9.4.1. Observations of ocean eddies
Figures 9.14 to 9.16 show currents averaged over many (about 20) years of observations. However, just as in the atmosphere
(in fact, even more so) the picture we have described of the general circulation of the ocean, while appropriate to the time-averaged flow, is inadequate for describing the instantaneous flow. There are large variations of currents and of surface height that, instantaneously, can mask the time-averaged picture.
The altimetric and drifter data can be analyzed to yield statistics of the time variability. For example, if the sea surface height is n(x, y, t), then we can write n(x, y, t) = n(x, y) + H (x, y, t), where the overbar denotes a long time average and tf is the instantaneous departure of n from the average. An impression of the magnitude of these variations and their geographical distribution can be obtained by mapping the variance of surface height about the time mean defined as
The global map is shown in Fig. 9.19 (bottom). There are distinct maxima of in regions of strong flow, such as in the western boundary currents of the Gulf Stream and Kuroshio and in the Antarctic Circumpolar Current (especially at the southern tip of Africa and south of Australia and New Zealand). In these regions, ah > 20 cm, so the temporal variance is not very much less than mean spatial variations seen in the top frame of Figure 9.19. Figure 9.22 shows the mean speed of the surface currents together with eddy speeds based on surface drifter data. Mean surface currents in the middle of the gyres are less than 10 cm s-1, but note how eddy speeds considerably exceed the mean almost everywhere. This implies that the instantaneous flow can be directed opposite to the time-mean flow. Note also how the tropical variability that is very evident in drifter observations, Fig. 9.22, is less evident in the surface elevation field measured by altimetry, Fig. 9.19. This is not surprising. Geostrophic balance, as expressed by Eq. 9-11, tells us that, for a given current amplitude, the height gradient vanishes as f ^ 0 near the equator. Of course this is true for both mean and eddy components of the flow.
These observations show us that the ocean is not steady and laminar; rather, just like the atmosphere, it is highly turbulent. Indeed, ocean eddies are dynamically analogous to the baroclinic eddies studied in GFD Lab XI, Section 8.2.2, and can be called ocean weather systems. However, as will be discussed in Section 10.5, because of the weaker vertical stratification in the ocean compared to the atmosphere, these eddies are typically only 100 km in lateral scale, much smaller than their meteorological counterparts. Ocean eddies typically have a lifetime of many months and intense eddies shed by boundary currents, such as the Gulf Stream, can survive for a year or so.
The SST distribution over the North Atlantic obtained from satellite observations, shown in Fig. 9.23, presents an instantaneous picture of ocean variability in the region of the Gulf Stream. There is a strong gradient of SST across the Gulf Stream, which exhibits large meanders and undulations, some of which break off to form closed eddies (''warm core rings'' to the north of the stream and ''cold core rings'' to the south). The eddies and rings have a scale of 100 km or so. In Section 10.5 their formation by the baroclinic instability of the thermal wind shear associated with the substantial temperature gradient across the Gulf Stream will be discussed. Fig. 9.21, a hydrographic section across the Gulf Stream, shows that there is a strong and systematic gradient of temperature, more or less coincident with the main current, extending all the way down to about 1 km depth. Recall the map of the mean currents at 700 m in the North Atlantic in Fig. 9.17. The mean spatial gradient of surface height seen in Fig. 9.19 and the surface currents observed by drifters, Fig. 9.15, is coincident with this temperature gradient.
High-resolution numerical simulations also give a vivid impression of what the real ocean is probably like; near-surface current speeds in a numerical model are shown in Fig. 9.24. Again, we see large-amplitude spatial variations characteristic of eddy motions in the western boundary and circumpolar currents. Just as in the observations, Fig. 9.22, there is also much eddy activity straddling the equator.
Was this article helpful?