What goes down must come up


A compilation of high-resolution measurements of ocean mixing collected over the past three decades reveals how deep ocean waters return to the surface — a process that helps to regulate Earth's climate.

Deep ocean circulation is fed by waters that become dense enough to sink into the ocean abyss in the North Atlantic Ocean and the Southern Ocean around Antarctica. These waters carry dissolved carbon away from the atmosphere and into the deep ocean, thereby playing a crucial part in modulating Earth's carbon budget and climate. Despite theoretical and observational efforts dating back to the beginning of the twentieth century, we are still struggling to understand how and where these waters return to the ocean surface — in other words, we know how ocean carbon is 'breathed in', but are still trying to figure out how it is 'exhaled'. Writing in the Journal of Physical Oceanography, Waterhouse et al.1 report remarkable progress in resolving this long-running detective story.

The first quantitative hypothesis for the return pathway of high-latitude waters was proposed in a seminal paper2 by the oceanographer Walter Munk in 1966. He speculated that dense bottom waters are mixed back up to the surface by breaking internal waves. To explain what this means, picture the ocean as a layer cake with colder — and therefore denser — layers at the bottom, and progressively warmer and lighter layers stacked on top. Internal waves are oscillations of these layers, analogous to the more familiar ocean waves that we see at the surface. Occasionally, internal waves overturn and break, much like surface waves on a beach, thereby mixing water from a denser layer into a lighter one and raising the potential energy of the ocean.

Direct in situ measurements in the years following Munk's paper, however, failed to detect enough mixing to bring back to the surface all of the high-latitude waters that sink into the abyss (calculated to require rates of approximately 30 × 106 cubic metres per second)3. Lacking alternative theories for the return pathway of bottom waters, oceanographers speculated that their measurements sampled areas of weak mixing and missed hotspots of intense mixing. An oceanographic gold rush to find the 'missing mixing' ensued.

Munk and fellow oceanographer Carl Wunsch quantified the amount of missing mixing on a global scale4 in 1998. They estimated that potential energy had to be supplied at a rate of approximately 0.4 terawatts (1 terawatt is 1012 watts) to continuously lift dense bottom waters to the ocean surface. During an internal-wave-breaking event, about 20% of the wave energy is converted into potential energy and lifts fluid, with the rest being dissipated by inconsequential small-scale motions. Internal waves would thus have to be generated at a rate of approximately 2 TW to mix bottom waters back to the surface.

At that time, it was thought that internal waves were mainly generated by variable surface wind at a rate of less than 1 TW. Munk and Wunsch4 suggested, and later work confirmed5, that internal waves are also generated by tidal forcing at a rate greater than 1 TW. More recently, it was shown that another roughly 0.5 TW is supplied by large-scale currents impinging on the bottom topography6. But just as global estimates of internal-wave generation finally seemed to be coming close to the approximately 2 TW required, in situ observations showed that internal waves tend to break close to ocean-bottom topography (the equivalent of beaches for surface waves), thus confining mixing to within a few hundred metres of the ocean bottom. So although the energy to support mixing was no longer lacking, the mixing was not delivered uniformly throughout the water column, as was needed to lift waters back to the surface.

The final piece of the puzzle was anticipated in 1998, when another seminal paper7 pointed out that most of the ocean waters above depths of 2,000 m come to the surface in the Southern Ocean, where winds known as the Roaring Forties, blowing around Antarctica, pull them to the surface along surfaces of constant density. The uplift process therefore requires no mixing. Only in the past few years have oceanographers been able to integrate Munk's hypothesis with the discovery of uplift in the Southern Ocean. The emerging view is that mixing brings bottom waters in all oceans up to about 2,000 m, the characteristic depth of the most prominent oceanic topographic features. The waters then flow at approximately the same depth all the way to the Southern Ocean, where the Roaring Forties lift them to the surface (Fig. 1).

Figure 1: An emerging model of deep-ocean circulation.

