Simulated Multiscale Variations in the Western Tropical Pacific: The Mindanao Dome Revisited

Tomoki Tozuka Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, Tokyo, Japan

Search for other papers by Tomoki Tozuka in
Current site
Google Scholar
PubMed
Close
,
Takashi Kagimoto Institute for Global Change Research/Frontier Research System for Global Change, Tokyo, Japan

Search for other papers by Takashi Kagimoto in
Current site
Google Scholar
PubMed
Close
,
Yukio Masumoto Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, and Institute for Global Change Research/Frontier Research System for Global Change, Tokyo, Japan

Search for other papers by Yukio Masumoto in
Current site
Google Scholar
PubMed
Close
, and
Toshio Yamagata Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, and Institute for Global Change Research/Frontier Research System for Global Change, Tokyo, Japan

Search for other papers by Toshio Yamagata in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Using a high-resolution OGCM result, multiscale variations of the Mindanao Dome (MD) are discussed. The MD is generated by local Ekman upwelling as the positive curl of the Asian winter monsoon increases over the western tropical Pacific, as discussed in the literature. It is shown, however, that the MD decays owing to the Pacific basinwide annual cycle; the warm anomaly that propagates from the eastern tropical Pacific plays an important role in the attenuation of the MD. Also, the negative wind stress curl associated with the northward shift of the ITCZ in the eastern Pacific and the easterly component of the trade winds in the central Pacific strengthens the warm anomaly.

The interannual variation of the MD is governed by variations in both the Ekman upwelling forced locally and the downwelling forced remotely in the east. At its northern fringe, the MD is attenuated by the westward propagating warm eddies on the North Equatorial Current, which are triggered by a negative wind stress curl and amplified through a baroclinic conversion process. This suggests the importance of oceanic internal variabilities in discussing climate variations in the tropical western Pacific.

Corresponding author address: Tomoki Tozuka, Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Email: tozuka@aos.eps.s.u-tokyo.ac.jp

Abstract

Using a high-resolution OGCM result, multiscale variations of the Mindanao Dome (MD) are discussed. The MD is generated by local Ekman upwelling as the positive curl of the Asian winter monsoon increases over the western tropical Pacific, as discussed in the literature. It is shown, however, that the MD decays owing to the Pacific basinwide annual cycle; the warm anomaly that propagates from the eastern tropical Pacific plays an important role in the attenuation of the MD. Also, the negative wind stress curl associated with the northward shift of the ITCZ in the eastern Pacific and the easterly component of the trade winds in the central Pacific strengthens the warm anomaly.

The interannual variation of the MD is governed by variations in both the Ekman upwelling forced locally and the downwelling forced remotely in the east. At its northern fringe, the MD is attenuated by the westward propagating warm eddies on the North Equatorial Current, which are triggered by a negative wind stress curl and amplified through a baroclinic conversion process. This suggests the importance of oceanic internal variabilities in discussing climate variations in the tropical western Pacific.

Corresponding author address: Tomoki Tozuka, Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Email: tozuka@aos.eps.s.u-tokyo.ac.jp

1. Introduction

The western tropical Pacific, where the Mindanao Dome (MD) is located, is considered very important for various reasons. First, it is well known that the heat content of the western tropical Pacific increases prior to the occurrence of El Niño. This warm pool plays an essential role in the turnabout of the El Niño–Southern Oscillation (ENSO) (Philander et al. 1984; Wyrtki 1985, 1989; Masumoto and Yamagata 1991a, hereafter MYa; Wallace et al. 1998; Neelin et al. 1998). Second, it is characterized by a complex near-surface wind-driven circulation of much oceanographic interest (Fig. 1) (Kendall 1989; Toole et al. 1990; Qiu and Joyce 1996; Lukas et al. 1996; Qu et al. 1998). The North Equatorial Current (NEC) meets the Philippine coast and bifurcates near 13°N at the surface (Toole et al. 1990); the bifurcation latitude shifts northward with increasing depth (Qu et al. 1997; Kagimoto and Yamagata 1997). The northern branch forms the origin of the Kuroshio, which is one of the major western boundary currents; its variation is now considered very important in the North Pacific decadal variation (Yasuda et al. 2000; Luo 2001). The southern branch forms the highly variable Mindanao Current (MC) (Wijffels et al. 1995); Toole et al. (1990) observed a dramatic two-fold change in the transport of the MC in six months with similar drastic changes in transport on a timescale of a few days. The MC flows southward into the Celebes Sea and retroflects eastward to form the North Equatorial Countercurrent (NECC), but a part of the MC separates from the southern tip of Mindanao and directly flows into the NECC. Some part of the MC is linked to the Indonesian Throughflow via Makassar Strait (Masumoto and Yamagata 1993, 1996; Gordon 1986; Inoue and Welsh 1993; Miyama et al. 1995; Ilahude and Gordon 1996; Godfrey 1996; Godfrey and Masumoto 1999). This suggests that the area of interest is also pivotal to the global thermohaline circulation (Broecker 1991; Shriver and Hurlburt 1997) and the global climate (Hirst and Godfrey 1993; Schneider 1998). To the south, the New Guinea Coastal Undercurrent (NGCUC) always flows northwestward along the coast of New Guinea, while the surface New Guinea Coastal Current (NGCC) changes direction (Lindstrom et al. 1987; Tsuchiya et al. 1989; Kuroda 2000). These two currents are linked to the Equatorial Undercurrent (EUC) and the South Equatorial Current (SEC). Lastly, since several water masses meet in this area (Bingham and Lukas 1994; Fine et al. 1994; Qu et al. 1999) just like water mass crossroads, a study of the mixing processes may contribute to the improvement of the Lagrangian view of the ocean circulation.

The cyclonic circulation composed of the NEC in the northern flank, the MC in the western flank, and the NECC in the southern flank is often associated with the cold Mindanao Eddy (Takahashi 1959; Masuzawa 1968; Lukas 1988), whose center is located approximately at 7°N, 130°E (see Fig. 1). Masumoto and Yamagata (1991b, hereafter MYb) call the above cold cyclonic circulation west of 160°E the Mindanao Dome, which covers an area much larger than the Mindanao eddy. The cold signal of the MD appears in both upper-layer heat content and subsurface temperature fields. These two terms, the MD and the Mindanao Eddy, are often confused (Udarbe-Walker and Villanoy 2001). The seasonal variation of the MD is studied in MYb using an ocean general circulation model (OGCM) for the first time. According to MYb, two necessary elements for the life cycle of the MD are the local Ekman upwelling induced by the positive curl of the Asian winter monsoon winds, which is important for generation, and westward propagation of downwelling Rossby waves excited in winter by the northeast trade winds farther eastward near 160°E, which is important for decay.

However, neither the MD nor the Mindanao Eddy were always found in past observations (Kashino et al. 1999). For example, while Lukas et al. (1991) identified the MD and the cyclonic eddy at 7°N, 129°E using paths of drifters deployed during the Western Equatorial Pacific Ocean Circulation Study III (WEPOCS III) from July to September 1988, Wijffels et al. (1995) did not find the cyclone at 8°N in the data obtained during the cruise in the spring of 1989. This suggests the existence of strong interannual variations in this particular region. However, so far no study has been done on the interannual variation of the MD or the Mindanao Eddy; this is one of the main motives of the present study.

Just south of the MD, there exists an anticyclonic eddy. Because of its location near Halmahera Island, it is called the Halmahera Eddy (Lukas et al. 1991; Inoue and Welsh 1993; Kashino et al. 1996). Since the model without Halmahera Island does not simulate the Halmahera Eddy, MYb concluded that this small island right on the equator is crucial to the existence of the Halmahera Eddy. Also, the circulation of the Halmahera Eddy plays a significant role in transporting South Pacific tropical water (SPTW) from the Southern Hemisphere to the Northern Hemisphere (Kashino et al. 1996). Therefore, it has a large influence on the water mass entering the Indonesian seas (Kashino et al. 1999; Morey et al. 1999). Although its importance is recognized, no extensive study has been done so far on variations and generation mechanisms of this particular anticyclonic eddy.

In this study, we analyze multiscale variations of the western Pacific low-latitude currents and the MD using a high-resolution OGCM and present a new result, particularly on the decaying process and the interannual variation of the MD. This paper is organized as follows. In the next section, a brief description of the model used in this study and its validation are presented. In section 3, seasonal variations of the western Pacific low-latitude currents and the MD are described, together with a discussion of our new result in contrast to numerical experiments performed by MYb. The interannual variation of the low-latitude currents is described in section 4 and, in particular, a detailed discussion of the interannual variation of the MD is given there. In section 5, we discuss the synoptic event in the spring of 1989 observed by Wijffels et al. (1995) and suggest an instability mechanism for the event. Conclusions are given in the final section.

2. Model description

The high-resolution Pacific basin model used in this study is that of Kagimoto (1999). The model is based on version 2 of the Modular Ocean Model (MOM2), which has been developed at the National Oceanic and Atmospheric Administration (NOAA)/Geophysical Fluid Dynamics Laboratory and the basic equations are given in Pacanowski (1996). Our model covers most of the Pacific basin from 30°S to 65°N, 105°E to 70°W. The horizontal resolution varies from ¼° in the western Pacific to ½° in the eastern Pacific. There are 30 levels in the vertical with 10 levels in the upper 100 m. The bottom topography adopted in this model is based on the 5-min Earth Topography (ETOPO5) dataset and is smoothed to make the numerical calculation stable (Killworth 1987). The lateral eddy viscosity and diffusivity are based on the formula given by Smagorinsky (1963), and vertical eddy viscosity and diffusivity are calculated using the parameterization of Pacanowski and Philander (1981).

The surface boundary condition employed in the present model is almost the same as that of Rosati and Miyakoda (1988). However, we adopted the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) 40-Year Reanalysis data (Kalnay et al. 1996), the cloud cover data estimated by the statistical method (Saito and Baba 1988), and the correction to the net surface heat flux to relax the sea surface temperature to observations (Smith et al. 1996) in 60 days. Also, the surface salinity is restored to the monthly mean climatology (Levitus et al. 1994) with the same relaxation time of 60 days. Near the northern and southern boundaries, the temperature and the salinity are relaxed to the monthly mean climatology within 3° from each boundary (Levitus and Boyer 1994; Levitus et al. 1994).

The model is spun up for 15 years using the NCEP–NCAR climatological data from 1979 to 1997 and then integrated for 19 years using the daily NCEP–NCAR reanalysis data (Kalnay et al. 1996). The simulation data is stored every five days and used for the present analysis. Since the high-resolution OGCM is driven by daily atmospheric data, we are able to obtain a finer image of the ocean both in space and time.

It is necessary to check the validity of the present model results prior to analyzing variations in the tropical region. Kagimoto (1999) used tide gauge data to validate the performance of the simulation; correlation coefficients of hindcast sea levels and observations are quite high (above 0.8) at most of the tide gauge stations, particularly in the tropical Pacific. We have calculated the oceanic heat content of the upper 300 m between 3°N and 10°N at 137°E (Fig. 2a). This special meridian is adopted because it crosses the MD region and also corresponds to one of the Japan Meteorological Agency (JMA) repeat (twice a year) hydrographic sections. The model heat content corresponds remarkably well to the observational data of the JMA. In addition, we note that both model results and observations show minima during the mature phase of El Niño and maxima during La Niña, as discussed in past literature (White et al. 1985; MYa). Also, biennial oscillations are seen in the two time series (Yasunari 1990; MYa,b; Meehl 1997). Since the model result is consistent with observations not only qualitatively but also quantitatively, we proceed further to study interannual variations using the present OGCM result.

In order to check a synoptic picture, we also created a mean current vector field at a depth of 17.5 m for July–September 1988 (Fig. 2b), which corresponds well to averaged drifter velocity vectors for the same period reported by Lukas et al. (1991). General flow patterns around the MD are reproduced well except in the Indonesian seas. The MD and the Halmahera Eddy exist in both the model and the observation. In addition, wavelike structures of the NECC are well captured. Inability to reproduce flow patterns in the Indonesian seas is due to the model configuration; the OGCM used in this study treats only the Pacific basin (cf. Masumoto and Yamagata 1993, 1996; Masumoto et al. 2001).

3. Seasonal variations

a. Western Pacific low-latitude currents

From the 19 years of simulation data, we have created monthly climatology data to discuss the seasonal variation in the Tropics. Figure 3 shows seasonal transport variations of the NEC (across 130°E between 8° and 18°N), the MC (across 8°N), and the Kuroshio (across 18°N) in the upper layer with sigma theta below 26.5. We choose this particular sigma theta value to calculate transports above the thermocline, although it is a little smaller than the value (26.7) adopted by Qu et al. (1998) to analyze the hydrographic data of the World Ocean Circulation Experiment (WOCE). Since the model thermocline is more diffuse than observed due to the model bias, we adopt the level at which the potential density gradient of the OGCM corresponds well to the observed one. As seen in Fig. 3, the NEC and the MC are strongest in May and weakest in fall. The seasonal transport variation of the Kuroshio in this study is consistent with Qu et al. (1998) as well as the model result of MYb, that is, it is strongest in late spring and weakest in fall.

The mean transport from the model is 65 Sv for the NEC, 33 Sv for the MC, and 31 Sv for the Kuroshio (Sv ≡ 106 m3 s−1). The sum of transports of the MC and the Kuroshio is almost equal to the transport of the NEC, as expected. This is due to the fact that the water pumped upward/downward from the subsurface in the area bounded by 8°N, 130°E, 18°N, and the Philippine coast is about two orders of magnitude smaller than the horizontal transports of each current. According to Qu et al. (1998), observed transports of the NEC, the MC, and the Kuroshio are 45 Sv, 28 Sv, and 13 Sv, respectively. Considering that more than half of the hydrographic data are sampled during boreal fall when all three currents take minimum values, the larger values of the mean model transports seem reasonable. In addition, we note that the model transport of the NEC is comparable to the geostrophic transport calculated by Qiu and Joyce (1992) from the JMA repeat hydrographic section data along 137°E. Although the model transport of the Kuroshio is twice as large as the transport estimated by Qu et al. (1998), we believe that the observation still suffers from a coarse spatial resolution; the Kuroshio shows the extremely narrow structure confined near the boundary. Also, note that estimates given by Qu et al. are relative transports and depend on the choice of reference levels.

b. Mindanao Dome

Figure 4 shows the annual march of the simulated subsurface temperature at a depth of 97 m. The cold core of the dome reaches maximum strength in late winter, but weakens from spring to early summer. The position and the seasonal variation of the model dome are consistent with the Levitus monthly climatology (cf. Antonov et al. 1998) shown in Fig. 5. Although the dome area is larger and the temperature of the core is colder by about 1°C throughout a year in the model, our core temperature is consistent with the objectively analyzed temperature field by Udarbe-Walker and Villanoy (2001).

