Appendix A Ocean Circulation

(10/15/01)
The chemistry and biology of the ocean are superimposed on the ocean's
circulation, thus it is important to review briefly the forces driving this circulation and
give some estimates of the transport rates. There are many reasons why it is important to
understand the basics of the circulation. Four examples are given as an illustration.
(1) Poleward-flowing, warm, surface, western boundary currents such as the Gulf
Stream and the Kuroshiro have a profound effect on the seasurface temperature (SST)
and the climate of land areas bordering the oceans. For example, the Gulf Stream
transports approximately 3.2 peta Watts (peta = 1015) of heat to the North Atlantic
(Hartmann, 1994), moderating the climate of northern Europe.
(2) The El-Nino Southern Oscillation (ENSO) phenomenon is an interannual
perturbation of the climate system characterized by aperiodic weakening of the
tradewinds and warming of the surface water in the central and eastern equatorial Pacific
Ocean. The impacts of ENSO are felt worldwide through disruption of atmospheric
circulation and weather patterns (McPhaden, 1993; Wallace et al, 1998)
(3) Atmospheric testing of nuclear bombs resulted in the contamination of the
surface of the ocean with various isotopes including 14C, 3H, 90Sr, 239Pu, and 240Pu
that have different chemistries and cource functions. These isotopes are slowly being
mixed through the ocean and can be used as radioactive dyes and clocks (e.g. Broecker
and Peng, 1982).
(4) The atmospheric CO2 concentration has been increasing since the beginning
of the industrial age, but the increase (~3.2 Gt C yr-1) is less than the sum of
anthropogenic emissions and deforestation (~ 7.0 Gt C yr-1) (Siegenthaler and Sarmiento,
1993). Some of the CO2 has gone into the ocean (~2 Gt C yr-1). All CO2 taken up by
the ocean is done so by the process of gas-exchange. Some of the excess CO2 has been
transported into the intermediate and deep water by the subduction of water masses.
Circulation replenishes the surface with water undersaturated with respect to the
anthropogenically perturbed CO2 levels.

In this Appendix we give a brief overview of what controls the density of

seawater and the vertical density stratification of the ocean. Surface currents, abyssal
circulation, and thermocline circulation will be considered individually. For more
thorough discussion of these and other aspects of physical oceanography see…

1. Density stratification in the ocean

The density of seawater is controlled by its salt content or salinity and its
temperature. Salinity is historically defined as the total salt content of seawater and the
units were given as grams of salt per kilogram of seawater or parts per thousand (‰).
Salinity was expressed on a mass of seawater basis because mass, rather than volume, is
conserved as the temperature and pressure change. In modern oceanography salinity is
determined as a conductivity ratio on the practical salinity scale and has no units
(Millero, 1993). For more details on the formal definition of salinity and on the
preparation of very accurate standards see the report by UNESCO (1981).
 The salinity of surface
seawater is controlled primarily by
the balance between evaporation
and precipitation. As a result the
highest salinities are found in the
so-called sub-tropical central gyre
regions centered at about 20° to 30°
North and South, where evaporation
is extensive but rainfall is minimal.
Surprisingly, they are not found at
the equator where evaporation is
also large, but so is rainfall. The
highest surface salinities, other than
evaporite basins, are found in the
Red Sea.
 The temperature of seawater is fixed at
the seasurface by heat exchange with the
atmosphere. The average incoming energy
from the sun at the earth's surface is about
four times higher at the equator than at the
poles. The average infrared radiation heat loss
to space is more constant with latitude.
As a result there is a net input of heat
to the earth's surface into the tropical regions
and this is where we find the warmest surface
seawater. Heat is then transferred
from low to high latitudes by winds in the atmosphere and by currents in the ocean. The
geothermal heat flux from the interior of the Earth is generally insignificant except in the
vicinity of hydrothermal vents at spreading ridges (references) and in relatively stagnant
locations like the abyssal northern North Pacific (Joyce, et al. 1986) and the Black Sea
(Murray et al., 1991).

Because the seawater signatures of temperature and salinity are acquired by

processes occurring at the air-sea interface we can also state that the density
characteristics of a parcel of seawater are determined when it is at the seasurface. This
density signature is locked into the water when it sinks. The density will be modified by
mixing with other parcels of water but if the density signatures of all the end member
water masses are known, this mixing can be unraveled and the proportions of the
different source waters to a given parcel can be determined.

To a first approximation the vertical density distribution of the ocean can be

described as a three layered structure. The surface layer is the region from the seasurface
to the depth having a temperature of about 10°C. The transition region where the
temperature decreases from 10°C to 4°C is called the thermocline. The deep sea is the
region below the thermocline.
 Because temperature (T) and salinity (S) are the main factors controlling density,
oceanographers use T-S diagrams to describe the features of the different water masses.
The average temperature and salinity of the world ocean and various parts of the ocean
are given in Fig. B-4 and Table B-3.

The North Atlantic

contains the warmest and saltiest
water of the major oceans. The
Southern Ocean (the region
around Antarctica) is the coldest
and the North Pacific has the
lowest average salinity.
Conventional T-S
diagrams for specific locations in
the individual oceans are shown
in Fig. B-5. The inflections in
the curves reflect the inputs of
water from different sources.
The linear regions represent
mixing intervals between these core sources. For example, in the Atlantic Ocean the
curves reflect input from Antarctic Bottom Water(AABW), North Atlantic Deep Water
(NADW), Antarctic Intermediate Water (AIW), Mediterranean Water (MW), and Warm
Surface Water (WSW).
2. Surface Currents

Surface ocean currents respond primarily to the climatic wind field. The
prevailing winds supply much of the energy that drives surface water movements.
This becomes
clear when charts of
the surface winds and
ocean surface currents
are superimposed.
The wind-driven
circulation occurs
principally in the
upper few hundred
meters and is therefore
primarily a horizontal
circulation although
vertical motions can
be induced when the
geometry of surface
circulation results in convergences (downwelling) or divergences (upwelling). The depth
to which the surface circulation penetrates is dependent on the water column
stratification. In the equatorial region the currents extend to 30-500 m while in the
circumpolar region where stratification is weak the surface circulation can extend to the
sea floor.
 The net direction of motion of the water is
not always the same as the wind, because other
factors come into play. These are shown
schematically in Fig. B-6. The
wind blowing across the seasurface drags the
surface along and sets this thin layer in motion.
The surface drags the next layer and the process
continues downward, involving successively
deeper layers. As a result of friction between the
layers each deeper layer moves more slowly than
the one above and its motion is deflected to the
right (clockwise) in the northern hemisphere by
the Coriolis force. If this effect is represented by
arrows (vectors) whose direction indicates
current direction and length indicates speed, the
change in current direction and speed with depth
forms a spiral. This feature is called the Ekman
spiral. If the wind blew continuously in one
direction for a few days a well developed Ekman spiral would develop. Under these
conditions the integrated net transport over the entire depth of the Ekman spiral would be
at 90° to the right of the wind direction (right in the northern hemisphere and left in the
southern hemisphere). Normally the wind direction is variable so that the actual net
transport is some angle less than 90°.
Ekman transport, changes in seasurface topography and the Coriolis force
combine to form geostrophic currents. Take the North Pacific for example. The
westerlies at ~40°N and the Northeast trades (~10°N) set the North Pacific Current and
North Equatorial Current in motion as a circular gyre. Because of the Ekman drift,
surface water is pushed toward the center of the gyre (~25°N) and piles up to form a
seasurface "topographic high". As a result of the elevated seasurface, water tends to flow
"downhill" in response to gravity. As it flows, however, the Coriolis force deflects the
water to the right (in the northern hemisphere). When the current is constant and results
from balance between the pressure gradient force due to the elevated seasurface and the
Coriolis force, the flow is said to be in geostrophic balance. The actual flow is then
nearly parallel to the contours of the elevated seasurface and clockwise. The seasurface
topography of the Pacific Ocean was determined by Tai and Wunsch (1983) from
satellite altimetry. The absolute elevation of the subtropical gyre can be clearly seen and
fits the schematic description given above.
As a result of these factors (wind, Ekman transport, Coriolis force) the surface
ocean circulation in the mid latitudes is characterized by clockwise gyres in the northern
hemisphere and the counterclockwise gyres in the southern hemisphere. The main
surface currents around these gyres for the world's oceans are shown in Fig. B-7. The
regions where Ekman transport tends to push water together (such as the subtropical
gyres) are called convergences. Divergences (such as the equator) result when surface
waters are pushed apart.
Total transport by the surface currents varies greatly and reflects the mean
currents and cross sectional area. Some representative examples will illustrate the scale.
The transport around the subtropical gyre in the North Pacific is about 70 Sv (1 Sverdrup
or Sv = 1 x 106 m3 s-1). The Gulf Stream, which is a major northward flow off the east
coast of North America, increases from 30 Sv in the Florida Straits to 150 Sv at 64°30'W,
or 2000 km downstream.

3. El Nino-Southern Oscillation
The equatorial Pacific is one of the best studied regions of ocean divergence
(Philander, 1990; McPhaden et al., 1998). This is because of the El Nino-Southern
Oscillation phenomenon. The region around the equator normally experiences strong
easterly trade winds that result in divergence from the equator and upwelling of colder,
nutrient rich water from below (Fig. B-8a). This “cold-tongue” typically extends from the
coast of South America to about the date line (180°). The trade winds also drive near-
equatorial surface flow westward as the South Equatorial Current (SEC). This piles up
warm surface water in the western Pacific to create a deep warm pool. As a result the
thermocline has a tilt and is deeper in the west than the east. The westward flow in the
surface SEC is partly compensated by a return flow to the east in the thermocline
(~150m) called the Equatorial Undercurrent (EUC).
There is a zonal atmospheric circulation system associated with this normal ocean
condition called the Walker Cell (Fig. B-8a). Evaporation rates are high over the warm
pool and warm moist air ascends to great heights (deep convection) producing extensive
cloud systems and rain. The Walker Cell is closed by westerly winds aloft and subsidence
in the high-pressure zone of the eastern Pacific.
 During El Nino (Fig B-8b) the trade winds weaken, and even reverse, in the
central and western Pacific resulting in a local eastward acceleration of the surface
currents. Westerly wind events in the western Pacific excite downwelling equatorial
Kelvin waves which propagate into the eastern equatorial Pacific where they depress the
thermocline (Kessler and McPhaden, 1995)(Fig B-8b). The winds in the eastern Pacific
are usually still easterly favoring upwelling, but because the thermocline is depressed,
warmer water is upwelled. The net result is migration of the “warm pool” and its
associated atmospheric deep convection from the western Pacific to east of the date line
(Fig B-8b). Anomalously warm sea surface temperatures occur from the coast of South
America to the date line.
Deep Convergence in the atmosphere is the main driving force for atmospheric
circulation through the release of latent heat at mid-tropospheric levels. The zonal shift in
the site of deep convection during El Nino affects atmospheric circulation and climate on
a global basis (Wallace et al., 1998). The variations in the upwelling also influence the
flux of CO2 from the ocean to the atmosphere (Feely et al., 1997) and the biological
characteristics of the region (Murray et al., 1994).

