instability) must occur in the near-Gulf Stream regionbut the difficulties of measurement in the strong ve-locities that occur there has precluded quantitative es-timates of the heat flux-terms. (Strong mooring mo-tions induced by the high velocities produce fictitioustemperature signals coherent with the velocity field.)If we rule out open-sea generation by meteorologyand open-sea baroclinic instability as mesoscale eddysources, we are left only with topographic generationand generation and radiation from strong boundary cur-rents as possible significant eddy sources. Both are dif-ficult to evaluate quantitatively. Schmitz (1978) andFu and Wunsch (1979) showed that the major effect oftopography seems to be the spin-down of the deepocean layers relative to the thermocline. Topographywould thus more closely resemble a sink of eddy energythan its source. At the present time, then, the strongboundary currents seem the most likely major sourceof eddy energy.These general results, coupled with the intense con-centration of low-frequency energy near the westernboundary currents, suggests that the eddy field mayindeed be dynamically unimportant except in the im-mediate vicinity of the boundary currents. The pres-ence of a strong recirculation on the flanks of the GulfStream System (Worthington, 1976; Stommel, Niiler,and Anati, 1978; Wunsch, 1978a; Reid, 1978; and seechapter 1) is associated with an intense barotropic eddyfield (Schmitz, 1978) and a large resident population ofGulf Stream rings (Richardson, Cheney and Worthing-ton, 1978). A discussion of the Gulf Stream Systemvariability may be found in chapter 4 and will thus notbe pursued here. But at this time the only strong evi-dence for the importance of the variability in the dy-namics of the mean flow is in the western boundarycurrent regions. Theories (Pedlosky, 1977) show thatthe interior field could be largely radiated from theGulf Stream System and from moving Gulf Streamrings (Flierl, 1977; see also chapter 18).The mesoscale and other spatial scales of variabilitybecome bound up with almost all other aspects ofoceanography, including biological, geological, andchemical as well as physical oceanography. To the ex-tent that the variability transports properties, nutrientsand other tracers can be expected to be carried along.Eddies will scour the bottom, confusing the interpre-tation of sediments, and indeed will carry sedimentswith them from one region to another, possibly quitecontrary to the overall time-mean flows. On the phys-ical side, the variability will interact with internalwaves and fine structure on the short-wavelength endof the spectrum, and there is probably an interactionbetween the mesoscale variability and the interannualfluctuations at the other end. We have not dealt di-rectly here with any of these questions; to do so wouldrequire a complete discussion of almost all aspects ofoceanography and would be premature in any case.11.3 Summary and ConclusionsThe extent to which the ocean undergoes large-scaleclimatological fluctuations remains uncertain in theface of difficult sampling problems and the short du-ration over which appropriate instrumentation has ex-isted. Given that large fluctuations exist at least in thesea-surface temperature field, we cannot really distin-guish changes imposed by the atmosphere with a staticocean response from those in which large-scale oceanicdynamics are directly involved. Study of the atmos-pheric spectra displayed here and elsewhere suggeststhat the forcing and response by the atmosphere is acomplex function of position and time scale. With allof the current emphasis on the question of climaticchanges, we have been able to do little more thandefine what we do not know.Study of the more accessible mesoscale is very muchin its infancy and generalizations are dangerous be-cause so little of the ocean has actually been ap-propriately sampled. At present, one can draw a fewqualitative conclusions that may survive futureobservational programs. The mesoscale eddy field ishighly inhomogeneous spatially-both in the horizon-tal and the vertical. Direct wind forcing and instabili-ties of the interior ocean are unlikely to provide muchenergy to the mesoscale. Much of the total mesoscaleenergy is found in the immediate proximity to thewestern boundary currents in the region where theyare going seaward. It seems, therefore, that these re-gions are the generators of much of the eddy field en-ergy. Detailed mechanisms are not yet known. Theintense recirculations that exist on both sides of theGulf Stream, the underlying eddy field (Luyten, 1977),and the formation and decay of Gulf Stream rings thereare probably related. They suggest an intimate relationbetween the generation of the eddy field, the mainte-nance of the eastward going jet, and the transition ofthe jet into the comparatively quiescent interior in acomplex dynamic linkage that we can only vaguelyunderstand. One would guess that the next 10 yearswill find many of the keys to these puzzles.The spectra that are displayed throughout this papersuggest that oceanic low-frequency variability doeshave some characteristic features independent of ge-ography. An eddy-containing band between 50 and 150days is a feature of most regions. Almost all the spectrashow an isotropic band at frequencies higher than theeddy-containing band with a slope not far from -2. (Inthe immediate vicinity of topography the isotropicband may show-more structure.) The interannual bandhas a distinct tendency to zonality, especially in the373Low-Frequency Variability of the Sea _ __ ·thermocline. These features are nearly independent ofthe remarkable fluctuations (two orders of magnitude)in the overall energy levels.Although the supporting evidence is more sparse, thedominant mode of baroclinic motion in the eddy-con-taining band is a simple vertical translation of the ther-mocline with roughly isotropic horizontal velocities;lower frequencies are more zonal and confined to theupper ocean (cf. Schmitz, 1978). The isotropic band
View Full Document
Unlocking...