the westerlies and southward in the easterlies, as expected. Similar considerations hold for a limited region of the surface stream function near 30°S.

Also disguised by the zonally averaged stream function shown in Figure 3a are the details of the mean sinking motion. Near the surface, there are three regions of sinking in the basin (Figure 5a). There is sinking along the entire northern boundary, joined to sinking along a considerable portion of the eastern boundary north of about 20°N. There is sinking in the southeast corner of the basin poleward of the slot. Finally, there is a sinking region on the eastern boundary in the Southern Hemisphere north of the slot: This is the sinking that corresponds to the shallow counterclockwise meridional cell in Figure 3a. Figure 5b shows that the only flow that reaches the bottom is the sinking along the northern wall, which is joined to intense sinking in the northeast corner of the basin, and a small region of sinking near the southeast corner.

Examination of the currents clarifies the movement of the water from those sinking regions in which water does not reach the bottom. Figure 4c shows the currents at 2000 m. The sinking water along the northern part of the eastern boundary feeds a westward current flowing along 50°N, which then becomes part of the deep southward-flowing western-boundary deep current. This broad current (about 11° of longitude) extends from about 1000 m down to the bottom while north of the equator, and from 1000 m down to only 4000 m from the equator to its termination at 20°S; it lies directly below the equally broad (in this coarse-resolution model) northward-flowing western-boundary current. Below 4000 m, the southward-flowing current turns eastward at the equator, leaving southward flow only above 4000 m. This remaining southward flow turns eastward at about 20°S and flows back across the basin to the eastern boundary.

The other region of shallow sinking (centered at 40°S just north of the slot on the eastern boundary) extends only to 1000 m. It feeds a westward-flowing current (Figure 4b), which ultimately turns northward above the southward-flowing western-boundary current.

The water that does reach the bottom along the entire northern boundary flows southward and westward upon reaching the bottom (Figure 4d). The negative vertical velocity, which goes to zero at the bottom, is accompanied by anticyclonic flow; the zero vertical velocity, which increases to positive values in the interior near-bottom upwelling regions, is accompanied (through the Sverdrup relation fwz= v) by cyclonic flow, in particular northward flow toward the northern sinking regions. The southern sinking water in the southeast corner of the basin reaches the bottom and flows northwestward. The near-bottom vertical velocity is downward and, again by the Sverdrup relation, implies northward motion.

The combination of the northward-flowing warm western-boundary current and the deeper southward-flowing cold western-boundary countercurrent produces a strong northward heat flux (Figure 6). This flux is consistent with the input of heat through the surface in low latitudes and the strong removal of heat at higher northern latitudes (Figure 9b). The vertically and zonally integrated meridional oceanic heat flux is northward north of 8°S; it has a peak value of 0.61 PW.

The eastward flow through the slot (the analog of the Antarctic Circumpolar Current) is relatively intense. The eastward flow, which extends below the sill to about 3000 m, is caused by westward-flowing deeper water's hitting the sill, upwelling, and returning eastward.

Barely evident in Figure 3a is the so called "Deacon Cell," a cell with rising motion at lower depths to the south of the re-entrant current, northward surface currents (consistent with westerly winds), and sinking still further north. This cell, which appears in more realistic ocean models that have geography and topography (Toggweiler et al., 1989), is responsible for bringing old 14C to the surface.

Experiment 2: Colder Southern Restoring Temperature

Figure 3b shows the zonally averaged stream function for Experiment 2, which has slightly colder restoring temperatures in the south (with 0.6°C maximum difference at the southern boundary). Compared to the results of Experiment 1, the Antarctic Cell has markedly increased in strength, and the apparent penetration of Antarctic Bottom Water has moved considerably northward. The northern thermohaline cell has weakened by 2 sverdrups.

Examination of the vertical velocities near the surface (Figure 7a) and near the bottom (Figure 7b) shows that in


Zonally and vertically integrated meridional heat flux for Experiment 1 (solid line, symmetric restoring temperature) and Experiment 2 (dashed line, slightly colder southern temperature). Units are petawatts.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement