Grantee: Potsdam Institute for Climate Impact Research, Potsdam, Germany
Researcher: Stefan Rahmstorf, Ph.D.GLOBAL & COMPLEX SYSTEMS
Grant Title: Currents of Change
Program Area: Centennial Fellowship
Grant Type: Research Award
Year Awarded: 1999Currents of Change
My work is 'Climate Modelling'. I simulate the currents of the world ocean and their effect on climate in a computer, and I develop theoretical concepts for understanding the role which the ocean plays in the climate system. On longer time scales (years to millennia) the ocean plays a dominant role in climate variations, due to its large thermal 'memory' (heat capacity). The component of the ocean circulation which is most important for climate is the thermohaline circulation or 'conveyor belt', which is driven by density (i.e. temperature and salinity) differences.
The Atlantic Ocean Conveyor Belt. The idea of an oceanic conveyor belt was formulated by Wallace Broecker, who thought of it as a global heat and salinity transport system connecting Atlantic and Pacific. The 'conveyor belt' picture has been scientifically fruitful, even though model studies of the past few years (including my own) have shown that the original concept needs modification in some important respects. There appears to be no such simple dynamical link between the Atlantic and Pacific circulations as Broecker thought. Nevertheless, within the Atlantic there is a coherent circulation system that maybe called the 'Atlantic conveyor belt'.
Its role in climate. The Atlantic Conveyor Belt consists of warm surface water flowing north and a cold deep branch (North Atlantic Deep Water, NADW) flowing southward; the deep water is formed by convection in the Nordic Seas and the Labrador Sea. This works like a central heating system, bringing some 1015 W towards the northern North Atlantic, warming the sea surface temperatures there by 4-5'C. Atmospheric temperatures are warmed even more, as the sea ice margin is pushed north by the warm currents, decreasing the amount of sunlight reflected back into space. Coupled ocean-atmosphere models show an annual mean air temperature cooling which affects most of the Northern Hemisphere and peaks at 10-20° C near Scandinavia if NADW formation is turned off.
Its stability. The stability of the Atlantic thermohaline circulation has been the focus of many systematic model studies in recent years, including my own. These have confirmed Stommel's idea of 1961 that there are two states which are stable under present climatic conditions, namely with and without deep water formation in the North Atlantic. Stommel described the positive salt advection feedback responsible for this strange behaviour: salinity in the high latitudes need to be high enough for deep water to form, but it is only high enough because the thermohaline circulation continually brings in salty water from the south. The system therefore self-maintaining. Strommel's theory is valid for a simple one-hemisphere system, but I have extended it to the cross-hemispheric flow of the Atlantic and shown that it leads to a simple stability diagram which describes the behaviour of complex ocean circulation models, and probably the real ocean. I discovered that there is a second positive feedback affecting the thermohaline circulation, which involves vertical convection and acts more locally. Correspondingly, there are two distinct mechanisms through the thermohaline circulation can change state: a fast convective instability, acting within a decade or less, and a slow advective spindown taking centuries.
Its past. The behaviour of the Atlantic Conveyor Belt during the last Ice Age has been the focus of much research. Reconstructions from sediment cores show that during the last glacial maximum, around 21,000 years ago, North Atlantic Deep Water formed south of Iceland and sank to intermediate depths only, while Antarctic bottom water pushed further northward than today. This picture was confirmed in the world's first coupled model simulation of ocean-atmosphere circulation of the last glacial maximum, performed by our group in Potsdam. This simulation showed that the change in Atlantic circulation may have significantly enhanced glacial cooling, by as much as 50% over the Northern Hemisphere. Another striking aspect of past thermohaline circulation is ' its high variability: ice and sediment cores show that surface climate and circulation changed in step throughout the last ice age, sometimes making drastic swings within a decade or so. While cause and effect are not clearly established yet, the positive feedbacks and instabilities discussed above are a prime candidate for explaining the sudden changes.
Its future. Given the past instability, the future of the Atlantic circulation in the changing climate of the next century is a natural concern. Model scenarios for a greenhouse world generally show a reduction in thermohaline circulation between 20% and 50% for a carbon dioxide doubling in the atmosphere, and if carbon dioxide levels rise further after that, the circulation may be halted altogether. This happens gradually over one or two centuries in these scenarios; rapid changes as seen during the last glacial have not been forecast so far. However, this does not mean that they are impossible. Due to their poor resolution, present climate models cannot capture the fast convective instability very well, as this process depends on regional details. Further research, both to understand past climate and to predict future climatic changes, is urgently required. In particular, a theoretical understanding of the mechanisms of change is needed, without which it is difficult to interpret the results of complex models.