In order to calculate the annual heat budget, we considered an artificial box (5°N–10°N, 127°–140°E) in the upper 110 m off the Philippine coast as shown in Fig. 5; this box (hereafter Box A) covers a major part of the MD. The depth of 110 m is chosen because the mean thermocline depth is about 110 m. As in MYb, we assume that the upwelling in this shallow box is composed of the local Ekman pumping and the remote effect. The remote effect is calculated as the difference between the modeled upwelling speed and the wind-induced Ekman pumping speed. Note that the remote effect term also includes nonlinear interactions. For example, downwelling associated with the Halmahera Eddy generated by internal instability is included. This will be discussed in more detail in section 5. Here, we note that the wind-induced pumping can be decomposed into two parts, such as wE = −curl τ/ρof + βτx/ρof2, where ρo is the density of seawater, f is the Coriolis parameter, and τ is the wind stress. Also, the convergence and diffusion term here is simply calculated as the difference between the tendency term and the surface heat flux term. The averaged temperature in Box A decreases from October to January, where the maximum cooling occurs in December (Fig. 6a). The cooling is associated with the divergence of heat transport, or the upwelling of cold water. This upwelling is caused by the Ekman pumping due to the positive local wind curl during this period (Figs. 6b and 7a). Thus, we confirm that the MD is generated by the cyclonic local wind curl of the Asian winter monsoon, as shown in MYb.

The rate of change of heat storage becomes positive from February through May. This is due to the surface heat flux and the convergence of the heat transport. We suggest that the latter is caused by the intrusion of a warm anomaly generated remotely. Figure 8 shows the time–longitude diagram of the model heat content anomaly (HCA) integrated above a depth of 300 m along 5°N. The warm anomaly originates in the eastern tropical Pacific in late spring to early summer. Fig. 7b shows that, as the remote forcing is the only term causing significant downwelling at the eastern boundary, the warm anomaly is due to the reflection of the downwelling Kelvin wave. Then, the warm anomaly is amplified as a result of Ekman pumping by the curl of the wind stress (Fig. 7c) associated with the northward shift of the intertropical convergence zone (ITCZ) (cf. Umatani and Yamagata 1991). As the northeast trade winds strengthen west of the date line during the boreal winter, the warm anomaly is further forced owing to the Ekman convergence due to the beta effect (Fig. 7a); this effect was discussed in MYb. The warm anomaly propagates farther to the west with a phase speed of about 50 cm s−1, which agrees with the observational value estimated by Mitchum and Lukas (1990). Finally, this warming reaches the MD and contributes to its decay; this is consistent with the heat budget analysis (Fig. 6a) since a large warming begins in March. This warm anomaly corresponds to annual Rossby waves that propagate westward along about 5°N across the whole Pacific basin as studied by Meyers (1979a,b) and Kessler (1990).

The seasonal evolution of the upper-ocean heat content in the western tropical Pacific (the MD region) is very different from that of the South China Sea, which is located west of the Philippines. According to Qu (2001), the mixed layer depth is shallower during spring and summer and is negatively correlated with the SST (higher SST corresponding to shallower mixed layer depth). This is similar to the situation in the eastern equatorial Pacific (Wang and Fu 2001) or the Kuroshio Extension region (Yasuda et al. 2000). On the other hand, SST and the mixed layer depth or upper-ocean heat content is positively correlated in the western tropical Pacific. This is due to the difference in the thermocline depth, which is much deeper in the western tropical Pacific than in the South China Sea.

Before discussing our new view on the seasonal march of the MD, we review two experiments performed by MYb. In their first run (CR1), winds after 15 September were held constant for ten months. On the other hand, in the second run (CR2), the seasonal change of wind stress was assumed only west of 170°E, and the wind stress was kept constant from 15 September for ten months east of the date line. The magnitude of the seasonal wind stress change was linearly increased to the actual value from the date line to 170°E. The time–longitude sections of the HCA integrated above a depth of 310 m at 5°N for CR1 and CR2 are shown in Fig. 9. Since there is no local Ekman upwelling associated with the Asian winter monsoon (Fig. 9a), the MD does not develop in CR1. The Wang et al. (2000) numerical experiment without the western Pacific monsoon does not show the existence of the MD either, thus further supporting our view on the generation mechanism. In contrast to CR1, the life cycle of the MD is almost perfectly simulated by the seasonal wind variation west of 170°E in CR2 (Fig. 9b). Therefore, MYb claim that the MD is generated by the positive wind stress curl of the Asian winter monsoon and is destroyed by downwelling Rossby waves generated around 160°E, both of which are west of 170°E. However, as discussed here, the downwelling Rossby wave from the eastern Pacific plays an important role in the decay process.

This apparent contradiction is mediated as follows. In our model results and CR2, the warm anomaly in the far eastern tropical Pacific is generated before 15 September (Figs. 8, 9b). Since it takes about 10 months to propagate across the whole Pacific basin, the warm anomaly destroys the MD in the following spring, even though there are no wind stress variations assumed east of the date line after 15 September (Fig. 9b). In addition, the intensified northeast trade winds during the boreal winter can intensify the accumulation of the surface water west of the date line off the equator owing to the Ekman convergence as captured by MYb. Thus, our new view on the attenuation process of the MD is consistent with results in MYb.

Wang et al. (2000) also performed another numerical experiment, in which only the annual variation of wind stress in the western Pacific monsoon region is retained. The modeled MD develops locally as expected, but the MD disappears by summer, although there is no westward intrusion of a warm anomaly from the east. This appears to be inconsistent with our study because the decay process cannot be associated with the local Ekman downwelling; their MD begins to decay in March when the wind stress curl over the MD region is still positive (Fig. 6b). It may be attributed to the large diffusivity they adopted, as inferred from the fact that their Kelvin wave reflection at the eastern boundary is significantly suppressed (Wang et al. 1995).

Interestingly, Box A cools again in June and July (Figs. 4 and 6a); the heat budget analysis shows a clear semiannual signal in the MD. This is due to cold anomaly propagation from the eastern Pacific. The cause of this semiannual signal is explained by the combination of the local Ekman upwelling by the Asian winter monsoon, which plays a major role in the development of the MD, and remotely forced Rossby waves, which destroy the MD in spring and cause cooling in the MD region in June and July. Thus, the evolution of the MD is a good indicator of the existence of a semiannual signal in the western tropical Pacific (cf. Wang et al. 2000); it is an outcome of the combination of local responses and basinwide remote influence.

4. Interannual variations

a. Western Pacific low-latitude currents

Figure 10 shows the transport variations of the Kuroshio, the MC, and the NEC, together with low-pass filtered (0.9 yr) time series to capture the interannual variation. The observational analysis (above the reference level at 2000 dbar, corresponding to the sigma theta below 26.7) of Qu et al. (1998) is also shown in the same figure. The low-pass filtered data is calculated using a padding method that adds mirror images of the time series on two sides of the finite-length data; this method is adopted to reduce influences of starting as well as ending on the time series analysis, as in Luo (2001). The model results indicate that both the NEC and the MC are weaker than normal during the mature phase of ENSO events and experience maximum transport in the year before and after the event in the 1980s, which is consistent with Qu et al. (1998).

It is clear from comparison with Fig. 3 that interannual variations of these currents are considerably larger than seasonal variations, as in observations (Lukas 1988). Table 1 reveals that all three currents have significantly larger standard deviations for interannual changes. This unique behavior of western Pacific low-latitude currents is possibly due to the existence of oceanic internal instabilities as well as ocean–atmosphere coupled phenomena such as ENSO. Overall, interannual variations of the modeled western Pacific low-latitude currents are very reasonable and further support the ability of our model.

b. Mindanao Dome

Figure 11 shows the subsurface temperature at a depth of 97 m in the beginning of February in each year from 1980 to 1997, when the MD reaches the mature phase on a climatological basis. Figure 12 shows the same field at the end of May, when the MD is usually decaying. These figures show surprisingly large interannual variations of the MD. Also, we have shown the time–longitude diagram of HCA reflecting the interannual variability in Fig. 13 to compare with the seasonal variation (Fig. 8).

