The Decade After Tomorrow?

Authors note: This post was mostly written before Christmas, but I then got distracted and didn't get the chance to finish it. I guess it's a bit less topical now but I don't like to waste hard-written words...

What does the the paper published in Nature last month by Bryden et al.[1] have in common with a recent Hollywood disaster movie? Answer: before The Day After Tomorrow descended into 3000km-wide superstorms sucking down air from the stratosphere and turning everybody into icicles (casually breaking the laws of thermodynamics in the process), both are concerned with a possible weakening of the thermohaline circulation in response to climate change.

However much we moan about the weather here in the UK, we should bear in mind that we are located roughly as far north as Hudson Bay. Which, it could be argued, is a tad cooler than here. The reason for this is that we (and the rest of Western Europe) benefit from the Gulf Stream, a strong current of warm surface water that flows from the Gulf of Mexico to the northeast Atlantic, bringing a lot of heat from the tropics with it. A large part of this north-eastward flow is wind-driven, but not all of it is. In the current climate system, heat loss and evaporation causes seawater in the northern Atlantic, particularly in the Nordic and Labrador Seas, to become so cold and salty, and therefore dense, that it sinks into the deep ocean and moves south back towards the equator below the surface currents, at depths of several thousand metres. About 20-30% of the total northward flow of water in the Gulf Stream is moving to replace the downwelling water in the North Atlantic as part of this 'thermohaline circulation'; the additional heat transfer from equatorial regions, and the penetration of warm water to much higher latitudes than would result from wind-driven flow alone, combine to produce much higher seasonal temperatures on our side of the Atlantic.

Looking back into the past, we can see a link between the strength of the thermohaline circulation and high latitude temperatures in the Northern Hemisphere. The sinking water in the Nordic and Labrador seas - often referred to as the North Atlantic Deep Water (NADW) has chemical properties which make it distinct from, for example, water from the oceans around Antarctica, or the Pacific. By taking sediment cores from the South Atlantic and chemically analysing shells or minerals that formed in ancient seas thousands of years ago, the penetration of the NADW into this region, which is greater when the thermohaline circulation is stronger, can be estimated (e.g. [2]). Such studies show that at the end of the last Ice Age, about 19,000 years ago, when glaciers covered much of northern Europe, the thermohaline circulation was much weaker than today, and strengthened over the next 10,000 years as we moved into the present (warm) Holocene period. More significantly, the sediment cores also indicate rapid oscillations in the penetration of NADW, where the thermohaline circulation suddenly got much stronger or weaker - oscillations which can be correlated to abrupt warming or cooling events in the northern hemisphere. These variations apparently occurred over very short timescales - certainly not three days, as a certain movie would have you believe, but in the course of decades, still extremely fast by geological standards.

The current warming trend is making surface waters in the North Atlantic less saline (and therefore less dense) by adding large amounts of fresh water - for example, from the melting of the Greenland icecap. This can potentially reduce NADW formation by weakening downwelling. Models of the thermohaline circulation suggest that once enough meltwater is added, the circulation abruptly collapses into a much weaker current. Furthermore, this change is not reversible - the density of surface waters has to increase substantially before vigorous NADW formation restarts.

Whilst this is all very worrying, up until now there has been no real indication that the increased supply of meltwater to the North Atlantic has been having the predicted effect. Now, however, Bryden et al. report the results of the latest of a series of east-west transects (starting in 1957, with the most recent in 1998 and 2004), measuring variations of water temperature and salinity with depth at various points, across the Atlantic Ocean at 25 degrees North. These measurements can be used to calculate the 'geostrophic flow': Coriolis forces caused by the Earth's rotation mean that north- or south- flowing currents in the Atlantic oceans generate east-west pressure gradients, which can be calculated from the temperature and salinity profiles (in case that made no sense whatsoever - not unlikely - this post at realclimate.org is a good starting point). Compared to the previous transects, results from the latest one in 2004 appear to show a 50% reduction in the southward flow of water between 3000 and 5000 m depth, which is sourced from downwelling water in the Norwegian Sea.

However, before you start investing wholesale in your uncle's woolly hat company, it's worth reading the small print. Because the current strengths are being estimated by an indirect method, the potential errors are quite large; in fact, the authors estimate that the error is about the same size as the apparent decrease in NADW flow. They also argue that because the observed reduction is associated with a particular component of the flow, rather than distributed at all levels, the effect is real, which is a fair point but not a conclusive one. The other problem is that there is virtually no data on the natural variability of the thermohaline circulation, so it is also possible that the most recent survey is just detecting this rather than a sustained reduction.

These caveats make it hard to really assess the significance of these results. But watch this space - the cruise which collected the data for this study was part of a programme to deploy a series of buoys along the 25N transect, which will continuously monitor the thermohaline circulation. This system will give a much better idea of the seasonal variability of the thermohaline circulation, and also reduce the errors in estimates of its strength. In a couple of years, maybe we'll know whether I'll retire to vineyards or polar bears in my back garden.

Newer developments:
The Case of the THC "Shutdown"
THC not as weak as we thought - most of the time
What, no polar bears?


[1] Bryden et al (2005). Nature 381, p655-657.
[2] Piotrowski et al. (2004). Earth and Planetary Science Letters 225, p205-220.