4. Thermocline Circulation

The transition region between the

surface and deep ocean is referred to as
the thermocline. This is also a pycnocline
zone where the density increases
appreciably with increasing depth. Most
of the density change results from the
decrease in temperature (hence thermocline).
A simple but physically realistic model based on lateral transport has evolved to
explain the origin of the thermocline. According to this view, the interior of the ocean is
ventilated by rapid mixing and advection along isopycnal surfaces (Jenkins, 1980). The
density surfaces that lie in the thermocline at 200 to 1000 m in the equatorial and mid-
latitude regions shoal and outcrop at high latitudes. The argument is that water acquires
its T and S (and chemical tracer) signature while at the sea surface and then sinks and is
transported horizontally as shown
in Fig. B-9. A map showing the
winter outcrops of isopycnal
surfaces in the North Atlantic
Ocean is shown in Fig. B-10.
Characteristic
values of the horizontal eddy
diffusion coefficient (K) are of the
order of 107 cm2/s. Assuming a
distance (L) of the order of 2000
km (30°) and assuming the
characteristic time is τ = L2/K, we
obtain a characteristic ventilation
time for the main thermocline of about 130 years.
 The horizontal isopycnal thermocline model is important for the problem of
determining the fate of the excess atmospheric CO2. The increase of CO2 in the
atmosphere is modulated by transport of excess CO2 from the atmosphere into the
interior of the ocean. The direct ventilation of the thermocline in its outcropping regions
at high latitudes plays an important role in removing CO2 from the atmosphere
(Siegenthaler and Sarmiento, 1993).
Nuclear bomb produced 14CO2 and 3H (as HTO) have been used to describe and
model this rapid thermocline ventilation (Ostlund et al., 1974; Sarmiento et al., 1982;
Fine et al., 1983). For example, changes in the distributions of tritium in the western
Atlantic between 1972 (GEOSECS) and 1981 (TTO) are shown in Fig. B-11 (Baes and
Mulholland, 1985). In the 10

years following the atmospheric bomb tests of the early 1960's, a massive penetration of
3H into the thermocline at all depths has occurred. Comparison of the GEOSECS and
TTO data, which have a 9 year time difference, clearly shows the rapid ventilation of the
North Atlantic and the value of such "transient" tracers. A similar distribution can be
seen in the distribution of man-made chlorofluorocarbons, which have been released over
a longer period (40 years) (Gammon et al., 1982).

5. Abyssal Circulation

The circulation of the deep ocean below the thermocline is referred to as abyssal
circulation. The currents are slow and difficult to measure but the pattern of circulation
can be clearly seen in the properties of the abyssal water. For example, the water of
lowest temperature in the water column is usually the densest and lies deepest. As a
result, charts of the bottom water temperature have been useful in describing the pattern
of the abyssal circulation (e.g., Mantyla, 1975; Mantyla and Reid, 1983). The
topography of the sea floor plays an important role in constraining the circulation and
much of the abyssal flow is funneled through passages such as the Denmark Straight,
Gibbs Fracture Zone, Vema Channel, Samoan Passage, and Drake Passage.

For a steady state ocean, a requirement of the heat balance is that the input of new
cold abyssal water (Antarctic Bottom Water and North Atlantic Deep Water) sinking in
the high latitude regions must be balanced by the input of heat by geothermal heating
(heat flow from the Earth), downward convection of relatively warm water (e.g., from the
Mediterranean) and downward diffusion of heat across the thermocline. A general mass
balance of the world's oceans requires that the water sinking in the polar regions must be
exactly balanced by the upwelling of water from the abyssal ocean to the surface water.
A combination of the mass and heat balances together with the forcing of the wind and
the effect of a rotating Earth determine the nature of the abyssal circulation.

6. The Ocean Conveyor Belt

The ocean conveyor-belt is one of the major elements of today’s ocean circulation
system (Broecker, 1997). A key feature is that it delivers an enormous amount of heat to
the North Atlantic and this has profound implications for past, present and probably
future climates.
 The conveyor-belt is shown schematically in Fig. B-12. Warm and salty surface
currents in the western North Atlantic (e.g. the Gulf Stream) transport heat to the
Norwegian-Greenland Seas where it is transferred to the atmosphere. This heat helps
moderate the climate of northern Europe. The cooling increases the density resulting in
formation of the now cold and salty North Atlantic Deep Water (NADW)(Worthington,
1970). The NADW travels south through the North and South Atlantic and then joins the
Circumpolar Current that travels virtually unimpeded in a clockwise direction around the
Antarctic Continent.
Deep water also forms along the margins of Antartica and feeds the Circumpolar
Current. The Weddell Sea, because of its very low temperature, is the main source of
Antarctic Bottom Water (AABW) which flows northward at the very bottom into the
South Atlantic, and then through the Vema Channel in the Rio Grande Rise into the
North Atlantic. It ultimately returns southward as part of the NADW.

