Chapter 13 - Deep Circulation in the Ocean

The direct forcing of the oceanic circulation by wind discussed in the last few chapters is limited mostly to the upper kilometer of the water column. Below a kilometer lies the vast water masses of the ocean extending to depths of 4 km - 5 km. The water is everywhere cold, with a potential temperature less than 4°C. The water mass is formed when cold, dense water sinks from the surface to great depths at high latitudes. It spreads out from these regions to fill the ocean basins. Deep mixing eventually pulls the water up through the thermocline over large areas of the ocean. It is this upwelling that drives the deep circulation. The vast deep ocean is usually referred to as the abyss, and the circulation as the abyssal circulation.

The densest water at the sea surface, water that is dense enough to sink to the bottom, is formed when frigid air blows across the ocean at high latitudes in winter in the Atlantic between Norway and Greenland and near Antarctica. The wind cools and evaporates water. If the wind is cold enough, sea ice forms, further increasing the salinity of the water because ice is fresher than sea water. Bottom water is produced only in these two regions. In other polar regions, cold, dense water is formed, but it is not quite salty enough to sink to the bottom.

At mid and low latitudes, the density, even in winter, is sufficiently low that the water cannot sink more than a few hundred meters into the ocean. The only exception are some seas, such as the Mediterranean Sea, where evaporation is so great that the salinity of the water is sufficiently great for the water to sink to intermediate depths in the seas. If these seas are can exchange water with the open ocean, the waters formed in winter in the seas spreads out to intermediate depths in the ocean.

Defining the Deep Circulation
Many terms have been used to describe the deep circulation. They include:

1. Abyssal circulation;
2. Thermohaline circulation;
3. Meridional overturning circulation; and
4. Global conveyor.

Thermohaline circulation seems to be the most widely used. Unfortunately, the concept has not been clearly defined, and the term is widely misused (Wunsch, 2002b).

It seems most sensible to to regard the thermohaline circulation as the circulation of temperature and salt. However, because the three-dimensional (3D) distributions and surface boundary conditions of temperature and salt are different, it should come as no surprise that one must separate the thermal circulation from the salt (or freshwater) circulation.
—Wunsch (2002b)

The meridional overturning circulation is better defined. It is the zonal average of the flow plotted as a function of depth and latitude. Plots of the circulation show where vertical flow is important, but they show no information about how circulation in the gyres influences the flow.

Following Wunsch (2002b), I define the deep circulation as the circulation of mass. Of course, the mass circulation also carries heat, salt, oxygen, and other properties. But the circulation of the other properties is not the same as the mass transport. For example, Wunsch points out that the North Atlantic imports heat but exports oxygen.

The deep circulation is mostly wind driven, but tidal mixing is also important. The wind enters several ways. It cools the surface and evaporates water, which determines where deep convection occurs. And, it produces turbulence in the deep ocean which mixes cold water upward.

chapter contents

13.1 Importance of the Deep Circulation

The deep circulation which carries cold water from high latitudes in winter to lower latitudes throughout the world has very important consequences.

1. The contrast between the cold deep water and the warm surface waters determines the stratification of the oceans. Stratification strongly influences ocean dynamics.
2. The volume of deep water is far larger than the volume of surface water. Although currents in the deep ocean are relatively weak, they have transports comparable to the surface transports.
3. The fluxes of heat and other variables carried by the deep circulation influences Earth's heat budget and climate. The fluxes vary from decades to centuries to millennia, and this variability is thought to modulate climate over such time intervals. The ocean may be the primary cause of variability over times ranging from years to decades, and it may have helped modulate ice-age climate.

Two aspects of the deep circulation are especially important for understanding Earth's climate and its possible response to increased carbon dioxide CO₂ in the atmosphere:

1. The ability of cold water to absorb CO₂ from the atmosphere, and
2. The ability of deep currents to modulate the heat transported from the tropics to high latitudes.

The Oceans as a Reservoir of Carbon Dioxide
The oceans are the primary reservoir of readily available CO₂, an important greenhouse gas. The oceans contain 40,000 GtC of dissolved, particulate, and living forms of carbon. The land contains 2,200 GtC, and the atmosphere contains only 750 GtC. Thus the oceans hold 50 times more carbon than the air. Furthermore, the amount of new carbon put into the atmosphere since the industrial revolution, 150 GtC, is less than the amount of carbon cycled through the marine ecosystem in five years. (1 GtC = 1 gigaton of carbon = 10¹² kilograms of carbon.) Carbonate rocks such as limestone, the shells of marine animals, and coral are other, much larger, reservoirs. But this carbon is locked up. It cannot be easily exchanged with carbon in other reservoirs.

More CO₂ dissolves in cold water than in warm water. Just imagine shaking and opening a hot can of Coke^TM. The CO₂ from a hot can will spew out far faster than from a cold can. Thus the cold deep water in the ocean is the major reservoir of dissolved CO₂ in the ocean.

New CO₂ is released into the atmosphere when fossil fuels and trees are burned. Very quickly, 48% of the CO₂ released into the atmosphere dissolves in the cold waters of the ocean, much of which ends up deep in the ocean.

Forecasts of future climate change depend strongly on how much CO₂ is stored in the ocean and for how long. If little is stored, or if it is stored and later released into the atmosphere, the concentration in the atmosphere will change, modulating Earth's long-wave radiation balance. How much and how long CO₂ is stored in the ocean depends on the deep circulation and the net flux of carbon deposited on the seafloor. The amount that dissolves depends on the temperature of the deep water, the storage time in the deep ocean depends on the rate at which deep water is replenished, and the deposition depends on whether the dead plants and animals that drop to the sea floor are oxidized. Increased ventilation of deep layers, and warming of the deep layers could release large quantities of the gas to the atmosphere.

The storage of carbon in the ocean also depends on the dynamics of marine ecosystems, upwelling, and the amount of dead plants and animals stored in sediments. But we won't consider these processes.

Oceanic Transport of Heat
The oceans carry about half the heat out of the tropics needed to maintain earth's temperature. Heat carried by the Gulf Stream and the north Atlantic drift warms the north Atlantic, keeping it ice free in winter, and it helps warm Europe. Norway, at 60°N is far warmer than southern Greenland or northern Labrador at the same latitude. Palm trees grow on the west coast of Ireland, but not in Newfoundland which is further south.

Wally Broecker (1987), working at Lamont-Doherty Geophysical Observatory of Columbia University, calls the oceanic component of the heat-transport system the Global Conveyor Belt. The basic idea is that the Gulf Stream carries heat to the far north Atlantic (Figure 13.1). There the surface water releases heat and water to the atmosphere. Some of the ocean water becomes sufficiently cold, salty, and dense that it sinks to the bottom in the Norwegian and Greenland Seas. It then flows saothward in very cold, bottom currents along western boundaries as a western boundary current. Some of the water remains at the surfae and returns to the south in cool surface currents such as the Labrador Current and the Portugal Current (see Figure 11.8).

Figure 13.1 The surface (red, orange, yellow) and deep (violet, blue, green) currents in the North Atlantic. The North Atlantic Current brings warm water northward where it cools. Some sinks and returns southward as a cold, deep, western-boundary current. Some returns southward at the surface. From Woods Hole Oceanographic Institution.

The deep bottom water from the north Atlantic is mixed upward in other regions and ocean, and eventually it makes its way back to the Gulf Stream and the North Atlantic. Thus most of the water that sinks in the north Atlantic must be replaced by water from the far south Atlantic. As this surface water moves northward across the equator and eventually into the Gulf Stream, it carries heat out of the south Atlantic. So much heat is pulled northward by the formation of north-Atlantic bottom water in winter that heat transport in the Atlantic is entirely northward, even in the southern hemisphere (figure 5.11). Much of the solar heat absorbed by the tropical Atlantic is shipped north to warm Europe and the northern hemisphere. Imagine then what might happen if the supply of heat is shut off. We will get back to that topic in the next section.

We can make a crude estimate of the importance of the conveyor-belt circulation from a simple calculation based on what we know about waters in the Atlantic compiled by Bill Schmitz (1996) in his wonderful summary of his life's work. The Gulf Stream carries 40 Sv of 18°C water northward. Of this, 14 Sv return southward in the deep western boundary current at a temperature of 2°C. The flow carried by the conveyor belt must therefore lose 0.9 petawatts (1 petawatt = 10¹⁵ watt) in the north Atlantic north of 24°N. Although the calculation is very crude, it is remarkably close to the value of 1.2 ± 0.2 petawatts estimated much more carefully by Rintoul and Wunsch (1991).

