How will man-made climate change affect
the ocean circulation? Is the present system of ocean currents stable,
and could it be disrupted if we continue to fill the atmosphere with greenhouse
gases? These are questions of great importance not only to the coastal
nations of the world. While the ultimate cause of anthropogenic climate
change is in the atmosphere, the oceans are nonetheless a vital factor.
They do not respond passively to atmospheric changes but are a very active
component of the climate system. There is an intense interaction between
oceans, atmosphere and ice. Changes in ocean circulation appear to have
strongly amplified past climatic swings during the ice ages, and internal
oscillations of the ocean circulation may be the ultimate cause of some
Our understanding of the stability
and variability of the ocean circulation has greatly advanced during the
past decade through progress in modelling and new data on past climatic
changes. I will not attempt to give a comprehensive review of all the new
findings here, but rather I will emphasise four key points.
Ocean currents have a profound influence
Covering some 71 per cent of the Earth
and absorbing about twice as much of the sun's radiation as the atmosphere
or the land surface, the oceans are a major component of the climate system.
With their huge heat capacity, the oceans damp temperature fluctuations,
but they play a more active and dynamic role as well. Ocean currents move
vast amounts of heat across the planet - roughly the same amount as the
atmosphere does. But in contrast to the atmosphere, the oceans are confined
by land masses, so that their heat transport is more localised and channelled
into specific regions.
The present El Niño event
in the Pacific Ocean is an impressive demonstration of how a change in
regional ocean currents - in this case, the Humboldt current - can affect
climatic conditions around the world. As I write, severe drought conditions
are occurring in a number of Western Pacific countries. Catastrophic forest
and bush fires have plagued several countries of South-East Asia for months,
causing dangerous air pollution levels. Major floods have devastated parts
of East Africa. A similar El Niño event in 1982/83 claimed nearly
2,000 lives and global losses of an estimated US$ 13 billion.
Another region that feels the influence
of ocean c?????eurrents particularly strongly is the North Atlantic. It is at
the receiving end of a circulation system linking the Antarctic with the
Arctic, known as 'thermohaline circulation' or more picturesquely as 'Great
Ocean Conveyor Belt' (Fig. 1). The Gulf Stream and its extension towards
Scotland play an important part in this system. The term thermohaline circulation
describes the driving forces: the temperature (thermo) and salinity (haline)
of sea water, which determine the water density differences which ultimately
drive the flow. The term 'conveyor belt' describes its function quite well:
an upper branch loaded with heat moves north, delivers the heat to the
atmosphere, and then returns south at about 2-3 km below the sea surface
as North Atlantic Deep Water (NADW). The heat transported to the northern
North Atlantic in this way is enormous: it measures around 1 PW, equivalent
to the output of a million power stations. If we compare places in Europe
with locations at similar latitudes on the North American continent, the
effect becomes obvious. Bodö in Norway has average temperatures of
-2°C in January and 14°C in July; Nome, on the Pacific Coast of
Alaska at the same latitude, has a much colder -15°C in January and
only 10°C in July. And satellite images show how the warm current keeps
much of the Greenland-Norwegian Sea free of ice even in winter, despite
the rest of the Arctic Ocean, even much further south, being frozen.
Figure 1. Europe's heating system.
This highly simplified cartoon of Atlantic currents shows warmer surface
currents (red) and cold north Atlantic Deep Water (NADW, blue). The thermohaline
circulation heats the North Atlantic and Northern Europe. It extends right
up to the Greenland and Norwegian Seas, pushing ba?????eck the winter sea ice
margin. Reproduced from Rahmstorf 1997.
We now have computer models that
give fairly realistic simulations of the ocean circulation, and these models
can be used to examine the effects of the currents on climate. For the
Atlantic 'conveyor belt' this task is made particularly straightforward
by a peculiarity of the climate system: there are two stable climate states,
one with the Atlantic conveyor, one without it. Just by using different
initial conditions, all else remaining the same, the models can come up
with either of these two different climates. This makes it easy to compare
what the world would look like without the ocean circulation that warms
Europe. Manabe and Stouffer 1988 were the first to analyse what happens
when the familiar conveyor circulation is absent in an ocean-atmosphere
circulation model. They found that the sea surface temperatures in the
northern North Atlantic dropped up to 7°C in this case. Air temperatures
dropped even more, up to 10°C over the Arctic seas near Scandinavia,
even though the root cause for the atmospheric cooling was the lower sea
surface temperatures. The reason for this amplification of the cooling
was the advance of sea ice, which reflects sunlight back into space and
thus led to further cooling. The air temperature changes in the model are
roughly consistent with the observed difference between Bodö and Nome,
confirming that this difference is indeed mainly caused by the warmth brought
north by the Atlantic ocean currents in the present climate.
Ocean currents were different in the
Painstaking detective work involving
sediments of the deep sea has enabled scientists to derive a wealth of
information on ocean currents of the distant ?????epast. What is just mud to
a lay-person, provides a valuable archive of past climate data to the expert.
Like tree rings or the annual layers of snow accumulated on glaciers, the
sediments at the sea bottom preserve information on the environmental conditions
from the time they were formed. It is even possible to distinguish between
conditions at the sea surface and at the bottom, as the prime source of
data are the shells of tiny organisms. Under the microscope, the abundance
of different species can be counted and identified as surface or bottom
dwellers. The chemical composition of their shells has been determined
by the temperature, salinity and nutrient content of the waters the organisms
lived in, which in turn reveals information on the ocean currents of the
Looking back over the oceanic records
of the past 100,000 years or so, it is striking how variable the currents
must have been. Only the last 8,000 years, i.e. most of the Holocene, were
a relatively stable period. Before then, throughout the last ice age, sudden
jumps and jolts occur in the record roughly every 1,000 years. These are
consistent between different sediment cores, and what is more, most of
the spikes in the oceanic conditions correspond to synchronous climate
shifts on land as recorded in the Greenland ice cap (Bond et al. 1993).
Some cold climate episodes started with a temperature drop over Greenland
of 5°C happening over a few decades or even less. The most plausible
explanation for these sudden climatic changes are rapid shifts or breakdowns
in the ocean currents of the North Atlantic. The exact timing and sequence
of events and the ultimate causes are still under investigation, but there
is widespread agreement that the 'conveyor belt' circulation of the Atlantic
played an active and dynamic role in the climatic roller coaster of the
Fairly detailed reconstructions
of the Atlantic ocean circulation (Labeyrie et al. 1992; Sarnthein et al.
1994) at the height of last Ice Age, the Last Glacial Maximum (LGM) around
21,000 years before present, show that North Atlantic Deep Water (NADW)
then formed south of Iceland (today much of it forms by convection to the
north of Iceland). It sank to intermediate depths only, and Antarctic Bottom
Water (AABW), which comes in from the south below the NADW, pushed further
north than it does today, filling most of the abyssal North Atlantic. Recently
the first coupled ocean-atmosphere simulation of glacial climate was performed
at the Potsdam Institute (Ganopolski et al. 1998), accurately reproducing
these features of the glacial ocean circulation (Fig. 2). Through a sensitivity
experiment using the present-day ocean heat transports (instead of taking
the glacial circulation changes into account), the authors were able to
demonstrate that the change in Atlantic ocean currents played a major role
in surface climate, amplifying the glacial cooling of the Northern Hemisphere
by 50%. In the North Atlantic the southward shift of deep water formation
sites and the corresponding advance of sea-ice led to an air temperature
drop of 20°C.
Figure 2. Stream functions of meridional
ocean transport in the Atlantic, for the present climate (left) and for
the last glacial maximum (right), from the coupled ocean-atmosphere model
simulation of Ganopolski et al. 1998.
Both model experiments and paleo-data
thus demonstrate that the ocean circulation has undergone important change?????es
in the past, and that these have led to major perturbations of the climate,
at least in the North Atlantic region. It is possible that there are other
regions of the globe, such as the Southern Ocean, in which the dynamic
ocean caused major climate variations, but until now they have not been
studied nearly as well as the North Atlantic.
The thermohaline circulation is a strongly
Understanding the role of the ocean
in climate change requires an understanding of the dynamics of ocean circulation
changes. Systematic computer simulations have led to important advances
in our knowledge of circulation dynamics in recent years. We have found
that there are two distinct mechanisms that can cause non-linear transitions
in the state of the Atlantic ocean circulation, a 'fast' and a 'slow' mechanism.
The slow mechanism is quite well understood, and can be described by a
simple stability diagram (discussed below).
The modelling studies have confirmed
Stommel's 1961 idea that there are two states which are stable under present
climatic conditions, namely with and without deep water formation in the
North Atlantic (see section 1). Stommel described the positive salt advection
feedback responsible for this strange behaviour: salinity in the high latitudes
needs 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 is therefore self-maintaining.The flow depends
on precariously balanced forces: cooling pulls in one direction, while
the input of freshwater from rain, snow, melting ice and rivers pulls in
the other. This freshwater threatens to reduce the salinity, and therefore
the density, of the surface waters; only by a conti?????enuous flushing away
of the freshwater and replenishing with salty water from the south does
the conveyor survive. If the flow slows down too much, there comes a point
where it can no longer keep up at all and the conveyor breaks down. This
'spin-down' takes many decades or even centuries: this is the 'slow' transition
Figure 3. Stability diagram showing
how the meridional transport in the Atlantic ('strength of the conveyor')
depends on the amount of freshwater (precipitation and river run-off minus
evaporation) entering the Atlantic. Note the bifurcation point S, beyond
which no North Atlantic Deep Water formation can be maintained. For a detailed
discussion, see Rahmstorf 1996.
A look at a simple stability diagram
shows how it works (Fig. 3). The key feature is that there is a definite
threshold for how much freshwater input the conveyor can cope with. Such
thresholds are typical for complex, non-linear systems. The diagram is
based on Stommel's theory, adapted for the Atlantic conveyor, but experiments
with global circulation models also show the same behaviour (Rahmstorf
1996). Different models locate the present climate at different positions
on the stability curve - for example, models with a rather strong conveyor
are located further left in the graph, and require a larger increase in
precipitation to push the conveyor 'over the edge'. The stability diagram
is a unifying framework that allows us to understand and compare different
computer models and experiments.
The model st?????eudies also revealed
another kind of threshold where the conveyor flow can change or break down
(Rahmstorf 1995). While the vulnerability in Stommel's theory arises from
the large-scale transport of salt by the conveyor, this second type of
threshold depends on the vertical mixing in the convection areas (e.g.
the Greenland Sea and the Labrador Sea). If the mixing is interrupted,
then the conveyor may break down completely in a matter of years, or the
locations of the convection sites may shift. This process is known as 'convective
instability', and is the 'fast' transition mechanism. We do not yet know
where the critical limits of convection are, nor what it would take to
set off such an event. Current climate models are not powerful enough to
resolve such regional processes accurately. Convective instability could
be the mechanism responsible for some of the very fast climatic changes
seen in the paleo-climate records. Both mechanisms are summarised in table
gradual (~100 y)
rapid (~10 y)
large-scale salt advection
local convection physics
basin-scale heat and freshwater
local forcing in convection region
conveyor winds down
shift of convection locations or
complete breakdown of conveyor
conveyor 'on' or 'off'
several equilibria with different
modelled quite well by climate
large uncertainty in forcing and
Table 1: Overview over properties
of the two instability mechanisms relevant to the Atlantic ocean circulation.
The ocean circulation may change in
Given the past instability of ocean
currents and our understanding of their non-linear behaviour, the future
of the Atlantic circulation in the changing climate of the next century
is a natural concern. Manabe and Stouffer 1993 published scenario simulations
with a coupled ocean- atmosphere model in which the carbon dioxide content
of the atmosphere was gradually increased to both twice and four times
the pre-industrial value and then kept constant (Fig.4). With a doubling
of CO2, the Atlantic conveyor circulation declined strongly but subsequently
recovered. If the CO2 content was increased fourfold, h?????eowever, the thermohaline
circulation in the model was interrupted completely. Other 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 (Rahmstorf
1997). A systematic sensitivity study with a simpler model revealed that
not only the total amount of carbon dioxide, but also its rate of increase
determines the effects on the ocean (Stocker and Schmittner 1997).
Figure 4. Time series of meridional
transport in the Atlantic for two greenhouse scenarios of Manabe and Stouffer
1993. Top panel: carbon dioxide forcing of the runs. For the scenario leading
to a quadrupling of carbon dioxide in the atmosphere, the thermohaline
ocean circulation winds down almost completely.
The circulation changes in all these
experiments happen on the slow, advective time scale over one or two centuries;
rapid changes as seen during the last glacial were not triggered in these
scenarios. This is the main reason why the effects on regional temperatures
are only moderate in these models; the reduced ocean heat transport then
falls in a time of strong greenhouse warming and is partly cancelled by
this. The effects of such circulation changes on marine ecosystems are
largely unexplored and will probably be serious. Furthermore, a weakened
circulation reduces the ability of the ocean to absorb carbon dioxide,
making the climate system even less forgiving of human emissions (Sarmiento
and Le Quéré 1996).
The lack of rapid circulation changes
in the model scenarios does not rule out that they could happen. Due to
poor resolution, present climate models cannot capture the fast convective
instability very well; this process depends on regional details. The latest
report of the Intergovernmental Panel on Climate Change (Houghton et al.
1995) concluded: "Future climate changes may also involve 'surprises'.
(...) Examples of such non- linear behaviour include rapid circulation
changes in the North Atlantic." This is still a valid conclusion today.
Bond, G., W. Broecker, S. Johnsen,
J. McManus, L. Labeyrie, J. Jouzel and G. Bonani, 1993: Correlations between
climate records from North Atlantic sediments and Greenland ice. Nature,
Ganopolski, A., S. Rahmstorf, V.
Petoukhov and M. Claussen, 1998: Simulation of modern and glacial climates
with a coupled global model of intermediate complexity. Nature, 391, 350-356.
Houghton, J. T., L. G. Meira Filho,
B. A. Callander, N. Harris, A. Kattenberg and K. Maskell, 1995: Climate
Change 1995. Cambridge University Press, Cambridge, 572 pp.
Labeyrie, L. D., J. C. Duplessy,
J. Duprat, A. Juillet-Leclerc, J. Moyes, E. Michel, N. Kallel and N. J.
Shackleton, 1992: Changes in the vertical structure of the North Atlantic
ocean between glacial and modern times. Quaternary Science Review, 11,
Manabe, S. and R. J. Stouffer, 1988:
Two stable equilibria of a coupled ocean-atmosphere model. J. Clim., 1,
Manabe, S. and R. J. Stouffer, 1993:
Century-scale effects of increased atmospheric CO2 on the ocean-atmosphere
system. Nature, 364, 215-218.
Rahmstorf, S., 1995: Bifurcations
of the Atlantic thermohaline circulation in response to changes in the
hydrological cycle. Nature, 378, 145-149.
Rahmstorf, S., 1996: On the freshwater
forcing and transport of the Atlantic thermohaline circulation. Clim. Dyn.,
Rahmstorf, S., 1997: Risk of sea-change
in the Atlantic. Nature, 388, 825- 826.
Sarmiento, J. L. and C. Le Quéré,
1996: Oceanic carbon dioxide uptake in a model of century-scale global
warming. Science, 274, 1346-1350.
Sarnthein, M., K. Winn, S. J. A.
Jung, J. C. Duplessy, L. Labeyrie, H. Erlenkeuser and G. Ganssen, 1994:
Changes in east Atlantic deepwater circulation over the last 30,000 years:
Eight time slice reconstructions. Paleoceanography, 9, 209-267.
Stocker, T. and A. Schmittner, 1997:
Influence of CO2 emission rates on the stability of the thermohaline circulation.
Nature, 388, 862-865.
Stommel, H., 1961: Thermohaline
convection with two stable regimes of flow. Tellus, 13, 224-230.