The circumpolar current is a blend of waters of North Atlantic (~47%) and

Antarctic margin (~53%) origin (Broecker, 1997). This current is referred to as the
Pacific Common Water and is the source of deep water to the Indian and Pacific Oceans.
Deep water does not form in a similar way in the North Pacific because the salinity is too
low (Warren, 1983). Pacific Common water enters the Pacific in the southwest corner
and flows north along the western boundary of the Tonga Trench. The abyssal circulation
model of Stommel (1958) and Stommel and Arons (1960) predicted that deep waters
flow most intensely along the western boundaries in all oceans and gradually circulate
into the interior with a cyclonic flow as allowed by topography. Most of the northward
abyssal flow passes from the southwest Pacific to the north central pacific through the
Samoan Passage, located west of Samoa. In the North Pacific, the abyssal flow splits and
goes west and east of the Hawaiian Islands. These flows meet again north of Hawaii
where they mix, upwell and flow back to the South Pacific at mid-depths.
The conveyor-belt is completed by return flow of surface water from the Pacific
to the Atlantic. There are two main paths of this return flow, which amounts to about 19
Sv. Some passes through the Indonesian Archepelgo, the Indian Ocean and around the tip
of South Africa via the Agulhas Current (Gordon, 1985). Some enters the South Atlantic
via the Drake Passage. Based on the salt budget, Broecker (1991) argued that the Drake
Passage route transports about 50% more than the Agulhas Current. Finally there is a
small transport (about 1 Sv) from the Pacific to the Atlantic through the Bering Strait
that is important for the Arctic salt balance.
The salt budget for the Atlantic, which is determined in part by the flux of fresh
water through the atmosphere, drives the conveyor belt and can explain how it has varied
in the past. At present there appears to be a net water vapor loss of about 0.32 Sv (greater
than the flow of the Amazon) from the Atlantic to the Pacific. The NADW transports
about 16.3 Sv of water with a salinity of 34.91. This is produced from 15 Sv of Gulf
Stream water with a salinity of 35.8, 1 Sv of transport from the Bering Straits with S = 32
and a net excess of river inflow and rainfall over evaporation of about 0.3 Sv (Zauker and
Broecker, 1992). It is easy to show that small changes in the freshwater budget can have a
significant impact. For example, if the excess of precipitation plus runoff over
evaporation increased by 50% to 0.45 Sv, the salt content of the NADW would decrease
to 34.59. In order to compensate for the resulting reduction in density the water would
have to be cooled by an additional 1.4°C and the conveyor would have to more than
double its flow to restore the salt balance (Broecker, 1997). Model simulations have also
shown that the oceans’ thermohaline circulation is extremely sensitive to freshwater input
(Manabe and Stauffer 1995).
Although the general abyssal circulation patterns are fairly well known, it is
difficult to quantify the rates of the various flows. Abyssal circulation is generally quite
slow and variable on short time scales. The calculation of the rate of formation of
abyssal water is also fraught with uncertainty. Probably the most promising means of
assigning the time dimension to oceanic processes
is through the study of the distribution of
radioactive chemical tracers. Difficulties
associated with the interpretation of radioactive
tracer distributions lie both in the models used,
non-conservative interactions, and the difference
between the time scale of the physical transport
phenomenon and the mean life of the tracer.
An example of the power of such tracers is
in the "dating" of abyssal water using 14C. 14C
has an atmospheric source and a half-life of 5720
years. Stuiver et al. (1983) measured the 14C
distribution in dissolved inorganic carbon in deep
samples from major ocean basins (Fig. B-13). This
data was used to calibrate a box model which
indicated that the replacement times for Atlantic,
Indian and Pacific Ocean deep waters (depths >
1500 m) are 275, 250, and 510 years respectively.
The present form of the conveyor belt
appears to have been initiated by closure of the Panamanian seaway between the North
and South American continents (Keigwin, 1982; Maier-Reimer et al., 1990). Geological
evidence indicates that gradual closing of the Isthmus of Panama lasted from 13 to 1.9
Myr ago. Paleoceanographic records indicate that closure was sufficient by 4.6 Myr ago
to cause a marked reorganization of ocean circulation (Burton et al., 1997; Haug and
Tiedemann, 1998). At this time the Gulf Stream intensified resulting in the transport of
warm water to high latitudes. As a result NADW formation intensified and increased
atmospheric moisture input to high latitudes which probably helped trigger the growth of
northern hemispheric ice-sheets.
There is strong evidence that the conveyor belt has switched regularly from one
mode of operation to another in the past. The associated changes in climate have been
large, abrupt and global (Denton and Hendy, 1994). The changes appear to be driven by
factors controlling the density of high-latitude North Atlantic surface water. These events
appear to have been triggered by an increase in ice berg input, mainly from Canada
(Bond et al., 1992). These ice bergs transport terrigenous debris across the North
Atlantic. When they melt they deposit a layer if ice rafted material on the sea floor. These
periodic events in the geological record are called Heinrich events (Broecker, 1994; Bond
et al., 1997). The input of fresh water reduces production of NADW thus shutting down
the present mode of the conveyor belt. The timing of these events has been perfectly
preserved in the sediments from as far away as the Santa Barbara Basin (Behl and
Kennett, 1996) and the glaciers in the Columbian Andes (Thompson et al, 199x). At the
time of these event the climate cools at high latitudes and globally.
The climate records in Greenland ice reveal that over the past 60 kyr conditions
switched back and forth between intense cold and moderate cold on a few 1000 yr time
scale. These so-called Dansgaard-Oeschger cycles are characterized by abrupt changes in
temperature, dust content, ice accumulation rate, methane concentration and CO2 content.
The onset of these cold events occurred on time scales as short as a few decades to a few
years (Alley et al., 1993). Each period of intense cold has been matched by an ice-rafting
or Heinrich event in North Atlantic sediments. As a result of the switch to a colder
climate, ice-berg production slows and the salinity of the north Atlantic surface water
slowly increases enabling NADW formation to occur again. The return to the warm phase
occurs much more slowly, over a 1000 yr time frame. These cyclic events appear to have
continued in the Holocene, although with much muted amplitude (Alley et al., 1997).
There is great concern that one of the effects of global warming could be
reduction in formation of NADW and associated reorganization of the conveyor-belt
circulation (Manabe and Stouffer, 1995). The consequences of global warming will be to
warm surface seawater and to intensify the hydrological cycle. Both factors will make it
more difficult to form deep water and could lead to an “anthropogenically” induced
climate shift. Paradoxically, global warming could result in climate cooling for northern
Europe.
Detailed Structure: Equatorial Pacific
Detailed Structure: Antarctic Polar Front
References:

Alley et al., 1993

Alley et al., 1997
Baes and Mulholland (1985)
Behl and Kennett, 1996
Bond et al (1997)
Broecker and Peng (1982)
Broecker (1991)
Broecker (1997)
Burton et al (1997)
Denton and Hendy (1994)
Feely et al (1997)
Fine et al (1983)
Gammon et al (1982)
Gordon (1985)
Hartmann (1994)
Haug and Tiedemann (1998)
Jenkins (1980)
Joyce et al (1986)
Kegwin (1982)
Kessler and McPhaden (1995)
Maier-Reimer et al (1990)
Manabe and Stauffer (1995)
Mantyla A.W. (1975) On the potential temperature in the abyssal Pacific Ocean. J. Mar.
Res., 33, 341-354.
Mantyla and Reid (1983)
McPhaden (1993)
McPhaden M.J. and others (1998) The tropical ocean-global atmosphere observing
system: A decade of progress. J. Gorphys. Res., 103, 14,169-14,240.
Millero (1993)
Murray et al (1991)
Murray et al (1994)
Ostlund et al (1974)
Philander (1990)
Sarmiento et al (1982)
Siegenthaler and Sarmiento (1993)
Stommel (1958)
Stommel and Arons (1960)
Stuiver et al (1983)
Tai and Wunsch (1983)
Thompson et al (199X)
UNESCO (1981)
Wallace et al (1998)
Warren (1983)
Zauker and Broecker (1992)