Note that if the water remained on the surface and returned as an eastern boundary current, it would be far warmer than the deep current when it returned southward. Hence, the heat transport would be much reduced. Furthermore, the warm water would not reach high latitudes in the North Atlantic, and they would freeze over in winter, leading to further cooling of the region.

The production of bottom water is influenced by the salinity of surface waters in the north Atlantic. It is also influenced by the rate of upwelling due to mixing in other oceanic areas. First, let's look at the influence of salinity.

Saltier surface waters form denser water in winter than less salty water. At first you may think that temperature is also important, but at high latitudes water in all ocean basins gets cold enough to freeze, so all ocean produce -2° C water at the surface. Of this, only the most salty will sink, and the saltiest water is in the Atlantic and under the ice on the continental shelves around Antarctica.

The production of bottom water is remarkably sensitive to small changes in salinity. Rahmstorf (1995), using a numerical model of the meridional overturning circulation, showed that a ±0.1 Sv variation of the flow of fresh water into the north Atlantic can switch on or off the deep circulation of 14 Sv. If the deep-water production is shut off during times of low salinity, the 1 petawatt of heat may also be shut off. Weaver and Hillaire-Marcel (2004) point out that the shutdown of the production of bottom water is unlikely, and if it did happen, it would lead to a colder Europe, not a new ice age, because of the higher concentrations of CO₂ now in the atmosphere.

I write may be shut off because the ocean is a very complex system. We don't know if other processes will increase heat transport if the deep circulation is disturbed. For example, the circulation at intermediate depths may increase when deep circulation is reduced.

The production of bottom water is also remarkably sensitive to small changes in mixing in the deep ocean. Munk and Wunsch (1998) calculate that 2.1 TW (terawatts = 10¹² watts) are required to drive the deep circulation, and that this small source of mechanical mixing drives a poleward heat flux of 2000 TW. Most of the energy for mixing comes from winds which can produce turbulent mixing throughout the ocean. Some energy comes from the dissipation of tidal currents, which depend on the distribution of the continents. Thus during the last ice age, when sea level was much lower, tides, tidal currents, tidal dissipation, winds, and deep circulation all differed from present values.

Role of the Ocean in Ice-Age Climate Fluctuations
What might happen when the production of deep water in the Atlantic is shut off? Information contained in the Greenland and Antarctic ice sheets and in north Atlantic sediments provide important clues.

Several ice core through the Greenland ice sheet and three through the Antarctic sheet provide a continuous record of atmospheric conditions over Greenland and Antarctica extending back more than 400,000 years before the present in some cores. Annual layers in the core are counted to get age. Deeper in the core, where annual layers are hard to see, age is calculated from depth. Occasional world-wide dustings of volcanic ash provide common markers in cores. Oxygen-isotope ratios in the ice give temperatures over parts of the northern hemisphere; bubbles in the ice give atmospheric CO₂ and methane concentration; pollen, chemical composition, and particles give information about volcanic eruptions, wind speed, and direction; thickness of annual layers gives snow accumulation rates; and isotopes of some elements give solar and cosmic ray activity (Alley, 2000).

Cores through deep-sea sediments in the north Atlantic made by the Ocean Drilling Program give information about sea-surface temperature and salinity above the core, the production of north Atlantic deep water, ice volume in glaciers, and production of icebergs.

1. The oxygen-isotope record in the ice cores show abrupt temperature variability over the past 100,000 years. Many times during the last ice age, temperatures near Greenland warmed rapidly over periods of 1-100 years, followed by gradual cooling over longer periods (Dansgaard et al., 1993). For example, roughly 11,500 years ago, temperatures over Greenland warmed by roughly 8°C in 40 years in three steps, each spanning 5 years (Alley, 2000). Such abrupt warming is called a Dansgaard/Oeschger event. Other studies have shown that much of the northern hemisphere warmed and cooled in phase with temperatures calculated from the ice core.
   
   
2. The climate of the past 8,000 years was constant with very little variability. Our perception of climate change is thus based on highly unusual circumstances. All of recorded history has been during a period of warm and stable climate.
   
   
3. Hartmut Heinrich and colleagues (Bond et al., 1992), studying the sediments in the north Atlantic noticed periods when coarse material was deposited on the bottom in mid-ocean. Only icebergs can carry such material out to sea, and the find indicated times when large numbers of icebergs were released into the north Atlantic. These are now called Heinrich events.
   