From 1980 to 1981, the evolution of the MD takes a normal course; the MD gradually develops from late October and reaches its maximum in February (Fig. 11). Since the westward propagating warm water from the eastern Pacific reaches the MD region (Fig. 13), the dome is weakened at the end of May (Fig. 12). The analysis of the heat budget and upwelling speed supports this picture (Figs. 14a,b). A strong upwelling event takes place in the core during early winter mainly by the local Ekman pumping as shown in section 3. Then, the downwelling associated with the remote forcing causes warming of the core from February on. However, development of the MD is constrained or delayed when the remote forcing introduces strong downwelling in this area during the formation period. This is the case in 1984/85. The average rate of downwelling due to the remote forcing from November 1984 to January 1985 is 7.7 × 10−4 cm s−1 (Fig. 15b), approximately twice as large as the climatological mean value. Actually, the remotely forced downwelling is stronger than the mean for the whole winter (Fig. 6b). Therefore, as seen in Fig. 11, the MD develops much slower than in a normal year. Figure 13 further supports this view; there exists strong westward propagation of warm water from the eastern Pacific. This picture also agrees with the conventional idea that warm water accumulates around the western tropical Pacific just prior to warm ENSO events (Wyrtki 1985; Yamagata and Masumoto 1989) as in 1981/82.

Since the remotely forced downwelling is very weak during spring and early summer in 1991, the MD survives longer than in a normal year (Fig. 15d). The downwelling associated with the remote forcing from February through May in 1991 is only about two-thirds of the climatological value for the same months. Therefore, the dome is well defined, even at the end of May (Fig. 12) when the dome is lessened in the normal year (Figs. 4, 5). This is also confirmed by the time–longitude diagram of HCA (Fig. 13). The warm anomaly from the eastern Pacific does not reach the western Pacific in 1991; it decays near 160°E. Similar behavior is also seen in 1987–88 and is consistent with Lukas et al. (1991), who show the existence of a large cyclonic gyre in the MD region using drifter data from July to September in 1988 (see Fig. 2b).

The importance of local Ekman pumping in the generation of the MD is further demonstrated by analyzing the 1986/87 event. The development of the MD is significantly delayed (Fig. 11) due to abnormally strong downwelling caused by local Ekman pumping before the formation period (Fig. 15c). However, extremely strong local Ekman upwelling after February induces cooling so that the MD develops. The MD is more developed in May (Fig. 12) than in February (Fig. 11).

The relationship between the Asian winter monsoon and the ENSO is very interesting. Since the 1982/83 event corresponds to an El Niño year, the sea temperature of the western tropical Pacific is colder than normal and the temperature difference between the Asian continent and the ocean is smaller. Therefore, the Asian winter monsoon, due essentially to the land–sea temperature contrast, is weaker than normal (cf. Hanawa et al. 1989), but is not weak enough to obstruct the development of the MD. Consequently, the MD is well developed in 1982/83 (Fig. 11). Although there is a general tendency for the seasonal signal to become weaker during the mature phase of El Niño, the MD shows distinct seasonal variations. Since the warm water moves to the east, the thermocline is shallower during the mature phase of El Niño. This leads to more active influence of subsurface cold water on the surface mixed layer, which is a key to the generation of the MD. The evolution of the dome in the El Niño event of 1991/92 is also similar (Fig. 15e).

There are some exceptions where the evolution of the MD cannot be explained by the aforementioned two necessary elements. In the 1993/94 winter, the strength of the local Ekman upwelling and the remotely forced downwelling are almost the same (Fig. 15f); this implies that the dome cannot evolve. However, the dome exists in the beginning of February (Fig. 11) because the dome is not destroyed in 1993, suggesting hysteresis. As shown in Fig. 13, the warm anomaly does not reach the MD region in 1993.

In 1982/83, it appears that the dome evolution takes a normal course in terms of the upwelling (Fig. 15a). However, even though the westward propagation of warm water from the eastern Pacific does not exist (Fig. 13), the MD disappears before May (Fig. 12). In this particular year, the northward movement of the warm Halmahera Eddy plays a major role in the decaying process of the MD. This synoptic event related to the Halmahera Eddy is described in detail in the next section.

c. Decay process of the Mindanao Dome at its northern fringe

Figure 16 shows the time–longitude diagram of the simulated sea surface height along 13°N, which is along the axis of the NEC and the northern flank of the MD. Large positive surface height signals, which propagate westward, originate near 160°W and evolve near 160°E. Here particularly, we discuss the 1992 event. Figure 17a shows the subsurface current vector field at a depth of 97 m in April 1992. There are a series of warm eddies lined up along the NEC in the western tropical Pacific. Eddies that propagate westward and reach the Philippine coast are responsible for decay of the MD at its northern flank.

In order to investigate the generation mechanism of these anticyclonic eddies, the wind stress field is first examined (Fig. 17b). The NEC axis is close to the zero wind stress curl line, as expected from the basic theory of wind-driven ocean circulation (Stommel 1965). We also note that there is an interesting structure near the Hawaiian Islands with an array of positive and negative wind stress curl patterns. In particular, a large negative curl region exists to the south of Hawaii between 11° and 17°N. This is due to a topographic effect of the Hawaiian Islands on the northeast trade winds (see Trenberth et al. 1990). Since the Ekman upwelling speed is given by wE = −curl (τ/ρof), substitution of ρo = 1025 kg m−2, f = 3.27 × 10−5 s−1 (at 13°N), and curl τ = −2.5 × 10−9 dyn cm−3 (the largest negative values of the curl in the NEC region, which is located south of the Hawaiian Islands) leads to the upwelling speed of wE = 7.5 × 10−4 cm s−1. This explains only 38 m of the total depression of isopycnal under the assumption that these eddies are generated in 2 months with a nonlinear eddy-shedding process (cf. Yamagata et al. 1990; Aiki and Yamagata 2000). Therefore, the regional wind stress effect alone cannot explain the oceanic isopycnal depression (about 100 m) associated with eddies.

One candidate for the amplification mechanism is the oceanic internal instability. Calculating eddy kinetic energy budget as in Masina et al. (1999) and Wells et al. (2000) (see the appendix for detail) may lead to understanding of the mechanism. Figure 18a shows the 3-month mean eddy kinetic energy (EKE) along 145°E for April 1992. A large EKE is found in the NEC region with a maximum of 1000 cm2 s−2 in the upper 100 m. Also, the EKE is very large in the equatorial region, which is associated with the eddy activity of the NECC and the EUC. Figures 18b and 18c show the baroclinic conversion and the barotropic conversion, respectively. The large baroclinic conversion is taking place in the NEC at the depth from 100 to 400 m (Fig. 18b), which corresponds to the area of large EKE. Compared to the baroclinic conversion, the barotropic conversion is negligible in the NEC region (Fig. 18c). These anticyclonic eddies, triggered by winds and enhanced through the baroclinic energy transfer, propagate westward.

Notice that the eddy activity in the NEC is most active in spring. When the subtropical gyre system spins up from winter to spring (Trenberth et al. 1990), the NEC becomes baroclinically unstable so that the baroclinic conversion takes place in the NEC. This picture is consistent with Fig. 3, in which the transport of the NEC becomes largest in spring. As discussed in Gill et al. (1974), the action of the large-scale wind field increases the available potential energy in the large-scale ocean circulation, which may be released in the form of the EKE via baroclinic conversion. This baroclinic instability associated with the seasonal march of the NEC either generates or amplifies anticyclonic eddies near 160°E. Although observations are still poor, we believe that the scenario developed here is worth pursuing.

