II Low Frequency Variability of the Sea Carl Wunsch 11 1 Introduction The purposeful study of the time dependent motion of the sea having periods longer than about 1 day is comparatively recent In the classic Handbuch of the early 1940s Sverdrup Johnson and Fleming 1942 one searches in vain for more than the most peripheral reference to temporal changes on the large scale one of the few examples is their figure 110 showing the California Current at two different times Until very recently the ocean was treated as though it had an unchanging climate with no large scale temporal variability The reason for this is compelling and plain until the electronics revolution of the past 30 years the major oceanographic observational tool was the Nansen bottle using slow uncomfortable ships it took essentially 100 years to develop a picture of the gross characteristics of the mean ocean The more recent period 1947 Sverdrup 1947 through about 1970 Stommel 1965 Veronis 1973b and see chapter 5 was one of the intensive development of the theory of large scale steady models of the ocean circulation The methods were initially analytic later numerical Most of these models were essentially low Reynolds number steady sluggish sticky climatic oceans In them the role if any of small scale time dependent processes is simply parameterized by a positive eddy coefficient Austauch implying a down the mean gradient flow of energy momentum heat etc The westwardintensification theories Stommel 1948 Munk 1950 imply that any strong influence of such eddy coefficients would be confined to the western boundaries and could be ignored in the interior ocean except possibly in the immediate vicinity of the eastward moving free jet Gulf Stream see Morgan 1956 The resulting models bear a remarkable resemblance to many of the gross features of the large scale mean ocean circulation see chapter 5 The culmination of these analytic and numerical models of the large scale circulation coincided with a number of developments that ultimately undermined the momentary confidence that the models represented the correct dynamics of the ocean circulation These developments were of two kinds instrumental and intellectual By 1970 instruments had been developed that made it possible to obtain time series measurements in the open sea for periods far longer than a ship could possibly remain in one location These instruments included moored current meters drifting neutrally buoyant floats pressure gauges and many others Gould 1976 and see chapter 14 An additional instrument was the computer which made it possible both to handle the large data sets generated by time series in 342 Carl Wunsch struments and to explore new ideas by nonanalytic means This computer impact has been felt of course in most branches of science The intellectual developments that shifted the focus from the mean circulation to the time dependent part were also of various kinds The analytic models seemingly had reached a plateau at which their increasingly intricate features e g essentially laminar boundary layers of higher and higher order as in Moore and Niiler 1975 seemed untestable and intuitively implausible outside the laboratory Physical oceanography is also to some extent a mirror of meteorology by 1970 most oceanographers were at least vaguely familiar with the picture of the atmosphere that had emerged over the previous decades In that fluid system the view of the role of eddies had shifted from a passive means of dissipating the mean flows through purely down gradient fluxes of momentum energy etc to a much more interesting and subtle dynamic linkage in which the mean flows the climate were in at least some parts of the system driven by the eddy fluxes Jeffreys 1926 Starr 1968 Lorenz 1967 Because many of the meteorological results would apply to any turbulent fluid there was reason to believe that the ocean could also exhibit such intimate dynamic linkages But we should note that even now much work is still directed at studying the mean circulation by essentially classical though improved means as if the variability were not dynamically important e g Schott and Stommel 1978 Wunsch 1978a Reid 1978 The extent to which such pictures of the mean circulation of the large scale tracers will survive complete understanding of variability dynamics is not now clear In this chapter we shall review what is known about the variability of the ocean The expression low frequency variability which is part of the title of this chapter is a vague one used in a variety of ways by oceanographers and encompassing a wide range of things Here we mean by it anything with a time scale longer than a day out to the age of the earth although we cannot really study by instrumental means phenomena with time scales longer than about 100 years In spatial scale it means phenomena ranging from some tens of kilometers to the largest possible global ocean oscillations We shall in common with recent practice also refer to the eddy field in the ocean This word is often prefixed by mesoscale and is used loosely to denote the subclass of variability encompassing motions occurring on scale of hundreds of kilometers with time scales of months and longer It is a convenient shorthand and is meant to imply neither any particular dynamics nor only flows with closed streamlines The equivalent Soviet term is synoptic scale There is little doubt that oceanographers were quite aware from the very beginning of time variability in the ocean Maury 1855 p 358 remarked that in drawing his charts he had disregarded numerous eddies and local currents which are found at sea He also notes in particular p 188 the highly variable equatorial currents of the Pacific Ocean Even earlier Rennel 1832 had quoted another observer C Blagden as referring to North Atlantic currents as casual Swallow 1976 Most of the astute observers who worked at sea since Maury were very conscious of the difficulties of drawing conclusions about the mean circulation in the presence of a highly time dependent field Figure 11 1 taken from Helland Hansen and Nansen 1909 clearly depicts what one suspects to be a time dependent eddy field Sverdrup et al 1942 make the statement that determining the mean is difficult in the presence of the time variations and that the closer the station pairs are together the greater is the requirement of simultaneity in hydrographic measurements This is of course a statement about the frequency wavenumber character
View Full Document
Unlocking...