This data set allows an investigation into the longitudinal, seasonal, and interannual evolution of the mean upper ocean zonal currents and water masses across the equatorial Pacific. The mean zonal velocity sections at the ten longitudes and along the equator are discussed first and then related to mean sections of potential temperature and salinity. Following this discussion of the means, the seasonal cycle and the linear regression against the SOI (a simple means of evaluating the ENSO cycle) are described at a western longitude, a central longitude, and an eastern longitude, as well as along the equator.
The mean sections of zonal velocity (Fig. 2) show the two major westward currents (unshaded regions), the SEC, and the Equatorial Intermediate Current (EIC). The surface-intensified SEC is split into a northern and a southern branch (Wyrtki, 1974a), herein referred to as SEC(N) and SEC(S). At 155°W the SEC(N) is centered at 2°N and the SEC(S) at 3°S. These branches are separated by a near-equatorial minimum in westward surface flow, or sometimes by eastward surface flow (McPhaden & Taft, 1988). The EIC (Delcroix & Henin, 1988) is most obvious in the western half of the basin. At 155°W it is centered around 1°N at 350 m. In the far western section, at 143°E, the zonal component of the New Guinea Coastal Undercurrent (NGCUC; Tsuchiya et al., 1989; Butt & Lindstrom, 1994), a western boundary current, is centered near 200 m at 2.5°S, just off the coast.
Fig. 2. Sections of mean zonal velocity (U) estimated at 10 nominal longitudes and along the equator. Locations marked in the bottom right corner of each section. Contour interval is 10 cm s–1, with heavy contours at 50 cm s–1 intervals. Eastward velocities are shaded.
The SEC(N) is strongest in the central Pacific, building in strength as it flows from 95°W to 140°W, but then declining almost to nothing by 165°E, and continuing to be weak to 143°E. Like the SEC(N), the SEC(S) is surface-intensified in the east where the thermocline is shallow, but it deepens considerably toward the dateline. The SEC(S) also builds in strength from the eastern to the central Pacific, but maintains its strength further west, weakening by 156°E. Both branches of the SEC have deep extensions in the east that extend far below the thermocline (Fig. 3). These deep branches are strongest at 110°W, but are visible from 140°W to 95°W.
Fig. 3. Sections of mean potential temperature (
)
estimated at 10 nominal longitudes and along the equator, following Fig. 2.
Contour interval is 1°C, with heavy contours at 5°C.
The EIC is very weak or nonexistent in the east, and in some sections it is apparent only as a minimum in eastward velocity near 350 m at the equator. By 155°W a robust EIC is centered near 350 m at 1°N, but to the west of that longitude it shifts equatorward. The current strengthens and deepens from 155°W to 165°E, becoming too deep to be fully sampled by the ADCP measurements, which tend to drop out below 350 m. The deep extension of the SEC(N) in the east may be connected to the EIC in the west.
The NGCUC is only evident at 143°W, as a subsurface velocity maximum in the south of the section. This current is not discussed extensively in other sections since it is not a transpacific current. However, the feature is significant to the general circulation, so some statistics are computed from the mean section. The NGCUC mean zonal volume transport is about –19 × 106 m3 s–1, with a velocity-weighted potential temperature of 18.85°C, salinity of 35.18, and potential density anomaly of 25.06 kg m–3.
Eastward currents (Fig. 2, shaded regions) include the EUC, the North Equatorial Countercurrent (NECC), and the SCCs. At 140°W the core of the EUC is located near 110 m at the equator and the NECC is a near-surface current centered around 7°N. The SCCs, or Tsuchiya Jets (Tsuchiya, 1975; Johnson & Moore, 1997), are deeper eastward currents, also known as the Northern Subsurface Countercurrent (NSCC) and the two branches of the Southern Subsurface Countercurrent (SSCC). At 140°W the NSCC is centered near 200 m at 4°N and the two SSCCs are centered near 200 m at 4.5°S and near 250 m at 7°S, respectively. In addition, the northern edge of the surface-intensified South Equatorial Countercurrent (SECC) can be seen in some of the sections in the west, for instance, near 8°S at 165°E.
The EUC starts out deep and relatively weak at 143°E, where it is displaced
to the north of the equator by the NGCUC (Fig. 2). The EUC strengthens to 156°E,
but then appears to weaken at 165°E before building up to full strength
between 155°W and 125°W, and then weakening considerably by 95°W.
