table of contents

Global Warming and the Thermohaline Circulation

originally posted: July 15, 2004
last updated: February 12, 2006

In a previous article I gave a short overview of the theory of global warming, discussing the evidence that the world is getting warmer, why it is getting warmer, and the likely effects. I think the contents of that article gives a pretty good summation of what most scientists today believe is happening with our global climate.

Some scientists, however, have an additional concern that I didn't discuss in that article. They worry that fairly modest amounts of global warming could shut down the thermohaline circulation. If that should happen, then the effects on the Earth's climate could be profound. This theory has gained some cultural currency. The disaster movie The Day After Tomorrow is based on a wildly exaggerated version of this scenario. The Pentagon has been busy working out strategies for what to do in this scenario (which should not be taken as evidence that they believe it is likely - they wargame out worst cases of all sorts of unlikely scenarios, partly to be prepared for anything, but mostly as training exercises for their officers).

The likelihood of this actually happening is difficult to assess. It's certainly possible, but I don't think you could even meaningfully attach a probability to it. We really don't know enough about the Earth's climate yet. So I don't think it's time yet to stock up on mukluks and canned Spam, but I think this scenario is a very instructive example of the complexity and potential instability of the Earth's climate. When humans make small changes, the results are not necessarily always small.

The Thermohaline Circulation

Ocean currents are driven and shaped by a variety of different forces, resulting in two major kinds of currents. In each of the ocean basins, there is a great circular surface current called a gyre. Gyres rotate clockwise in the northern hemisphere and counterclockwise in the southern hemisphere. The North Atlantic Gyre, for instance, begins with the Gulf Stream travelling north along the east coast of the United States. Then, under the name "North Atlantic Current," it turns eastward toward Europe. When it turns south past Spain and Northwestern Africa, we call it the "Canary Current". Before it reaches the equator, it turns west again, becoming the "North Equatorial Current," and then starting north again as the Gulf Stream. These gyres are driven by wind and shaped into loops by coriolis forces resulting from the rotation of the earth. They effect only the surface waters of the ocean.

Thermohaline Circulation in the Atantic
The other major type of ocean current is the thermohaline circulation. This is often called the "conveyor belt", because it is not just a surface current - it travels one way along the surface, and then dives down and returns along the ocean floor, before rising up to complete the loop.

At right I've sketched the Atlantic portion of the thermohaline circulation. Let's start on the red line, a north-going surface current. As it enters the North Atlantic, it joins the North Atlantic Gyre. As these surface waters drift north, some of the heat they picked up at the equator seeps into the atmosphere, where the prevailing westerly winds carry it out to Europe. As the water moves north, not only does it cool off, but it grows saltier. This is because as water evaporates from the surface, it leaves the salt behind, so the concentration of salt steadily increases.

Up north near Iceland and Greenland, the cooling and saltification (that's "thermo-" and "-haline") processes reach a critical point. The surface water is saltier than the deep water and thus denser. The only thing holding it up was the fact that it was warmer than the deep water, and thus less dense, but as its heat seeps away and its saltiness increases, the surface water suddenly sinks. This happens in two relatively small regions, one in the Labrador Sea between Greenland and Canada, the other in the Norwegian Sea, north of Iceland, between Greenland and Scandinavia.

This cold, salty water then heads back south, as a deep ocean current, shown by the blue line in my sketch. It moves slowly, about 0.1 meters per second (0.22 miles per hour), but it is a vast current, carrying about as much water as 100 Amazon Rivers. After about 200 years, it reaches the Antartic, where it joins the general westward current around Antartica. It veers north into the Indian Ocean. Some upwells there. Some proceeds into the Pacific Ocean, where after about 1000 years crawling along the ocean floors, it rises back to the surface. The areas where it rises in the Indian and Pacific Oceans are quite large, in contrast to the rather small sink zones. The deep cold water is rich in carbon dioxide and nutrients that have drifted down into it from all the oceans of the world, so these upwelling zones tend to be rich in sea life.

Once on the surface, the current returns around Africa into the Atlantic closing the loop as the red line in my sketch. Note that the surface portion of the thermohaline circulation merges into the gyre currents, but the deep cold waters and the warm surface waters do not normally mix much.

This whole globe-spanning circulation is driven by the two sink zones in the North Atlantic. There are two other similar sink zones in the world, near Antartica in the Ross and Wendell Seas, but their effects are more local. Due to differences in geography, there are no similarly significant sink zones in the North Pacific. It's not entirely clear why. One theory is that the high mountain ranges on the west coast of North America cause much rain to fall there, draining into the ocean and reducing the saltiness enough so no sinkage occurs.

Historical Changes

Scientists believe that the thermohaline circulation has sometimes stopped. The most recent and best documented time when some scientists believe that might have happened was during a time period called the "Younger Dryas".

Greenland Temperatures
The graph at right show temperature data for the last 25,000 years drawn from ice cores in Greenland. The left side of the graph represents the last ice age. We can see how temperatures gradually increased as the ice age ended, rising up to modern levels. But about 12,700 years ago there was a sudden reversal. Temperatures dropped abruptly back down to ice age levels, falling about 10 degrees Fahrenheit within a period of 50 years. It stayed cold for about 1300 years before rising again, just as abruptly.

This abrupt temperature drop appears not only in the Greenland ice records. Data about ancient climates has been collected in many ways, including by studying pollen remains in bog sediments, moraines left by advancing and retreating glaciers and plankton remains in ocean bottom sediments. Temperatures in northwest Europe probably dropped 7 degrees Fahrenheit within a few decades. More moderate effects were felt world wide. The Younger Dryas may have caused the droughts in the middle east that motivated the invention of agriculture. It may have contributed to the extinction of many large mammals in North America.

We know that as the ice age ended, the the North American ice sheets gradually melted. Before the onset of the Younger Dryas, a large fresh water meltwater lake was formed in North America. Just at the time the Younger Dryas began, the ice dam separating this lake from the North Atlantic broke, and a huge amount of fresh water poured into the ocean.

The effect of this would have been to suddenly and significantly reduce the saltiness of the surface waters in the North Atlantic. The theory is that this would reduce their density, and they would no longer sink. Since the sinkage in the North Atlantic drives the whole thermohaline circulation, it all would have stopped. Among the notable effects would be a cooling in Europe, which is normally warmed substantially by the current.

During the 50,000 years of the last ice age, there were perhaps 20 climate shifts that appear similar in nature. Some believe that during the ice age the thermohaline circulation toggled on and off repeatedly. During off cycles, ice caps expanded, removing fresh water and increasing the salinity of the northern oceans until the circulation restarted, warming the northern oceans and remelting the ice until salinity dropped enough to stop the circulation again.

None of this is certain. There is some significant contradictory evidence. For example, a recent study of sediment data failed to find evidence of a slowdown of currents during the Younger Dryas. So while most scientists agree that a shutdown of the thermohaline circulation is possible, it is not clear how frequently it has happened in the past, or how big a climate change would be required to trigger it.

Influence of Global Warming on Thermohaline Circuation

It is possible that global warming could effect the thermohaline circulation. There are several ways in which this happens. First as things get warmer, the north flowing surface waters don't cool off as much. This effect is likely to be exaggerated because it is likely that with global warming, the north will warm more than the south. Greenland, in particular, is likely to experience greater than average temperature increases. Second, global warming is likely to increase rainfalls in the northern region, so there will be more fresh water raining directly into the North Atlantic or running in from the surrounding continents, so the saltiness of the north flowing surface currents will be reduced. Third, there is likely to be melting of polar ice caps, and the Greenland ice sheet, which would further freshen the surface ocean waters.

So, if the surface waters become less salty and less cold, they are clearly going to be less inclined to sink. The thermohaline circulation would slow, and, beyond some point, suddenly stop. Within a period of three to ten years, one or both of the sink points could simply stop functioning, and the whole conveyor could halt.

This abrupt change happens because the conditions that make the thermohaline circulation possible are, in part, caused by the thermohaline circulation. It is due to the thermohaline circulation that the North Atlantic is so warm, and it is, in part, this warmth that causes so much evaporation and increase in salinity. So the thermohaline circulation may be partly self-sustaining. If it slows beyond a certain point, then instead of slowing further, it completely stops.

This also means the thermohaline circulation does not restart easily. One might think that the cooling caused by the shutdown would counteract the warming that stopped it in the first place, causing it to start right back up. As the historical record seems to show, this does happen, but not quickly. Models indicate that it could take about 500 years to recover from a 30% reduction of the thermohaline circulation. How long it would take to restart after a complete shutdown is hard to even guess.

There is some evidence that something of this sort is going on now. We have data showing that the North Atlantic has been growing steadily less salty over the last 40 years. Measurements of the deep ocean currents moving south from the Nordic seas show a reduction of 20% since 1950. These deep waters have been growing fresher over the last 3 or 4 decades.

