4.6. Abyssal Circulation

The theory for abyssal circulation has had the same dynamical basis as the Sverdrup balance; the source of vertical velocity has been assumed to be general upwelling in response to point sources of water at high latitudes [ Stommel and Arons, 1960]. Recently observations for the deep North Pacific have suggested that the cyclonic flow predicted by this theory is actually reversed and in at least two distinct large anticyclonic cells [ Roemmich and McCallister, 1989; Talley and Joyce, 1992] (Fig. 1) in addition to an intense northern cyclonic circulation [ Warren and Owens, 1988]. Johnson and Toole [1993] show similarly anticyclonic flow in the deep tropics in the North Pacific. On the other hand, deep circulation in some smaller and eastern basins is apparently cyclonic [ Reid, 1986, 1989, 1994] (Fig. 2). Numerical models [ Semtner and Chervin, 1992; Fujio et al., 1992; Ishizaki, 1994] show a complex of anticyclonic and cyclonic deep circulations in the Pacific; there is a general sense of more anticyclonic than cyclonic circulations particularly in the North Pacific. Holloway [personal communication] suggests that the deep circulation may be a very strong function of topography, in the sense of rectification of topographic Rossby waves which would create cyclonic flow around the deep basins. Rhines and MacCready [1989] suggest that deep circulation [ Stommel and Arons, 1960] may be reversed simply by filling from the bottom a bowl whose cross-section increases upward.

An important abyssal tracer signature in the Pacific is primordial helium released from hydrothermal vents; the helium sections in the Geosecs atlas [ Ostlund et al., 1987] show two plumes on either side of the equator, with the stronger one in the South Pacific. The vents are located on spreading ridges, which center around 2500 meters depth throughout the world. Lupton and Craig [1981] demonstrated that the source of the westward-extending plume of helium 3 is the East Pacific Rise, with a particularly strong source at 15S. Craig [personal communication] observed that the plume in the North Pacific is robust. Lupton [personal communication] has mapped the plume extending southwestward from the Juan de Fuca fracture zone near 45N. The vents are also a heat and silica source, and the effect of venting on these properties in the Pacific has been demonstrated by Reid [1982b, 1986], Talley and Joyce [1992], and Talley and Johnson [1994].

The dynamical effect of the heat source from the hydrothermal vents on ocean circulation has been considered, starting with Stommel's [1982] model of flow on a plane driven by a hydrothermal source. The `` plane'' is a simplification of the spherical coordinates appropriate for the earth; Cartesian coordinates are used, with the effect of sphericity included simply as a variation in Coriolis parameter with north-south distance. Speer and Rona [1989] modeled the entrainment of surrounding water by the plumes, and showed the height to which the plumes should rise with different background stratification. Hautala and Riser [1993] showed that flow in the neighborhood of the strong East Pacific Rise source is consistent with a hydrothermal source, including circulation around the topography and vertical circulation. Joyce and Speer [1987] modeled the effect of hydrothermal sources on basin-wide circulation, showing in a more quantitative way the effect of superimposing hydrothermal flow on the basic Stommel and Arons [1960] abyssal circulation. Hautala and Riser [1989] showed the circulation in the abyssal layer which results from a superposition of general abyssal upwelling, wind driving, the hydrothermal sources, and the effect of topography. Both of these models suggest that the westward flow from the East Pacific Rise can be geothermally driven. Helfrich and Battisti [1991] on the other hand suggest that the circulation effect of distinct sources might be more local, with plumes which break into smaller eddies. Talley and Johnson [1994] suggested that while the westward flow in the South Pacific might be enhanced by the hydrothermal source, the principal driving force might not be hydrothermal because the symmetry about the equator of the large-scale temperature pattern at about 2500 meters (Fig. 3) suggests a fairly symmetric circulation pattern with respect to the equator. The suggested circulation pattern is very similar to that suggested by tracers in the Atlantic as well, and it seems unlikely that hydrothermal forcing produces these patterns in both oceans and both hemispheres.

The strength of the abyssal circulation is probably best measured in the western boundary currents and at choke points for the flow. One of the principal choke points is Samoan Passage, through which Wüst [1929] showed northward abyssal flow. Mantyla and Reid [1983] show clearly, with data of unquestionable quality, that this is the main conduit for northward flow. Recently Taft et al. [1991] and Johnson et al. [1994] have estimated the transport through the passage, at approximately 5 Sv (5 x 10 m sec) below 1.2C. Observations from a coherent current meter array in the passage show persistent northward flow with significant transport fluctuations, with a mean transport of 6 Sv. [Rudnick, personal communication].

The location and intensity of abyssal upwelling is an open matter at this point. Roemmich and McCallister [1989] inferred a large amount of upwelling near the western boundary in the North Pacific (Fig. 4), largely based on the persistent westward flow measured directly at 152E [ Niiler et al., 1985]