Therefore, the weakening of the MD at its northern fringe is due to the intrusion of warm eddies formed in the NEC. Although the baroclinic instability of the NEC and the direct negative vorticity input by the wind stress are crucial to this process, the latter may explain why most of these eddies formed in the NEC are anticyclonic.

5. A synoptic event

According to Wijffels et al. (1995), a warm eddy is observed in the MD region during a cruise in the spring of 1989. They suggest that this eddy is formed near 5°N in the confluence zone of the SEC and the MC because the oxygen and salinity in the core of the eddy resemble those of the SPTW. Since our model captures this warm eddy, we focus on this interesting synoptic event in this section.

Figure 19 shows the upper 300-m HCA from 11 February to 2 April 1989 with 10-day intervals. A zonally elongated warm eddy exists along 5°N in February and then splits into two eddies in March. Two warm eddies gradually expand northward and, by the beginning of April, the MD region is occupied by meridionally elongated warm eddies. These results are consistent with the observational result reported by Wijffels et al. (1995) in that the MD region is occupied by warm eddies in the spring of 1989. In addition, high salinity water of Southern Hemispheric origin occupies the core of the eddy in our model (Fig. 20). The maximum value of about 35.2 psu in the core at the depth of around 140 m is equivalent to the value obtained in the observation (see Fig. 10 of Wijffels et al. 1995). The high salinity water of Southern Hemispheric origin normally remains south of the NECC, but during this event the water is transported as far north as 10°N. The eddy is a deep structure with salinity higher than in the surrounding region at a depth between 50 and 250 m, and the meridional velocity signal extends below the depth of 500 m. Similar events in which pinched-off warm anticyclonic eddies carry salty water to the north take place about 10 times during our 19 years of model run, and all events take place in spring and summer. Since a higher-resolution model sheds eddies more frequently, as in observations (Qiu et al. 1999; Masumoto et al. 2001), we suggest that in reality more anticyclonic eddies may be pinched off to transport South Pacific water to the north. Godfrey and Wilkin (1995) also suggest the possibility of interhemispheric water transport by pinched-off warm eddies. Similar phenomena in the Atlantic are North Brazil Current Retroflection eddies (Johns et al. 1990; Richardson et al. 1994; Nof and Pichevin 1996) and the Agulhas Current Retroflection eddies (Lutjeharms and Gordon 1987; de Ruijter et al. 1999; Wells et al. 2000).

Figure 21a shows the model EKE along 131°E, which passes cores of the Halmahera and Mindanao Eddies. As expected, there are two strong peaks above 3000 cm2 s−2 in the EKE, which correspond to the Mindanao and Halmahera Eddies (Fig. 18a). Although the baroclinic conversion always amplifies the Halmahera Eddy (Fig. 21b), the barotropic conversion term changes sign as time elapses. During early winter, when the Halmahera Eddy develops, the barotropic conversion term is positive and acts as the main source for the large EKE (Fig. 21c). Later in the winter, it remains positive in the Mindanao Eddy, while it becomes negative for the Halmahera Eddy (Fig. 21d). Thus, the barotropic conversion plays a major role to strengthen the Halmahera Eddy during its growing phase. Once the eddy matures, however, the term starts strengthening the mean flow of the NECC. The instability analysis for the seasonal cycle also gives exactly the same mechanism (not shown).

6. Conclusions

Using data from a high-resolution OGCM, we have clarified various multiscale variations in the MD region. The model result is consistent with various observational data in terms of the heat content along 137°E, synoptic current patterns, and transport variations of western Pacific low-latitude currents.

From the present study, we have proposed a new scenario for the seasonal variation of the MD (Fig. 22a). It is the interaction between the local seasonal cycle and the basin-scale seasonal cycle that determines the evolution of the MD. The MD develops locally in the western tropical Pacific during boreal winter owing to the local Ekman upwelling induced by the Asian winter monsoon. The warm anomaly, which is amplified in association with the northward migration of the ITCZ in the eastern tropical Pacific in the previous year and by the Ekman downwelling associated with zonal wind in the central tropical Pacific, propagates westward and attenuates the MD. We suggest for the first time that propagation of a warm anomaly from the eastern tropical Pacific plays an important role in the attenuation of the MD. In order to understand the mechanism, even for the seasonal variation of the western tropical Pacific including the MD, we need to take into account the whole Pacific basin. Since the interannual variation of the MD is also governed by changes in the local Ekman pumping and the remotely forced downwelling, as shown in Fig. 22b, we emphasize that the understanding of the seasonal cycle is very important in understanding the interannual variation (cf. Rasmusson et al. 1999). Also, it is interesting to note that the northward migration of the ITCZ, which is crucial to the maintenance of the matured Costa Rica Dome (Umatani and Yamagata 1991), plays an important role in the decay process of the MD. Warm eddies in the NEC, which are triggered by the negative wind stress curl south of the Hawaiian Islands and amplified through the baroclinic conversion process, contribute to the decay process of the MD at its northern flank.

We have discussed, in particular, 1988/89 event described in Wijffels et al. (1995). The present model result suggests that the warm eddy found in the MD region originates south of the NECC and carries high salinity water of Southern Hemispheric origin. We have confirmed that the Halmahera Eddy provides the medium of the interhemispheric water exchange. The instability analysis shows that the Halmahera Eddy is generated by both barotropic and baroclinic conversion in the model. However, once it matures, it strengthens the mean flow of the NECC through the barotropic conversion process. Further observations of the subsurface temperature and salinity fields using subsurface floats and other in situ observations will help us clarify the importance of the Halmahera Eddy in terms of the water mass transport and interhemispheric water exchange.

Because of lack of systematic, high-density oceanic measurements, it is rather difficult to compare the above model results suggesting the importance of oceanic internal variabilities with observations in a precise way. Thus, the role of the present study will lie in its usefulness for designing an observational system in this particular region of the world oceans.

Acknowledgments

We are indebted to the Japan Meteorological Agency (JMA) for the repeat hydrographic section data and to NCEP–NCAR for the wind stress data used in the present study. This study benefited from fruitful discussion with Gary Meyers. We are also grateful to Toshiyuki Hibiya and Masahide Kimoto for valuable suggestions. Useful comments made by two anonymous reviewers helped us to improve our manuscript.

REFERENCES

  • Aiki, H., and T. Yamagata, 2000: Successive formation of planetary lenses in the intermediate layer. Geophys. Astrophys. Fluid Dyn., 92 , 129.

    • Search Google Scholar
    • Export Citation
  • Antonov, J., S. Levitus, T. P. Boyer, M. Conkright, T. O'Brien, and C. Stephens, 1998: Temperature of the Pacific Ocean. World Ocean Atlas 1998, Vol. 2,. NOAA Atlas NESDIS 28, 166 pp.

    • Search Google Scholar
    • Export Citation
  • Bingham, F. M., and R. Lukas, 1994: The southward intrusion of North Pacific Intermediate Water along the Mindanao coast. J. Phys. Oceanogr., 24 , 141154.

    • Search Google Scholar
    • Export Citation
  • Broecker, W., 1991: The great ocean conveyor. Oceanography, 4 , 7989.

  • de Ruijter, W. P. M., P. J. van Leeuwen, and J. R. E. Lutjeharms, 1999: Generation and evolution of Natal pulses: Solitary meanders in the Agulhas Current. J. Phys. Oceanogr., 29 , 30433055.