The maximum EUC velocity occurs at 125°W. The EUC shoals steadily eastward
to 110°W, with its center in the thermocline, near = 17°C
(Fig. 3). This behavior is in agreement with a velocity climatology derived
from TAO moored current meter and ADCP data (Yu & McPhaden, 1999). Their
data were not resolved sufficiently spatially to show the EUC weakening near
165°E, but similar behavior does emerge from models (Fig. 8 of Vintzileos,
Delecluse, & Sadourny, 1999) and observations suggest there is a reversal
in the zonal pressure gradient at the core of the EUC near these longitudes
(Fig. 2 of Johnson et al., 2000). On the other hand the EUC weakening at 165°E
could be an artefact caused by insufficient sampling. Around 140°W, where
the trade winds are at their peak, the entire equatorial current system (including
the EUC, the SEC(N) and the NECC) is most intense, in qualitative agreement
with surface drifter climatological velocities (Reverdin, Frankignoul, Kestenare, & McPhaden,
1994; Johnson, 2001).
The NECC shifts northwards from 4°N at 143°E to 6°N at 165°E and then to about 8°N at 125°W (Fig. 2). The velocity maximum in the NECC starts out strong at 143°E, weakens to 165°E, re-builds to 140°W before weakening significantly and shifting southwards at 110°W. It once again rebuilds and shifts back north at 95°W. This pattern is once more consistent with surface velocities from drifters. The pattern in the eastern and central Pacific is also expected from the Sverdrup relation (Kessler, 2002). At many latitudes the NECC velocity maximum is slightly subsurface, which is the result of a combination of two factors: firstly, the westward Ekman tendency at the surface from southerly winds, and secondly, a nearsurface geostrophic tendency toward westward flow induced by northward surface warming and freshening across the NECC.
The SCCs also evolve from west to east (Fig. 2). In the west, the SCCs appear as deep lobes on the EUC with velocity maxima approaching 30 cm s–1. The SCCs shift poleward and upward to the east. By 140°W they are fully separated from the EUC. In synoptic sections the SCCs are still strong in the central Pacific (Rowe et al., 2000), but because they are surrounded by westward flow and meander significantly, probably influenced by tropical instability waves (TIWs; Baturin & Niiler, 1997), they are weak in the mean. By 110°W the SSCCs are quite shallow and have shifted significantly poleward, with the southern SSCC lying south of 8°S. By 95°W the NSCC has all but disappeared, likely having turned north under the Costa Rica Dome (Johnson & McPhaden, 1999; Kessler, 2002). The NSCC and NECC are sometimes difficult to separate, with only a minimum in eastward velocity found between them.
Finally, a portion of the South Equatorial Countercurrent (SECC; Reid, 1959, 1961; Merle, Rotschi, & Voituriez, 1969; Wyrtki, 1974a; Eldin, 1983; Kessler & Taft, 1987; Delcroix, Eldin, & Henin, 1987) is visible at the southern edge of the sections at 165°E to 170°W in the mean. However, the sampling of this current is insufficient for discussion here.
The shape of the thermocline is related to these currents through the geostrophic balance (Fig. 3). Isotherms slope up toward the equator below the SEC, as the thermocline sharpens approaching ±2° latitude. Inside of this range, the thermocline spreads vertically around the zonal velocity maximum of the EUC, as geostrophy requires. There is a poleward thermocline uplift in the northern portion of the sections that is geostrophically associated with the NECC, and a slight poleward uplift in the very southern portion of the western sections that is associated with the SECC. The sharp subthermocline meridional fronts in temperature, found poleward of the tropical 13°C thermostad, are associated with the SSCC and NSCC. The thermostad builds poleward to the east as these currents, which flank it, move poleward under the shoaling thermocline to conserve potential vorticity (Johnson & Moore, 1997). Finally, isotherms pinch equatorward below this thermostad, near 11°C, associated with a relative minimum in eastward flow in the east and the westward flowing EIC in the west.
In the equatorial region the thermocline shoals eastward and thereby provides the zonal pressure gradient on the equator that drives the EUC (Fig. 3). In the western equatorial Pacific the trade winds pile up a thick layer of warm surface water above a deep thermocline, the warm pool (Webster & Lukas, 1992), while in the eastern equatorial Pacific the thermocline rises to the surface, creating the cold tongue (Bryden & Brady, 1985). There is a subtle reversal of isotherm slopes below about 15°C between 156°E and 165°E (and hence in zonal pressure gradient) that is concordant with a reduction in the strength of the EUC at 165°E (Fig. 2). The overall mean picture along the equator is very similar to a climatology from TAO moorings (Yu & McPhaden, 1999). The thermocline also sharpens to the east as it shoals, especially off the equator (Fig. 3).
The salinity field (Fig. 4) shows the influence of heavy rain under the intertropical convergence zone (ITCZ) in the east and the warm pool in the west. Fresh surface values are found in these regions. In addition, equatorward advection of water subducted in the subtropics is evident as tongues of high salinity southern water and low salinity northern water. The tongues extend toward the equator at the level of the EUC. The high-salinity tongue in the pycnocline extends northward and westward from the southeast Pacific, and is associated with equatorward thermocline flow that feeds the EUC (Johnson & McPhaden, 1999), driven by the zonal pressure gradient resulting from the thermocline slope. The strong meridional salinity front within the thermocline on the equator reflects equatorial convergence of these salty southern and fresher northern waters, feeding the EUC. An isolated salinity maximum is found at a depth of about 80 m just south of the equator east of 155°W. This feature is the result of eastward advection of this salty water in the EUC, which is surrounded by the westward advection of fresher water in both branches of the SEC. Fed from the western and central Pacific, this feature is remarkably persistent, flowing eastward in the EUC over 5000 km (Fig. 4). In the west, there is a chimney of salty water near the equator that appears to extend up into the mixed layer. This feature is better discussed in the context of the seasonal cycle below.
Fig. 4. Sections of mean salinity (S) estimated at 10 nominal longitudes and along the equator, following Fig. 2. Contour interval is 0.1, with heavy contours at 0.5.
Seasonal and ENSO cycle variations of zonal velocity, temperature, and salinity are described along the equator and at three representative longitudes: in the western Pacific at 165°E, in the central Pacific at 155°W, and in the eastern Pacific at 110°W. At each longitude, the zonal current variations are described, and these variations are then related to the potential temperature and salinity fields. A set of eight zonal sections at the equator is shown for each of these properties (Figs. 5–7). The top six sections are bimonthly presentations of the properties, evaluated in the middle of the month noted. These sections start with mid-February in the upper left and proceed down to mid-June, then go on to mid-August in the upper right proceeding down to mid-December. The bottom two sections show the perturbations imposed on the mean during mild El Niño conditions (SOI = –1.0) and mild La Niña conditions (SOI = +1.0). A similar convention is followed for meridional sections at the western, central, and eastern Pacific longitudes (e.g., Figs. 8–10).
Fig. 5. Zonal sections of zonal velocity (U) along the equator estimated for 6 months and both phases of the ENSO cycle. Six months of a SOI = 0 (normal) year are shown in the upper six panels as described in the text. SOI = –1 (El Niño) is shown in the bottom left panel and SOI = +1 (La Niña) in the bottom right panel. Details follow Fig. 2.
Fig. 6. Zonal sections of potential temperature ()
along the equator estimated for 6 months and both phases of the ENSO cycle.
Details follow Fig. 3.
Fig. 7. Zonal sections of salinity (S) along the equator estimated for 6 months and both phases of the ENSO cycle. Details follow Fig. 4.
Fig. 8. Meridional sections of zonal velocity (U) at 165°E estimated for 6 months and both phases of the ENSO cycle. Details follow Fig. 2.
Fig. 9. Meridional sections of potential temperature ()
at 165°E estimated for 6 months and both phases of the ENSO cycle. Details
follow Fig. 3.
Fig. 10. Meridional sections of salinity (S) at 165°E estimated for 6 months and both phases of the ENSO cycle. Details follow Fig. 4.
There were not sufficient degrees of freedom to include semi-annual harmonics with any confidence in this analysis. TAO mooring data do suggest that near-surface velocities in the western Pacific have a significant semi-annual signal (Yu & McPhaden, 1999), which has been missed here. Fortunately, these data also suggest that the semi-annual signal on the equator is small at depth in the west and throughout the water column in the east. A surface drifter velocity climatology is in agreement with the TAO mooring data on the importance of the semi-annual harmonic on the equator in the western Pacific (Reverdin et al., 1994). In addition, this climatology suggests significant semi-annual energy in both branches of the SEC in the central Pacific. The lack of a semi-annual harmonic in the present work should be borne in mind, especially in these regions.
The isopycnal averaging framework allows more structure than one might expect (below the mixed layer) given the annual harmonics and the linear regression against the SOI. Since fits are made for both the potential isopycnal depths and for the properties (zonal velocity, potential temperature, and salinity) on isopycnals, there are twice as many degrees of freedom as for a similar fit on isobaths. The depth of an isopycnal and the values of properties on that isopycnal can vary in- or out-of-phase. This potential independence in the phase and amplitude of isopycnal depth and isopycnal properties is why the data have been presented using bimonthly sections, rather than by discussing the harmonic amplitudes and phases.
At 165°E the eastward-flowing EUC peaks in June (Figs. 8 and 5), but variations in isothermal spreading about the equator associated with the seasonal cycle are not obvious (Fig. 9). Fresher surface salinities may lag the eastward surface velocities by a month (Figs. 7 and 5), perhaps a signature of eastward advection of the fresh pool from the west during this season. In contrast, the westward-flowing EIC and SEC(N) both peak in December. The eastward-flowing NECC also peaks in December, consistent with the thermocline trough between the SEC(N) and NECC being strongest during this time (Fig. 9). The SEC(S) is maximum in February–April at 165°E (Fig. 8). The current also shows a relatively weak seasonal cycle in the central Pacific (Fig. 11), and is maximum in October–December at 110°W (Fig. 14).
Fig. 11. Meridional sections of zonal velocity (U) at 155°W estimated for 6 months and both phases of the ENSO cycle. Details follow Fig. 2.
In February–April a chimney of salty water appears to extend upward from the subsurface salinity maximum near the equator in the western Pacific (Figs. 10 and 7). This chimney results from some combination of equatorial upwelling, westward advection of saltier water from the east (Figs. 5 and 7), seasonal changes in local precipitation, and deep mixing from strong wind events in November–March (Cronin & McPhaden, 1998).
At 165°E El Niño weakens the EUC at its core (Figs. 8 and 5), but a strong eastward surface flow develops on the equator (Cronin, McPhaden, & Weisberg, 2000). The thermocline becomes shallower and relatively diffuse in the west (Figs. 6 and 9) as the warm pool is drained by eastward advection (Picaut & Delcroix, 1995). Surface salinities are fresh during El Niño as precipitation migrates east with the warm pool (Figs. 7 and 10; Delcroix & Picaut, 1998). In addition a strong NECC is present around 5°N (Fig. 8; Kessler & Taft, 1987). The EIC also appears to be stronger than normal. In contrast, La Niña conditions see both a strong EUC and a strong SEC, while the NECC is pushed northward to about 7°N. In other words, during La Niña the equatorial current system is strengthened and looks more like the central Pacific than the western Pacific as the easterly trade winds migrate westward. The trades drive both the equatorial SEC (frictionally) and the EUC through the zonal pressure gradient. In addition, during La Niña there is a relatively sharp and deep thermocline as the warm pool builds in the west, in addition to salty surface conditions as the convection shifts west and advection in the SEC carries saltier central Pacific waters westward.
At 155°W the EUC peaks in June (Figs. 11 and 5), and a contrast in the spreading of the thermocline about the equator between June and December is seen (Figs. 12 and 6). Highest EUC speeds occur when the current is at its shallowest (Fig. 5), which occurs when equatorial easterly trade winds are weakest. The very strong shears above the EUC and the interaction between the eastward zonal pressure gradient and the westward surface stress indicate the complexity of the mix of influences on the EUC (Philander & Pacanowski, 1980; McPhaden & Taft, 1988; Wacogne, 1990). At 155°W there is no surface expression of eastward equatorial flow in any season (assuming that extrapolating the ADCP shear from below 20 m up to the surface adequately describes the surface velocity). Perhaps as a result, there is little change in the salinity field on the equator in the central Pacific (Figs. 13 and 7). The SEC(N) appears strongest in December, and the NECC peaks in October. This slight phase difference agrees with results from island sea level data that span the entire SEC(N) (Wyrtki, 1974a). However, this shift has not been noted in other studies, which could not resolve the near-equatorial SEC fully (Kessler & Taft, 1987). The thermocline is very sharp, with a strong seasonal cycle in the strength of the trough between the SEC(N) and the NECC as these currents wax and wane. The most prominent cycle in salinity at 155°W is found in the north, under the ITCZ, where very fresh surface values occur in October–December.