   
4. The correlation of Greenland temperature with iceberg production is related to the meridional overturning circulation. When icebergs melted, the surge of fresh water increased the stability of the water column shutting off the production of North Atlantic Deep Water. The shut-off of deep-water formation greatly reduced the transport of warm water in the north Atlantic, producing very cold northern hemisphere climate (Figure 13.2). The melting of the ice pushed the polar front, the boundary between cold and warm water in the north Atlantic further south than its present position. The location of the front, and the time it was at different positions can be determined from analysis of bottom sediments.
   

   Figure 13.2 Periodic surges of icebergs during the last ice age appear to have modulated temperatures of the northern hemisphere by lowering the salinity of the far north Atlantic and reducing the meridional overturning circulation. Data from cores through the Greenland ice sheet (1), deep-sea sediments (2,3), and alpine-lake sediments (4) indicate that: Left: During recent times the circulation has been stable, and the polar front which separates warm and cold water masses has allowed warm water to penetrate beyond Norway. Center: During the last ice age, periodic surges of icebergs reduced salinity and reduced the meridional overturning circulation, causing the polar front to move southward and keeping warm water south of Spain. Right: Similar fluctuations during the last interglacial appear to have caused rapid, large changes in climate. The Bottom plot is a rough indication of temperature in the region, but the scales are not the same. From Zahn (1994).

   
   
5. When the meridional overturning circulation shuts down, heat normally carried from the south Atlantic to the north Atlantic becomes available to warm the southern hemisphere. This explains the Antarctic warming.
   
   
6. The switching on and off of the meridional overturning circulation has large hysteresis (Figure 13.3). The circulation has two stable states. The first is the present circulation. In the second, deep water is produced mostly near Antarctica, and upwelling occurs in the far north Pacific (as it does today) and in the far north Atlantic. Once the circulation is shut off, the system switches to the second stable state. The return to normal salinity does not cause the circulation to turn on. Surface waters must become saltier than average for the first state to return (Rahmstorf, 1995)
   
   

   Figure 13.3 The meridional-overturning circulation is part of a non-linear system. The circulation has two stable states near 2 and 4. The switching of north Atlantic from a warm, salty regime to a cold, fresh regime and back has hysteresis. This means that as the warm salty ocean in an initial state 1 freshens, and becomes more fresh than 2 it quickly switches to a cold, fresh state 3. When the area again becomes salty, it must move past state 4 before it can switch back to 1.

   
   
7. Heinrich events seem to precede the largest Dansgaard/Oeschger events (Stocker and Marchal, 2000). Here's what seems to happen. The Heinrich event shuts off the Atlantic deep circulation which leads to a very cold North Atlantic. This is followed about 1000 years later by a Dansgaard/Oeschger event with rapid warming. Dansgaard/Oeschger-Heinrich tandem events have global influence, and they seem to be related to warming events seen in Antarctic ice cores. Temperatures changes in the two hemispheres are out of phase. When Greenland warms, Antarctica cools.
   
   
8. A weakened version of this process with a period of about 1000 years may be modulating present-day climate in the north Atlantic, and it may have been responsible for the Little Ice Age from 1100 to 1800.

The relationship between variations in salinity, air temperature, and deep-water formation is not yet well understood. For example, we don't know what causes the ice sheets to surge. Surges may result from warmer temperatures caused by increased water vapor from the tropics (a greenhouse gas) or from an internal instability of a large ice sheet. Nor do we know exactly how the oceanic circulation responds to changes in the deep circulation or surface moisture fluxes. Recent work by Wang, Stone and Marotzke (1999), who used a numerical model to simulate the climate system, shows that the meridional overturning circulation is modulated by moisture fluxes in the southern hemisphere.

The oceans play a key role in the development of the ice ages. Every 100,000 years for the past million years, ice sheets have advanced over the continents. Shackleton (2000) finds that the 100,000-year period of Earth's orbital eccentricity, deep-sea temperature, and atmospheric carbon-dioxide concentration are well correlated over the 100,000-year cycle. He also finds that ice-sheet volume lagged behind CO₂ changes in the atmosphere, implying that ice sheets changed as a result of CO₂ changes, not the other way around.

chapter contents