Part OneGeneralOceanCirculation1.1 IntroductionIDeep Circulationof theWorld OceanBruce A. WarrenHistorically, the deep circulation of the ocean has beenviewed from the perspective of property fields, mainlythe distributions of temperature, salinity, density, anddissolved-oxgyen concentration. The practical reasonfor not considering velocity measurements as well, ofcourse, was a technical incapacity for making themuntil very recently. On the whole, this was probablynot a bad thing: not merely because the property dis-tributions are as interesting in themselves as the mo-tion field, but also because the scalar fields are so muchmore stable than the velocity vectors-allowing spotmeasurements from different areas even years apart tobe combined into coherent pictures that tell a gooddeal about general patterns of deep flow, albeit indi-rectly. The slight differences between correspondinghydrographic sections in the atlases by Fuglister (1960)and by Wust and Defant (1936), when compared withthe fluctuations much larger than the means of deepvelocities observed by the MODE Group (1978), forexample, demonstrate how much easier it is to obtainstatistically significant information pertinent to theoverall global deep circulation from water-propertydata than from current measurements.On the other hand, the information gained from theproperty fields allows only a limited view of the deepmotions, at very best some kind of long-term average.Although oceanographers have usually been mindful ofvariability in the deep flow, even if only to accomplisheddy mixing, it seems extremely unlikely that anyoneimagined the highly energetic low-frequency meso-scale motions that current records have revealed. In-stead, because of the stability of the property fields, itwas stationarity rather than variability that was em-phasized, however implicitly, in the circulation pic-tures derived from them. That stability also was surelythe basis for the conceptual structure of water typesand masses that has been so enormously useful in sum-marizing and comprehending the temperature-salinitystructure of the ocean and in identifying features inthe property fields that can be exploited as tracers forthe flow. Without velocity information, though, suchdescriptions of oceans have sometimes degeneratedinto taxonomic sterility (naming something doesn'texplain it), and perhaps sometimes there has been tooelemental a character ascribed to water masses (as ifthey were truly building blocks rather than names forfeatures), leading to pictures of the ocean more sugges-tive of rigid geological strata than of the real motionfield that forms the distributions.Plainly there cannot be a satisfying description ofthe deep ocean circulation that does not meld stationdata with current records. It does not seem to me,though, that such a description is yet possible. Far too6Bruce A. Warrenfew current records have been obtained to describe thedeep, low-frequency motions in a global sense; it isonly in the western North Atlantic that one can evencontemplate making a basin-wide description. More-over, we simply have not learned how to combine thestable station data with the fluctuating-velocity rec-ords to tell a story that is both consistent and inform-ative. For example, Reid, Nowlin, and Patzert (1977)reported a record (Cato 2) from a current meter mooredon the South American continental slope in the coreof the North Atlantic Deep Water: for 2 weeks thedaily-averaged velocity vectors were directed south-westward, parallel to the isobaths, as one would haveexpected in this particular deep western boundary cur-rent; but then the flow abruptly changed direction andwent eastward for nearly 2 weeks. Nevertheless, withdue regard for different density and accuracy of obser-vations, the high-salinity core of the current lookedvery much as had been depicted by Fuglister (1960) andWuiist and Defant (1936). How are we to approach thesetwo different sets of data, to reconcile the variabilityof the one to the steadiness of the other, and to learnsomething significant from their combination aboutthat boundary current?Finally, it is not at all clear what effect the low-frequency velocity fluctuations have on the long-termmean flow. There is enough theoretical reason (e.g.,Rhines, 1977) to suspect that their role in its dynamicsmay be substantial, but measurements of deep Rey-nolds stresses are meager. In fact, values reported bySchmitz (1977) from the Sargasso Sea well south of theGulf Stream actually favor a negligible contribution tothe vorticity balance there, but those measurementsare far too few to give a general characterization of thedeep open ocean.Consequently, although I recognize its incomplete-ness, the following account of the deep circulation ofthe world ocean is undertaken mainly from the tradi-tional perspective of hydrographic station data, withreference to current measurements only where theyseem helpful in estimating velocities and transports ofthe prominent currents. The emphasis is on mean ther-mohaline circulation. What has been learned about thelow-frequency motions is. described in detail byWunsch in chapter 11 of this volume.In section 1.2, I have attempted a historical reviewof what seem to me to be the important events anddates in the development of ideas about the deep cir-culation, from the first deep temperature measure-ments through Sverdrup's comprehensive synthesis inchapter XV of The Oceans (Sverdrup, Johnson, andFleming, 1942). In section 1.3, I have discussed thedynamical ideas of Stommel and his colleagues thatled to the overthrow of a substantial part of Sverdrup'spicture, and its revision in contemporary thinking withdynamically consistent models of circulation. Section1.4 is an account of the sinking processes that supplywater to the deep ocean from the surface layer. Section1.5 is a consideration of how well the kinds of deep-circulation patterns envisioned in dynamical theorystand up to observation; it is necessarily mainly a di-gest of the evidence for deep western boundary currentsin the world ocean. Finally,
View Full Document
Unlocking...