Fig. 12. Meridional sections of potential temperature ()
at 155°W estimated for 6 months and both phases of the ENSO cycle. Details
follow Fig. 3.
Fig. 13. Meridional sections of salinity (S) at 155°W estimated for 6 months and both phases of the ENSO cycle. Details follow Fig. 4.
At 155°W the ENSO cycle of variation in the equatorial Pacific current system can be seen (Fig. 11) as the trades weaken during El Niño and strengthen during La Niña. The EUC and SEC are both weaker during El Niño compared to La Niña. As at 165°E, the NECC shifts southward during El Niño and northward during La Niña. Finally, as expected, the cold tongue is suppressed during El Niño but enhanced during La Niña (Figs. 12 and 6).
While the meridional shifts of the NECC and the amplitude modulation of the SEC(N) with the ENSO cycle are in accord with previous studies, the modest NECC during El Niño is not (Wyrtki, 1974b; Taft & Kessler, 1991). The sparse temporal and spatial sampling of the sections (Johnson et al., 2000) may not resolve NECC amplitude changes associated with the SOI in the central Pacific. The near-equatorial currents are strong during the boreal fall, winter and La Niña. Their strength supports the generation of TIWs in the central and eastern Pacific (Baturin & Niiler, 1997). These waves can be sufficiently vigorous that they may alias the analysis in these regions. Examination of individual sections suggests that TIWs are aliasing the currents in the central Pacific, but not so much the eastern Pacific. During the La Niña phase of the SOI, it appears that in the central Pacific TIWs have been sampled preferentially in phases that reinforce the SEC(N) and NECC.
At 110°W the relationships among zonal velocity, temperature, and salinity in the EUC are all evident in the seasonal cycle. The EUC peaks in strength around April, when it also surfaces (Figs. 14 and 5; Philander & Pacanowski, 1980). This far east, the thermocline is extremely sharp and shallow (Figs. 15 and 6). The equatorial spreading of the thermocline associated with the EUC is noticeably stronger during April than in October, when equatorial SST is lowest and the SEC is strongest. The laterally isolated salinity maximum within the thermocline just south of the equator (Figs. 16 and 7) is strongest when the EUC velocity is a maximum, or slightly thereafter. This lagged correlation of maxima in EUC salinity and velocity agrees with results from slightly further to the east (Lukas, 1986).
Fig. 14. Meridional sections of zonal velocity (U) at 110°W estimated for 6 months and both phases of the ENSO cycle. Details follow Fig. 2.
Fig. 15. Meridional sections of potential temperature ()
at 110°W estimated for 6 months and both phases of the ENSO cycle. Details
follow Fig. 3.
Fig. 16. Meridional sections of salinity (S) at 110°W estimated for 6 months and both phases of the ENSO cycle. Details follow Fig. 4.
In the north, the NECC is minimum in strength near 110°W (Fig. 2), with no eastward surface flow in December–February and a surface expression that peaks in August (Fig. 14). The thermocline trough between the SEC and NECC is only apparent from August to October, again concordant with the zonal velocity (Fig. 15). Interestingly, while the NECC is weak at 110°W, the NSCC and SSCC are strong. These countercurrents also display a noticeable seasonal cycle at 110°W, being stronger in the first half of the year than in the second. At 110°W surface salinities under the ITCZ are freshest in October (Fig. 16), when the ITCZ is strongest. In addition, surface salinities are fresh as far south as 7°S in December and February, advected west from fresh coastal waters by the SEC, with westward surface velocities that peak a few months earlier (Fig. 14).
As at 155°W, the current system at 110°W is spun up during La Niña, when compared to El Niño (Fig. 14). The cold tongue is much stronger during La Niña and quite weak during El Niño (Fig. 15). Surface salinities are generally fresher during El Niño than during La Niña (Fig. 16). These fresh salinities are at least partially a product of local precipitation associated with eastward migration of warm sea surface temperatures and convection and part a result of the reduced trade winds (Ando & McPhaden, 1997).
Return to previous section or go to next section