BACKGROUND OCEANOGRAPHY and CLIMATOLOGY
INTRODUCTION
Climate is generally considered to be the long-term average of weather. One might say
somewhat flippantly that climate is what you expect, and weather is what you get. Factors
typically taken into consideration when characterizing climate include average temperature, the
range of temperatures, and average precipitation. One may also consider factors such as
humidity, wind speeds, snow and ice, photoperiod, and so forth. Broadly speaking one can
divide the Earth’s climate into three zones based on latitude: polar, temperate, and tropical.
However, climatic regimes can also be characterized in many other ways based on a variety of
factors: for example, maritime (influenced by the ocean), continental (typical of the interior of
large land masses and far from the influence of the ocean), alpine (high altitude – above the tree
line), arid (dry).
Most scientists now agree that human activities are causing the climate of the Earth to
change and that the changes, now subtle, will become much more apparent during the next
several centuries. The effects of projected climate changes on the human population are likely to
be profound (Patz et al., 2005). Impacts will include, inter alia, changes in temperature and
precipitation and associated effects on agricultural productivity, sea level rise, and a spread
toward higher latitudes of the prevalence of tropical diseases such as yellow fever and malaria
(Laws, 2007). By far the most important cause of anthropogenic effects on climate has been the
release of carbon dioxide (CO2) into the atmosphere as a result (primarily) of fossil fuel burning
and deforestation. Because CO2 is a greenhouse gas (i.e., it effectively traps infrared radiation
that would otherwise escape to outer space), its presence in the atmosphere helps to warm the
Earth. If human beings burn most of the remaining fossil fuels (coal, oil, and natural gas) over
the course of the next 100-200 years, the concentration of CO2 in the atmosphere will likely rise
1
to 1,900 parts per million by volume (ppmv)1 from its current value of 380 ppmv (Caldeira and
Wickett, 2003), enough to raise global temperatures by 10oC (Berner, 1994). The ocean has the
potential to absorb virtually all of this anthropogenic CO2, but the response time of the ocean is
very slow, on the order of 10,000 years, because the air-sea boundary is a considerable limiting
factor to gaseous exchange. More efficient use of fossil fuels will not change this picture.
Because the response time of the ocean is so long, it makes little difference whether the fossil
fuels are burned over the course of the next 100 years or the next 300 years. Either way, the CO2
concentration in the atmosphere would rise to 1,900 ppmv. In this chapter we review some of
the basic information needed to understand the climate of the Earth, the variations of climate
from one region of the globe to another, and the impact of the ocean on climate and climate
change, and hence its potential for impacting human health.
THE CLIMATE OF THE EARTH OVER GEOLOGIC TIME
To put our discussion in context, it is important to realize that over geologic time the
climate of the Earth has in fact changed dramatically. Despite geological evidence for oxygenproducing photosynthesis as early as 3.5 billion years ago (e.g., widespread deposits of oxidized
iron called Banded Iron Formations) the Earth’s atmosphere appears to have remained devoid of
oxygen for roughly another 1.5 billion years. Most of the oxygen produced by photosynthetic
processes was apparently consumed by reactions with (primarily) ferrous iron and (secondarily)
sulfide in seawater (Schlesinger, 1997). Following this so-called rusting of the oceans it was
possible for oxygen to diffuse into the atmosphere, but atmospheric O2 concentrations
comparable to present values (21%) were probably not reached until the Silurian, roughly 430
million years ago. Initially much of the oxygen released to the atmosphere was apparently
consumed by reactions with reduced minerals such as pyrite (FeS2), resulting in fluvial transfer of
Fe2O3 to the ocean. This process of terrestrial weathering is evidenced by the accumulation of
the so-called Red Beds, deposits of Fe2O3 alternating with layers of other lithogenous ocean
sediments. Consistent with this scenario is the fact that the earliest occurrence of Red Beds
roughly coincides with the latest Banded Iron Formation deposits (Schlesinger, 1997).
1
One ppmv is one liter of CO2 in one million liters of air. Since air behaves very much like an ideal gas, 1,900
ppmv is equivalent to 1,900 molecules of CO2 for every million molecules of N2 plus O2, the principal components
of the Earth’s atmosphere.
2
There is good reason to believe that atmospheric O2 levels have not fluctuated outside the
15–35% range since the Silurian (Berner et al., 1989). At O2 concentrations less than 15% fires
would not burn (Lovelock, 1979), and at concentrations greater than about 25% even wet organic
matter would burn freely (Watson et al., 1978). The principal mechanism responsible for the
stability of atmospheric O2 concentrations appears to be the negative feedback between O2
concentrations and the long-term burial of organic matter in sedimentary rocks (Schlesinger,
1997).
Particularly noteworthy from the standpoint of current global climate change issues is the
fact that atmospheric CO2 concentrations during Phanerozoic time (approximately the last
570 million years)
Figure1. Ratio of atmospheric CO2 in times past to the present concentration (RCO2) as
determined from the Geocarb II model (Berner, 1994)
3
have generally been higher than current values, perhaps by as much as a factor of 20–25 during
the Cambrian (Berner et al., 2001). The impact of these elevated CO2 concentrations on the
climate of the Earth has been profound (Fig. 1). Since the formation of the solar system the
luminosity of the Sun has increased by about 43 percent, a result of the Sun’s slow expansion
associated with the conversion of hydrogen to helium in its core (Sagan et al., 1997). In the
absence of greenhouse gases to trap infrared radiation, the Earth would have been fully glaciated
until roughly 1 billion years ago, but geological evidence indicates that there has been abundant
liquid water on the Earth’s surface for more than 3 billion years (Sagan et al., 1997). Ammonia
may have accounted for much of the greenhouse effect in the reducing atmosphere of the early
Earth (Sagan et al., 1972; Sagan, 1977), but once atmospheric O2 levels rose to 21%, ammonia
concentrations were probably far too low to provide much of a greenhouse effect. At the present
time water vapor accounts for about 95% of the total greenhouse effect, CO2 for 3.6%, N2O for
about 1%, and CH4 for 0.4%. In the absence of an atmosphere, the Earth’s surface temperature
would average about 255oK or –18oC. The fact that the Earth’s surface temperature averages
about 288oK or 15oC is largely attributable to the fact that greenhouse gases are rather opaque to
infrared radiation. At the beginning of the Phanerozoic eon the solar constant was about 5% less
than it is today. Had atmospheric CO2 concentrations been the same then as now, the Earth’s
surface temperature would have averaged about 2oC (Berner, 1994).
In addition to climatic effects associated with variations in atmospheric CO2
concentrations, the Earth has experienced very dramatic climatic changes manifested by the
advance and retreat of continental ice sheets and polar ice caps. Continental drift is certainly one
factor that has influenced the ice age cycle; the movement of Antarctica to the South Pole is a
case in point. The most recent ice age began roughly 40 million years ago with the accumulation
of ice on Antarctica, but intensified during the Pleistocene with the development of continental
ice sheets in the Northern Hemisphere. During the Pleistocene ice age there was a cyclical
advance and retreat of the Northern Hemisphere ice sheets that is most commonly attributed to
variations in the eccentricity, axial tilt, and precession of the Earth’s orbit around the Sun. This
explanation of glacial/interglacial periodicity was initially advanced by the Serbian geophysicist
Milutin Milanković, but did not gain widespread acceptance until studies of deep-sea sediments
4
during the 1960s and 1970s produced evidence consistent with so-called Milankovitch cycles
(Hays et al., 1976). These cycles are clearly apparent in the record of atmospheric CO2 in the
Vostok ice core (Fig. 2). Evident in this figure is a systematic pattern of atmospheric CO2
Figure 2. Atmospheric CO2 concentrations during the past 420,000 years based on the
composition of air entrapped in the Vostok ice core (Barnola et al., 1999)
variation from roughly 180 to 280 ppmv. Low CO2 concentrations are associated with glacial
periods, the most recent of which have been the Wisconsinan (15–70 thousand years ago) and
Illinoian (125–200 thousand years ago). High CO2 concentrations are associated with
interglacial periods, the most recent of which have been the Eemian (115–130 thousand years
ago) and Holocene (11,500 years ago to present). The record clearly implicates CO2 as an
amplifier of the effect of orbital forcing on the glacial/interglacial cycle.
As noted, climate change at the present time is largely associated with the accumulation
of CO2 in the atmosphere due to fossil fuel burning and deforestation. Fossil fuel burning, which
currently releases about seven billion tons of carbon to the atmosphere each year, is generally
5
blamed for roughly 70% of anthropogenic CO2 emissions. Much of the rest is attributed to
deforestation, because of the decrease in the uptake of CO2 by plants (Raven et al., 1999).
Although roughly half of the anthropogenic CO2 released to the atmosphere is absorbed by the
oceans and continental vegetation, the rest accumulates in the atmosphere. The result is clearly
apparent in Fig. 3, which documents the rise in atmospheric CO2 concentrations by roughly 100
ppmv during the last two centuries.
Figure 3. Atmospheric CO2 concentrations since 1000 AD estimated from ice core data
and monitoring of CO2 at Mauna Loa (Etheridge et al., 2006; Keeling et al., 2006).
CONTROLS ON THE CLIMATE OF THE EARTH
Understanding the general characteristics of the Earth’s climate requires a modest amount
of information and an understanding of a few important concepts. The first important piece of
information is the fact that the radiant energy from the Sun is not equally distributed over the
6
surface of the Earth. Equatorial latitudes receive much more energy than polar latitudes, and as a
result the atmosphere near the surface of the Earth is much warmer near the equator than near the
poles. Heating air causes it to expand, become less dense, and rise (a phenomenon routinely used
by hot air balloon enthusiasts). Cooling air causes it to sink. Because equatorial latitudes receive
more solar energy than the poles, the differential heating of the Earth-atmosphere system causes
air to rise near the equator and to descend near the poles. One might imagine that the atmosphere
would therefore move directly north and south, rising at the poles and sinking at the equator, as
shown in Fig. 4.
Figure 4. Cross section of the Earth showing the pattern of circulation of the lower
atmosphere that might be expected from differential heating of the Earth-atmosphere
system by the Sun.
In fact, atmospheric circulation is not so simple. Although air tends to rise near the
equator, as it moves poleward it radiates heat into outer space and eventually cools and sinks at
about 30o latitude. Similarly, cold air that sinks at the poles tends to be warmed as it flows along
the surface of the Earth toward the equator and to rise near 60o latitude. The vertical circulation
of the atmosphere, in simplified terms, consists of three circulation cells as shown in Fig. 5. The
subtropical and temperate-latitude circulation cells are referred to as Hadley cells and Ferrel cells,
7
respectively, after the scientists who discovered them. The high-latitude cells are called polar
cells.
Figure 5. Meridional circulation that results from differential heating of the Earthatmosphere system by the Sun. Note that the vertical scale of circulation cells is greatly
exaggerated. The vertical extent of the cells is approximately 10 km.
THE EFFECT OF THE EARTH’S ROTATION
In most respects Fig. 5 is an accurate characterization of the overall meridional (northsouth) circulation of the atmosphere, but it is an oversimplification. The real circulation pattern
is neither as uniform nor as continuous as Fig. 5 implies. The figure suggests, for example, that
surface winds would blow directly toward the equator in tropical and subtropical latitudes and
directly toward the poles in temperate latitudes. This is only partly true.
If we were to slice up the Earth along its latitude lines, we would get a series of rings, the
largest at the equator and diminishing in size toward the poles. Because the Earth is rotating as a
solid body, a point on a large ring moves faster than a point on a small ring. At 30o latitude, for
example, the circumference of our latitudinal ring would be about 34,600 kilometers. A point on
the Earth’s surface at that latitude is moving toward the east at a rate of 34,600 kilometers per
8
day, or 1,442 kilometers per hour. At 29o latitude, the surface of the Earth is moving faster, at
1,458 kilometers per hour, because the circumference of a cross section there is 35,000
kilometers.
If there are no other zonal (east-west) forces acting on it, a mass of air flowing toward the
equator across the surface of the Earth will appear to be deflected toward the west, because the
underlying Earth is moving faster toward the east the closer to the equator the air travels (see Fig.
6). The surface winds that blow from about 30o toward the equator are referred to as the Trade
Winds. Because winds are customarily named on the basis of the direction from which (rather
than to which) they are flowing, these winds are known as the Northeast Trades in the Northern
Hemisphere and the Southeast Trades in the Southern Hemisphere.
Figure 6. The effect of the rotation of the Earth on a parcel of air initially at a latitude of
30o and moving at a speed of 8 m s-1 directly toward the equator (Trade Winds) or directly
away from the equator (Westerlies). No east-west forces are assumed to act on the parcel
of air. By the time the air has moved 1o, its direction has changed by about 45o. In the
Trade Wind zone the parcel of air acquires a westerly component, while in the region of
the Westerlies it acquires an easterly component. The effect of the Earth’s rotation is
always to divert the air to the right of its direction of motion in the northern hemisphere
and to the left in the southern hemisphere.
Now consider the air that sinks at 30o and flows toward the poles. Since at higher
latitudes the surface of the Earth is moving to the east more slowly than at 30o, this air will
9
acquire an apparent eastward motion. The surface winds between 30o and 60o are more complex
and unstable than the Trade Winds, but they consistently have a west-to-east component, and
hence are known as the Westerlies. Because surface winds between the poles and 60o are moving
toward the equator, they are affected by the Earth’s rotation in the same way as the Trade Winds,
blowing out of the northeast in the Northern Hemisphere and the southeast in the Southern
Hemisphere (see Fig. 7).
Figure 7. Direction of surface winds resulting from the combined effects of the Coriolis
force and meridional cell circulation.
Once again though, the situation is more complicated. The continental landmasses
influence the flow of the wind, and because the land is unevenly distributed between the northern
and southern hemispheres, the winds do not blow in an entirely symmetrical manner with respect
to the equator. In fact, the entire wind system shown in Fig. 1-7 is shifted about 5-10o to the
north. In addition, in temperate latitudes surface winds tend to circulate about high-pressure
ridges and low-pressure troughs, and shifts in the positions of these ridges and troughs can
produce important climatological effects.
Finally, the difference in the heat capacity of the continents and oceans causes seasonal
temperature differentials to develop between them. Because it takes a great deal of heat to warm
a mass of water, and because the upper mixed layer of the ocean is large (typically it extends to
tens of meters in the summer and perhaps hundreds of meters in the winter), the temperature of
the ocean remains relatively constant compared to the temperature of the continents. During the
10
summer the continents are warmer than the ocean, and during the winter they are cooler. The
exchange of heat between the Earth and atmosphere therefore causes the air over the continents to
be warmer and less dense than the air over the surrounding oceans during the summer. During
the winter the conditions are reversed. As the continental air warms and rises during the summer,
air overlying the surrounding ocean is drawn in to replace it. In the winter, the cool, dense air
over the continents tends to sink and flow towards the surrounding ocean. The winds associated
with this seasonal circulation pattern are referred to as monsoon winds and are best developed
over India, Southeast Asia, and Australia.
THE EFFECT OF SURFACE WINDS AND THE CORIOLIS FORCE ON OCEAN CURRENTS
Because the Earth is a rotating sphere, it appears to an observer on Earth that a force is
always pushing the wind to the right of the direction of motion in the northern hemisphere and to
the left in the southern hemisphere (e.g., Fig. 6). This force is called the Coriolis force, and it
affects the oceans as well as the atmosphere. The Coriolis force is directly proportional to the
speed of motion and to the sine of the latitude. The force is zero at the equator and a maximum at
the poles.
One would expect that ocean currents would flow in the same direction as the surface
winds, but they rarely do. Just as landmasses affect the flow of winds, they impose some
constraints on the direction in which ocean currents can flow. Virtually all coastal current
systems flow parallel to the coast, regardless of the direction in which the wind is blowing. But
even in the open ocean, surface currents do not tend to move in the same direction as the wind.
Again, this is due to the Coriolis force, which causes those currents to flow at an angle to the
right of the wind in the northern hemisphere and to the left of the wind in the southern
hemisphere. The transport of currents at an angle to the wind is referred to as Ekman transport,
after the Scandinavian oceanographer who explained the phenomenon theoretically.
The combination of the Coriolis force and Ekman transport causes ocean surface currents
in the region of the Trade Winds to flow almost exactly due west across the ocean basins, while
in the vicinity of the Westerlies the flow is due east. When these transoceanic surface currents
encounter continental landmasses, they may either turn and flow parallel to the coastline or
completely reverse direction and flow back across the ocean basin. In the former case, they are
called boundary currents; in the latter case countercurrents. The major current systems driven by
11
the Trade Winds and Westerlies in the Pacific Ocean are shown in Fig. 8. The transoceanic
currents to the north of the equator are the North Pacific Current and the North Equatorial
Current, and the corresponding boundary currents are the California and Kuroshio currents. The
Figure 8. The Pacific Ocean subtropical gyre current systems. Note that the current
gyres are not symmetric with respect to the equator. The Equatorial Countercurrent
actually flows between about 4o and 10o N latitude.
analogous current systems in the South Pacific are the West Wind Drift, the South Equatorial
Current, the Peru Current, and the East Australia Current, respectively. The South Equatorial
Current actually extends to about 4oN, and much of the flow in the West Wind Drift is actually
circumpolar, since there are no continental landmasses to impede it between roughly 55o and
65oS. The Equatorial Countercurrent flows from west to east across the Pacific between
approximately 4o and 10oN. Another eastward-flowing countercurrent, called the Equatorial
Undercurrent, is at the equator at depths of approximately 100-200 meters. Obviously neither the
Equatorial Countercurrent nor the Equatorial Undercurrent is driven directly by the wind. The
Equatorial Countercurrent, in particular, would seem to be flowing into the teeth of the prevailing
Trade Winds, but it flows through a region of light and variable winds called the Doldrums,
which offers little resistance. The more-or-less continuous current system consisting of the
12
California, North Equatorial, Kuroshio, and North Pacific currents is called the North Pacific
subtropical gyre, and its counterpart in the South Pacific is the South Pacific subtropical gyre.
Table I compares the major boundary currents in the Atlantic and Pacific oceans. The
Table I. Comparison of major boundary current systems in the Atlantic and Pacific Oceans
North Atlantic
North Pacific
South Atlantic
South Pacific
Subtropical gyre current systems
Western
Gulf Stream and
boundary current
North Atlantic
Kuroshio
Brazil
East Australia
California
Benguela
Peru
Current
Eastern boundary Canary
current
Subolar gyre current systems
Western
Labrador
Oyashio
boundary current
Eastern boundary North Atlantic
current
Alaska
Drift
poleward flowing boundary currents (Gulf Stream, Kuroshio, Brazil, East Australia, North
Atlantic Drift, and Alaska) are particularly important from the standpoint of climate because they
transport large amounts of heat from low latitudes to high latitudes. The impact of the heat
transported by the combined Gulf Stream/North Atlantic Drift current system, for example,
warms northwestern Europe by an annual average of as much as 5-10oC (Manabe et al., 1988;
Rahmstorf et al., 1999). There is no subpolar gyre current system in the Southern Hemisphere,
since there are no continental landmasses to block the West Wind Drift, a circumpolar current
system that forms the southern boundary of the subtropical gyres in both the Atlantic, Pacific,
and Indian ocean basins.
An important point about the subtropical and subpolar gyres is the fact that Coriolis forces
tend to push water toward their interior and exterior, respectively. This fact is apparent from an
examination of Fig. 1-8, taking into account the fact that the Coriolis force pushes to the right of
the direction of motion in the northern hemisphere and to the left in the southern hemisphere.
13
The result is that the sea surface is actually somewhat higher to the right of a current system
flowing in the northern hemisphere and to the left of a current system flowing in the southern
hemisphere. In a steady state situation, the force of gravity acting on the tilted sea surface exactly
balances the Coriolis force. When this happens, the current is said to be in geostrophic balance,
and the current is characterized as a geostrophic current. The difference in sea surface height
(SSH) across the Gulf Stream, for example, is about one meter, with SSH being higher to the
interior of the North Atlantic subtropical gyre (Kelly et al., 1999).
Similar considerations influence the circulation of the atmosphere, but with the caveat that
the analogues of high and low SSH are high and low atmospheric pressure, respectively. Thus in
the northern hemisphere winds tend to blow in a clockwise direction around a region of
high pressure and in a counterclockwise direction around a region of low pressure. In each case
the pressure gradient force is in the opposite direction of the Coriolis force. In the southern
hemisphere the circulation is in the opposite sense because the Coriolis force pushes to the left of
the direction of motion. Thus a satellite image of a cyclone or hurricane (extreme low pressure
system) in the northern hemisphere always reveals a pattern of counterclockwise circulation (Fig.
1-9). In the southern hemisphere cyclonic winds blow in a clockwise sense. Appropriately
enough, the circulation of winds or currents around any region of low pressure or low SSH is
characterized as cyclonic circulation (i.e., counterclockwise in the northern hemisphere and
clockwise in the southern hemisphere). The circulation of winds or currents around any region of
high pressure or high SSH is characterized as anti-cyclonic circulation.
With this introduction, it is straightforward to understand some of the major patterns of
the climate of the Earth. As the Trade Winds blow across the tropical ocean they pick up both
heat and water vapor. Because warm, moist air is less dense than cold, dry air2, this air tends to
rise where the Northeast and Southeast Trade Winds converge. This region is known as the
intertropical convergence zone or ITCZ. As the air rises the water vapor condenses and falls as
rain. The ITCZ is therefore characterized by excess precipitation over evaporation. Once the air
has risen to an altitude of roughly three kilometers it is transported to higher latitudes by Hadley
2
Air behaves very much like an ideal gas, for which PV = nRT. The number of moles of air (n) per unit volume
(V), therefore equals P/(RT). At constant pressure (P), n/V is inversely proportional to the absolute temperature (T).
Water (H2O) has a molecular weight of 18. N2 and O2, the principal gases in air, have molecular weights of 28 and
32, respectively. When water displaces nitrogen and oxygen, the average molecular weight of the gases in the air
decreases. Therefore warm, moist air is less dense than cold, dry air because there are fewer molecules per unit
volume in warm air and because the average molecular weight of the molecules is lower in moist air.
14
Figure 9. Hurricane Katrina in the Gulf of Mexico.
cell circulation (Fig. 5). Having lost most of its water vapor to condensation, the air is now dry,
and as it moves poleward it radiates heat into outer space. As the air approaches a latitude of
roughly 30o it becomes sufficiently dense (i.e., cold and dry) that it begins to sink. The climate
15
near 30o is therefore characterized by very low humidity and an excess of evaporation over
precipitation. Most of the major desert areas of the world (the Sahara Desert in northern Africa,
the Namib and Kalahari deserts in southern Africa, the Great Victoria Desert in Australia, the
Arabian Desert, and the Great Desert of the southwestern United States and northern Mexico) are
all found near 30o latitude.3
In the polar gyre systems air moving over the ocean toward the equator picks up heat and
water vapor as do the Trade Winds in the tropics. The combination of increased temperature and
humidity causes the air to rise at roughly 60o latitude. Like the ITCZ, the region near 60o latitude
is also characterized by an excess of precipitation over evaporation. When the air rises to an
altitude of roughly 3 km it moves either toward the poles (polar cell circulation) or toward the
equator (Ferrel cell circulation). Having lost most of its water vapor, it now loses heat to outer
space via radiation and eventually sinks near the poles or near 30o latitude. We can now
understand why the climate of the Earth is wet near the equator and 60o and dry near 30o and the
poles. It is no accident, for example, that rain forests are found in the tropics. Superimposed on
this pattern precipitation and evaporation is a meridional4 temperature gradient, warm at the
equator and cold at the poles.
This analysis can also account for some of the general features of atmospheric pressure at
the surface of the Earth. Keeping in mind that cold, dry air is more dense than warm, moist air,
we can easily see that sea level pressure will be relatively low near the equator and 60o latitude
and relatively high near 30o and the poles. The lowest sea level pressure tends to be found near
the equator (warm, moist air) and the highest near the poles (cold, dry air).
In the tropics an important east-west asymmetry in both precipitation and sea level
pressure is also apparent across the major ocean basins. The explanation is apparent from an
examination of Fig. 10. The Trade Winds blow both toward the equator and toward the west.
Atmospheric pressure is relatively high and the climate cool and dry along the coast of northern
Peru.
3
One major desert that does not fit this pattern is the Gobi Desert at approximately 40-45oN latitude. It cannot be
attributed to the sinking of cool, dry air in the subtropics. However, it does lie in the region of the Westerlies, one
manifestation of Ferrel cell circulation, and its location places it in the rain shadow of some very high mountain
ranges.
4
Along a meridian or line of constant longitude.
16
Figure 10. The Walker cell circulation cycle over the Pacific Ocean. The vertical scale is
exaggerated, the height of the circulation cell being about 15 km. This atmospheric
circulation pattern tends to produce low atmospheric pressure and a warm, moist climate
over Indonesia.
For reasons already noted, they become warm and moisture-laden as they move from east to west
over the tropical ocean. The result is an east-west asymmetry in sea level pressure and
precipitation near the equator, with the lowest pressure and greatest precipitation at the western
edge of the ocean basin. At the western edge of the ocean basin, part of the rising air mass moves
back toward the east. As it moves, it radiates heat into the surrounding atmosphere and
eventually cools and sinks near the eastern edge of the ocean basin. This circulation pattern is
called a Walker cell, after British mathematician Sir Gilbert Walker, who made major
contributions to our understanding of tropical meteorology in the first half of the 20th century.
Because the air that sinks near the equator near the eastern edge of the ocean basin has
lost heat as well as water vapor, it tends to be denser than the air that rises along the equator in
the west. Consequently there is a small east-west difference in sea level pressure between the
eastern and western sides of ocean basins in the Trade Wind zone.
The pressure differentials associated with Walker Cell and Hadley Cell circulation are
both manifestations of the impact of the Trade Winds on climate. Within the Trade Wind zone,
the pressure will be highest near the eastern side of ocean basins at 30o latitude and lowest at the
17
equator near the western side of ocean basins. In the Pacific Ocean this pressure differential is
known as the Southern Oscillation Index (SOI). One common measure of the SOI is the sea level
pressure difference between Easter Island (27oS) and Darwin, Australia (12oS).
THE OCEAN AND CLIMATE CHANGE
Now that we have a basic understanding of how the oceans influence climate, let’s
consider the issue of climate change. We will consider two kinds of climate change, one with a
relatively short-term periodicity, the El Niño Southern Oscillation (ENSO) cycle, and the other
with a much longer time constant, the thermohaline circulation of the ocean. We will begin with
the ENSO cycle.
El Niño was originally the name given to a dramatic shift in weather and sea conditions
off the coast of Peru. Because of the tendency of the change to begin near Christmas, it was
given the name El Niño, literally “the child” in Spanish. The changes observed included a
warming of the ocean and, in extreme cases, torrential rains in a region normally characterized by
very dry conditions.5 At one time El Niño was regarded as an abnormal event. However,
scientists currently view El Niño as simply one phase of a natural cycle, the El Niño Southern
Oscillation or ENSO cycle, that occurs every several years and is no more usual or unusual than
the conditions during any other phase of the cycle. Furthermore, they now recognize that the
changes in climate observed during El Niño years along the coast of Peru are simply a local
manifestation of a much larger phenomenon that is driven by interactions between the ocean and
atmosphere in the subtropics.
The history of El Niños has been reconstructed from as early as 1525 using proxy
information, and the record indicates that they occur about every four years, with strong events
separated by an average of ten years. Unfortunately for purposes of prediction, the interval
between El Niños is very irregular. It is not uncommonly six or seven years, but some events
have been separated by as little as one year. The most recent El Niños occurred in 1957-58
(strong), 1965 (moderate), 1969 (weak), 1972-73 (strong), 1976 (moderate), 1982-83 (very
5
The normally dry weather reflects the fact Peru lies in the rain shadow of the Andes Mountains and that the sea
surface temperature is cool for the latitude (e.g., 12 oS for Lima), a reflection of the cold water transported by the
Peru current (Fig. 8) and the fact that the Southeast Trade Winds and Ekman transport induce upwelling of cold
water along the coast.
18
strong), 1986-87 (strong), 1991-92 (very strong), 1993 (weak), 1994 (weak), 1997-98 (very
strong), and 2002-03 (weak).
El Niño conditions are triggered by a movement of warm water from the western Pacific
to the eastern Pacific via the Equatorial Countercurrent and Undercurrent. The water is
transported largely in the form of so-called Kelvin waves. Kelvin waves and similar waves
known as Rossby waves are internal waves (they have their maximum amplitude below the
surface of the ocean) whose dynamics are affected by the Coriolis force. Their wavelengths are
on the order of thousands of kilometers, and their effects can be felt across an entire ocean basin.
Kelvin waves cross the Pacific in two to three months. As their warm water reaches the coast of
South America, it flows over the cooler water of the Peru Current system. The result is an
elevation of sea level (Fig. 11) and an increase in sea-surface temperature. Some of the warm
water flows north along the coast. Some flows south and causes El Niño conditions off the coasts
Figure 11. The response of sea level in the equatorial Pacific Ocean to the 1972 El Niño.
Note that sea level was high in the western Pacific (Solomon Islands) preceding El Niño,
but had dropped dramatically by the end of 1972 as water flowed toward the east along the
Equatorial Countercurrent and Undercurrent. Sea level was relatively low in the eastern
Pacific (Galapagos Islands) preceding El Niño but rose by almost 30 cm as water arrived
from the western Pacific. Redrawn from Wyrtki (1979).
of Ecuador and Peru. As sea level rises and warm water accumulates in the eastern equatorial
Pacific, air-sea interactions generate Rossby waves that move westward across the Pacific. The
19
time they take to cross the ocean is strongly dependent on latitude; it is about nine months near
the equator and four years at a latitude of 12o. When the Rossby waves reach the western Pacific,
they travel toward the equator in the form of coastal Kelvin waves. Upon reaching the equator
they turn east and begin another crossing of the Pacific. When this second set of Kelvin waves
reaches the eastern Pacific, sea level is lowered, the sea-surface temperature declines, and
conditions along the coast of Peru return to “normal”. Since roughly 1985 these “normal”
conditions have come to be known as La Niña (literally “the girl” in Spanish). However, the airsea interactions associated with the lowered sea-surface temperatures intensify the Trade Winds,
and this shift in the winds sends Rossby waves westward across the Pacific. Upon reaching the
western Pacific, these waves travel toward the equator as coastal Kelvin waves and then return to
the east along the equator. This final set of equatorial Kelvin waves raises the sea level in the
eastern Pacific and completes the El Niño cycle. The entire process is illustrated in Fig. 12.
Figure 12. The wave system that constitutes the negative feedback mechanism in the El Niño
cycle. Equatorial Kelvin waves (EK) travel west to east across the Pacific Ocean raising sea
levels. When they reach the coastline of South America they propagate poleward and are clearly
identifiable as coastal Kelvin waves (CK) at latitudes higher than 5o. Air-sea interactions
associated with the arrival of warm water in the eastern equatorial Pacific cause the Trade Winds
to slacken. This shift in the winds sends a series of off-equatorial Rossby waves (R) that lower
sea levels back across the Pacific. These Rossby waves reach the western Pacific and propagate
toward the equator in the form of coastal Kelvin waves (CK) that also lower sea levels. The
Kelvin waves reach the equator, turn east, and move back across the Pacific as sea-level-lowering
equatorial Kelvin waves. The equatorial Kelvin waves require about 2-3 months to cross the
Pacific, but the off-equatorial Rossby waves require anywhere from a few months to a few years.
A complete El Niño cycle requires that the Pacific be crossed by two sets of Rossby waves and
Kelvin waves, one set raising sea levels in the direction they are moving and the other lowering
them. Hence a complete El Niño cycle typically requires 3-5 years.
20
AIR-SEA INTERACTIONS
Because of the exchange of heat between the atmosphere and ocean, changes in seasurface temperature in the eastern Pacific can have a significant effect on the intensity of the
Trade Wind system. When the eastern Pacific warms during an El Niño year, the Walker cell
circulation is slowed because the temperature difference between the eastern and western Pacific
is reduced. Thus, the speed of the equatorial Trade Winds, and consequently the speed of both
the South Equatorial and North Equatorial currents, decreases. A decline in the strength of the
equatorial Trades allows more warm water to flow from the western to the eastern Pacific,
further reducing the temperature differential between the eastern and western Pacific. On the
other hand, when the eastern Pacific is cool, the Walker cell circulation is increased, because
there is a greater temperature differential between the eastern and western Pacific. The Trade
Winds become stronger, and the North and South Equatorial Currents intensify. The
strengthening of the Trade Winds opposes the transport of warm water via the Equatorial
Countercurrent and Undercurrent, further increasing the temperature difference between the
eastern and western Pacific. Those air-sea interactions are an example of what is known as a
positive feedback loop. They tend to reinforce El Niño or La Niña conditions, whichever
condition prevails.
The reason there is an oscillation between El Niño and La Niña conditions is the negative
feedback loop created by the movement of the Kelvin and Rossby waves across the Pacific.
During El Niño conditions, the eastern equatorial Pacific warms and the Trade Winds slacken.
The change in Trade Wind intensity generates off-equatorial Rossby waves that lower sea levels
in the western Pacific. Ultimately these lower sea levels generate Kelvin waves that travel back
east and lower sea levels in the eastern Pacific.
One implication of this analysis of air-sea interactions is that the Southern Oscillation
Index may provide a useful predictor of forthcoming El Niños. The index is high (the pressure
differential is large) when the Trade Winds are strong (La Niña conditions). The index is low
(the pressure differential is small) when the Trade Winds are weak (El Niño conditions). Figure
13 shows the behavior of the Southern Oscillation Index and sea-surface temperatures off the
coast of Peru for the period from 1968 to 1985. The El Niños of 1972-73, 1976, and 1982-83 are
all apparent as increases in sea-surface temperature of at least 2oC above long-term monthly
averages over a period of several months, and each El Niño is associated with a drop in the
21
Southern Oscillation Index of at least 8 millibars (mb). A drop of greater than 4 mb is usually a
sign that an El Niño is approaching.
Figure 13. Three-month running mean variations in the Southern Oscillation Index (top) and
sea-surface temperature (SST) off the coast of Chimbote, Peru (bottom) from 1968 to 1985.
