The view that much of the energy of ocean tides is dissipated in deep water, rather than in shallow coastal seas, now finds observational support. Curiously, the results bear upon our understanding of climate change.
The Moon is receding from the Earth at about 4 centimetres a year, as measured1 by laser reflectors left there by astronauts. What does this motion have to do with the ocean circulation? By Kepler's laws, the recession implies that there is a continuing loss of energy in the Earth–Moon system of 3×1012 watts, or 3 terawatts, mostly in the ocean. But where in the ocean does this energy go, and what are its effects? On page 775 of this issue, Egbert and Ray2 produce evidence for the seemingly lunatic conclusion3 that dissipation of tidal energy in the deep sea, and the resulting mixing, are controlling features of the overall ocean circulation.
In the conventional picture of the oceans, there is a wind-driven upper circulation that gives rise to massive, near-surface flows such as the Gulf Stream and the Kuroshio and Antarctic Circumpolar currents. Superimposed upon this circulation is one often labelled, unhappily, the ‘thermohaline’ circulation. This is supposedly driven by surface-ocean density contrasts arising from temperature and salt variations produced by strong atmospheric cooling and wind-induced evaporation. In this process, dense water sinks at high latitudes through convection, driving a ‘meridional overturning circulation’, which many believe dominates the heat and freshwater budgets of the climate system. (The terminological problem with ‘thermohaline’ circulation arises because, for example, half the heat transport in the North Pacific Ocean is in the wind-driven upper circulation4.)
As formulated in almost all models of the Earth's climate, both theoretical and numerical, the dense water sinking in the meridional overturning circulation at high latitudes then flows, close to the ocean bottom, throughout the world, returning to the surface by a uniform upwelling through the ‘interior’ ocean. Under the simple assumption that a uniform upwelling of cold water is balanced by a uniform downward mixing of warmer water throughout the water column, a steady state is achieved. Almost all numerical models of the ocean and climate systems represent this process through spatially constant vertical ‘eddy’-mixing coefficients, as do the textbook theories.
It has become evident, however, that the actual circulation is much more subtle and interesting. Consideration of the stability and energetics of a fluid being heated and cooled at the surface5 shows that the resulting motion would be extremely weak — a ‘diffusive creep’. Such a fluid system is stable, and in a steady state it cannot produce the vigorous flow we observe in the deep oceans. There cannot be a primarily convectively driven circulation of any significance3,6.
Furthermore, as early as about 1970 it was clear that the picture of uniform vertical mixing was not correct in the upper ocean7,8,9. Both direct measurements of turbulence and dye-diffusion experiments10,11 have shown the weakness of open-ocean mixing all the way down to the sea floor. Instead, measurements confirm an old hypothesis7: that the ocean is mixed primarily at its boundaries, including the mid-ocean ridges that rise from the sea floor, where the local mixing rates are orders of magnitude larger than those in the ocean interior10,11. The reliance of almost all numerical circulation models on uniform interior-ocean mixing calls into question inferences about the physics of the circulation based on them. Only in the past two years have models that eliminate such mixing12 finally started to appear.
Surprisingly, it was only recently recognized that the need for an energy source to sustain the vertical mixing (lifting dense water through lighter) has important consequences. The difficulties of driving fluid motions by surface heating and evaporation mean that a mechanical source of energy must control not only the directly wind-driven flows, but also the deep-water components of the meridional overturning circulation. There are only two candidates for such a source: winds and tides.
For over 75 years13,14 it was thought that the tides dissipated almost entirely by friction in the shallow seas above the continental shelves. But Munk and I concluded3 that about half of the power required to return the deep waters to the surface was coming from mixing driven primarily by dissipation of tidal energy — principally lunar, but with a minor solar component — in the deep ocean (Fig. 1). Now, by fitting a dynamical model to satellite altimetric measurements of the tides, Egbert and Ray2 have produced an observational estimate of 1 terawatt of open-ocean tidal dissipation. Their numbers are not definitive, but they are in agreement with the energy values required by the deep upwelling, and with the total — shallow (about 2 terawatts) plus deep — energy losses implied by the lunar recession.
If the hypothesis of tide- and wind-driven controls on the rates at which the ocean transports heat and fresh water survives further tests, there are several implications. One is that it brings into question the extent to which uniform-mixing models of the ocean circulation could either reproduce the present-day circulation or predict responses to external changes. The hypothesis also suggests that the rate-limiting15 factor for oceanic heat transport is not primarily the surface density gradient imposed on the ocean; rather, it is the strengths and patterns of the wind, and the distributions of the tides.
What would be the consequences if the hypothesis is correct? One is that the atmospheric wind patterns would have to be known in considering past and future climate change. The other is that changes in tidal distributions and the consequent mixing would need to be understood over geological time. During the Last Glacial Maximum, the sea level was about 130 metres lower than today. This configur- ation removed much of the present regions of shallow-water energy dissipation and changed the deep-ocean tides, presumably affecting oceanic heat transport. Over longer periods in the past, the entire continental configuration was different, with radically different tidal distributions and mixing. It appears that the tides are, surprisingly, an intricate part of the story of climate change, as is the history of the lunar orbit.
Dickey, J. O. et al. Science 265, 482–490 (1994).
Egbert, G. D. & Ray, R. D. Nature 405, 775–778 (2000).
Munk, W. & Wunsch, C. Deep-Sea Res. 45, 1976–2009 (1998).
Bryden, H. L., Roemmich, D. H. & Church, J. A. Deep-Sea Res. 38, 297– 324 (1991).
Sandström, J. W. Ann. Hydr. Mar. Met. 6 (1908).
Huang, R. X. J. Phys. Oceanogr. 29, 727–746 (1999).
Munk, W. H. Deep-Sea Res. 13, 707–730 (1966).
Osborn, T. R. & Cox, C. S. Geophys. Fluid Dyn. 3 , 321–345 (1972).
Gregg, M. C. J. Geophys. Res. 92, 5249–5286 (1987).
Ledwell, J. R. et al. Nature 403, 179–182 (2000).
Polzin, K., Toole, J. M., Ledwell, J. R. & Schmitt, R. W. Science 276, 93–96 ( 1997).
Marotzke, J. J. Phys. Oceanogr. 27, 1713–1728 (1997).
Jeffreys, H. Phil. Trans. R. Soc. Lond. A 221, 239– 264 (1920).
Taylor, G. I. Phil. Trans. R. Soc. Lond. A 220, 1– 33 (1919).
Marotzke, J. & Scott, J. R. J. Phys. Oceanogr. 29, 2962–2970 (1999).
Munk, W. & Wunsch, C. Oceanography 10, 132–134 (1997).
About this article
Journal of Fluid Mechanics (2019)
Climate Dynamics (2016)
Biomixing due to diel vertical migrations of zooplankton: Comparison of computational fluid dynamics model with observations
Ocean Modelling (2016)
Journal of Geophysical Research: Oceans (2015)
Soft Matter (2014)