continuous body of salt water that is contained in enormous basins on the Earth's surface.

When viewed from space, the predominance of the oceans on the Earth is readily apparent. The oceans and their marginal seas cover nearly 71 percent of the Earth's surface, with an average depth of 3,795 metres (12,450 feet). The exposed land occupies the remaining 29 percent of the planetary surface and has a mean elevation of only 840 metres (2,756 feet). Actually, all the elevated land could be hidden under the oceans and the Earth reduced to a smooth sphere that would be completely covered by a continuous layer of seawater 2,686 metres deep. This is known as the sphere depth of the oceans and serves to underscore the abundance of water on the Earth's surface.

The Earth is unique in the solar system because of its distance from the Sun and its period of rotation. These combine to subject the Earth to a solar radiation level that maintains the planet at a mean surface temperature of 16° C (61° F), which varies little over annual and night-day cycles. This mean temperature allows water to exist on the Earth in all three of its phases—solid, liquid, and gaseous. No other planet in the solar system has this feature. The liquid phase predominates on the Earth. By volume, 97.957 percent of the water on the planet exists as oceanic water and associated sea ice. The gaseous phase and droplet water in the atmosphere constitute 0.001 percent. Fresh water in lakes and streams makes up 0.036 percent, while groundwater is 10 times more abundant at 0.365 percent. Glaciers and ice caps constitute 1.641 percent of the Earth's total water volume.

Each of the above is considered to be a reservoir of water. Water continuously circulates between these reservoirs in what is called the hydrologic cycle, which is driven by energy from the Sun. Evaporation, precipitation, movement of the atmosphere, and the downhill flow of river water, glaciers, and groundwater keep water in motion between the reservoirs and maintain the hydrologic cycle.

The large range of volumes in these reservoirs and the rates at which water cycles between them combine to create important conditions on the Earth. If small changes occur in the rate at which water is cycled into or out of a reservoir, the volume of a reservoir changes. These volume changes may be relatively large and rapid in a small reservoir or small and slow in a large reservoir. A small percentage change in the volume of the oceans may produce a large proportional change in the land-ice reservoir, thereby promoting glacial and interglacial stages. The rate at which water enters or leaves a reservoir divided into the reservoir volume determines the residence time of water in the reservoir. The residence time of water in a reservoir, in turn, governs many of the properties of that reservoir.

This article focuses on the oceanic reservoir of the world. It discusses in general terms the properties of this body of water and the processes that occur within it and at its boundaries with the atmosphere and the crust of the Earth. The article also delineates the major features of the ocean basins, along with those of the continental margins and shorelines. Considered, too, are the economic aspects of the oceans, including some of the environmental problems linked with the utilization of marine resources.

For specifics concerning the relationship of the oceans to the other reservoirs of the Earth's waters, see hydrosphere. See also biosphere for coverage of the life-forms that populate the marine environment. Information about the nature, scope, and methods of oceanography and marine geology are provided in hydrologic sciences (Earth sciences).

If area-volume analyses of the oceans are to be made, then boundaries must be established to separate individual regions. In 1921 Erwin Kossina, a German geographer, published tables giving the distribution of oceanic water with depth for the oceans and adjacent seas. This work was updated in 1966 by H.W. Menard and S.M. Smith. The latter only slightly changed the numbers derived by Kossina. This was remarkable, since the original effort relied entirely on the sparse depth measurements accumulated by individual wire soundings, while the more recent work had the benefit of acoustic depth soundings collected since the 1920s. This type of analysis, called hypsometry, allows quantification of the surface area distribution of the oceans and their marginal seas with depth.

The distribution of oceanic surface area with 5° increments of latitude shows that the distribution of land and water on the Earth's surface is markedly different in the Northern and Southern hemispheres. The Southern Hemisphere may be called the water hemisphere, while the Northern Hemisphere is the land hemisphere. This is especially true in the temperate latitudes.

This asymmetry of land and water distribution between the Northern and Southern hemispheres makes the two hemispheres behave very differently in response to the annual variation in solar radiation received by the Earth. The Southern Hemisphere shows only a small change in surface temperature from summer to winter at temperate latitudes. This variation is controlled primarily by the ocean's response to seasonal changes in heating and cooling. The Northern Hemisphere has one change in surface temperature controlled by its oceanic area and another controlled by its land area. In the temperate latitudes of the Northern Hemisphere, the land is much warmer than the oceanic area in summer and much colder in winter. This situation creates large-scale seasonal changes in atmospheric circulation and climate in the Northern Hemisphere that are not found in the Southern Hemisphere.

The Atlantic exhibits the largest change in surface area and volume when its marginal seas are subtracted. This indicates that the Atlantic has the greatest area of bordering seas, many of which are shallow.

Portions of this curve describe the area of the Earth's surface that exists between elevation (altitude and azimuth) or depth increments. On land, little of the Earth's total area—only about 4 percent—is at elevations above 2,000 metres. Most of the land, 25.3 percent of the total Earth, is between 0 and 2,000 metres. About 13.6 percent of the total land area is at higher elevations, with 86.4 percent between 0 and 2,000 metres when the areas are determined relative to land area only. In the oceans, the percentages of the area devoted to depth increments yield information about the typical structure and shape of the oceanic basins. The small depth increment of 0–200 metres occupies about 5.4 percent of the Earth's total area or 7.6 percent of the oceans' area. This approximates the world's area of continental (continental shelf) shelves, the shallow flat borderlands of the continents that have been alternately covered by the oceans during interglacial stages and uncovered during glacial periods (see below the section Continental margins: Continental shelf (ocean)).

At depths between 200 and 1,000 metres and between 1,000 and 2,000 metres, an area only slightly larger—6.02 percent of the Earth's total area or 8.5 percent of the oceans' area—is found. These depths are related to the regions of the oceans that have very steep slopes where depth increases rapidly. These are the continental slope regions that mark the true edge of the continental landmasses. Marginal seas of moderate depths and the tops of seamounts, however, add their area to these depth zones when all the oceans are considered. The majority of the oceanic area lies between 4,000 and 5,000 metres.

The continental shelf region varies immensely from place to place. The seaward boundary of the continental shelf historically is determined by the 100-fathom, or 200-metre, depth contour. However, 85 fathoms, or 170 metres, is a closer approximation. The true boundary at any given location is marked by a rapid change in slope of the seafloor known as the shelf break. This change in slope may be nearly at the coastline in areas where crustal plates converge, as along the west coast of North and South America, or it may be located more than 1,000 kilometres seaward of the coast, as off the north coast of Siberia. The average width of the shelf is about 75 kilometres, and the shelf has an average slope of about 0.01°, a slope that is barely discernible to the human eye. Seaward of the shelf break, the continental slope is inclined by about 4°.

The huge volume of water contained in the oceans (and seas), 137 × 10⁷ cubic kilometres, has been produced during the geologic history of the Earth. There is little information on the early history of the Earth's waters. However, fossils dated from the Precambrian (Precambrian time) some 3.3 billion years ago show that bacteria and cyanobacteria (blue-green algae) existed, indicating the presence of water during this period. Carbonate (carbonate rock) sedimentary rocks (sedimentary rock), obviously laid down in an aquatic environment, have been dated to 1 billion years ago. Also, there is fossil evidence of primitive marine algae and invertebrates from the outset of the Cambrian Period some 540 million years ago.

The presence of water on the Earth at even earlier times is not documented by physical evidence. It has been suggested, however, that the early hydrosphere formed in response to condensation from the early atmosphere. The ratios of certain elements on the Earth indicate that the planet formed by the accumulation of cosmic dust and was slowly warmed by radioactive and compressional heating. This heating led to the gradual separation and migration of materials to form the Earth's core, mantle, and crust. The early atmosphere is thought to have been highly reducing and rich in gases, notably in hydrogen, and to include water vapour.

The Earth's surface temperature and the partial pressures of the individual gases in the early atmosphere affected the atmosphere's equilibration with the terrestrial surface. As time progressed and the planetary interior continued to warm, the composition of the gases escaping from within the Earth gradually changed the properties of its atmosphere, producing a gaseous mixture rich in carbon dioxide (CO₂), carbon monoxide (CO), and molecular nitrogen (N₂). Photodissociation (i.e., separation due to the energy of light) of water vapour into molecular hydrogen (H₂) and molecular oxygen (O₂) in the upper atmosphere allowed the hydrogen to escape and led to a progressive increase of the partial pressure of oxygen at the Earth's surface. The reaction of this oxygen with the materials of the surface gradually caused the vapour pressure of water vapour to increase to a level at which liquid water could form. This water in liquid form accumulated in isolated depressions of the Earth's surface, forming the nascent oceans. The high carbon dioxide content of the atmosphere at this time would have allowed a buildup of dissolved carbon dioxide in the water and made these early oceans acidic and capable of dissolving surface rocks that would add to the water's salt content. Water must have evaporated and condensed rapidly and accumulated slowly at first. The required buildup of atmospheric oxygen was slow because much of this gas was used to oxidize methane, ammonia, and exposed rocks high in iron. Gradually, the partial pressure of the oxygen gas in the atmosphere rose as photosynthesis by bacteria and photodissociation continued to supply oxygen. Biological processes involving algae increased, and they gradually decreased the carbon dioxide content and increased the oxygen content of the atmosphere until the oxygen produced by biological processes outweighed that produced by photodissociation. This, in turn, accelerated the formation of surface water and the development of the oceans. (For further details on the formation and development of the oceans, see below Chemical and physical properties of seawater: Chemical evolution of the oceans (ocean).)

The chemical composition of seawater is influenced by a wide variety of chemical transport mechanisms. Rivers add dissolved and particulate chemicals to the oceanic margins. Wind-borne particulates are carried to mid-ocean regions thousands of kilometres from their continental source areas. Hydrothermal solutions that have circulated through crustal materials beneath the seafloor add both dissolved and particulate materials to the deep ocean. Organisms in the upper ocean convert dissolved materials to solids, which eventually settle to greater oceanic depths. Particulates in transit to the seafloor, as well as materials both on and within the seafloor, undergo chemical exchange with surrounding solutions. Through these local and regional chemical input and removal mechanisms, each element in the oceans tends to exhibit spatial and temporal concentration variations. Physical mixing in the oceans (thermohaline and wind-driven circulation; see below Circulation of the ocean waters (ocean)) tends to homogenize the chemical composition of seawater. The opposing influences of physical mixing and of biogeochemical input and removal mechanisms result in a substantial variety of chemical distributions in the oceans.

A variety of elements essential to the growth of marine organisms, as well as some elements that have no known biological function, exhibit nutrient-like behaviour broadly similar to nitrate and phosphate. Silicate (silicate mineral) is incorporated into the hard structural parts of certain types of marine organisms (diatoms and radiolarians) that are abundant in the upper ocean. Dissolved silicate concentrations range between less than 1 micromole/kg (1 × 10⁻⁶ mole/kg) in surface waters to approximately 180 micromoles/kg (1.8 × 10^-4 mole/kg) in the deep North Pacific. The concentration of zinc, a metal essential to a variety of biological functions, ranges between approximately 0.05 nanomole/kg (5 × 10⁻¹¹ mole/kg) in the surface ocean to as much as 6 nanomoles/kg (6 × 10⁻⁹ mole/kg) in the deep Pacific. The distribution of zinc in the oceans is observed to generally parallel silicate distributions. cadmium, though having no known biological function, generally exhibits distributions that are covariant with phosphate and concentrations that are even lower than those of zinc.

Many elements, including the essential trace metals iron, cobalt, and copper, show surface depletions but in general exhibit behaviour more complex than that of phosphate, nitrate, and silicate. Some of the complexities observed in elemental oceanic distributions are attributable to the adsorption of elements on the surface of sinking particles. Adsorptive processes, either exclusive of or in addition to biological uptake, serve to remove elements from the upper ocean and deliver them to greater depths. The distribution patterns of a number of trace elements are complicated by their participation in oxidation-reduction (oxidation–reduction reaction) (electron-exchange) reactions. In general, electron-exchange reactions lead to profound changes in the solubility and reactivity of trace metals in seawater. Such reactions are important to the oceanic behaviour of a variety of elements, including iron, manganese, copper, cobalt, chromium, and cerium.

The processes that deliver dissolved, particulate, and gaseous materials to the oceans ensure that they contain, at some concentration, very nearly every element that is found in the Earth's crust and atmosphere. The principal components of the atmosphere, nitrogen (78.1 percent), oxygen (21.0 percent), argon (0.93 percent), and carbon dioxide (0.035 percent), occur in seawater in variable proportions, depending on their solubilities and oceanic chemical reactions. In equilibrium with the atmosphere, the concentrations of the unreactive gases, nitrogen and argon, in seawater (0° C, salinity 35) are 616 micromoles/kg and 17 micromoles/kg, respectively. For seawater at 35° C, these concentrations would decrease by approximately a factor of two. The solubility behaviours of argon and oxygen are quite similar. For seawater in equilibrium with the atmosphere, the ratio of oxygen and argon concentrations is approximately 20.45. Since oxygen is a reactive gas essential to life, oxygen concentrations in seawater that are not in direct equilibrium with the atmosphere are quite variable. Although oxygen is produced by photosynthetic organisms at shallow, sunlit ocean depths, oxygen concentrations in near-surface waters are established primarily by exchange with the atmosphere. Oxygen concentrations in the oceans generally exhibit minimum values at intermediate depths and relatively high values in deep waters. This distribution pattern results from a combination of biological oxygen utilization and physical mixing of the ocean waters. Estimates of the extent of oxygen utilization in the oceans can be obtained by comparing concentrations of oxygen with those of argon, since the latter are only influenced by physical processes. The physical processes that influence oxygen distributions include, in particular, the large-scale replenishment of oceanic bottom waters (bottom water) with cold, dense, oxygen-rich waters sinking toward the bottom from high latitudes. Due to the release of nutrients that accompanies the consumption of oxygen by biological debris, dissolved oxygen concentrations generally appear as a mirror image of dissolved nutrient concentrations.

Processes involving dissolved (solution) and particulate organic carbon are of central importance in shaping the chemical character of seawater. Marine organic carbon principally originates in the uppermost 100 metres of the oceans where dissolved inorganic carbon is photosynthetically converted to organic materials. The “rain” of organic-rich particulate materials, resulting directly and indirectly from photosynthetic production, is a principal factor behind the distributions of many organic and inorganic substances in the oceans. A large fraction of the vertical flux of materials in the uppermost waters is converted to dissolved substances within the upper 400 metres of the oceans. Dissolved organic carbon (DOC) accounts for at least 90 percent of the total organic carbon in the oceans. Estimates of DOC appropriate to the surface of the open ocean range between roughly 100 and 500 micromoles of carbon per kilogram of seawater. DOC concentrations in the deep ocean are 5 to 10 times lower than surface values. DOC occurs in an extraordinary variety of forms, and, in general, its composition is controversial and poorly understood. Conventional techniques have indicated that, in surface waters, about 15 percent of DOC can be identified as carbohydrates and combined amino acids. At least 1–2 percent of DOC in surface waters occurs as lipids and 20–25 percent as relatively unreactive humic substances. The relative abundances of reactive organic substances, such as amino acids and carbohydrates, are considerably reduced in deep ocean waters. Dissolved and particulate organic carbon in the surface ocean participates in diel cycles (i.e., those of a 24-hour period) related to photosynthetic production and photochemical transformations. The influence of dissolved organic matter on ocean chemistry is often out of proportion to its oceanic abundance. Photochemical reactions involving DOC can influence the chemistry of vital trace nutrients such as iron, and, even at dissolved concentrations on the order of one nanomole/kg (1 × 10^-9 mole/kg), dissolved organic substances in the upper ocean waters are capable of greatly altering the bioavailability of essential trace nutrients, as, for example, copper and zinc.

Although the oceans constitute an enormous reservoir, human activities have begun to influence their composition on both a local and a global scale. The addition of nutrients (through the discharge of untreated sewage or the seepage of soluble mineral fertilizers, for example) to coastal waters results in increased phytoplankton growth, high levels of dissolved and particulate organic materials, decreased penetration of light through seawater, and alteration of the community structure of bottom-dwelling organisms. Through industrial and automotive emissions, lead concentrations in the surface ocean have increased dramatically on a global scale compared with preindustrial levels. Certain toxic organic compounds, such as polychlorinated biphenyls (polychlorinated biphenyl) (PCBs), are found in seawater and marine organisms and are attributable solely to the activities of humankind. Although most radioactivity in seawater is natural (approximately 90 percent as potassium-40 and less than 1 percent each as rubidium-87 and uranium-238), strontium-90 and certain other artificial radioisotopes have unique environmental pathways and potential for bioaccumulation. Among the most dramatic influences of human activities on a global scale is the remarkable increase of carbon dioxide levels in the atmosphere. Atmospheric carbon dioxide levels are expected to double by the middle of the 21st century, with potentially profound consequences for global climate and agricultural patterns. It is thought that the oceans, as a great reservoir of carbon dioxide, will ameliorate this consequence of human activities to some degree. (For more specific information on this subject, see hydrosphere: Impact of human activities on the hydrosphere: Buildup of greenhouse gases (hydrosphere).)

The chemical history of the oceans has been divided into three stages. The first is an early stage in which the Earth's crust was cooling and reacting with volatile or highly reactive gases of an acidic, reducing nature to produce the oceans and an initial sedimentary rock mass. This stage lasted until about 3.5 billion years ago. The second stage was a period of transition from the initial to essentially modern conditions, and it is estimated to have ended 2 to 1.5 billion years ago. Since that time it is likely that there has been little change in seawater composition.

At an initial crustal temperature of about 600° C, almost all these compounds, including water (H₂O), would be in the atmosphere. The sequence of events that occurred as the crust cooled is difficult to construct. Below 100° C all the H₂O would have condensed, and the acid gases (gas) would have reacted with the original igneous (igneous rock) crustal minerals to form sediments (sedimentary rock) and an initial ocean (marine sediment). There are at least two possible pathways by which these initial steps could have been accomplished.

One pathway assumes that the 600° C atmosphere contains, together with other compounds, water (as vapour), carbon dioxide, and hydrochloric acid in the ratio of 20:3:1 and would cool to the critical temperature of water. The water vapour therefore would have condensed into an early hot ocean. At this stage, the hydrochloric acid would be dissolved in the ocean (about 1 mole per litre), but most of the carbon dioxide would still be in the atmosphere with about 0.5 mole per litre in the ocean water. This early acid ocean would react vigorously with crustal minerals, dissolving out silica and cations and creating a residue that consisted principally of aluminous clay minerals that would form the sediments of the early ocean basins. This pathway of reaction assumes that reaction rates are slow relative to cooling. A second pathway of reaction, which assumes that cooling is slow, is also possible. In this case, at a temperature of about 400° C most of the water vapour would be removed from the atmosphere by hydration reactions with pyroxenes and olivines. Under these conditions, water vapour would not condense until some unknown temperature was reached, and the Earth might have had at an early stage in its history an atmosphere rich in carbon dioxide and no ocean: the surface would have been much like that of present-day Venus.

The pathways described are two of several possibilities for the early surface environment of the Earth. In either case, after the Earth's surface had cooled to 100° C, it would have taken only a short time geologically for the acid gases to be used up in reactions involving igneous rock minerals. The presence of bacteria and possibly algae in the fossil record of rocks older than 3 billion years attests to the fact that the Earth's surface had cooled to temperatures lower than 100° C by this time and that the neutralization of the original acid gases had taken place. If most of the degassing of primary volatile substances from the Earth's interior occurred early, the chloride released by reaction of hydrochloric acid with rock minerals would be found in the oceans and seas or in evaporite deposits, and the oceans would have a salinity and volume comparable to those that they have today.

This conclusion is based on the assumption that there has been no drastic change in the ratios of volatiles released through geologic time. The overall generalized reaction indicative of the chemistry leading to formation of the early oceans can be written in the form: primary igneous rock minerals + acid volatiles + H₂O → sedimentary rocks + oceans + atmosphere. Notice from this equation that if all the acid volatiles and H₂O were released early in the history of the Earth and in the proportions found today, then the total original sedimentary rock mass produced would be equal to that of the present time, and ocean salinity and volume would be near what they are now. If, on the other hand, degassing were linear with time, then the sedimentary rock mass would have accumulated at a linear rate, as would oceanic volume. However, the salinity of the oceans would remain nearly the same if the ratios of volatiles degassed did not change with time. The most likely situation is that presented here—namely, that major degassing occurred early in Earth history, after which minor amounts of volatiles were released episodically or continuously for the remainder of geologic time. The salt content of the oceans based on the constant proportions of volatiles released would depend primarily on the ratio of sodium chloride (NaCl) locked up in evaporites to that dissolved in the oceans. If all the sodium chloride in evaporites were added to the oceans today, the salinity would be roughly doubled. This value gives a sense of the maximum salinity the oceans could have attained throughout geologic time.

One component missing from the early terrestrial surface was free oxygen because it would not have been a constituent released from the cooling crust. As noted earlier, early production of oxygen was by photodissociation of water in the atmosphere as a result of absorption of ultraviolet (ultraviolet radiation) light. The reaction is 2H₂O + hν → O₂ + 2H₂, in which hν represents a photon of ultraviolet light. The hydrogen produced would escape into space, and the O₂ would react with the early reduced gases by reactions such as 2H₂S + 3O₂ → 2SO₂ + 2H₂O. Oxygen production by photodissociation gave the early reduced atmosphere a start toward present-day conditions, but it was not until the appearance of photosynthetic organisms approximately 3.3 billion years ago that it was possible for the accumulation of oxygen in the atmosphere to proceed at a rate sufficient to lead to today's oxygenated environment. The photosynthetic reaction leading to oxygen production may be written 6CO₂ + 6H₂O + hν → C₆H₁₂O₆ + 6O₂, in which C₆H₁₂O₆ represents sugar.

The nature of the rock record from the time of the first sedimentary rocks (about 3.5 billion years ago) to approximately 2 to 1.5 billion years ago suggests that the amount of oxygen in the atmosphere was significantly lower than today and that there were continuous chemical trends in the sedimentary rocks formed and, more subtly, in oceanic composition. The source rocks of sediments during this time were likely to be more basaltic (basalt) than would later ones; sedimentary detritus was formed by the alteration of these rocks in an oxygen-deficient atmosphere and accumulated primarily under anaerobic marine conditions. The chief difference between reactions involving mineral-ocean equilibriums at this time and at the present time was the role played by ferrous iron. The concentration of dissolved iron in the present-day oceans is low because of the insolubility of oxidized iron oxides. During the period (Precambrian time) 3.5 to 1.5 billion years ago, oxygen-deficient environments were prevalent; these favoured the formation of minerals containing ferrous iron (reduced state of iron) from the alteration of basaltic rocks. Indeed, the iron carbonate siderite and the iron silicate (silicate mineral) greenalite, in close association with chert (chert and flint) and the iron sulfide pyrite, are characteristic minerals that occur in middle Precambrian iron formations (those about 1.5 to 2.4 billion years old). The chert originally was deposited as amorphous silica; equilibrium between amorphous silica, siderite, and greenalite at 25° C and one atmosphere total pressure requires a carbon dioxide pressure of about 10^-2.5 atmosphere, or 10 times the present-day value.

The oceans at this time can be thought of as the solution resulting from an acid leach of basaltic rocks, and because the neutralization of the volatile acid gases was not restricted primarily to land areas as it is presently, much of this alteration may have occurred by submarine processes. The atmosphere at the time was oxygen-deficient; anaerobic depositional environments with internal carbon dioxide pressures of about 10^-2.5 atmosphere were prevalent, and the atmosphere itself may have had a carbon dioxide pressure near 10^-2.5 atmosphere. If so, the pH of early ocean water was lower than that of modern seawater, the calcium concentration was higher, and the early ocean water was probably saturated with respect to amorphous silica (about 120 parts per million 【ppm】).

To simulate what might have occurred, it is helpful to imagine emptying the Pacific basin, throwing in great masses of broken basaltic material, filling it with hydrochloric acid so that the acid becomes neutralized, and then carbonating the solution by bubbling carbon dioxide through it. Oxygen would not be permitted into the system. The hydrochloric acid would leach the rocks, resulting in the release and precipitation of silica and the production of a chloride ocean containing sodium, potassium, calcium, magnesium, aluminum, iron, and reduced sulfur species in the proportions present in the rocks. As complete neutralization was approached, aluminum could begin to precipitate as hydroxides and then combine with precipitated silica to form cation-deficient aluminosilicates. The aluminosilicates, as the end of the neutralization process was reached, would combine with more silica and with cations to form minerals like chlorite, and ferrous iron would combine with silica and sulfur to make greenalite and pyrite. In the final solution, chlorine would be balanced by sodium and calcium in roughly equal proportions, with subordinate potassium and magnesium; aluminum would be quantitatively removed, and silicon would be at saturation with amorphous silica. If this solution were then carbonated, calcium would be removed as calcium carbonate, and the chlorine balance would be maintained by abstraction of more sodium from the primary rock. The sediments produced in this system would contain chiefly silica, ferrous iron silicates, chloritic minerals, calcium carbonate, calcium magnesium carbonates, and minor pyrite.

If the hydrochloric acid added were in excess of the carbon dioxide, the resultant ocean would have a high content of calcium chloride, but the pH would still be near neutrality. If the carbon dioxide added were in excess of the chlorine, calcium would be precipitated as the carbonate until it reached a level approximately that of the present oceans—namely, a few hundred parts per million.

If this newly created ocean were left undisturbed for a few hundred million years, its waters would evaporate and be transported onto the continents (in the form of precipitation); streams would transport their loads into it. The sediment created in this ocean would be uplifted and incorporated into the continents. Gradually, the influence of the continental debris would be felt, and the pH might shift slightly. Iron would be oxidized out of the ferrous silicates to produce iron oxides, but the water composition would not vary a great deal.

The primary minerals of igneous rocks (igneous rock) are all mildly basic compounds. When they react in excess with acids such as hydrochloric acid and carbon dioxide, they produce neutral or mildly alkaline solutions plus a set of altered aluminosilicate and carbonate reaction products. It is highly unlikely that ocean water has changed through time from a solution approximately in equilibrium with these reaction products, which are clay minerals and carbonates.

The oceans probably achieved their modern characteristics 2 to 1.5 billion years ago. The chemical and mineralogical compositions and the relative proportions of sedimentary rocks of this age differ little from their Paleozoic counterparts (those dating from about 570 to 245 million years ago). The fact that the acid sulfur gases had been neutralized to sulfate by this time is borne out by calcium sulfate deposits of late Precambrian age (roughly 570 million to 1.6 billion years old). Chemically precipitated ferric oxides in late Precambrian sedimentary rocks indicate available free oxygen, whatever its percentage. The chemistry and mineralogy of middle and late Precambrian shales is similar to that of Paleozoic shales. Thus, it appears that continuous cycling of sediments like those of the present time has occurred for 1.5 to 2 billion years and that these sediments have controlled oceanic composition.

It was once thought that the saltiness (salt) of the modern oceans simply represents the storage of salts derived from rock weathering and transported to the oceans by fluvial processes. With increasing knowledge of the age of the Earth, however, it was realized that, at the present-day rate of delivery of salts to the oceans or even at much reduced rates, the total salt content and the mass of individual salts in the oceans could be attained in geologically short-time intervals compared to the Earth's age. The total mass of salt in the oceans can be accounted for at present-day rates of stream delivery in about 12 million years. The mass of dissolved silica in ocean water can be doubled in only 20,000 years by addition of stream-derived silica; to double sodium would take 70 million years. It then became apparent that the oceans were not simply an accumulator of salts, but as water evaporated from the oceans, along with some salt, the introduced salts must be removed in the form of minerals. Thus, the concept of the oceans as a chemical system changed from that of a simple accumulator to that of a steady-state system in which rates of inflow of materials into the oceans equal rates of outflow. The steady-state concept permits influx to vary with time, but it would be matched by nearly simultaneous and equal variation of efflux. Calculations of rates of addition of elements to the oceanic system and removal from it show that for at least 100 million years the oceanic system has been in a steady state with approximately fixed rates of major element inflow and outflow and, thus, fixed chemical composition.

A discussion of salinity, the salt content of the oceans, requires an understanding of two important concepts: (1) the present-day oceans are considered to be in steady state, receiving as much salt as they lose (see above), and (2) the oceans have been mixed over such a long time period that the composition of sea salt is everywhere the same in the open ocean. This uniformity of salt content results in oceans in which the salinity varies little over space or time.

The range of salinity observed in the open ocean is from 33 to 37 grams of salt per kilogram of seawater or parts per thousand (⁰/₀₀). For the most part, the observed departure from a mean value of approximately 35⁰/₀₀ is caused by processes at the Earth's surface that locally add or remove fresh water. Regions of high evaporation have elevated surface salinities, while regions of high precipitation have depressed surface salinities. In nearshore regions close to large freshwater sources, the salinity may be lowered by dilution. This is especially true in areas where the region of the ocean receiving the fresh water is isolated from the open ocean by the geography of the land.

Areas of the Baltic Sea may have salinity values depressed to 10⁰/₀₀ or less. Increased salinity by evaporation (vaporization) is accentuated where isolation of the water occurs. This effect is found in the Red Sea, where the surface salinity rises to 41⁰/₀₀. Coastal lagoon salinities in areas of high evaporation may be much higher. The removal of fresh water by evaporation or the addition of fresh water by precipitation does not affect the constancy of composition of the sea salt in the open sea. A river draining a particular soil type, however, may bring to the oceans only certain salts that will locally alter the salt composition. In areas of high evaporation where the salinity is driven to very high values, precipitation of particular salts may alter the composition too. At high latitudes where sea ice forms seasonally, the salinity of the seawater is elevated during ice formation and reduced when the ice melts.

At depth in the oceans, salinity (halocline) may be altered as seawater percolates into fissures associated with deep-ocean ridges and crustal rifts involving volcanism. This water then returns to the ocean as superheated water carrying dissolved salts from the magmatic material within the crust. It may lose much of its dissolved load to precipitates on the seafloor and gradually blend in with the surrounding seawater, sharing its remaining dissolved substances.

Salt concentrations as high as 256⁰/₀₀ have been found in hot but dense pools of brine trapped in depressions at the bottom of the Red Sea. The composition of the salts in these pools is not the same as the sea salt of the open oceans.

The salinities of the open oceans found at the greater depths are quite uniform in both time and space with average values of 34.5 to 35⁰/₀₀. These salinities are determined by surface processes such as those described above when the water, now at depth, was last in contact with the surface.

The average distribution of the surface salinity of the open oceans is depicted in Figure 3. This figure shows the response of the salinity of the ocean surface to the latitudinal variation in exchange of water between the oceans and the atmosphere. It also shows the impact of major ocean currents (ocean current) that displace surface water from one latitudinal zone to another. The northward displacement of subtropical water of higher salinity by the Gulf Stream and the North Atlantic Current is evident.

The intertropical convergence, with its high precipitation centred about 5° N, supports the tropical rain forests of the world and leaves its imprint on the oceans as a latitudinal depression of surface salinity. At approximately 30°–35° N and 30°–35° S, the subtropical zones called the horse latitudes (horse latitude) are belts of high evaporation that produce major deserts and grasslands on the continents and cause the surface salinity to rise. At 50°–60° N and 50°–60° S, precipitation again increases.

Mid-ocean surface temperatures vary with latitude in response to the balance between incoming solar radiation and outgoing long-wave radiation. There is an excess of incoming solar radiation at latitudes less than approximately 45° and an excess of radiation loss at latitudes higher than approximately 45°. Superimposed on this radiation balance are seasonal (season) changes in the intensity of solar radiation and the duration of daylight hours due to the tilt of the Earth's axis to the plane of the ecliptic and the rotation of the planet about this axis. The combined effect of these variables is that average ocean surface temperatures are higher at low latitudes than at high latitudes. Because the Sun, with respect to the Earth, migrates annually between the tropic of Cancer and the tropic of Capricorn, the yearly change in heating of the Earth's surface is small at low latitudes and large at mid- and higher latitudes.

Water has an extremely high heat capacity, and heat is mixed downward during summer surface-heating conditions and upward during winter surface cooling. This heat transfer reduces the actual change in ocean surface temperatures over the annual cycle. In the tropics the ocean surface is warm year-round, varying seasonally about 1° to 2° C. At mid-latitudes the mid-ocean temperatures vary about 8° C over the year. At the polar latitudes the surface temperature remains near the ice point of seawater—about −1.9° C.

In winter, prevailing winds (wind) carry cold air masses off the continents in temperate and subarctic latitudes, cooling the adjacent surface seawater below that of the mid-ocean level. In summer, the opposite effect occurs, as warm continental air masses move out over the adjacent sea. This creates a greater annual range in sea surface temperatures at mid-latitudes on the western sides of the oceans of the Northern Hemisphere but has only a small effect in the Southern Hemisphere as there is little land present. Instead, the oceans of the Southern Hemisphere act to control the air temperature, which in turn influences the land temperatures of the temperate zone and reduces the annual temperature range over the land.

Currents carry water having the characteristics of one latitudinal zone to another zone. The northward displacement of warm water to higher latitudes by the Gulf Stream of the North Atlantic and the Kuroshio (Japan Current) of the North Pacific creates sharp changes in temperature along the current boundaries or thermal fronts, where these northward-moving flows meet colder water flowing southward from higher latitudes. Cold water currents flowing from higher to lower latitudes also displace surface isotherms from near constant latitudinal positions. At low latitudes the trade winds act to move water away from the lee coasts of the landmasses to produce areas of coastal upwelling of water from depth and reduce surface temperatures.

Temperatures in the oceans decrease with increasing depth. There are no seasonal changes at the greater depths. The temperature range extends from 30° C at the sea surface to −1° C at the seabed. Like salinity, the temperature at depth is determined by the conditions that the water encountered when it was last at the surface. In the low latitudes the temperature change from top to bottom in the oceans is large. In high temperate and Arctic regions, the formation of dense water at the surface that sinks to depth produces nearly isothermal conditions with depth.

Areas of the oceans that experience an annual change in surface heating have a shallow wind-mixed layer of elevated temperature in the summer. Below this nearly isothermal layer 10 to 20 metres thick, the temperature decreases rapidly with depth, forming a shallow seasonal thermocline (i.e., layer of sharp vertical temperature change). During winter cooling and increased wind mixing at the ocean surface, convective overturning and mixing erase this shallow thermocline and deepen the isothermal layer. The seasonal thermocline re-forms when summer returns. At greater depths, a weaker nonseasonal thermocline is found separating water from temperate and subpolar sources.

Below this permanent thermocline, temperatures decrease slowly. In the very deep ocean basins, the temperature may be observed to increase slightly with depth. This occurs when the deepest parts of the oceans are filled by water with a single temperature from a common source. This water experiences an adiabatic temperature rise as it sinks. Such a temperature rise does not make the water column unstable because the increased temperature is caused by compression, which increases the density of the water. For example, surface seawater of 2° C sinking to a depth of 10,000 metres increases its temperature by about 1.3° C. When measuring deep-sea temperatures, the adiabatic temperature rise, which is a function of salinity, initial temperature, and pressure change, is calculated and subtracted from the observed temperature to obtain the potential temperature. Potential temperatures are used to identify a common type of water and to trace this water back to its source.

The unit of heat called the gram calorie is defined as the amount of heat required to raise the temperature of one gram of water 1° C. The kilocalorie, or food calorie, is the amount of heat required to raise one kilogram of water 1° C. heat capacity is the amount of heat required to raise one gram of material 1° C under constant pressure. In the International System of Units (SI), the heat capacity of water is one kilocalorie per kilogram per degree Celsius. Water has the highest heat capacity of all common Earth materials; therefore, water on the Earth acts as a thermal buffer, resisting temperature change as it gains or loses heat energy.

The heat capacity of any material can be divided by the heat capacity of water to give a ratio known as the specific heat of the material. Specific heat is numerically equal to heat capacity but has no units. In other words, it is a ratio without units. When salt is present, the heat capacity of water decreases slightly. Seawater of 35⁰/₀₀ has a specific heat of 0.932 compared to 1.000 for pure water.

Pure water freezes at 0° C and boils at 100° C under normal pressure conditions. When salt is added, the freezing point is lowered and the boiling point is raised. The addition of salt also lowers the temperature of maximum density below that of pure water (4° C). The temperature of maximum density decreases faster than the freezing point as salt is added.

At 30⁰/₀₀ salinity, the temperature of maximum density is lower than the initial freezing point of saltwater. Therefore, a maximum density is never achieved, as seawater of this salinity is cooled because freezing occurs first. At 24.70⁰/₀₀ salinity, the freezing point and the temperature of maximum density coincide at −1.332° C. At salinities typical of the open oceans, which are greater than 24.7⁰/₀₀, the freezing point is always higher than the temperature of maximum density.

When water changes its state, hydrogen bonds between molecules are either formed or broken. Energy is required to break the hydrogen bonds, which allows water to pass from a solid to a liquid state or from a liquid to a gaseous state. When hydrogen bonds are formed, permitting water to change from a liquid to a solid or from a gas to a liquid, energy is liberated. The heat energy input required to change water from a solid at 0° C to a liquid at 0° C is the latent heat of fusion and is 80 calories per gram of ice. Water's latent heat of fusion is the highest of all common materials. Because of this, heat is released when ice forms and is absorbed during melting, which tends to buffer air temperatures as land and sea ice form and melt seasonally.

When water converts from a liquid to a gas, a quantity of heat energy known as the latent heat of vaporization is required to break the hydrogen bonds. At 100° C, 540 calories per gram of water are needed to convert one gram of liquid water to one gram of water vapour under normal pressure. Water can evaporate (vaporization) at temperatures below the boiling point, and ice can evaporate into a gas without first melting in a process called sublimation. Evaporation below 100° C and sublimation require more energy per gram than 540 calories. At 20° C about 585 calories are required to vaporize one gram of water. When water vapour condenses (condensation) back to liquid water, the latent heat of vaporization is liberated. The evaporation of water from the surface of the Earth and its condensation in the atmosphere constitute the single most important way that heat from the Earth's surface is transferred to the atmosphere. This process is the source of the power that drives hurricanes and a principal mechanism for cooling the surface of the oceans. The latent heat of vaporization of water is the highest of all common substances.

The density of a material is given in units of mass per unit volume and expressed in kilograms per cubic metre in the SI system of units. In oceanography the density of seawater has been expressed historically in grams per cubic centimetre. The density of seawater is a function of temperature, salinity, and pressure. Because oceanographers require density measurements to be accurate to the fifth decimal place, manipulation of the data requires writing many numbers to record each measurement. Also, the pressure effect can be neglected in many instances by using potential temperature. These two factors led oceanographers to adopt a density unit called sigma-t (σ_t). This value is obtained by subtracting 1.0 from the density and multiplying the remainder by 1,000. The σ_t has no units and is an abbreviated density of seawater controlled by salinity and temperature only.

During the fall (autumn) a lake is cooled at its surface, the surface water sinks, and convective overturn proceeds as the density of the surface water increases with the decreasing temperature. By the time the surface water reaches 4° C, the temperature of maximum density for fresh water, the density-driven convective overturn has reached the bottom of the lake, and overturn ceases. Further cooling of the surface produces less dense water, and the lake becomes stably stratified with regard to temperature-controlled density. Only a relatively shallow surface layer is cooled below 4° C. When this surface layer is cooled to the ice point, 0° C, ice is formed as the latent heat of fusion is extracted. In a deep lake the temperature at depth remains at 4° C. In the spring the surface water warms up and the ice melts. A shallow convective overturn resumes until the lake is once more isothermal at 4° C. Continued warming of the surface produces a stable water column.

In seawater in which the salinity exceeds 24.7⁰/₀₀, convective overturn also occurs during the cooling cycle and penetrates to a depth determined by the salinity and temperature-controlled density of the cooled water. Since no density maximum is passed, the thermally driven convective overturn is continuous until the ice point is reached where sea ice (pack ice) forms with the extraction of the latent heat of fusion. Since salt is largely excluded from the ice in most cases, the salinity of the water beneath the ice increases slightly and a convective overturn that is both salt- and temperature-driven continues as sea ice forms.

The continuing overturn requires that a large volume of water be cooled to a new ice point dictated by the salinity increase before additional ice forms. In this manner, very dense seawater that is both cold and of elevated salinity is formed. Such areas as the Weddell Sea in Antarctica produce the densest water of the oceans. This water, known as Antarctic Bottom Water, sinks to the deepest depths of the oceans. The continuing overturn slows the rate at which the sea ice forms, limiting the seasonal thickness of the ice. Other factors that control the thickness of ice are the rate at which heat is conducted through the ice layer and the insulation provided by snow on the ice. Seasonal sea ice seldom exceeds about two metres in thickness. During the warmer season, melting sea ice supplies a freshwater layer to the sea surface and thereby stabilizes the water column (see below Ice in the sea (ocean)).

Surface processes that alter the temperature and salinity of seawater drive the vertical circulation of the oceans. Known as thermohaline circulation, it continually replaces seawater at depth with water from the surface and slowly replaces surface water elsewhere with water rising from deeper depths (see below Circulation of the ocean currents: Thermohaline circulation (ocean)).

Water is transparent to the wavelengths of electromagnetic radiation that fall within the visible spectrum and is opaque to wavelengths above and below this band. However, once in the water, visible light is subject to both refraction and attenuation.

Light rays that enter the water at any angle other than a right angle are refracted (i.e., bent) because the light waves travel at a slower speed in water than they do in air. The amount of refraction, referred to as the refractive index, is affected by both the salinity and temperature of the water. The refractive index increases with increasing salinity and decreasing temperature. This relationship allows the refractive index of a sample of seawater at a constant temperature to be used to determine the salinity of the sample.

Some of the Sun's (sunlight) radiant energy is reflected at the ocean surface and does not enter the ocean. That which penetrates the water's surface is attenuated by absorption and conversion to other forms of energy, such as heat that warms or evaporates water, or is used by plants to fuel photosynthesis. Sunlight that is not absorbed can be scattered by molecules and particulates suspended in the water. Scattered light is deflected into new directional paths and may wander randomly to eventually be either absorbed or directed upward and out of the water. It is this upward scattered light and the light reflected (reflection) from particles that determine the colour of the oceans, as seen from above.

Water molecules, dissolved salts, organic substances, and suspended particulates combine to cause the intensity of available solar radiation to decrease with depth. Observations of light attenuation in ocean waters indicate that not only does the intensity of solar radiation decrease with depth but also the wavelengths present in the solar spectrum (spectroscopy) are not attenuated at the same rates. Both short wavelengths (ultraviolet) and long wavelengths (infrared) are absorbed rapidly and are not available for scattering. Only blue-green wavelengths penetrate to any depth, and because the blue-green light is most available for scattering, the oceans appear blue to the human eye. Changes in the colour of the ocean waters are caused either by the colour of the particulates in suspension and dissolved substances or by the changing quality of the solar radiation at the ocean surface as determined by the angle of the Sun and atmospheric conditions. In the clearest ocean waters only about 1 percent of the surface radiation remains at a depth of 150 metres. No sunlight penetrates below 1,000 metres.

There are many ways of measuring light attenuation in the oceans. A common method involves the use of a Secchi disk, a weighted round white disk about 30 centimetres in diameter. The Secchi disk is lowered into the ocean to the depth where it disappears from view; its reflectance equals the intensity of light backscattered from the water. This depth in metres divided into 1.7 yields an attenuation, or extinction, coefficient for available light as averaged over the Secchi disk depth. The light extinction coefficient, x, may then be used in a form of Beer's law, I_z = I₀e^xz, to estimate I_z, the intensity of light at depth z from I₀, the intensity of light at the ocean surface. This method gives no indication of the attenuation change with depth or the attenuation of specific wavelengths of light.

A photocell may be lowered into the ocean to measure light intensity at discrete depths and to determine light reduction from the surface value or from the previous depth value. The photocell may sense all available wavelengths or may be equipped with filters that pass only certain wavelengths of light. Since I_z and I₀ are known, changing light intensity values may be used in Beer's law to determine how the attenuation coefficient changes with depth and quality of light. Measurements of this type are used to determine the level of photosynthesis as a function of radiant energy level with depth and to measure changes in the turbidity of the water caused by particulate distribution with depth.

These data indicate that water is a good absorber of solar radiation.

Water is an excellent conductor of sound, considerably better than air. The attenuation of sound by absorption and conversion to other energy forms is a function of sound frequency and the properties of water.

The attenuation coefficient, x, in Beer's law, as applied to sound, where I_z and I₀ are now sound intensity values, is dependent on the viscosity of water and inversely proportional to the frequency of the sound and the density of the water. High-pitched sounds are absorbed and converted to heat faster than low-pitched sounds. Sound velocity (speed of sound) in water is determined by the square root of elasticity divided by the water's density. Because water is only slightly compressible, it has a large value of elasticity and therefore conducts sound rapidly. Since both the elasticity and density of seawater change with temperature, salinity, and pressure, so does the velocity of sound.

In the oceans the speed of sound varies between 1,450 and 1,570 metres per second. It increases about 4.5 metres per second per each degree C increase and 1.3 metres per second per each 1⁰/₀₀ increase in salinity. Increasing pressure also increases the speed of sound at the rate of about 1.7 metres per second for an increase in pressure of 100 metres in depth, which is equal to approximately 10 bars, or 10 atmospheres.

The greatest changes in temperature and salinity with depth that affect the speed of sound are found near the surface. Changes of sound speed in the horizontal are usually slight except in areas where abrupt boundaries exist between waters of different properties. The effects of salinity and temperature on sound speed are more important than the effect of pressure in the upper layers. Deeper in the ocean, salinity and temperature change less with depth, and pressure becomes the important controlling factor.

In regions of surface dilution, salinity increases with depth near the surface, while in areas of high evaporation salinity decreases with depth. Temperature usually decreases with depth and normally exerts a greater influence on sound speed than does the salinity in the surface layer of the open oceans. In the case of surface dilution, salinity and temperature effects on the speed of sound oppose each other, while in the case of evaporation they reinforce each other, causing the speed of sound to decrease with depth (depth finder). Beneath the upper oceanic layers the speed of sound increases with depth.

If a sound wave (sonic pulse) travels at a right angle to these layers, as in depth sounding, no refraction occurs; however, the speed changes continuously with depth, and an average sound speed for the entire water column must be used to determine the depth of water. Variations in the speed of sound cause sound waves to refract when they travel obliquely through layers of water that have different properties of salinity and temperature. Sound waves (transverse wave) traveling downward and moving obliquely to the water layers will bend upward when the speed of sound increases with depth and downward when the speed decreases with depth. This refraction of the sound is important in the sonar detection of submarines because the actual path of a sound wave must be known to determine a submarine's position relative to the transmitter of the sound. Refraction also produces shadow zones that sound waves do not penetrate because of their curvature.

At depths of approximately 1,000 metres, pressure becomes the important factor: it combines with temperature and salinity to produce a zone of minimum sound speed. This zone has been named the SOFAR (sound fixing and ranging) channel. If a sound is generated by a point source in the SOFAR zone, it becomes trapped by refraction. Dispersed horizontally rather than in three directions, the sound is able to travel for great distances. Hydrophones lowered to this depth many kilometres from the origin of the sound are able to detect the sound pulse. The difference in arrival time of the pulse at separate listening posts may be used to triangulate the position of the pulse source.

Formation of sea ice was briefly discussed above (see Density of seawater and pressure (ocean)). Sea ice formation is a thermal physical property of water and plays a role in driving convective overturn in the oceans. It does so by increasing the density of the seawater under the forming ice and thereby helps to drive convective overturn.

There are two types of ice in the seas: sea ice, which is ice formed by the freezing of seawater, and ice that has come from land, such as icebergs and ice islands.

From an initial stage of so-called frazil crystals (floating needles and platelets) and sludge composed of them, sea ice grows to a compact aggregate of crystals of pure ice with pockets of seawater entrapped between them. Because of this composition, the salinity of sea ice is lower than that of the seawater from which it has grown. The initial sea-ice salinity may vary between 2 and 20 parts per thousand; the more rapid the freezing, the saltier the ice, as brine can be trapped in cavities in the forming ice and become isolated from the seawater.

After sea ice has formed, a process of salt removal by drainage of part of the enclosed brine sets in, because the cells in which it is contained are not completely isolated. Old ice has very low salinity, on the order of 1 part per thousand or less.

The growth rate of sea ice depends on surface temperature, the depth of snow cover, and the heat flux in the underlying water. In the central Arctic, the thickness of an ice cover formed in one growing season is about two metres. If the ice is not broken up or melted each season, it finally reaches an equilibrium thickness of about three to four metres in five to eight years, when the annual ablation (loss by any means) at the top and the bottom equals the annual growth. In the Antarctic (Antarctica), perennial sea ice is found only in the Weddell Sea and a narrow strip around the continent. Most of the Antarctic sea ice is seasonal and reaches a thickness of about 1.5 metres by the end of October.

The high albedo (or reflectivity) of sea ice and its snow cover (80 percent, compared to 5–10 percent for liquid water), the insulation characteristics of ice and snow, and the latent heat of fusion combine to affect the heat budget of the oceans during both freezing and thawing.

The boundaries of the sea ice are highly variable. In the Norwegian (Norwegian Sea) and Greenland seas (Greenland Sea), deviations of 300 kilometres north or south of the average position are not uncommon. The estimated mean areas of sea ice at the end of the summer and at the end of the winter in the Arctic are 9 million square kilometres (3.5 million square miles) and 12 million square kilometres, respectively. In the Antarctic, the corresponding values are 4 million square kilometres and 20 million square kilometres. The mean total volume of sea ice on Earth is 40,000 to 50,000 cubic kilometres (9,600 to 12,000 cubic miles), and the total amount of freezing and melting that occurs each year has been estimated at 30,000 cubic kilometres.

In the Arctic, it is possible to distinguish three regimes of sea ice: the great inner core, the permanent polar cap of sea ice (the Arctic pack), which covers about six million square kilometres; around this the true drift ice or pack ice; and the landfast ice, which is present during nine months of the year, when it fringes the shores of the Arctic Ocean out to the 22-metre depth line. Large amounts of pack ice drift southward each year. The ice discharge through the gap between Greenland and Spitsbergen is estimated to be 3,000 cubic kilometres per year. On the west side of the North Atlantic, the pack ice reaches approximately latitude 45° N in winter and spring. On the east side, along the Norwegian coast, the sea remains open up to 73° N.

Ice islands, of which a number have been found drifting in Arctic waters, are heavy sheets of ice that are far thicker than sea ice. Their thickness may amount to 50 metres, 5 metres of which project above water. The surface area of the largest known ice island is about 1,000 square kilometres; others are far smaller. Ice islands consist of a kind of glacierlike snow ice. The majority probably have been formed by the breaking of the shelf ice that borders the north coast of Ellesmere Island. The first ice island reported has undergone little change in configuration since its detection in 1946.

Icebergs (iceberg) are formed by the calving (detaching of parts) of glaciers or of inland ice that reaches the sea. The main sources of icebergs in the northern seas are the valley glaciers of Greenland, which produce some 12,000 to 15,000 sizable icebergs annually. Almost as many are calved by the glaciers reaching the sea on the eastern seaboard as by those on the west coast, but the icebergs deriving from the east side do not travel much farther south than Cape Farvel, the southern tip of Greenland. The icebergs of the west coast, on the other hand, after traveling northward and across to the other side of Baffin Bay, are carried far south, along Baffin Island and Labrador, by the Labrador Current. It is estimated that about 1 in every 20 icebergs derived from west Greenland ends up south of Newfoundland (48° N), the greatest numbers arriving there in April, May, and June.

The icebergs of the Antarctic derive from an ice barrier, or shelf ice, a layer of ice that stretches out from the inland ice into the ocean. It rests on the bottom near shore, but farther out to sea it floats on the water. Because of their origin, the Antarctic icebergs are much longer than they are high, occasionally measuring some tens of kilometres in length. For this reason they are called table bergs.

The frequency with which icebergs occur in the Southern Ocean does not vary much with the season in contrast to the North Atlantic occurrences. Generally speaking, October and November are the months in which they are most numerous in the south because of the release of the bergs from the pack ice in the southern spring. They reach farthest north from November to February. The average northern boundary for icebergs is about 40° S in the Atlantic Ocean, between 40° and 50° S in the Indian Ocean, and about 50° S in the Pacific. At least several thousands of them are adrift every year in the southern seas.

The general circulation of the oceans defines the average movement of seawater, which, like the atmosphere, follows a specific pattern. Superimposed on this pattern are oscillations of tides and waves, which are not considered part of the general circulation. There also are meanders and eddies that represent temporal variations of the general circulation. The ocean circulation pattern exchanges water of varying characteristics, such as temperature and salinity, within the interconnected network of oceans and is an important part of the heat and freshwater fluxes of the global climate. Horizontal movements are called currents, which range in magnitude from a few centimetres per second to as much as 4 metres per second. A characteristic surface speed is about 5 to 50 centimetres per second. Currents diminish in intensity with increasing depth. Vertical movements, often referred to as upwelling and downwelling, exhibit much lower speeds, amounting to only a few metres per month. As seawater is nearly incompressible, vertical movements are associated with regions of convergence and divergence in the horizontal flow patterns.

Ocean circulation derives its energy at the sea surface from two sources that define two circulation types: (1) wind-driven circulation forced by wind stress on the sea surface, inducing a momentum exchange, and (2) thermohaline circulation driven by the variations in water density imposed at the sea surface by exchange of ocean heat and water with the atmosphere, inducing a buoyancy exchange. These two circulation types are not fully independent, since the sea-air buoyancy and momentum exchange are dependent on wind speed. The wind-driven circulation is the more vigorous of the two and is configured as large gyres that dominate an ocean region. The wind-driven circulation is strongest in the surface layer. The thermohaline circulation is more sluggish, with a typical speed of one centimetre per second, but this flow extends to the seafloor and forms circulation patterns that envelop the global ocean.

Maps of the general circulation at the sea surface are constructed from a vast amount of data obtained from inspecting the residual drift of ships after course direction and speed are accounted for in a process called dead reckoning. This information is amplified by satellite-tracked drifters at sea. The pattern is nearly entirely that of wind-driven circulation.

Deep-ocean circulation consists mainly of thermohaline circulation. The currents are inferred from the distribution of seawater properties, which trace the spreading of specific water masses. The distribution of density or field of mass is also used to estimate the deep currents. Direct observations of subsurface currents are made by deploying current meters from bottom-anchored moorings and by setting out neutral buoyant instruments whose drift at depth is tracked acoustically.

The general circulation is governed by the equation of motion (motion, equation of), one of Sir Isaac Newton's fundamental laws of mechanics applied to a continuous volume of water. This equation states that the product of mass and current acceleration equals the vector sum of all forces that act on the mass. Besides gravity, the most important forces that cause and affect ocean currents are horizontal pressure-gradient forces, Coriolis forces, and frictional forces. Temporal and inertial terms are generally of secondary importance to the general flow, though they become important for transient features of meanders and eddies.

The hydrostatic pressure, p, at any depth below the sea surface is given by the equation p = gρz, where g is the acceleration of gravity, ρ is the density of seawater, which increases with depth, and z is the depth below the sea surface. This is called the hydrostatic equation, which is a good approximation for the equation of motion for forces acting along the vertical. Horizontal differences in density (due to variations of temperature and salinity) measured along a specific depth cause the hydrostatic pressure to vary along a horizontal plane or geopotential surface, a surface perpendicular to the direction of the gravity acceleration. Horizontal gradients of pressure, though much smaller than vertical changes in pressure, give rise to ocean currents.

In a homogeneous ocean, which would have a constant potential density, horizontal pressure differences are possible only if the sea surface is tilted. In this case, surfaces of equal pressure, called isobaric surfaces, are tilted in the deeper layers by the same amount as the sea surface. This is referred to as the barotropic field of mass. The unchanged pressure gradient gives rise to a current speed independent of depth. The oceans of the world, however, are not homogeneous. Horizontal variations in temperature and salinity cause the horizontal pressure gradient to vary with depth. This is the baroclinic field of mass, which leads to currents that vary with depth. The horizontal pressure gradient in the ocean is a combination of these two mass fields.

The tilt, or topographic relief, of the isobaric surface marking sea surface (defined as p = 0) can be constructed from a three-dimensional density distribution using the hydrostatic equation. Since the absolute value of pressure is not known at any depth in the ocean, the sea surface slope is presented relative to that of a deep isobaric surface; it is assumed that the deep isobaric surface is level. Since the wind-driven circulation attenuates with increasing depth, an associated decrease of isobaric tilt with increasing depth is expected. Representation of the sea surface relief relative to a deep reference surface is a good representation of the absolute shape of the sea surface. The total relief of the sea surface amounts to about two metres, with “hills” in the subtropics and “valleys” in the polar regions. This pressure head drives the surface circulation.

The rotation of the Earth about its axis causes moving particles to behave in a way that can only be understood by adding a rotational dependent force. To an observer in space, a moving body would continue to move in a straight line unless the motion were acted upon by some other force. To an Earth-bound observer, however, this motion cannot be along a straight line because the reference frame is the rotating Earth. This is similar to the effect that would be experienced by an observer standing on a large turntable if an object moved over the turntable in a straight line relative to the “outside” world. An apparent deflection of the path of the moving object would be seen. If the turntable rotated counterclockwise, the apparent deflection would be to the right of the direction of the moving object, relative to the observer fixed on the turntable. This remarkable effect is evident in the behaviour of ocean currents. It is called the Coriolis force, named after Gustave-Gaspard Coriolis, a 19th-century French engineer and mathematician. For the Earth, horizontal deflections due to the rotational induced Coriolis force act on particles moving in any horizontal direction. There also are apparent vertical forces, but these are of minor importance to ocean currents. Because the Earth rotates from west to east about its axis, an observer in the Northern Hemisphere would notice a deflection of a moving body toward the right. In the Southern Hemisphere, this deflection would be toward the left. At the equator there would be no apparent horizontal deflection.

It can be shown that the Coriolis force always acts perpendicular to motion. Its horizontal component, C_f, is proportional to the sine of the geographic latitude (θ, given as a positive value for the Northern Hemisphere and a negative value for the Southern Hemisphere) and the speed, c, of the moving body. It is given by C_f = c (2ω sin θ), where ω = 7.29 × 10⁻⁵ radian per second is the angular velocity of the Earth's rotation.

Movement of water through the oceans is slowed by friction, with surrounding fluid moving at a different velocity. A faster-moving fluid layer tends to drag along a slower-moving layer, and a slower-moving layer will tend to reduce the speed of a faster-moving layer. This momentum transfer between the layers is referred to as frictional forces. The momentum transfer is a product of turbulence that moves kinetic energy to smaller scales until at the centimetre scale it is dissipated as heat. The wind blowing over the sea surface transfers momentum to the water. This frictional force at the sea surface (i.e., the wind stress) produces the wind-driven circulation. Currents moving along the ocean floor and the sides of the ocean also are subject to the influence of boundary-layer (boundary layer) friction. The motionless ocean floor removes momentum from the circulation of the ocean waters.

For most of the ocean volume away from the boundary layers, which have a characteristic thickness of 100 metres, frictional forces are of minor importance, and the equation of motion for horizontal forces can be expressed as a simple balance of horizontal pressure gradient and Coriolis force. This is called geostrophic balance.

On a nonrotating Earth, water would be accelerated by a horizontal pressure gradient and would flow from high to low pressure. On the rotating Earth, however, the Coriolis force deflects the motion, and the acceleration ceases only when the speed, c, of the current is just fast enough to produce a Coriolis force that can exactly balance the horizontal pressure-gradient force. This geostrophic balance is given as dp/dn = ρc2ω sin θ, where dp/dn is the horizontal pressure gradient. From this balance, it follows that the current direction must be perpendicular to the pressure gradient because the Coriolis force always acts perpendicular to the motion. In the Northern Hemisphere this direction is such that the high pressure is to the right when looking in current direction, while in the Southern Hemisphere it is to the left. This type of current is called a geostrophic current. The simple equation given above provides the basis for an indirect method of computing ocean currents. The relief of the sea surface also defines the streamlines (paths) of the geostrophic current at the surface relative to the deep reference level. The hills represent high pressure, and the valleys stand for low pressure. Clockwise rotation in the Northern Hemisphere with higher pressure in the centre of rotation is called anticyclonic motion. Counterclockwise rotation with lower pressure in its centre is cyclonic motion. In the Southern Hemisphere the sense of rotation is the opposite, because the effect of the Coriolis force has changed its sign of deflection.

The wind exerts stress on the ocean surface proportional to the square of the wind speed and in the direction of the wind, setting the surface water in motion. This motion extends to a depth of about 100 metres in what is called the Ekman layer, after the Swedish oceanographer V. Walfrid Ekman, who in 1902 deduced these results in a theoretical model constructed to help explain observations of wind drift in the Arctic. Within the oceanic Ekman layer the wind stress is balanced by the Coriolis force and frictional forces. The surface water is directed at an angle of 45° to the wind, to the right in the Northern Hemisphere and to the left in the Southern Hemisphere. With increasing depth in the boundary layer, the current speed is reduced, and the direction rotates farther away from the wind direction following a spiral form, becoming antiparallel to the surface flow at the base of the layer where the speed is ¹/₂₃ of the surface speed. This so-called Ekman spiral may be the exception rather than the rule, as the specific conditions are not often met, though deflection of a wind-driven surface current at somewhat smaller than 45° is observed when the wind field blows with a steady force and direction for the better part of a day. The average water particle within the Ekman layer moves at an angle of 90° to the wind; this movement is to the right of the wind direction in the Northern Hemisphere and to its left in the Southern Hemisphere. This phenomenon is called Ekman transport, and its effects are widely observed in the oceans.

Since the wind varies from place to place, so does the Ekman transport, forming convergence (convergence and divergence) and divergence zones of surface water. A region of convergence forces surface water downward in a process called downwelling, while a region of divergence draws water from below into the surface Ekman layer in a process known as upwelling. Upwelling and downwelling also occur where the wind blows parallel to a coastline. The principal upwelling regions of the world are along the eastern boundary of the subtropical ocean waters, as, for example, the coastal region of Peru and northwestern Africa. Upwelling in these regions cools the surface water and brings nutrient-rich subsurface water into the sunlit layer of the ocean, resulting in a biologically productive region. Upwelling and high productivity also are found along divergence zones at the equator and around Antarctica. The primary downwelling regions are in the subtropical ocean waters—e.g., the Sargasso Sea in the North Atlantic. Such areas are devoid of nutrients and are poor in marine life.

The vertical movements of ocean waters into or out of the base of the Ekman layer amount to less than one metre per day, but they are important since they extend the wind-driven effects into deeper waters. Within an upwelling region, the water column below the Ekman layer is drawn upward. This process, with conservation of angular momentum on the rotating Earth, induces the water column to drift toward the poles. Conversely, downwelling forces water into the water column below the Ekman layer, inducing drift toward the equator. An additional consequence of upwelling and downwelling for stratified waters is to create a baroclinic field of mass (see above). Surface water is less dense than deeper water. Ekman convergences have the effect of accumulating less dense surface water. This water floats above the surrounding water, forming a hill in sea level and driving an anticyclonic geostrophic current that extends well below the Ekman layer. Divergences do the opposite; they remove the less dense surface water, replacing it with denser, deeper water. This induces a depression in sea level with a cyclonic geostrophic current.

The ocean current pattern produced by the wind-induced Ekman transport is called the Sverdrup transport, after the Norwegian oceanographer H.U. Sverdrup, who formulated the basic theory in 1947. Several years later (1950), the American geophysicist and oceanographer Walter H. Munk and others expanded Sverdrup's work, explaining many of the major features of the wind-driven general circulation by using the mean climatological wind stress distribution at the sea surface as a driving force.

Wind stress induces a circulation pattern that is similar for each ocean. In each case, the wind-driven circulation is divided into large gyres that stretch across the entire ocean: subtropical gyres extend from the equatorial current system to the maximum westerlies in a wind field near 50° latitude, and subpolar gyres extend poleward of the maximum westerlies (see below). The depth penetration of the wind-driven currents depends on the intensity of ocean stratification: for those regions of strong stratification, such as the tropics, the surface currents extend to a depth of less than 1,000 metres. Within the low-stratification polar regions, the wind-driven circulation reaches all the way to the seafloor.

At the equator the currents are for the most part directed toward the west, the North Equatorial Current in the Northern Hemisphere and the South Equatorial Current in the Southern Hemisphere. Near the thermal equator, where the warmest surface water is found, there occurs the eastward-flowing Equatorial Counter Current (equatorial countercurrent). This current is slightly north of the geographic equator, drawing the northern fringe of the South Equatorial Current to 5° Ν. Τhe offset to the Northern Hemisphere matches a similar offset in the wind field. Τhe east-to-west wind across the tropical ocean waters induces Ekman transport divergence at the equator, which cools the surface water there.

At the geographic equator a jetlike current is found just below the sea surface, flowing toward the east counter to the surface current. This is called the Equatorial Undercurrent. It attains speeds of more than 1 metre per second at a depth of nearly 100 metres. It is driven by higher sea level in the western margins of the tropical ocean, producing a pressure gradient, which in the absence of a horizontal Coriolis force drives a west-to-east current along the equator. The wind field reverses the flow within the surface layer, inducing the Equatorial Undercurrent.