Monthly variations are the difference between the value for a given month and the long-term
average value for that month. During this period. El Niños occurred in 1972-73 (strong), 1976
(moderate), and 1982-83 (very strong). The El Niños of 1972-73, 1976, and 1982-83 are all
apparent as increases in temperature of at least 2oC over a period of several months, and each El
Niño is associated with a drop in the Southern Oscillation Index of at least 8 millibars (mb).
Recognition of the connection between the Southern Oscillation Index and El Niño has
given rise to the acronym ENSO – El Niño Southern Oscillation. The ENSO cycle is understood
to consist of an irregular meteorological oscillation characterized by two extreme conditions, a
warm phase (El Niño) and a cool phase (La Niña), driven by exchanges of heat and water
between the ocean and atmosphere in the tropical Pacific.
SHUTDOWN OF THE NORTH ATLANTIC CONVEYER BELT
Not all of the water transported to the North Atlantic by the North Atlantic Current and North
Atlantic Drift is returned via the Labrador Current (Table I). Instead, evaporation of water vapor
from these warm currents causes the salinity of their surface waters to increase and the
temperature to decrease. Sea ice formation is not a factor, but during the winter the combined
22
effect of increased salinity and decreased temperature causes some of the water transported by
these currents to sink to depths of 2-4 km in the Greenland Sea and Labrador Sea off Greenland.
In the Southern hemisphere bottom waters are formed along Antarctic ice shelves during
the time of sea ice formation in the winter. The fact that surface waters sink to depths of several
kilometers results from the surface waters’ being very cold and saline, but the mechanism
responsible for creating these conditions differs somewhat in the North Atlantic and Southern
Ocean. In the Southern Ocean surface waters sink to the bottom due to an increase in salinity
associated with the formation of sea ice.6
Because the formation and movement of water masses at intermediate and bottom depths
in the ocean are driven by temperature and salinity effects, the deep water current system is
referred to as the ocean’s thermohaline circulation. Once formed, bottom waters remain
submerged for roughly 1,000 years, but they eventually return to the surface. From there, surface
currents transport them back to the regions of deep and bottom water formation in the North
Atlantic and Southern Ocean, respectively. The grand pattern of surface and bottom water
circulation in the ocean is referred to as the ocean’s conveyer belt.
The analogs of the Gulf Stream and the North Atlantic Drift in the North Pacific Ocean
are the Kuroshio Current and North Pacific Current, respectively, but there is no analogous
formation of bottom water. Why does bottom water form in the North Atlantic but not in the
North Pacific? The answer is that of the major ocean basins the North Atlantic has the highest
salinity and the North Pacific the lowest. The low salinity of the North Pacific relative to the
North Atlantic is primarily the result of differences in rainfall. Precipitation on the Pacific and
Atlantic Ocean averages about 120 and 80 cm per year, respectively (Gross, 1982). The result is
that surface waters at high latitudes in the North Pacific are less saline than underlying waters,
and cooling of surface waters during the winter is insufficient to make them denser than the more
saline waters beneath them. In the North Atlantic on the other hand the salinity gradient is very
small, and cooling during the winter is sufficient to cause surface waters to sink to depths of
several kilometers.
This comparison underscores the importance of freshwater inputs in determining whether
bottom water is formed. In the Southern Ocean bottom waters are formed because freshwater is
6
Sea ice contains very little salt compared to the water from which it was formed. The liquid brines that remain
after sea ice forms are literally at the freezing point of seawater and are hypersaline due to the exclusion of salt from
the ice.
23
effectively removed by the formation of sea ice during the winter. In the North Atlantic, deep
waters are formed in the winter because freshwater and heat are removed by evaporation. In the
North Pacific freshwater and heat are also removed by evaporation, but the effect of evaporation
on the density of the surface waters is more than offset by the input of freshwater from rainfall.
Since global warming will warm the ocean’s surface waters and accelerate the hydrologic cycle,
it is reasonable to ask what impact global warming may have on the thermohaline circulation.
Figure 14 illustrates the nature of the problem. Freshwater forcing is here defined to be
the net effect of surface exchange, wind-driven ocean currents, and thermohaline circulation.
Figure 14. Relationship between freshwater forcing in the North Atlantic and the rate of
formation of North Atlantic Deep Water. One Sverdrup (Sv) = 106 m3s-1 (Rahmstorf, 2000).
When freshwater forcing is in the range zero to roughly 0.13 Sverdrup (Sv)7, two very different
but stable modes of the Atlantic thermohaline circulation are possible, one in which there is no
deep water formation and the other in which North Atlantic Deep Water (NADW) is formed at
rates ranging between roughly 11 and 22 Sv. Although the Atlantic is a net evaporative basin
(i.e., net surface exchange of freshwater is negative) the overall freshwater forcing is believed to
be positive at the present time but almost certainly less than 0.05 Sv (Rahmstorf, 2000). Hence
either of two modes of NADW formation is compatible with the present rate of freshwater
forcing, and an increase on the order of 0.1 Sv in freshwater forcing could cause the system to
7
One Sverdrup = 106 m3s-1 or 3.2x104 km3y-1.
24
undergo the transition indicated by the (a) arrow. Once the system settles into that mode, it will
remain there until freshwater forcing drops below zero, at which point the system transitions back
to the current mode as indicated by the (b) arrow.
The ocean contains about 1.3x109 km3 of water. Under current conditions deep and
bottom water is formed in the Southern Ocean and North Atlantic at a combined rate equal to
about 0.1% of this volume per year or about 43 Sv (Broecker, 1997). About 47% of this deep
water formation occurs in the North Atlantic, i.e., the NADW flow is about 20 Sv (Broecker,
1997). Based on Fig. 14 this would imply that freshwater forcing is roughly 0.02 Sv, and an
increase of about 0.1 Sv in freshwater forcing would indeed be necessary to shut down the North
Atlantic component of the conveyer belt. Is there any evidence that this has happened in the
past?
The short answer to this question is yes. During the most recent glacial period
(Wisconsinan) there was a series of brief warm periods known as Dansgaard-Oeschger events
and extreme cold periods known as Heinrich events. The best known of the Heinrich events is
the Younger Dryas cold event, which lasted from roughly 12,700 to 11,500 years ago and
immediately preceded the transition to the present Holocene interglacial. Many
paleoclimatologists believe that the Younger Dryas was triggered by the draining of about
9.5x103 km3 of water from Lake Agassiz8 through the St. Lawrence River into the Atlantic Ocean
(Perkins, 2002). Similar emptying of large lakes formed along the edge of northern hemisphere
ice sheets9 may have triggered other Heinrich events. The resultant influx of freshwater was
presumably sufficient to shut down the North Atlantic Drift and NADW formation (Fig. 14). The
associated drop in heat transport to the North Atlantic and Europe would have produced a
dramatic transition to frigid conditions in Europe and the accumulation of sea ice in the North
Atlantic. Eventually, however, conditions along the ice edge during winter months may have led
to the formation of bottom water by the same mechanism currently operative in the Southern
Ocean (see above). With the formation of NADW thus renewed, the transport of heat by the
8
Lake Agassiz was an immense lake, larger than the area of the present-day Great Lakes combined, and covered
much of Manitoba, Ontario, Saskatchewan, and northern Minnesota and North Dakota. It appears to have formed
~13,000 years ago and was fed by glacial runoff. At various times it discharged to the south through the Mississippi
River system or to the northwest through the Mackenzie River. The event that triggered drainage of about 85% of
Lake Agassiz’s volume through the St. Lawrence River about 12,700 years ago was apparently the failure of an ice
dam. Modern remnants of Lake Agassiz include, inter alia, Lake Winnipeg, Lake Winnipegosis, Lake Manitoba,
and Lake of the Woods.
9
For example, large ice-dammed lakes that are known to have formed in the Siberian Altai Mountains.
25
North Atlantic Drift would have returned, eventually leading to the next Dansgaard-Oeschger
event. Thus during glacial periods such as the Wisconsinan, a plausible mechanism exists to
explain alternating Dansgaard-Oeschger and Heinrich events.
One might naively assume that abrupt drainages of ice-dammed lakes would not be a
factor during interglacial periods, but this is not entirely true. During the Younger Dryas the
Laurentide Ice Sheet moved south again, eventually blocking the outflow of Lake Agassiz
through the St. Lawrence River. Lake Agassiz refilled with glacial meltwater and eventually
merged with another meltwater lake, Lake Ojibway. During the early years of the Holocene
interglacial the combined volume of the two lakes is estimated to have been about 2x105 km3,
about 60% more than the combined volume of all the world’s lakes today (Barber et al., 1999).
As the Holocene climate warmed, the ice dam again failed, this time over the Hudson Bay.
Geological studies indicate that most of the enormous volume of the combined meltwater lakes
drained into the Labrador Sea within one year, a flux of roughly 6 Sv (Barber et al., 1999). It is
likely that this influx of freshwater completely blocked formation of deep water in the Labrador
Sea and may have significantly reduced formation of NADW in the Greenland Sea as well. The
result, once again, was a dramatic reduction in the transport of heat to the North Atlantic and
Europe. The failure of the Hudson Bay ice dam occurred about 8,470 years ago and led to a cold
event that lasted roughly 400 years.
The cold event of ~8,200 years ago is the most recent climate change attributed to large
influxes of freshwater to the North Atlantic, but it is by no means the most recent Holocene
climate change. Both Bond et al., and DeMenocal et al. have argued persuasively that climate
during both glacial and interglacial periods is modulated by a cycle with a period of 1,500 ± 500
years (Bond et al., 1997; deMenocal et al., 2000). Although the ultimate mechanism responsible
for producing this modulation is unknown, the process appears to be independent of high-latitude
ice sheets and involves “large-scale ocean and atmosphere reorganizations that were completed
within decades or centuries, perhaps less” (deMenocal et al., 2000, p. 2201). The most recent
manifestation of this climate cycle was the Little Ice Age, which lasted for a period of several
hundred years following the so-called Medieval Warm Period (Fig. 15) and was associated
with bitterly cold winters in North America and Europe (Fig. 16). The fact that such climatic
changes can occur by mechanisms we do not currently understand raises serious concerns about
26
our ability to predict the impact of global warming on the dynamics of ocean/atmospheric
interactions and future climate.
Figure 15. Reconstruction of global temperature anomalies during the last 1,000 years. Source:
<http://en.wikipedia.org/wiki/Image:1000_Year_Temperature_Comparison.png>.
Figure 16. A Scene on the Ice by Hendrick Avercamp was inspired by the harsh winter of
1608 in Europe. Source: <http://en.wikipedia.org/wiki/Image:SCENEONICE.jpg>.
27
One very obvious concern is whether global warming could shut down the formation of
bottom water in the North Atlantic and thereby trigger a prolonged period of cooling. Based on
computer simulations, Rahmstorf has argued that a shutdown of the North Atlantic conveyer is
unlikely to occur through temperature effects alone (Rahmstorf, 2000). A large influx of
freshwater is a much more likely trigger, and as noted by Rahmstorf, “The location of the
freshwater perturbation is also important – a rule of thumb is: the closer to the deep water
formation regions, the more effective it is” (Rahmstorf, 2000, p. 251). Gregory et al., have
argued that the Greenland icecap will begin to melt if air temperatures rise more than 2.7oC and
that a temperature increase of 8oC would cause most of the Greenland icecap to melt within 1,000
years (Gregory et al., 2004). Is this likely to happen, and if so, would the influx of freshwater be
sufficient to shut down the North Atlantic conveyer?
If the entire Greenland icecap were to melt, sea level would rise by about seven meters
(Gregory et al., 2004). Since the surface area of the ocean is 3.6x1014 m2, the volume of water
added to the ocean by melting the Greenland icecap would be 2.5x1015 m3. If this amount of
freshwater were added to the ocean over a period of 1,000 years, the average flux would be 0.08
Sv. Based on Fig. 1-14 and the foregoing discussion, this might be insufficient to literally shut
down the formation of NADW, but it would certainly reduce the rate of formation, perhaps by as
much as 30-40%. An important caveat to this argument is that melting of the Greenland icecap
would almost certainly not result in a steady flux of freshwater into the North Atlantic Ocean for
1,000 years. The flux might be substantially less than 0.08 Sv for extended periods of time and
substantially greater than 0.08 Sv during other times.
Is there any reason to believe that the temperature over Greenland will increase by as
much as 8oC? The Intergovernmental Panel on Climate Change (IPCC) projections indicate that
by the end of the 21st century atmospheric CO2 concentrations will have increased to 710 ppmv
and temperatures will have risen by 1.4-5.8oC.10 What happens after that? Caldeira and Wickett
have addressed this question with the use of a computer simulation model in which they assume
that we continue to burn fossil fuels until there is literally nothing left (Fig. 17) (Caldeira and
Wickett, 2003). Their model says that atmospheric CO2 concentrations will rise to a peak of
~1,900 ppmv
10
The IPCC Web site is <http://www.ipcc.ch/>.
28
Figure 17. (a) Atmospheric CO2 emissions, historical atmospheric CO2 levels and predicted CO2
concentrations, together with changes in ocean pH based on horizontally averaged chemistry. (b)
Estimated maximum change in surface ocean pH as a function of final atmospheric CO2 pressure,
and the transition time over which this CO2 pressure is linearly approached from 280 p.p.m. A,
glacial−interglacial CO2 changes; B, slow changes over the past 300 Myr; C, historical changes in
ocean surface waters; D, unabated fossil-fuel burning over the next few centuries. Reprinted by
permission from Macmillan Publishers Ltd: [Nature] (Caldeira and Wickett (2003), copyright
2003).
around the year 2300 and then very slowly decline. Based on Berner’s GEOCARB II model, an
increase in atmospheric CO2 from 380 to 1,900 ppmv would increase average global
temperatures by about 9.7oC (Berner, 1994). The temperature rise would be substantially greater
at high northern latitudes, because the melting of Arctic sea ice would substantially reduce the
albedo of the Arctic Ocean. So there is a distinct possibility that burning fossil fuels until there is
literally nothing left will melt the Greenland icecap and raise sea level by seven meters. What
then?
There are several issues to consider. First, the icecap will require roughly 1,000 years to
melt. The rise in sea level will therefore average about 7 mm per year. Second, a complete
shutdown of NADW formation will require several centuries (Rahmstorf, 2000). Although most
of the anthropogenic CO2 added to the atmosphere will eventually be taken up by the ocean, the
process of air/sea exchange will require thousands of years to effect a significant drawdown of
atmospheric CO2 concentrations. Caldeira and Wickett’s model, for example, indicates that
atmospheric CO2 concentrations will decline from 1,900 ppmv in the year 2300 to ~1,500 ppmv
by the year 3000 (Caldeira and Wickett, 2003). Thus the global warming caused by the rise in
atmospheric CO2 concentrations will remain in effect for centuries. As noted by Rahmstorf, “A
29
serious cooling of the North Atlantic region (including northwestern Europe) results only in the
longer term, when greenhouse gases decline again and the circulation remains in the ‘off’ mode”
(Rahmstorf, 2000, p. 253). One major uncertainty in the long-term climate change forecasts
concerns the role of the El Niño Southern Oscillation (ENSO) cycle. Currently freshwater export
from the Atlantic increases by about 0.1 Sv during El Niño versus La Niña years, and “in one
model, increased El Niño frequency resulting from global warming draws enough water vapor
from the subtropical Atlantic across into the Pacific to cancel out the weakening effects on the
thermohaline circulation” (Rahmstorf, 2000, p. 252). It is therefore possible that after melting of
the Greenland icecap the increased frequency of El Niño events associated with global warming
would drive freshwater forcing of the North Atlantic to the left of transition (b) in Fig. 1-14 and
turn on the North Atlantic conveyer, if indeed it had been turned off.
30
REFERENCES
Barber, D.C., Dyke, A., Hillaire-Marcel, C., Jennings, A.E., Andrews, J.T., Kerwin, M.W.,
Bilodeau, G., McNeely, R., Southon, J., Morehead, M.D., Gagnon, J.-M., 1999. Forcing
of the cold event of 8,200 years ago by catastrophic drainage of Laurentide lakes. Nature
400, 344-348.
Barnola, J.M., Raynaud, D., Lorius, C., Barkov, N.I., 1999. Historical CO2 record from the
Vostok ice core. In: Trends: a compendium of data on global change, Carbon Dioxide
Information Analysis Center, Oak Ridge National Laboratory, US Department of Energy:
Oak Ridge. (Available online at http://cdiac.esd.ornl.gov/trends/co2/vostok.htm),
Berner, R.A., 1994. Geocarb II: A revised model of atmospheric CO2 over Phanerozoic time.
American Journal of Science 294, 56-91.
Berner, R.A., Canfield, D.E., 1989. A new model for atmospheric oxygen over Phanerozoic time.
American Journal of Science 289, 333-361.
Berner, R.A., Kothavala, Z., 2001. Geocarb III: A revised model of atmospheric CO2 over
Phanerozoic time. American Journal of Science 301, 182-204.
Bond, G., Showers, W.J., Cheseby, M., Lotti, R., Almasi, P., deMenocal, P., Priore, P., Cullen,
H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in North Atlantic
Holocene and glacial climates. Science 278, 1257-1266.
Broecker, W.S., 1997. Thermohaline circulation, the Achilles heel of our climate system: Will
man-made CO2 upset the current balance? Science 278, 1582-1588.
Caldeira, K., Wickett, M.E., 2003. Anthropogenic carbon and ocean pH. Nature 425, 365.
deMenocal, P., Ortiz, J., Guilderson, T., Sarnthein, M., 2000. Coherent high- and low-latitude
climate variability during the Holocene warm period. Science 288, 2198-2202.
Etheridge, D.M., Steele, L.P., Langenfelds, R.L., Francey, R.J., Barnola, J.M., Morgan, V.I.,
2006. Historical CO2 records from the Law Dome DE08, DE08-2, and DSS ice cores. In:
31
Division of Atmospheric Research, CSIRO, Aspendale, Victoria, Australia (Available
online at http://cdiac.ornl.gov/trends/co2/lawdome.html),
Gregory, J.M., Huybrechts, P., Raper, S.C.B., 2004. Threatened loss of the Greenland ice-sheet.
Nature 428, 616.
Gross, M.G., 1982. Oceanography: A View of the Earth. Prentice-Hall, New York.
Hays, J.D., Imbrie, J., Shackleton, N.J., 1976. Variations in the earth's orbit: pacemaker of the ice
ages. Science 194, 1121-1132.
Keeling, C.D., Whorf, T.P., 2006. Atmospheric CO2 concentrations (ppmv) derived from in situ
air samples collected at Mauna Loa Observatory, Hawaii. In: Carbon Dioxide Research
Group, Scripps Institution of Oceanography (Available online at
http://cdiac.ornl.gov/ftp/trends/co2/maunaloa.co2),
Kelly, K.A., Singh, S., Huang, R.X., 1999. Seasonal variations of sea surface height in the Gulf
Stream region. Journal of Physical Oceanography 29, 313-327.
Laws, E.A., 2007. Climate change, oceans, and human health. Ocean Yearbook 21.
Lovelock, J.E., 1979. Gaia: A New Look at Life on Earth. Oxford University Press, Oxford.
Manabe, S., Stouffer, R.J., 1988. Two stable equilibria of a coupled ocean-atmosphere model.
Journal of Climate 1, 841-866.
Patz, J.A., Campbell-Lendrum, D., Holloway, T., Foley, J.A., 2005. Impact of regional climate
chnge on human health. Nature 438, 310-317.
Perkins, S., 2002. Once upon a lake. Science News 162, 283.
Rahmstorf, S., 2000. The thermohaline ocean circulation: A system with dangerous thresholds?
Climate Change 46, 247-256.
Rahmstorf, S., Ganopolski, A., 1999. Long-term global warming scenarios computed with an
efficient coupled climate model. Climatic Change 43, 353-367.
32
Raven, J.A., Falkowski, P.G., 1999. Oceanic sinks for atmospheric CO2. Plant, Cell &
Environment 741-755.
Sagan, C., 1977. Reducing greenhouses and the temperature of Earth and Mars. Nature 269, 224226.
Sagan, C., Chyba, C., 1997. The early faint sun paradox: Organic shielding of ultraviolet-labile
greenhouse gases. Science 276, 1217-1221.
Sagan, C., Mullen, G., 1972. Earth and Mars: Evolution of atmospheres and surface temperature.
Science 177, 52-56.
Schlesinger, W.H., 1997. Biogeochemistry: An Analysis of Global Change. Academic Press, San
Diego.
Watson, A.J., Lovelock, J.E., Margulis, L., 1978. Methanogenesis, fires, and the regulation of
atmospheric oxygen. Biosystems 10, 293-298.
Wyrtki, K., 1979. An abnormal event in the ocean atmosphere system. La Recherche 10, 12121220.
33