    • Search Google Scholar
    • Export Citation
  • Fine, R. A., R. Lukas, F. M. Bingham, M. J. Warner, and R. H. Gammon, 1994: The western Pacific: A water mass crossroads. J. Geophys. Res., 99 , 2506325080.

    • Search Google Scholar
    • Export Citation
  • Gill, A. E., J. S. A. Green, and A. J. Simmons, 1974: Energy partition in the large-scale ocean circulation and the production of mid-ocean eddies. Deep-Sea Res., 21 , 499528.

    • Search Google Scholar
    • Export Citation
  • 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 , 1221712237.

    • Search Google Scholar
    • Export Citation
  • Godfrey, J. S., and J. L. Wilkin, 1995: Reply. J. Phys. Oceanogr., 25 , 15681570.

  • Godfrey, J. S., and Y. Masumoto, 1999: Diagnosing the mean strength of the Indonesian throughflow in an ocean general circulation model. J. Geophys. Res., 104 , 78897895.

    • Search Google Scholar
    • Export Citation
  • Gordon, A. L., 1986: Interocean exchange of thermocline water. J. Geophys. Res., 91 , 50375046.

  • Hanawa, K., Y. Yoshikawa, and T. Watanabe, 1989: Composite analyses of wintertime wind stress vector fields with respect to SST anomalies in the western North Pacific and the ENSO events. Part I: SST composite. J. Meteor. Soc. Japan, 67 , 385400.

    • Search Google Scholar
    • Export Citation
  • Hirst, A. C., and J. S. Godfrey, 1993: The role of the Indonesian Throughflow in a global ocean model. J. Phys. Oceanogr., 23 , 10571086.

    • Search Google Scholar
    • Export Citation
  • Ilahude, A. G., and A. L. Gordon, 1996: Thermocline stratification within the Indonesian seas. J. Geophys. Res., 101 , 1240112409.

  • Inoue, M., and S. E. Welsh, 1993: Modeling seasonal variability in the wind-driven upper-layer circulation in the Indo–Pacific region. J. Phys. Oceanogr., 23 , 14111436.

    • Search Google Scholar
    • Export Citation
  • Johns, W. E., T. N. Lee, F. Schott, R. J. Zantopp, and R. H. Evans, 1990: The North Brazil Current retroflection: Seasonal structure and eddy variability. J. Geophys. Res., 95 , 2210322120.

    • Search Google Scholar
    • Export Citation
  • Kagimoto, T., 1999: Numerical study on transport variations of the Kuroshio. Ph.D. thesis, The University of Tokyo, 75 pp.

  • Kagimoto, T., and T. Yamagata, 1997: Seasonal transport variations of the Kuroshio: An OGCM simulation. J. Phys. Oceanogr., 23 , 403418.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors. 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kashino, Y., M. Aoyama, T. Kawano, N. Hendiarti, S. Y. Anantasena, K. Muneyama, and H. Watanabe, 1996: The water masses between Mindanao and New Guinea. J. Geophys. Res., 101 , 1239112400.

    • Search Google Scholar
    • Export Citation
  • Kashino, Y., H. Watanabe, B. Herunadi, M. Aoyama, and D. Hartoyo, 1999: Current variability at the Pacific entrance of the Indonesian Throughflow. J. Geophys. Res., 104 , 1102111035.

    • Search Google Scholar
    • Export Citation
  • Kendall, T. R., 1989: Fluctuations of transports and sea level in the western boundary region of the tropical Pacific Ocean. J. Oceanogr. Soc. Japan, 45 , 279287.

    • Search Google Scholar
    • Export Citation
  • Kessler, W. S., 1990: Observations of long Rossby waves in the northern tropical Pacific. J. Geophys. Res., 95 , 51835217.

  • Killworth, P. D., 1987: Topographic instabilities in level model OGCMs. Ocean Modelling, 75 , (unpublished manuscripts),. 912.

  • Kuroda, Y., 2000: Variability of currents off the northern coast of New Guinea. J. Oceanogr., 56 , 103116.

  • Levitus, S., and T. P. Boyer, 1994: Temperature. World Ocean Atlas, Vol. 4. NOAA Atlas NESDIS 4, 117 pp.

  • Levitus, S., R. Burgett, and T. P. Boyer, 1994: Salinity. World Ocean Atlas, Vol. 3:. NOAA Atlas NESDIS 3, 99 pp.

  • Lindstrom, E., R. Lukas, R. Fine, E. Firing, S. Godfrey, G. Meyers, and M. Tsuchiya, 1987: The Western Equatorial Pacific Ocean Circulation Study. Nature, 330 , 533537.

    • Search Google Scholar
    • Export Citation
  • Lukas, R., 1988: Interannual fluctuations of the Mindanao Current inferred from sea level. J. Geophys. Res., 93 , 67446748.

  • Lukas, R., E. Firing, P. Hacker, P. L. Richardson, C. A. Collins, R. Fine, and R. Gammon, 1991: Observations of the Mindanao Current during the Western Equatorial Pacific Ocean Circulation Study. J. Geophys. Res., 96 , 70897104.

    • Search Google Scholar
    • Export Citation
  • Lukas, R., T. Yamagata, and J. P. McCreary, 1996: Pacific low-latitude western boundary currents and the Indonesian Throughflow. J. Geophys. Res., 101 , 1220912216.

    • Search Google Scholar
    • Export Citation
  • Luo, J-J., 2001: A study on long-term climate variations in the Pacific. Ph.D. thesis, University of Tokyo, 130 pp.

  • Lutjeharms, J. R. E., and A. L. Gordon, 1987: Shedding of an Agulhas ring observed at sea. Nature, 325 , 138140.

  • Masina, S., S. G. H. Philander, and A. B. G. Bush, 1999: An analysis of tropical instability waves in a numerical model of the Pacific Ocean. 2. Generation and energetics of the waves. J. Geophys. Res., 104 , 2963729661.

    • Search Google Scholar
    • Export Citation
  • Masumoto, Y., and T. Yamagata, 1991a: On the origin of a model ENSO in the Western Pacific. J. Meteor. Soc. Japan, 69 , 197207.

  • Masumoto, Y., . 1991b: Response of the western tropical Pacific to the Asian winter monsoon: The generation of the Mindanao Dome. J. Phys. Oceanogr., 21 , 13861398.

    • Search Google Scholar
    • Export Citation
  • Masumoto, Y., . 1993: Simulated seasonal circulation in the Indonesian seas. J. Geophys. Res., 98 , 1250112509.

  • Masumoto, Y., . 1996: Seasonal variations of the Indonesian throughflow in a general ocean circulation model. J. Geophys. Res., 101 , 1228712293.

    • Search Google Scholar
    • Export Citation
  • Masumoto, Y., T. Kagimoto, M. Yoshida, M. Fukuda, N. Hirose, and T. Yamagata, 2001: Intraseasonal eddies in the Sulawesi Sea simulated in an ocean general circulation model. Geophys. Res. Lett.,28, 1631–1634.

    • Search Google Scholar
    • Export Citation
  • Masuzawa, J., 1968: Second cruise for CSK Ryofu Maru, January to March 1968. Oceanogr. Mag., 20 , 173185.

  • Meehl, G. A., 1997: The south Asian monsoon and the tropospheric biennial oscillation. J. Climate, 10 , 19211943.

  • Meyers, G., 1979a: Annual variation in the slope of the 14°C isotherm along the equator in the Pacific Ocean. J. Phys. Oceanogr., 9 , 885891.

