The central remaining physical oceanographic problem of the eastern tropical Pacific is the three-dimensional interconnection among the zonal currents of mid-Pacific as they impinge on, or depart from, the American coast. Wyrtki realized this early on, and drew a schematic (Fig. 5a) that is not so different than one drawn today (Fig. 5b), except in added detail; however that is mostly a statement that we are still unable to produce a quantitatively realistic picture. The interconnections among the EUC, SEC and Peru upwelling remain difficult to diagnose with presently available data, while model descriptions appear to be highly dependent on the resolutions and physical parameterizations in the various models.
We know, as Wyrtki did, that some of the EUC upwells into the SEC, while another part continues southeastward at thermocline depth to feed the Peru upwelling (Roden, 1962; Lukas, 1986) but are not much further along in putting numbers to this than was possible in 1966 (see Sloyan et al., 2003, for an estimate from an inverse calculation). In general the eastern origin of the SEC remains almost as shrouded in mystery as it was in Wyrtki's time. How much of the SEC comes from equatorial upwelling, how much from the NECC, and how much from the Peru coast? We can assume that this division varies both seasonally and interannually, but can barely speculate on the answers. A further puzzle concerns the very thick westward flows on both sides of the EUC, that appear to be deep extensions of the SEC branches (Fig. 6, middle panel). (Although these appear only weakly in the 90°W section (Fig. 6, bottom), they exist down to the 900 m reference level at both 110°W and 90°W in this data set.) Rowe et al. (2000) hypothesized that these could be the westward limb of deep cyclonic gyres fed by the two Tsuchiya Jets. This has been difficult to confirm with the very sparse data coverage available.
The vertical circulation in the eastern tropical Pacific has local and basin-scale origins and consequences. The large-scale winds constrain the thermocline to tilt down to the west and be shallowest along the coast of the Americas (Fig. 9 of Fiedler and Talley, 2006). Local regions of upwelling-favorable winds (the Central American wind jets and alongshore winds on the coast of South America) can therefore open a window through which thermocline (and subthermocline) water can reach the surface and interact with the atmosphere. Although it has been known for centuries that the upwelled water is nutrient-rich and thus has tremendous economic and ecological importance, the mechanisms that control and produce variability of east Pacific upwelling remain not well diagnosed. The processes occurring off Central America have been discussed in several papers as reviewed in Sections 4.1 and 4.2 above, but upwelling along the Peru-Chile coast poses many unanswered questions. In particular, the depth of upwelling's reach, and the dependence of this depth on the background stratification, is poorly understood. Since the stratification can change in response to either Kelvin wave propagation or water mass advection from the equator, while the upwelling velocity itself is driven by local winds, variability of SST due to upwelling is produced by a mix of local and remote influences that have not been clearly identified or quantified.
The basin-scale impacts of coastal upwelling are even less well understood. The continuity of low SST from the coast of South America to the equatorial cold tongue, both in the mean and annual cycle (see, for example, Fig. 4 of Fiedler and Talley, 2006) might be taken to indicate that cold upwelled water advects from the coast out along the equator. This would provide an oceanic connection between the two cold regions and a mechanism by which equatorial SST, with its planetary impacts through the ENSO cycle, could be influenced by South American coastal wind variability. However, as noted in Section 4.3 above, this is by no means proven, and some evidence indicates that the Peru and equatorial upwelling regions are separately driven by their regional winds. In this case there might still be a connection through an atmospheric bridge.
A potentially important impact that is beginning to be investigated is the relation between South American coastal upwelling and the persistent stratocumulus cloud decks that are a principal feature of the climate of a very large region of the southeast Pacific (Klein and Hartmann, 1993; Yu and Mechoso, 1999; Kollias et al., 2004; Xu et al., 2005). These cloud decks have been one of the worst-simulated aspects of global coupled climate models, and are thought to be a major contribution to fundamental biases that confuse the interpretation of climate change (Davey et al., 2002; Bretherton et al., 2004). We noted in Section 4.3 above that the shading of the ocean due to stratus cools SST, inducing a positive feedback by chilling the lower atmospheric boundary layer, which fosters the continuation and growth of the stratus. Thus one could hypothesize that the initial trigger is the coastal upwelling, which, though it has a narrow scale of perhaps 50–100 km, could cool the atmosphere in the region immediately adjacent to the coast. Since upwelling is due to offshore Ekman advection of surface water, this would tend to push the cold SST offshore, increasing the area of cloud coverage and potentially driving this whole system. However, there are other influences that foster the stratus decks, especially the general subsidence of the atmosphere under the subtropical highs; recent diagnoses have suggested that this is episodically amplified by atmospheric communication from convection over interior South America (Colbo and Weller, submitted for publication). It has proven difficult, however, to quantify the various influences on the delicate balances that produce the stratus decks, and a field study to study these processes is being planned (VOCALS Scientific Working Group, 2005).
As mentioned in Section 4.3, water property evidence implies that EUC and much of the thermostad water below it originates in the southwest Pacific (see Fiedler and Talley, 2006). The Indonesian Throughflow represents a transport of approximately 10–15 Sv; this amount of intermediate water entering the South Pacific from the Southern Ocean leaves the North Pacific as surface water. Thus the water mass must both have its potential vorticity modified to change hemispheres and it must be upwelled, so that this transport can be effected. Both the water mass transformation and upwelling presumably occur in the eastern tropical Pacific, as some combination of equatorial, Costa Rica Dome and Peru-Chile upwelling. This suggests that these distinctive features of the eastern tropical Pacific play a material part in the global ocean circulation.
I thank Paul Fiedler and Miguel Lavín for encouraging me to take on this review, and for many helpful comments during its preparation. David Enfield, Maria Donoso and Pilar Cornejo at NOAA/AOML kindly provided their quality-controlled XBT data that is the foundation of this study. Antonio Badan-Dangon prodded me to open my mind concerning the CRCC. Emilio Beier carefully read the manuscript and saved me from some embarrassing errors. My understanding of the development of ideas about the circulation of the eastern tropical Pacific was greatly aided by three scientists who were early and ongoing contributors to the exploration and interpretation of the region: Gunnar Roden, Bruce Taft and Warren Wooster. They generously shared stories, analyses and documents, for which I am very grateful. This is a contribution to the scientific agenda of the Eastern Pacific Consortium of the InterAmerican Institute for Global Change Research. Support was also provided by the Protected Resources Division of NOAA Fisheries, Southwest Fisheries Science Center.
Return to previous section or go to next section
