anomaly of opposite sign is following the same trajectory, but 180° out of phase spatially. This is just the signal described by MMR, but it is produced here by totally white forcing as opposed to the RW forcing used in their experiment. As we saw above, the signal was also produced by our version of the RW forcing provided the magnitude of the high southern- latitude forcing was large enough.
The form of the signal in the Pacific (Figure 8) is somewhat like that found in the Atlantic. The two PCs form a rather simple attractor (Figure 6) and are in quadrature again, suggesting a propagating signal in the salinity field. The time scale is roughly 300 years, as found above. A major difference between the two oceans is that the signal does not penetrate in strength to the same depth it was found to in the Atlantic. The signal appears to propagate from one end of the ocean to the other, but is closely confined to the upper layers of the water column and has more spatial structure than in the Atlantic.
The simple correlation between the ACC transport fluctuations (Figure 2) and the variability of the salinity along the Atlantic-Pacific sections is shown in Figure 9. In the Atlantic, the plus-minus signature of the correlation field is indicative of the propagating signal discussed above. Note that major reductions in the strength of the ACC accompany reduced salinity over most of the South Atlantic to depths of the order of 1000 m. This is countered by increased salinity in the North Atlantic and in the deeper parts of the
entire Atlantic. The signature of the ACC variability in the Pacific is identical to that in the Atlantic, but only in the North Pacific. Most of the Pacific, even to the greatest depths, is more or less in phase with the ACC signal. However, the absolute values of the correlation in the Pacific are less than those found in the Atlantic.
The horizontal structure of the low-frequency MMR signal is investigated in Figure 10, which shows the first and second EOFs of the Atlantic salinity anomaly at a depth of 700 m from the WW run. The associated PCs (not shown) suggest motion such that EOF1 leads EOF2. The first EOF shows the North and South Atlantic having anomalies of opposite sign. But note the negative anomaly extending into the Gulf of Mexico region. The second EOF shows the positive anomaly that occupied the North Atlantic has rotated and moved clockwise such that it is now off Africa. The negative anomaly has also moved clockwise so that it now covers most of the North Atlantic. These clockwise motions suggest that the anomalies are being influenced mainly by advection. Much the same type of anomaly motion is seen at 2000 m also.
The simple correlation between the ACC index and the salinity anomalies at 700 m and 2000 m is given in Figure 11. The entire Indian, North Pacific, and North Atlantic