A history of thermocline theory

Abstract

The focus of my attention in this essay is to give the reader some insight into the development of the theoretical ideas concerning the thermocline from the very personal point of view of someone involved in that development. It is not a review paper of the full scientific history of the subject or a detailed scientific discussion of the problem. Rather, I want to present the more personal side of the history of the attempt to reach an understanding of the phenomenon of the thermocline. In that special sense this is a personal memoir of my involvement in that quest and I hope to capture the flavor of that experience for others. It culminated for me in the development of the theory of the Ventilated Thermocline, a theory that evolved in collaboration with Henry Stommel and Jim Luyten and so the story is one person's view of what was a collegial effort. The presence of the thermocline, the region of rapid increase of density with depth in the first one or two kilometers of the subtropical oceans, has been recognized for a long time.Yet looking back at the older literature it is hard to find a reference to the problem of the thermocline, that is, in explaining the maintenance of this sharp density gradient or its cause. The thermocline itself is bowl-shaped in the meridional plane and contains fairly entirely the wind-driven currents of the major subtropical gyres so there is much to explain. It is provocative to note that the very word does not even appear in the index of that great tome, The Oceans, by Sverdrup, Johnson, and Fleming (1942) which was for a long time the definitive work on oceanography although the word appears a few times1 in the text. More frequently, the thermocline is referred to there as a "discontinuity layer" or "transition layer." This makes the thermocline appear as a curious secondary feature separating zones of greater oceanographic interest. Similarly in Defant's two-volume treatise on physical oceanography (1961), the thermocline is noted largely as a transition zone between what are supposed to be oceanographic analogues of the atmosphere's troposphere and stratosphere. The existence of such a transition layer is taken, misleadingly, as a natural process similar to the seasonal thermocline in lakes. Indeed, the first use of the word "thermocline" appears in the limnology literature of the late nineteenth century, a fact for which I am indebted to my colleague Bruce Warren. It does seem like one of those examples, not uncommon in science, in which a phenomenon is noted but that it might be a problem requires first that it be posed as a problem, i.e., why should there be a permanent thermocline in the ocean? It is a bit like becoming used to looking at the Rocky Mountains and taking their presence for granted. It may not be natural for most people in their neighborhood to ask why they are there but once the question is asked it is easy to realize the importance of finding the answer. Of course it is a danger of amateur historiography to overlook past insights into a scientific problem that contributed to a background understanding but which often enter into oblivion because these insights were not united to a powerful enough method to obtain a complete solution to the problem. So, for example, the insightful work of Montgomery (1938), in which it is pointed out that the water mass properties at depth in the subtropical gyre can be traced back along an isopycnal to the properties of surface water, was an idea clearly influenced by the meteorological interest of the time in isentropic analysis of atmospheric motions suggested by Rossby (1937). This idea is seen in even more explicit form in the interpretive work of Iselin (1936, 1939) whose much reproduced schematic shows explicitly how water from the mixed layer sliding down along isopycnal surfaces sets the vertical distribution of density (and so implicitly also its horizontal distribution) in the subtropical gyres. These very profound insights were, however, innocent of a powerfully enough unifying dynamical foundation that could go beyond a quasi-diagnostic explanation to provide the basis for a theoretical understanding of thermocline structure. Interestingly enough, the key ingredient has turned out to be potential vorticity conservation and this was much on the minds of Rossby and Montgomery who had collaborated on its application to atmospheric motions. A similar application to oceanic motion had to wait another 40 years.2 In the years that followed there were few attempts to deal with the structure of the thermocline except for some linear models (Stommel andVeronis, 1957; Pedlosky, 1969; Gill, 1985) that attempted to describe the capture of the wind-driven motion to an upper region of the ocean while accepting an already specified background stratification as a starting point for the linearization of the theory. Such treatments, interesting in their own right, are clearly incapable of addressing the fundamental question of the density structure that defines the thermocline. They do, however, have the virtue of indicating the beta-effect as being the important dynamical mechanism for limiting the depth of penetration below the upper sea surface of the wind-driven motion. The principal issue is easily stated. The ocean is heated nonuniformly but persistently at the sea surface. How is it, then, that the signature of that forcing penetrates to only about 20% of the total ocean depth even though the heating has gone on for millions of years? Clearly, a dynamical process is required to trap the thermal signal to the upper ocean, but what specifically is the operating mechanism? © 2006 Springer Science+Business Media, Inc., All rights reserved.