Dense waters sink into the abyss at high latitudes north and south (downward arrows). The bottom waters are lifted up to depths of about 2,000 metres by mixing processes (wiggly arrows), and return to high latitudes at these intermediate depths, eventually rising to the surface via the Southern Ocean (southward and upward pointing arrows), closing the circulation loop; shading indicates the extent of the Southern Ocean. Horizontal arrows at the surface indicate the path of waters back to high latitudes. North of the Southern Ocean, the red line indicates the heights of the tallest topographic features below which mixing is strong. In the Southern Ocean, the red line indicates the topography of the Drake Passage, to illustrate topography at latitudes at which deep water is pulled to the surface by winds (the Roaring Forties). Solid blue regions indicate the deepest points at each latitude, based on a 0.25°-resolution bathymetry data set. Waterhouse et al.1 confirm that this scenario is consistent with available observations of ocean mixing. (Adapted from ref. 9.)

In this new scenario, the potential energy required from mixing is about half that estimated by Munk and Wunsch (the ocean is on average about 4,000 m deep, and mixing lifts the waters up to only half that depth), and it needs to be supplied in the bottom 2,000 m, the characteristic height of the major ocean ridges and sea mountains. Thus, there is no shortage of energy to support mixing, and the mixing is delivered close to the bottom topography, where it is needed. Problem solved? Not quite. In situ observations show that the intensity of bottom mixing is highly variable, being strong where topography is rough and bottom flows are fast, and weak elsewhere. Mapping this heterogeneity on a global scale is the next challenge in the quest to track the return journey of abyssal waters to the surface.

Enter Waterhouse et al.1, who have gathered the largest compilation of in situ measurements of mixing so far, using them to test whether the new scenario is consistent with all available observations. They confirm that internal waves are indeed generated along the major ridges and sea mountains in the Atlantic, Pacific and Indian oceans. Most importantly, they show that about 70% of the waves break close to the ocean bottom, whereas the remaining 30% propagate away from their generation sites and end up breaking against the continental slopes. They conclude that abyssal waters make their way to the surface along the steep slopes of mid-oceanic ridges and continents, where mixing is strong.

The authors did not address the question of whether mixing is confined to depths below approximately 2,000 m — instead, they lumped together all measurements below 1,000 m. Future work must address this, because the answer is crucial for understanding and modelling the partitioning of carbon between the atmosphere and oceans. It was recently suggested8 that the drop in atmospheric carbon dioxide concentrations recorded in ice cores from glacial periods is connected to the vertical profiles of ocean mixing. In the present climate, abyssal waters release carbon to the atmosphere when they return to the surface in the Southern Ocean. But in glacial climates, a large fraction of the Southern Ocean was covered by ice, thus trapping carbon in the ocean. This trapping was possible because strong mixing was confined to the ocean bottom, and waters could not be lifted to the surface at ice-free latitudes. Similarly, the present vertical profile of mixing will control the long-term rate (on millennial timescales) at which the ocean takes up the anthropogenic carbon we are releasing into the atmosphere.


  1. 1

    Waterhouse, A. F. et al. J. Phys. Oceanogr. 44, 1854–1872 (2014).

    ADS  Article  Google Scholar 

  2. 2

    Munk, W. H. Deep Sea Res. Oceanogr. Abstr. 13, 707–730 (1966).

    ADS  Article  Google Scholar 

  3. 3

    Wunsch, C. & Ferrari, R. Annu. Rev. Fluid Mech. 36, 281–314 (2004).

    ADS  Article  Google Scholar 

  4. 4

    Munk, W. & Wunsch, C. Deep Sea Res. I 45, 1977–2010 (1998).

    Article  Google Scholar 

  5. 5

    Garrett, C. & Kunze, E. Annu. Rev. Mar. Sci. 39, 57–87 (2007).

    ADS  Google Scholar 

  6. 6

    Nikurashin, M. & Ferrari, R. Geophys. Res. Lett. 38, L08610 (2011).

    ADS  Article  Google Scholar 

  7. 7

    Toggweiler, J. R. & Samuels, B. J. Phys. Oceanogr. 28, 1832–1852 (1998).

    ADS  Article  Google Scholar 

  8. 8

    Ferrari, R. et al. Proc. Natl Acad. Sci. USA 111, 8753–8758 (2014).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Lumpkin, R. & Speer, K. J. Phys. Oceanogr. 37, 2550–2562 (2007).

    ADS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Raffaele Ferrari.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Ferrari, R. What goes down must come up. Nature 513, 179–180 (2014). https://doi.org/10.1038/513179a

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing