oceanography.jpg - 6879 Bytes

S330 Oceanography

Block 3: Ocean Circulation


Coriolis deflection
Types of surface flows
Ocean currents
The Southern Ocean and global climate
Tides and Currents tutorial
Underwater weather and waterfalls
Hurricane science
Subsurface flow of water masses






Atmospheric circulation in mid-latitudes
Vertical convection in the atmosphere

Easterly waves and tropical cyclones
A brief look ahead


Frictional coupling with the ocean
Ekman motion


Pressure gradients in the ocean
Barotropic and baroclinic conditions
Determination of geostrophic current velocities
Pressure, density and dynamic topography


Kinetic energy spectra


Early observations and theories

Why is there a Gulf Stream
The equations of motion
Investigating the ocean through computer modelling

The Gulf Stream system
Geostrophic flow in the Gulf Stream
Insights from MODE
Measuring the Gulf Stream using water characteristics
Gulf Stream 'rings'
Other methods of current measurement
Modelling the circulation of the North Atlantic




The Equatorial Undercurrent
Upwelling in low latitudes

Monsoon winds over the Indian Ocean
The current system of the Indian Ocean

Oceanic wave guides and Kelvin waves
Rossby waves


The Arctic Sea
The Southern Ocean


Solar radiation
The heat-budget equation

Practical application of the principles of conservation and continuity

Upper and intermediate water masses
Deep and bottom water masses

Mixing in the ocean
Temperature-salinity diagrams


The global thermohaline conveyor
The World Ocean Circulation Experiment
Oceanography in the 21st century: predicting climatic change

Back to OU

inthegreeny.jpg - 1215 Bytes



Circulation in both the oceans and the atmosphere is driven by energy from the Sun and modified by the Earth's rotation.

The radiation balance of the Earth-ocean-atmosphere system is positive at low latitudes and negative at high latitudes. Heat is redistributed from low to high latitudes by means of wind systems in the atmosphere and current systems in the ocean. There are two principal components of the ocean circulation: wind-driven surface currents and the density-driven (thermohaline) deep circulation.

Air and water masses moving over the surface of the Earth are only weakly bound to it by friction and so are subject to the coriolis force. The Coriolis force acts at right angles to the direction of motion, so as to deflect winds and currents to the right in the Northern Hemisphere and to the left in the Southern Hemisphere; the deflections are significant because winds and currents travel relatively slowly. The Coriolis force is zero at the Equator and increases to a maximum at the poles.


The global wind system acts to redistribute heat between low and high latitudes.

Whinds blow from regions of high pressure to regions o flow pressure, but they are also affected by the Coriolis force, to an extent that increases with increasing latitude. Because of the differing thermal capacities of continental masses and oceans, wind patterns are greately influenced by the geographical distribution of land and sea.

In mid-latitudes, the predominant weather systems are cyclones (low pressure centres or depressions) and anticyclones (high pressure centres). At low latitudes, the atmospheric circulation consists essentially of the spiral Hadley cells, of which the Trade Winds form the lowermost limb. The Intertropcial Convergence Zone, where the wind systems of the two hemisphere meet, is generally associated with the zone of maximum sea-surface temperature in the vicinity of the Equator.

Heat is transported polewards in the atmosphere as a result of warm air moving into cooler latitudes. It is also transported as latent heat: heat used to convert water to water vapour is released when the water vapour condenses (eg in cloud formation) in a cooler environment. Over the tropical oceans, turbulent atmospheric convection transports large amounts of heat from the sea-surface high into the atmosphere, leading to the formation of cumulus and (especially) cumulonimbus clouds. An extreme expression of this convection of moisture-laden air is the generation of tropical cyclones.

Interaction between the atmosphere and the overlying ocean is most intense - ie atmosphere-ocean 'coupling' is closest - in the tropics.


The global surface current pattern to some extent reflects the surface wind field, but ocean currents are constrained by continental boundaries and current systems are often characterized by gyral circulations.

Maps of wind and current flows of necessity represent average conditions only; at any one time the actual flow at a given point might be markedly different from that shown.

The frictional force caused by the action of wind on the sea-surface is known as the wind stress. Its magnitude is proportional to the square of the wind speed; it is also affected by the roughness of the sea-surface and conditions in the overlying atmosphere.

Wind stress acting on the sea-surface generates motion in the form of waves and currents. The surface current is typically 3% of the wind speed. Motion is transmitted downwards through frictional coupling caused by turbulence. Because flow in the ocean is almost always turbulent, the coefficient of friction that is important for studies of current flow is the coefficient of eddy viscosity. Typical values are 10-5 to 10-1 for Az and 10 to 105m2s-1 for Ah.

Moving water tends towards a state of equilibrium. Flows adjust to the forces acting on them so that eventually those forces balance one another. Major forces that need to be considered with respect to moving water are wind sress at the sea-surface, internal friction (ie eddy viscosity), the Coriolis force and horizontal pressure gradient force; in some situations, friction with the sea-bed and/or with the coastal boundaries also needs to be taken into account. Although deflection by the Coriolis force is greater for slower-moving parcels of water, the magnitude of the force increases with speed, being equal to mfu.

Ekman showed theoretically that under idealized conditions the surface current resulting from wind stress will be 45 deg cum sole of the wind, and that the direction of the wind-induced current will rotate cum sole with depth, forming the Ekman spiral current pattern. An important consequence of this is that the mean flow of the wind-driven (or Ekman) layer is 90 deg to the right of the wind in the Northern Hemisphere and 90 deg to the left of the wind in the Southern Hemisphere.

When the forces that have set water in motion cease to act, the water will continue to move until the energy supplied has been dissipated, mainly by internal friction. During this time, the motion of the water is still influenced by the Coriolis force, and the rotational flows that result are known as inertia currents. The period of rotation of an inertia current varies with the Coriolis parameter f = 2 Ωsin φ, and hence with latitude, φ.

The currents that result when the horizontal pressure gradient force is balanced by the Coriolis force are known as geostrophic currents. The horizontal pressure gradient force may result only from the slope of the sea-surface, and in these conditions isobaric and isopycnic surfaces are parallel and conditions are described as barotropic. When the water is not homogeneous, but instead there are lateral variations in temperature and salinity, part of the variation in pressure at a given depth level results from the density distribution in the overlying water. In these situations, isopycnic surfaces slope in the opposite direction to isobaric surfaces; thus, isobars and isopycnals are inclined to one another and conditions are described as baroclinic.

In geostrophic flow, the angle of slope (θ) of each isobaric surface may be related to u, the speed of the geostrophic current in the vicinity of that isobaric surface, by the gradient equation: tan θ = fu/g. In barotropic flow, the slope of isobaric surfaces remains constant with depth, as does the velocity of the geostrophic current. In baroclinic conditions, the slope of isobaric surfaces follows the sea-surface less and less with increasing depth, and the velocity of the geostrophic current becomes zero at the depth where the isobaric surface is horizontal. The types of geostrophic current that occur in the two situations are sometimes known as 'slope currents' and 'relative currents', respectively. In the oceans, flow is often a combination of the two types of flow, with the relative current superimposed on a slope current.

In baroclinic conditions, the slopes of the isopycnals are very much greater than the slopes of the isobars. As a result, the gradient equation may be used to construct a relationship which gives the average velocity of the geostrophic current flowing between two hydrographic stations in terms of the density distributions at the two stations. This relationship is known as the geostrophic equation or (in its full form) as Helland-Hansen's equation. It is used to determine relative current velocities (ie velocities relative to a selected depth or isobaric surface, at which it may be aassumed that current flow is negligible) at right angles to the section. This method provides information about average conditions only, and is subject to certain simplifying assumptions. Nevertheless, much of what is known about oceanic circulation has been discovered through geostrophic calculations.

Departures of isobaric surfaces from the horizontal (ie from an equipotential surface) may be measured in terms of units of work known as dynamic metres. Variations in the dynamic height of an isobaric surface (including the sea-surface) are known as dynamic topography. On a map of dynamic topography, geostrophic flow is parallel to the contours of the dynamic height in such a direction that the 'highs' are on the right in the Northern Hemisphere and on the left in the Southern Hemisphere. Dynamic topography represents departures of an isobaric surface from the (marine) geoid, which itself has a relief of the order of 100 times that of dynamic topography.

Surface wind stress gives rise to vertical motion of water, as well as horizontal flow. In particular, cyclonic wind systems lead to a lowered sea-surface, raised thermocline and divergence and upwelling, while anticyclonic wind systems give rise to a raised sea-surface, lowered thermocline and convergence and downwelling. Relatively small-scale linear divergences and convergences occur as a result of Langmuir circulation in the upper ocean.

Flow in the ocean occurs over a wide range of time-scales and space-scales. The general circulation, as represented by the average position and velocity of well-established currents such as the Gulf Stream, is known as the 'mean flow' or 'mean motion'.

Most of the energy of the ocean, both kinetic and potential, derives ultimately from solar energy. The potential energy stored in the ocean is about 100 times its kinetic energy, and results from isobars and isopyncals being displaced from their position of least energy (parallel to the geoid) as a result of wind stress or changes in the density distribution of the ocean. If the ocean were at rest and homogeneous, all isobaric and isopycnic surfaces would be parallel to the geoid. The ocean's kinetic energy is that associated with motion in ocean currents including tidal currents (plus surface waves). The kinetic energy associated with the current is proportional to the square of the current speed, and for any given area of ocean, the total kinetic energy associated with the range of periods/frequencies may be represented by a kinetic energy density spectrum.

The ocean is full of eddies. They originate from perturbations in the mean flow, and their formation has the overall effect of transferring energy from the mean flow. There is effectively a 'cascade' of energy through (generally) smaller and smaller eddies, until it is eventually dissipated as heat (through molecular viscosity). Mesoscale eddies, which have length scales of 50-200 km and periods of one to a few months, represent the ocean's 'weather' and contain a significant proportion of the ocean's energy. Current flow around most mesoscale eddies is in approximate geostrophic equilibrium. They are known to form from meanders in intense frontal regions like the Gulf Stream and the Antarctic Circumpolar Current, but may form in other ways too. Eddies of various sizes are generated by interaction of currents with the bottom topography, islands, coasts or other currents or eddies, or as a result of horizontal wind shear.


The subtropical gyres are characterized by intense western boundary currents and diffuse eastern boundary currents. In the North Atlantic, the western boundary current is the Gulf Stream, and the easter boundary current is the Canary Current.

Exploration of the east coast of America, followed by colonization, trading and whaling, led to the western North Atlantic in generaly, and the Gulf Stream in particular, being charted earlier than most other areas of ocean. Some of the most notable charts were made by De Brahm and by Franklin and Folger (late eighteenth century) and Maury (mid-nineteenth century); Maury was also the first to encourage systematic collection and recording of oceangoraphic and meteorological data.

The Gulf Stream consists of water that has come from equatorial regions (largely via the Gulf of Mexico) and water that has recirculated within the subtropical gyre. The low-latitude origin of much of the water means that the Gulf Stream has warm surface waters, although the warm core becomes progressively eroded by mixing with adjacent waters as the Stream flows north-east.

The prevailing Trade Winds cause sea-levels to be higher in the western part of the Atlantic basin than in the eastern prt, and the resulting 'head' of water in the Gulf of Mexico provides a horozontal pressure gradient acting downstream. The flow leaving the Straits of Florida therefore has some of the characteristics of a jet.

The Gulf Stream follows the continental slope as far as Cape Hatteras where it moves into deeper water and has an increasing tendency to form eddies and meanders; the flow also becomes more filamentous, with cold counter-currents. Beyond the Grand Banks, the current becomes even more diffuse and is generally known as the North Atlantic Current.

Flow in the Gulf Stream is in approximate geostrophic equilibrium, and the strong lateral gradients in temperature and salinity mean that the flow is baroclinic. Confidence in practical application of the geostrophic method was greatly increased when it was successfully used to calculate geostrophic current velocities in the Straits of Florida.

The fast, deep currents in the Gulf Stream are associated with the steep downward slope of the isotherms and isopycnals towards the Sargasso Sea. The Gulf Stream may be regarded as a ribbon of high velocity water forming a front between the warm Sargasso Sea water and the cool waters over the continental margin. This is a frontal region with large lateral variations in density, and so is subject to wave-like perturbations known as 'baroclinic instabilities'. These lead to the formation of mesoscale eddies, especially downstream of Cape Hatteras. Eddies formed from meanders with an anticyclonic tendency are known as warm-core eddies, and those formed from meanders with cyclonic tendency are called cold-core eddies. The importance of mesoscale eddies only began to be appreciated in the 1970s, as a result of MODE and othe rsimilar projects.

One of the tools for studying fluid flow is the concept of vorticity, or the tendency to rotate. All objects on the surface of the Earth of necessity share the component of the Earth's rotation appropriate to the latitude: this is known as planetary vorticity and, because vorticity is defined as 2 x angular velocity, is equal to 2Ωsin φ and given the same symbol as the Coriolis parameter, f. Rotatory motion relative to the Earth is known as relative vorticity, ζ. The vorticity of a fluid parcel relative to fixed space - its absolute vorticity - is given by f + ζ. In the absence of external influences, potential vorticity (f + ζ)/D (where D is the depth of the water parcel) remains constant. Away from coastal waters and other regions of strong velocity shear, f is much greater than ζ. By convention, an anticlockwise rotatory tendency is described as positive, and a clockwise one as negative (regardless of hemisphere).

Early theories about oceanic circulation were restricted because the effects of the Earth's rotation - the Coriolis force and hence geostrophic currents - were not appreciated. During the 20th century, ideas about large-scale ocean circulation developed dramatically. Strommel demonstrated that the intensification of the western boundary currents of stubtropical gyres is a consequence of the increase in the Coriolis parameter with latitude, and that western intensification can be explained in terms of vorticity balance.

Sverdrup showed that when horizontal pressure gradient forces, caused by sea-surface slopes, are taken into account, the total wind-driven, meridional (north-south) flow is proportional to the torque, or curl, of the wind stress. Sverdrups' ideas were extended by Munk who used real wind data (rather than sinusoidally varying wind distributions, as used by Stommel and Sversdrup) and allowed for frictional forces resulting from turbulent mixing in both vertical and horizontal directions. The circulation pattern he derived bears a close resemblance to that of the real oceans.

The equations of motion - ie the mathematical equations used to investigate water movements in the oceans - are simply Newton's Second Law, force - mass x acceleration, applied in each of the three coordinate directions. They are most easily solved by considering equalibrium flow, in which there is no acceleration. When this is done for the equation of motion in a vertical direction, it becomes the hydrostatic equation.

Another important principle governing flow in the oceans is the principle of continuity, which expresses the fact that the mass (and, because water is virtually incompressible, volume) of water moving into a region per unit time must equal that leaving it per unit time.

Today, computer modelling is a valuable oceanographic tool. Process models (eg the simulations of the North Atlantic circulation by Sverdrup, Munk and Stommel) are not attempts to replicte the real situation in all its complexity, but experimental mathematical constructs intended to reveal the fundamental factors that determine the ocean circulation. Models may also be used in a predictive mode.

Direct methods of measuring currents divide into two categories. Langranian methods provide information about circulation patterns and involve tracking the motion of an object on the assumption that its motion represents that of the water surrounding it. Eulerian methods are those in which the measuring instrument is held in a fixed position, and the current flow past that point is measured (usually by a rotary current meter attached to a mooring). Increasingly, current velocities are being measured using Acoustic Doppler Current Profilers.

Circulation patterns can also be inferred from the distribution of properties of the water, such as temperature or chlorophyll content. Satellite images can provide near-instantaneous (synoptic) pictures of the sea-surface, and avoid the problems that arise from interpolating between widely spaced measurements, perhaps also taken over a long period of time.

Satellite radar altimetry may be used to determine variations in the height of the sea-surface (ie departures from the geoid), and hence effectively provide a direct measure of the dynamic topography of the sea-surface.

The eastern boundary currents of the subtropical gyres are associated with coastal upwelling which occurs in response to equatorward longshore winds. Areas of divergence and upwelling are characterized by cooler than normal surface waters, a raised thermocline and, because nutrients are continually being supplied to the photic zone, high primary productivity.

The most important factor affecting wintertime climatic conditions over the northern Altantic Ocean and the Nordic Seas is the state of the North Atlantic Oscillation (NAO). This is a continual oscillation in the difference in atmospheric pressure between the Iceland Low and the Azores High. A strong NAO (positive NAO index) leads to strong westerlies, and warm wet winters in north-west Europe. In the ocean, the variability associated with the NAO has a roughly decadal time-scale.


The major components of equatorial current systems are weatward-flowing North ANd South Equatorial Currents, oneo or more eastward-flowing Counter-Currents (surface and subsurface), and an eastward-flowing Equatorial Undercurrent, which is generally centred on the Equator. Flow in the North and South Equatorial Currents is partly directly driven by the Trade Winds and is partly geostrophic flow.

The equatorial current system is best developed in the Pacific Ocean, where the surface waters are under the cumulative influence of the prevailing Trade Winds over the greatest distances. In the Atlantic, the equatorial circulation is affected by the shape of the ocean basin and, indirectly, by the effect of the continental masses on the ITCZ. In the Indian Ocean, the circulation is nonsoonal, most resenbling that in the other tropical oceans in the northern winter.

The Intertropical Convergence Zone is generally displaced north of the Equator so that the South-East Trade Winds blow across it. As a result, divergence of surface waters, and upwelling, occur just south of the Equator. There is a convergence of surface waters at about 4 deg N.

The prevailing easterly winds over the tropical ocean cause the sea-surface to slope up (and the thermocline to slope down) towards the west. As a result, there is an eastward horizontal pressure gradient force and the Equatorial Counter-Current(s) flow(s) down this gradient towards the east in zones of small westward wind stress (the Doldrums).

This eastward horizontal pressure gradient force also drives the Equatorial Undercurrent, which flows in the thermocline below the mixed surface layer. The Equatorial Undercurrent is a rivvon of fast-flowing water, many hundred times wider than it is thick. It is generally aligned along the Equtor, although it may have long-wavelength undulations; if it is diverted away from the Equator, the Coriolis force turns it equatorwards again. The Equatorial Undercurrent has a significant volume transport, particularly in the Pacific.

In the Pacific and the Atlantic, extensive areas of upwelling occur just south of the Equator, in association with the South Equatorial Current. There is also coastal upwelling along the eastern boundaries - either year-round or seasonal - as a result of the Trade Winds blowing along the shore.

Surface divergence and upwelling may occur below the ITCZ because it is a region of low pressure and cyclonic winds. When the ITCZ is over certain regions of doming isotherms (apparently associated with flow in subsurface counter-currents), the doming intensifies and 'protrudes' into the thermocline. These thermal domes seem to be a feature of the eastern sides of oceans. In the Pacific and the Atlantic, all types of upwelling occur most readily on the eastern side of the ocean, because there the thermocline is at its shallowest, and the mixed layer at its thinnest.

The winds over the Indian Ocean change seasonally as a result of the differential heatin gof the ocean and the Asian landmass. During the North-East Monsoons (northern winter), the winds are from Asia and are dry and cool' during the stronger South-West Monsoon, the winds carry moisture from the Arabian Sea to the Indian subcontinent. Because the winds over the equatorial zone change over the course of the year, so does the direction of the sea-surface slope along the Equator. As a result, in the Indian Ocean the Equatorial Undercurrent is only a seasonal feature of the circulation.

The most dramatic seasonal change in the surface circulation of the Indian Ocean is the reversal of the Somali Current which flows south-westwards during the North-East Monsoon but is a major western boundary current during the South-West Monsoon. AT that time of the year, the North Equatorial Current reverses and becomes the South-West Monsoon Current. During the South-West Monsoon, there are regions of intense upwelling on the western side of the ocean, off Somalia and Oman.

The Agulhas Current is the next most powerful boundary current, second only to the Gulf Stream. Its retroflection off the tip of southern Africa is a source of eddies, many of which are carried into the Atlantic.

The ocean can respond to the winds in distant places by means of large=scale disturbances that travel as waves. These waves may propagate along the surface (barotropic waves) or along a region of sharp density gradient such as the thermocline (baroclinic waves); surface waves, in particular, travel very fast. Two of the most important types of waves are Kelvin waves and Rossby (or planetary) waves. Rossby waves result from the need for potential vorticity to be conserved and, relative to the flow, only travel westwards. Kelvin waves may travel eastwards along the Equator (as a double wave) or along coasts (with the coast to the right in the Northern Hemisphere and to the left in the Southern Hemisphere). In these cases, the Equator and the coast, respectively, are acting as wave guides. Because of the equatorial wave guide, the ocean in low latitudes can respond much more rapidly to changes in the overlying wind than can the ocean at higher latitudes.

El Nino or ENSO events are climatic fluctuations centred in the tropical Pacific, in which the east-west slopes in the sea-surface and thermocline collapse, and warm water spreads across the tropical Pacific, along with areas of vigorous convection and heavy rainfall. During El Nino events, the difference in pressure between the South Pacific High and the Indonesian Low is less than usual (ie the Southern Oscillation Index is large and negative), and the Trade Winds are weaker than usual. When the Southern Oscillation Index is large and positive (ie conditions are an extreme version of the 'normal' situation), there is said to be a La Nina.

Unlike the tropical Pacific and Atlantic, the Indian Ocean does not have well defined climatic oscillations, but it does have an anomalous mode in which ocnditions along the Equator become more like those in the other two oceans (warm in the west, cool in the east).

As a result of the contrasting distributions of land and sea in northern and southern high latitudes, both the type of ice cover and the current pattern of the two regions are very different. A large porportion of Arctic pack ice is several years old, while most Antarctic ice is renewed yearly. The main circulatory pattern in the ARctic Sea is an anticyclonic gyre with cross-basin flow between the Bering Sraits and the Fram Strait, where the outflow becomes the East Greenland Current.

The Great Salinity Anomaly was a pulse of low salinity water which, between 1968 and 1981-2, travelled westwards round Greenland, around the Labrador Sea, and the subpolar gyre, and then back to the north-east Altantic and the Norwegian and Greenland Seas.

The major current feature of the Southern Ocean is the Antarctic Circumpolar Current (ACC), which, by virtue of its great depth, has an enormous volume transport. The cumulative influence of the westerly wind stress acting on the ACC is balanced mainly by frictional forces generated by the interaction of the ACC flow along fronts in the Antarctic Polar Frontal Zone, and these current jets often form meanders and eddies. The Antarctic Polar Frontal Zone is a region where surface water converges and sinks; the Antarctic Divergence, between the Antarctic Curcumpolar Current and the Antarctic Polar Current, is a region of upwelling.

The Antarctic Circumpolar Wave is a wave-like progression of maxima and minima of atmospheric pressure, meridional wind stress, sea-surface temperature and sea-ice extent, which travels eastwards around the Antarctic continent. The wave has two wavelengths end-to-end and, as it has a period of 4-5 years, it takes 8-10 years to travel all the way around the Antarctic continent. Its importance for the global climate is not yet understood.


The temperature and salinity of water in the ocean are determined while that water is at the surface. The temperature of surface water is determined by the relative sizes of the different terms in the oceanic heat budget equation; the salinity is determined by the balance between evaporation and precipitation (E-P) and, at high latitudes, by the freezing and melting of ice.

The heat-budget equation for a part of the ocean is:

Qs + Qv = Qb + Qh + Qe + Qt

whereQs is the amount of heat reaching the sea-surface as incoming short-wave radiation, Qv is heat advected into the region in currents, Qb is the heat lost from the sea-surface by long-wave (back) radiation, Qh is the heat lost from teh sea-surface by conduction/convection, Qe is the net amount of heat lost from the sea-surface by evaporation, and Qt is the net amount of heat available to raise the temperature of the water.

The net radiation balance (Qs - Qb) is largely controlled by variations in Qs; these depend partly on the latitudinal variation in incoming solar radiation but also on the amount of cloud and water vapour in the atmosphere. The amount of long-wave radiation emitted from the sea-surface depends on its temperature. However, Qb is the net loss of heat from the sea-surface, and so is also affected by cloud cover, and by the water vapour content, etc., of the overlying air. Qs - Qb is generally positive.

Qh an dQe depend upon the gradients of temperature and water content, respectively, of the air above the sea-surface. Both Qh and Qe generally represent a loss of heat from the sea, and both are greatly enhanced by increased atmospheric turbulence above the sea-surface.

Within 10-15 deg of the Equator, there is a net gain of heat by the sea all year, but outside these latitudes, the winter hemisphere experiences a net heat loss. The patterns of heat loss in the two hemispheres are very different, with Qh playing a greater role in winter cooling in the Northern Hemisphere. This is largely a result of cold, dry continental air moving over the warm western boundary currents, which also increases evaporative heat losses, Qe.

The formation of ice at the sea-surface greatly influences the local heat budget; in particular, it leads to an increase in the albedo and a substantial decrease in Qs, while Qb is not much affected. Thus, once formed, ice tends to be maintained.

The principle of conservation of salt, combined with the principle of continuity, may be used to make deductions about teh volume transports into an dout of semi-enclosed bodies of water or, alternatively (if these are known), about the evaporation-precipitation balance in the region concerned.

Water masses are bodies of water that are identifiable because they have certain combinations of physical and chemical characteristics. The properties most used to identify water masses are temperature (strictly potential temperature) and salinity, because away from the sea-surface they may only be changed through mixing, ie. they are conservative properties. Deep and bottom water masses, and 'thick' upper water masses, are formed in regions of convergence, and where deep convection results from the destabilization of surface waters through cooling and/or increase in salinity. Whether a water mass can form in this way depends not on the absolute density of surface waters but on their density relative to that of underlying water.

Central water masses are 'thick' upper water masses that form in winter in the subtropical gyres. They are characterized by relatively high temperatures and relatively high salinities. the water mass that forms in the Sargasso Sea has a remarkably uniform temperature of about 18 deg C. This '18 deg C water' is an example of a mode water.

The most extensive intermediate water mass is Antarctic Intermediate Water (AAIW), which forms in the Antarctic Polar Frontal Zone. Like other intermediate water masses formed in subpolar regions (eg Labrador Sea Water) AAIW is characterized by relatively low temperature and relatively low salinity. Although Labrador Sea Water is referred to as an intermediate water mass, in some years it may be formed down to depths of 2000 m or more, and will contribute significantly to North Atlantic Deep Water. In contrast to AAIW and Labrador Sea Water, Mediterranean Water is characterized by relatively high temperature and relatively high salinity.

North Atlantic Deep Water is formed in winter, mainly through cooling of surface waters and ice-formation in the Greenland Sea. As in the Labrador Sea and the Mediterranean, the near-surface waters are more easily destabilized because isopycnals bow upwards as a result of cyclonic circulation. Deep convection seems to occur in small, well-defined regions ('chimneys') and produces a dense water mass that mixes at depth with a cold, highly saline outflow from the Arctic Sea. The resulting water mass ciruclates and accumulates in the deep basins of the Greenland and Norwegian Sea. Intermittently, it overflows the Greenland-Scotland ridge, and cascades down into the deep Atlantic, mixing with the overlying water masses. The densest overflow water eventually reaches the Labrador Sea where it mixes with overlying labrador Sea Water to produce a slightly less dense variety of North Atlantic Deep Water than that which flows south at depth in the eastern basin of the Atlantic. There seems to be a 'sea-saw' relationship between the intensity of convection in the Labrador Sea and that in the Greenland Sea; this may be related to the North Atlantic Oscillation.

Antarctic Bottom Water is the most widespread water mass in the oceans. Where are two types, the much more voluminous 'Circumpolar AABW', which flows north from the lower levels in the Antarctic Circumpolar Current, and the extremely cold 'true' AABW, which forms around the Antarctic continent at various locations on the shelf. Ice-formation plays an important part in its production, particularly in coastal polynyas; cooling of surface waters by winds is also important, especially in 'open ocean' polynyas. Most of this extremely dense water never escapes the Antarctic region, but some formed in the Weddell Gyre spills over the Scotia Ridge and flows north in the western Atlantic and in the Indian Ocean. Like other deep water masses, it is affected by the Coriolis force and so flows as deep western boundary currents (as predicted by Stommel).

The deep water mass with the largest volume is Pacific and Indian Ocean Common Water. It is a mixture, being about half Antarctic Bottom Water, and half Antarctic Intermediate Water plus North Atlantic Deep Water.

Dense water tends to sink along isopycnic surfaces, and mixing along isopycnic surfaces occurs with a minimum expenditure of energy. However, mixing does occur between water of different densities. In certain circumstances, mixing also occurs as a result of double-diffusive processes occurring on a molecular scale (eg salt fingering). Generally, different water masses are envisaged as mixing at their boundaries, but 'meddies' are an example of eddies of one water mass being carried into the body of an adjacent water mass; eddies of Indian Ocean water, formed at the Agulhas retroflection, are another example.

Temperature-salinity diagrams may be constructed for different locations in the ocean. For a given location, the shape of the temperature-salinity plot is determined by the different water masses present and the extent to which they have mixed together; also (as in the case of central water masses) by variations within a particular water mass. Assuming that mixing occurs through turbulent processes only, temperature-salinity diagrams may be used to determine the proportions of different water masses contributing to the water at a given depth. Sequences of temperature-salinity diagrams may be used to trace the least-mixed (or core) layer of a water mass, as it spreads through the ocean.

For convenience, density (ρ) is usually written in terms of σ (sigma), where σ = ρ - 1000; temperature-salinity diagrams show equal-σ contours. Temperature-salinity diagrams now generally use potential temperature (θ) rather than in situ temperature (T) (hence also σθ rather than σt), because θ is corrected for adiabatic heating as a result of compression and so is more truly a conservative property than is T. Comparison of the trend of θ-S curve with contoures of σθ may be used to evaluate the degree of stability of a water column, but using θ instead of T does not compensate for the direct effect of pressure on seawater density. This effect may be significant, especially as compressability is affected by temperature and salinity.

Because it is a conservative property, potential vorticity may be used to track water masses. Dissolved oxygen and silica concentration, which are non-conservative properties, may be used to identify as well as track water masses. The 'age' of a water mass (the time since leaving the sea-surface) is indicated by its concentration of dissolved oxygen, and may be determined with some accuracty from the concentrations of certain other substances that enter the ocean at the surface. These include radioactive isotopes such as 14C (radiocarbon) and 3H (tritium) - which are formed both naturally and in nuclear tests - and chlorofluorocarbons, ie CFCs. Using radiocarbon data, residence times for water in teh deep oceans have been estimated to be of the order of 250-500 years.

The 'global chermohaline conveyor' is a schematic representation of the vertical circulation (or 'meridional overturning circulation') of the global ocean. It encapsulates the idea of global convection being driven by the sinking of dense (cold, high salinity) water in the norhtern North Atlantic, with net heat transport in the topmost 1000m or so of the ocean being southward in the Pacific and Indian Oceans and northward in the Atlantic Ocean. The overall net transport of freshwater within the oceans is believed to be in the opposite direction, ie southward in the Atlantic and northward in the Indian and Pacific Oceans.

The observational phase of the World Ocean Circulation Experiment (WOCE) took place during the 1990s. WOCE was designed to investigate large-scale fluxes of heat and freshwater, ocean variability and climate (both now the subject of the CLIVAR programme), and mechnaisms of wter-mass formation. The enormous amount of data collected during WOCE, combined with improvements in modelling techniques, have added greatly to our understanding of the ocean and its role in the global climate system. In the future, the rate of supply of oceanographic data will increase still further, through (for example) Argo floats, underwater autonomous vehicles and new satellites. This improved supply of data will enable assimilation of data into models to be an even more useful tool than it is at present. Plans for a multidisciplinary, Global Ocean Observing System (GOOS) are well advanced, and the programme may become operational around 2010.