The likelihood of a shutdown as a result of global warming is difficult to assess. Typical estimates say it would require a global temperature increase of 4 to 5 degrees Celsius (7 to 9 degrees Fahrenheit) to trigger such an event. This is substantially more then is expected within the next century, so most scientists consider it unlikely. The more likely scenario for the next century is simply a reduction in the thermohaline circulation, though the magnitude of that reduction is hard to predict.

Some scientists, especially those who think there was no shutdown of the thermohaline circulation during the Younger Dryas do not think the temporarture changes resulting from global warming could be enough to shut down the the thermohaline circulation. After all, global warming would probably release less fresh water than were released at the start of the Younger Dryas. So the estimates above should probably be considered to be the on the pessimistic side of the spectrum.

Effects of a Thermohaline Circulation Failure

The effects of a shutdown, if it occured, would likely be most profound in Europe. If you look at map, you'll see that Europe is far north of other parts of the world with comparable temparate climates. Europe's unusual warmth depends on the warm surface currents that bring heat up from the equator. There have been widely circulated predictions that failure of the thermohaline circulation would cause average temperatures in Europe to drop 10 to 20 degrees Fahrenheit. Iceland would be glaciated down to sea level, requiring it to be abandoned. Ireland would have a climate more like Spitzbergen. Northern Europe would not be uninhabitable, but it would be more like Siberia than New York.

These predictions probably overstate the case. Though global warming could shut down the thermohaline circulation, it would not shut down the North Atlantic gyre. The forces driving that, including winds and expansion of water at the equator, would continue, although their pattern and strength could alter. Thus a shutdown of the thermohaline circulation should not completely halt the flow of the Gulf Stream, which is what the predictions above presume. However the Gulf Stream's influence on European climate is only about half as much as that of the thermohaline circulation, and without the influence of the thermohaline circulation the Gulf Stream might shift further southward, lessening it's influence on Europe.

There are more optimistic predictions say that if the thermohaline circulation failed, then the cooling in Greenland would be around 14 degrees Fahrenheit, Europe would only cool about 4 degrees, which would conveniently counteract the global warming that triggered the shutdown in the first place. Different models give very different results.

Though it is hard to be sure what the likelihood of a shutdown of the thermohaline circulation is, it seems almost certain that some slowdown is going to happen. This would have still more moderate effects.


When human beings assess risk, we tend to give undue weight to spectacularly bad outcomes, even if they are unlikely ones. The possibility of a thermohaline circulation shutdown is probably in that category. There are much more probable effects of global warming whose combined effects should be quite sufficient to inspire us to action. The possibility of another Younger Dryas event, though spectacular, is sufficiently unlikely, at least in the near future, that it really shouldn't be allowed alter our assessment of the risk of global warming very much.

But this scenario does demonstrate some important facts about the Earth's climate relevant to global warming. First, the Earth's climate has a history of dramatic fluxuations. The kinds of climate changes most scientists expect to result from global warming are actually fairly modest compared to what has happened in the past. We should not doubt the capacity of the Earth's climate to undergo such changes.

Second, the Earth's climate is very complex. Predictions that the Earth will be a few degrees warmer do not mean that every part of the earth will get warmer by that amount. Any such change will actually alter weather patterns. Some parts of the world will get warmer, some colder. Winds will change. Patterns of rainfall will change. The frequency of storms will change. Predicting the exact effects is difficult, but predicting that they won't be uniform is a done deal.

Third, local climate matters. You might think that, of all the things we have to worry about, the weather in Greenland would be pretty low on the list. After all, essentially nobody lives there. But, in fact, it happens to have profound effects on the whole world's climate. (Not to mention the importance of the fisheries there to the world food supply.)

Fourth, there are cliffs you can walk off. Take one step, things change just a bit. Take another step, things change just a bit more. Take another step, and the ground drops out from under you. Sometimes small changes have small effects. Sometimes small changes can cause the climate to flip into whole new patterns of operation, with profound worldwide effects. Just because the bit of global warming that has already be observed has had only moderate effects, that doesn't mean that continuing on the same course will do the same.

To quote Wallace Broecker, "The record of events that transpired during the last glacial period sends us the clear warning that by adding greenhouse gases to the atmosphere, we are poking an angry beast". I'm not sure that it is angry, but it certainly is a beast.

view 3 reader comments
Originally created with Backtalk version 1.4.5 / Wasabi version 1.0.3 - Copyright 1996-2005, Jan Wolter and Steve Weiss