Sea Res. I, 57, 1–13, doi:10.1016/j.dsr.2009.10.008.
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., 49, 1549–1572, doi:10.1016/S0967-0645(01)00158-8.
Reynolds, R. W., N. A. Rayner, T. M. Smith, D. C. Stokes, and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 1609–1625, doi:10.1175/1520-0442(2002)015<1609:AIISAS>2.0.CO;2.
Rosati, A., and K. Miyakoda, 1988: A general circulation model for upper ocean simulation. J. Phys. Oceanogr., 18, 1601–1626, doi:10.1175/1520-0485(1988)018<1601:AGCMFU>2.0.CO;2.
Taylor, K. E., R. J. Stouffer, and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485–498, doi:10.1175/BAMS-D-11-00094.1.
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.
Vecchi, G. A., and D. E. Harrison, 2004: Interannual Indian rainfall variability and Indian Ocean sea surface temperature anomalies. Earth’s Climate: The Ocean-Atmosphere Interaction, Geophys. Monogr., Vol. 147, Amer. Geophys. Union, 189–211.
Wunsch, C., 1977: Response of an equatorial ocean to a periodic monsoon. J. Phys. Oceanogr., 7, 497–511, doi:10.1175/1520-0485(1977)007<0497:ROAEOT>2.0.CO;2.
Wyrtki, K., 1973: An equatorial jet in the Indian Ocean. Science, 181, 262–264, doi:10.1126/science.181.4096.262.
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, doi:10.1175/1520-0442(2002)015<0864:SAMOSI>2.0.CO;2.
Yokoi, T., T. Tozuka, and T. Yamagata, 2008: Seasonal variation of the Seychelles Dome. J. Climate, 21, 3740–3754, doi:10.1175/2008JCLI1957.1.
Yokoi, T., T. Tozuka, and T. Yamagata, 2012: Seasonal and interannual variations of the SST above the Seychelles Dome. J. Climate, 25, 800–814, doi:10.1175/JCLI-D-10-05001.1.
Yoshida, K., 1959: A theory of the Cromwell Current and of the equatorial upwelling. J. Oceanogr. Soc. Japan, 15, 154–170.
Yuan, D., and W. Han, 2006: Roles of equatorial waves and western boundary reflection in the seasonal circulation of the equatorial Indian Ocean. J. Phys. Oceanogr., 36, 930–944, doi:10.1175/JPO2905.1.