    • Search Google Scholar
    • Export Citation
  • Meyers, G., . 1979b: On the annual Rossby wave in the tropical North Pacific Ocean. J. Phys. Oceanogr., 9 , 663674.

  • Mitchum, G. T., and R. Lukas, 1990: Westward propagation of annual sea level and wind signals in the western Pacific Ocean. J. Climate, 3 , 11021110.

    • Search Google Scholar
    • Export Citation
  • Miyama, T., T. Awaji, K. Akitomo, and N. Imasato, 1995: Study of seasonal transport variations in the Indonesian seas. J. Geophys. Res., 100 , 2051720541.

    • Search Google Scholar
    • Export Citation
  • Morey, S. L., J. F. Shriver, and J. J. O'Brien, 1999: The effect of Halmahera on the Indonesian Throughflow. J. Geophys. Res., 104 , 2328123296.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., D. S. Battisti, A. C. Hirst, F. F. Jin, Y. Wakata, T. Yamagata, and S. Zebiak, 1998: ENSO theory. J. Geophys. Res., 103 , 1426114290.

    • Search Google Scholar
    • Export Citation
  • Nof, D., and T. Pichevin, 1996: The retroflection paradox. J. Phys. Oceanogr., 26 , 23442358.

  • Pacanowski, R. C., 1996: MOM2 documentation, user's guide and reference manual. NOAA/GFDL Ocean Tech. Rep. 3.1. [Available from GFDL, Princeton University, Princeton, NJ 08542.].

    • Search Google Scholar
    • Export Citation
  • Pacanowski, R. C., and S. G. H. Philander, 1981: Parameterization of vertical mixing in numerical models of tropical oceans. J. Phys. Oceanogr., 11 , 14431451.

    • Search Google Scholar
    • Export Citation
  • Philander, S. G. H., T. Yamagata, and R. C. Pacanowski, 1984: Unstable air–sea interaction in the Tropics. J. Atmos. Sci., 41 , 604613.

    • Search Google Scholar
    • Export Citation
  • Qiu, B., and T. Joyce, 1992: Interannual variability in the mid- and low-latitude western North Pacific. J. Phys. Oceanogr., 22 , 10621079.

    • Search Google Scholar
    • Export Citation
  • Qiu, B., M. Mao, and Y. Kashino, 1999: Intraseasonal variability in the Indo-Pacific Throughflow and the regions surrounding the Indonesian seas. J. Phys. Oceanogr., 29 , 15991618.

    • Search Google Scholar
    • Export Citation
  • Qu, T., 2001: Role of ocean dynamics in determining the mean seasonal cycle of the South China Sea surface temperature. J. Geophys. Res., 106 , 69436955.

    • Search Google Scholar
    • Export Citation
  • Qu, T., T. Kagimoto, and T. Yamagata, 1997: A subsurface countercurrent along the east coast of Luzon. Deep-Sea Res., 44 , 413423.

  • Qu, T., H. Mitsudera, and T. Yamagata, 1998: On the western boundary currents in the Philippine Sea. J. Geophys. Res., 103 , 75377548.

    • Search Google Scholar
    • Export Citation
  • Qu, T., . 1999: A climatology of the circulation and water mass distribution near the Philippine coast. J. Phys. Oceanogr., 29 , 14881505.

    • Search Google Scholar
    • Export Citation
  • Rasmusson, E. M., M. Chelliah, and C. F. Ropelewski, 1999: The observed climate of the 20th century. Modeling the Earth's Climate and its Variability, W. R. Holland, S. Joussaume, and F. David, Eds., Elsevier, 567 pp.

    • Search Google Scholar
    • Export Citation
  • Richardson, P. L., G. E. Hufford, R. Limebumer, and W. S. Brown, 1994: North Brazil Current retroflection eddies. J. Geophys. Res., 99 , 50815093.

    • Search Google Scholar
    • Export Citation
  • Rosati, A., and K. Miyakoda, 1988: A general circulation model for upper ocean simulation. J. Phys. Oceanogr., 18 , 16011626.

  • Saito, K., and A. Baba, 1988: Statistical relation between relative humidity and the GMS observed cloud amount. J. Meteor. Soc. Japan, 66 , 187192.

    • Search Google Scholar
    • Export Citation
  • Schneider, N., 1998: The Indonesian Throughflow and the global climate system. J. Climate, 11 , 676689.

  • Shriver, J. F., and H. E. Hurlburt, 1997: The contribution of the global thermohaline circulation to the Pacific to Indian Ocean Throughflow via Indonesia. J. Geophys. Res., 102 , 54915511.

    • Search Google Scholar
    • Export Citation
  • Smagorinsky, J., 1963: General circulation experiments with the primitive equations. I. The basic experiments. Mon. Wea. Rev., 91 , 99164.

    • Search Google Scholar
    • Export Citation
  • Smith, T. M., R. W. Reynolds, R. E. Livezey, and D. C. Stokes, 1996: Reconstruction of historical sea surface temperature using empirical orthogonal functions. J. Climate, 9 , 14031420.

    • Search Google Scholar
    • Export Citation
  • Stommel, H., 1965: The Gulf Stream: A Physical and Dynamical Description. 2d ed., University of California Press, 248 pp.

  • Takahashi, T., 1959: Hydrographical researches in the western equatorial Pacific. Mem. Faculty Fish. Kagoshima Univ., 7 , 141147.

  • Toole, J. M., R. C. Millard, Z. Wang, and S. Pu, 1990: Observations of the Pacific North Equatorial Current bifurcation at the Philippine coast. J. Phys. Oceanogr., 20 , 307318.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., W. G. Large, and J. G. Olson, 1990: The mean annual cycle in global ocean wind stress. J. Phys. Oceanogr., 20 , 17421760.

    • Search Google Scholar
    • Export Citation
  • Tsuchiya, M., R. Lukas, R. A. Fine, E. Firing, and E. Lindstrom, 1989: Source waters of the Pacific Equatorial Undercurrent. Progress in Oceanography, Vol. 23, Pergamon, 101–147.

    • Search Google Scholar
    • Export Citation
  • Udarbe-Walker, M. J. B., and C. L. Villanoy, 2001: Structure of potential upwelling areas in the Philippines. Deep-Sea Res., 48 , 14991518.

    • Search Google Scholar
    • Export Citation
  • Umatani, S., and T. Yamagata, 1991: Response of the eastern tropical Pacific to meridional migration of the ITCZ: The generation of the Costa Rica Dome. J. Phys. Oceanogr., 21 , 346363.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., E. M. Rasmusson, T. P. Mitchell, V. E. Kousky, E. S. Sarachik, and H. von Storch, 1998: On the structure and evolution of ENSO-related climate variability in the tropical Pacific: Lessons from TOGA. J. Geophys. Res., 103 , 1424114259.

    • Search Google Scholar
    • Export Citation
  • Wang, B., and X. Fu, 2001: Processes determining the rapid reestablishment of the equatorial Pacific cold tongue/ITCZ complex. J. Climate, 14 , 22502265.

    • Search Google Scholar
    • Export Citation
  • Wang, B., T. Li, and P. Chang, 1995: An intermediate model of the tropical Pacific Ocean. J. Phys. Oceanogr., 25 , 15991616.

  • Wang, B., R. Wu, and R. Lukas, 2000: Annual adjustment of the thermocline in the tropical Pacific Ocean. J. Climate, 13 , 596616.

  • Wells, N. C., V. O. Ivchenko, and S. E. Best, 2000: Instabilities in the Agulhas Retroflection Current system: A comparative model study. J. Geophys. Res., 105 , 32333241.

    • Search Google Scholar
    • Export Citation
  • White, W. B., G. A. Meyers, J. R. Donguy, and S. E. Pazan, 1985: Short-term climatic variability in the thermal structure of the Pacific Ocean during 1972–82. J. Phys. Oceanogr., 15 , 917935.

    • Search Google Scholar
    • Export Citation
  • Wijffels, S., E. Firing, and J. Toole, 1995: The mean structure and variability of the Mindanao Current at 8°N. J. Geophys. Res., 100 , 1842118435.

    • Search Google Scholar
    • Export Citation
  • Wyrtki, K., 1985: Water displacements in the Pacific and the genesis of El Niño cycles. J. Geophys. Res., 90 , 71297132.

  • Wyrtki, K., . 1989: Some thoughts about the west Pacific warm pool. Proc. Western Pacific International Meeting and Workshop on TOGA–COARE, Noumea, New Caledonia, ORSTOM, 99–109.

    • Search Google Scholar
    • Export Citation
  • Yamagata, T., and Y. Masumoto, 1989: A simple ocean–atmosphere coupled model for the origin of a warm El Niño–Southern Oscillation event. Philos. Trans. Roy. Soc. London, A329 , 225236.

    • Search Google Scholar
    • Export Citation
  • Yamagata, T., K. Sakamoto, and M. Arai, 1990: Locally-induced nonlinear modes and multiple equilibria in planetary fluids. Pure Appl. Geophys., 133 , 733748.

    • Search Google Scholar
    • Export Citation
  • Yasuda, I., T. Tozuka, M. Noto, and S. Kouketsu, 2000: Heat balance and regime shifts of the mixed layer in the Kuroshio Extension. Progress in Oceanography, Vol. 47, Pergamon, 257–278.

    • Search Google Scholar
    • Export Citation
  • Yasunari, T., 1990: Impact of the Indian monsoon on the coupled atmosphere–ocean system in the tropical Pacific. Meteor. Atmos. Phys., 44 , 2942.

    • Search Google Scholar
    • Export Citation

APPENDIX

Instability Analysis

The horizontal equations of motion are written as
i1520-0485-32-5-1338-ea1
Decomposing each variable into a time independent part and its deviation, such as
i1520-0485-32-5-1338-ea2
and applying the above decomposition, we obtain the equation for the mean field
i1520-0485-32-5-1338-ea5
where
i1520-0485-32-5-1338-ea6
and the equation for the time deviation,
i1520-0485-32-5-1338-ea7
Multiplying (A7) by v′, we obtain the equation for the eddy kinetic energy as in the following:
i1520-0485-32-5-1338-ea8
Using the hydrostatic approximation and the continuity equation, the pressure work term on the right-hand side is rewritten as
vpvpwpzw
When the third term on the right-hand side (−w′) (which represents work done by buoyancy) is positive, it converts available potential energy into eddy kinetic energy (i.e., baroclinic conversion). On the other hand, when the horizontal deformation work term [second term on the right-hand side of (A8)] is positive, it represents the conversion of mean kinetic energy into eddy kinetic energy (i.e., barotropic conversion).

Fig. 1.
Fig. 1.

Enlarged map of the western tropical Pacific with a schematic diagram of the horizontal circulation and the Mindanao Dome.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 2.
Fig. 2.

(a) Oceanic heat content of the upper 300 m at 3°–10°N, 137°E for the OGCM (solid line) and observations along 137°E by the JMA (dots). (b) (top) Mean current vector field at a depth of 17.5 m for Jul–Sep 1988. (bottom) Drifter velocity vectors averaged by 1° squares, Jul–Sep 1988. ADCP current observations are included. Gridded data have been smoothed. (Adapted from Lukas et al. 1991)

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 3.
Fig. 3.

Time series of transport (sigma theta below 26.5) of the NEC, the MC, and the Kuroshio. All values are in Sv.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 4.
Fig. 4.

Annual march of the model subsurface temperatures at a depth of 97 m. Contour interval is 1°C. Temperatures less than 22°C are shaded.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 5.
Fig. 5.

Annual cycle of the subsurface temperatures at a depth of 100 m from Antonov et al. (1998) monthly climatology. Contour interval is 1°C. Temperatures less than 22°C are shaded. Box A (5°–10°N, 127°–140°E) is shown in the Jan temperature field.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 6.
Fig. 6.

(a) Monthly heat budget analysis for Box A. The rate of change of heat storage is determined by the convergence of heat transport and diffusion, and the surface flux. (b) Monthly upwelling speed for Box A. “Upwelling” (solid) denotes modeled upwelling speed at a depth of 120 m, “Curl” (dotted) denotes the Ekman pumping by wind stress curl, “Beta” (dashed) denotes Ekman pumping by zonal wind, and “Remote” (dash–dotted) denotes the remotely forced upwelling. Positive values mean upwelling.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 7.
Fig. 7.

Ekman pumping by wind curl, Ekman pumping by beta effect (zonal wind), total Ekman pumping, remote forcing, and total upwelling for (a) Jan, (b) May, and (c) Jul. Areas with negative values, which correspond to downwelling, are shaded. Contour interval is 0.0002 cm s−1.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

i1520-0485-32-5-1338-f702
Fig. 7.
Fig. 7.

(Continued)

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 8.
Fig. 8.

Time–longitude diagram of model HCA from 130°E to 80°W along (a) 5°N and (b) the equator. Contour interval is 3 × 108 J m−2. Negative anomalies are shaded.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 9.
Fig. 9.

Time–longitude diagram of the heat content anomaly integrated above a depth of 310 m at 5°N for (a) CR1 and (b) CR2 of MYb. Negative anomalies are shaded. The contour interval is 5 × 103 cal cm−2 [adapted from Figs. 12 and 13 of MYb].

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 10.
Fig. 10.

Interannual variations of transports (sigma theta below 26.5) for (a) the Kuroshio, (b) the MC, and (c) the NEC, from 1979 to 1997 (solid line). Also shown are low-pass filtered (0.9 yr) data (dashed line) and observation data of Qu et al. (1998) (dotted line). All values are in Sv.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 11.
Fig. 11.

Subsurface temperature at a depth of 97 m in the beginning of Feb from 1980 to 1997. Contour interval is 1°C. Temperatures less than 22°C are shaded.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 12.
Fig. 12.

As in Fig. 11 but for the end of May. Contour interval is 1°C. Temperatures less than 22°C are shaded.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 13.
Fig. 13.

As in Fig. 8 but from 1980 to 1997, and contour interval is 1 × 109 J m−2.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 14.
Fig. 14.

(a) Heat budget analysis, and (b) upwelling speed of Box;thA for 1980/81.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 15.
Fig. 15.

Upwelling speed of Box A for (a) 1982/83, (b) 1984/85, (c) 1986/87, (d) 1990/91, (e) 1991/92, and (f) 1993/94.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 16.
Fig. 16.

Time–longitude section of the sea surface height along 13°N from 1979 to 1997.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 17.
Fig. 17.

(a) Current vector field at a depth of 97 m for Jan 1992. (b) The wind stress fields of Jan and Apr 1992. Areas with negative wind stress curl are shaded. Contour interval is 5 × 10−9 dyn cm−3.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2

Fig. 18.
Fig. 18.

(a) Meridional section of the EKE along 145°E for Jun 1992 (3-month mean). Contour interval is 100 cm2 s−2. (b) As in (a), except for the baroclinic conversion. Contour interval is 2 × 10−4 cm2 s−3. (c) As in (a), except for the barotropic conversion. Contour interval is 2 × 10−4 cm2 s−3.

Citation: Journal of Physical Oceanography 32, 5; 10.1175/1520-0485(2002)032<1338:SMVITW>2.0.CO;2