Equatorial circulation undergoes variations following the irregular periods of roughly three to eight years of the Southern Oscillation (i.e., fluctuations of atmospheric pressure over the tropical Indo-Pacific region). Weakening of the east-to-west wind during a phase of the Southern Oscillation allows warm water in the western margin to slip back to the east by increasing the flow of the Equatorial Counter Current. Surface water temperatures and sea level decrease in the west and increase in the east. This event is called El Niño. The combined El Niño/Southern Oscillation effect has received much attention because it is associated with global-scale climatic variability (see below Impact of ocean-atmosphere interactions on weather and climate: El Niño/ Southern Oscillation and climatic change (ocean)). In the tropical Indian Ocean, the strong seasonal winds of the monsoons induce a similarly strong seasonal circulation pattern.

These are anticyclonic circulation features. The Ekman transport within these gyres forces surface water to sink, giving rise to the subtropical convergence near 20°–30° latitude. The centre of the subtropical gyre is shifted to the west. This westward intensification of ocean currents was explained by the American meteorologist and oceanographer Henry M. Stommel (Stommel, Henry Melson) (1948) as resulting from the fact that the horizontal Coriolis force increases with latitude. This causes the poleward-flowing western boundary current to be a jetlike current that attains speeds of two to four metres per second. This current transports the excess heat of the low latitudes to higher latitudes. The flow within the equatorward-flowing interior and eastern boundary of the subtropical gyres is quite different. It is more of a slow drift of cooler water that rarely exceeds 10 centimetres per second. Associated with these currents is coastal upwelling that results from offshore Ekman transport.

The strongest of the western boundary currents is the Gulf Stream in the North Atlantic Ocean. It carries about 30 million cubic metres of ocean water per second through the Straits of Florida and roughly 80 million cubic metres per second as it flows past Cape Hatteras off the coast of North Carolina, U.S. Responding to the large-scale wind field over the North Atlantic, the Gulf Stream separates from the continental margin at Cape Hatteras. After separation, it forms waves or meanders that eventually generate many eddies (eddy) of warm and cold water. The warm eddies, composed of thermocline water normally found south of the Gulf Stream, are injected into the waters of the continental slope off the coast of the northeastern United States. They drift to the southeast at rates of approximately five to eight centimetres per second, and after a year they rejoin the Gulf Stream north of Cape Hatteras. Cold eddies of slope water are injected into the region south of the Gulf Stream and drift to the southwest. After two years they reenter the Gulf Stream just north of the Antilles Islands. The path that they follow defines a clockwise-flowing recirculation gyre seaward of the Gulf Stream. (For additional details on the Gulf Stream, see below Impact of ocean-atmosphere interactions on weather and climate: The Gulf Stream and Kuroshio systems (ocean).)

Among the other western boundary currents, the Kuroshio of the North Pacific is perhaps the most like the Gulf Stream, having a similar transport and array of eddies. The Brazil and East Australian currents are relatively weak. The Agulhas Current has a transport close to that of the Gulf Stream. It remains in contact with the margin of Africa around the southern rim of the continent. It then separates from the margin and curls back to the Indian Ocean in what is called the Agulhas Retroflection. Not all the water carried by the Agulhas returns to the east; about 10 to 20 percent is injected into the South Atlantic Ocean as large eddies that slowly migrate across it.

The subpolar gyres are cyclonic circulation features. The Ekman transport within these features forces upwelling and surface water divergence. In the North Atlantic the subpolar gyre consists of the North Atlantic Current at its equatorward side and the Norwegian Current that carries relatively warm water northward along the coast of Norway. The heat released from the Norwegian Current into the atmosphere maintains a moderate climate in northern Europe. Along the east coast of Greenland is the southward-flowing cold East Greenland Current. It loops around the southern tip of Greenland and continues flowing into the Labrador Sea. The southward flow that continues off the coast of Canada is called the Labrador Current. This current separates for the most part from the coast near Newfoundland to complete the subpolar gyre of the North Atlantic. Some of the cold water of the Labrador Current, however, extends farther south.

In the North Pacific the subpolar gyre is composed of the northward-flowing Alaska Current, the Aleutian Current (also known as the Subarctic Current), and the southward-flowing cold Oyashio Current (Oya Current). The North Pacific Current forms the separation between the subpolar and subtropical gyres of the North Pacific.

In the Southern Hemisphere, the subpolar gyres are less defined. Large cyclonic flowing gyres lie poleward of the Antarctic Circumpolar Current and can be considered counterparts to the Northern Hemispheric subpolar gyres. The best-formed is the Weddell Gyre of the South Atlantic sector of the Southern Ocean (see above). The Antarctic coastal current flows toward the west. The northward-flowing current off the east coast of the Antarctic Peninsula carries cold Antarctic coastal water into the circumpolar belt. Another cyclonic gyre occurs north of the Ross Sea.

The Southern Ocean links the major oceans by a deep circumpolar belt in the 50°–60° S range. In this belt flows the Antarctic Circumpolar Current from west to east, encircling the globe at high latitudes. It transports 125 million cubic metres of seawater per second over a path of about 24,000 kilometres and is the most important factor in diminishing the differences between oceans. The Antarctic Circumpolar Current is not a well-defined single-axis current but rather consists of a series of individual filaments separated by frontal zones. It reaches the seafloor and is guided along its course by the irregular bottom topography. Large meanders and eddies develop in the current as it flows. These features induce poleward transfer of heat, which may be significant in balancing the oceanic heat loss to the atmosphere above the Antarctic region farther south.

The general circulation of the oceans consists primarily of the wind-driven currents. These, however, are superimposed on the much more sluggish circulation driven by horizontal differences in temperature and salinity—namely, the thermohaline circulation. The thermohaline circulation reaches down to the seafloor and is often referred to as the deep, or abyssal, ocean circulation. Measuring seawater temperature and salinity distribution is the chief method of studying the deep-flow patterns. Other properties also are examined; for example, the concentrations of oxygen, carbon-14, and such synthetically produced compounds as chlorofluorocarbons are measured to obtain resident times and spreading rates of deep water.

In some areas of the ocean, generally during the winter season, cooling or net evaporation causes surface water to become dense enough to sink. Convection penetrates to a level where the density of the sinking water matches that of the surrounding water. It then spreads slowly into the rest of the ocean. Other water must replace the surface water that sinks. This sets up the thermohaline circulation. The basic thermohaline circulation is one of sinking of cold water in the polar regions, chiefly in the northern North Atlantic and near Antarctica. These dense water masses spread into the full extent of the ocean and gradually upwell to feed a slow return flow to the sinking regions. A theory for the thermohaline circulation pattern was proposed by Stommel and Arnold Arons in 1960.

In the Northern Hemisphere, the primary region of deep water formation is the North Atlantic; minor amounts of deep water are formed in the Red Sea and Persian Gulf. A variety of water types contribute to the so-called North Atlantic Deep Water. Each one of them differs, though they share a common attribute of being relatively warm (greater than 2° C) and salty (greater than 34.9 parts per thousand) compared with the other major producer of deep and bottom water, the Southern Ocean (0° C and 34.7 parts per thousand). North Atlantic Deep Water is primarily formed in the Greenland (Greenland Sea) and Norwegian seas (Norwegian Sea), where cooling of the salty water introduced by the Norwegian Current induces sinking. This water spills over the rim of the ridge that stretches from Greenland to Scotland, extending to the seafloor to the south as a convective plume. It then flows southward, pressed against the western edge of the North Atlantic. Additional deep water is formed in the Labrador Sea. This water, somewhat less dense than the overflow water from the Greenland and Norwegian seas, has been observed sinking to a depth of 3,000 metres within convective features referred to as chimneys. Vertical velocities as high as 10 centimetres per second have been observed within these convective features. A third variety of North Atlantic Deep Water is derived from net evaporation within the Mediterranean Sea. This draws surface water into the Mediterranean through the Strait of Gibraltar. The mass of salty water formed within the Mediterranean exits as a deeper stream. It descends to depths of 1,000 to 2,000 metres in the North Atlantic Ocean, forming the uppermost layer of North Atlantic Deep Water. The outflow in the Strait of Gibraltar reaches as high as 2 metres per second, but its total transport amounts to only 5 percent of the total North Atlantic Deep Water formed. The outflow of the Mediterranean plays a significant role in boosting the salinity of North Atlantic Deep Water.

The blend of North Atlantic Deep Water, with a total formation rate of 15 to 20 × 10⁶ cubic metres per second, quickly ventilates the Atlantic Ocean, resulting in a residence time of less than 200 years. The deep water spreads away from its source along the western side of the Atlantic Ocean and, on reaching the Antarctic Circumpolar Current, spreads into the Indian and Pacific oceans. The sinking of North Atlantic Deep Water is compensated for by the slow upwelling of deep water, mainly in the Southern Ocean, to replenish the upper stratum of water that has descended as North Atlantic Deep Water. North Atlantic Deep Water exported to the other oceans must be balanced by the inflow of upper-layer water into the Atlantic. Some water returns as cold, low-salinity Pacific water through the Drake Passage in the form of what is known as Antarctic Intermediate Water, and some returns as warm salty thermocline water from the Indian Ocean around the southern rim of Africa.

Remnants of North Atlantic Deep Water mix with Southern Ocean water to spread along the seafloor into the North Pacific Ocean. Here, it upwells to a level of 2,000–3,000 metres and returns to the south lower in salinity and oxygen but higher in nutrient concentrations as North Pacific Deep Water. This North Pacific Deep Water is eventually swept eastward with the Antarctic Circumpolar Current. Modification of deep water in the North Pacific is the direct consequence of vertical mixing, which carries into the deep ocean the low salinity properties of North Pacific Intermediate Water. The latter is formed in the northwestern Pacific Ocean. Because of the immenseness of the North Pacific and the extremely long residence time (more than 500 years) of the water, enormous quantities of North Pacific Deep Water can be produced by vertical mixing.

Considerable volumes of cold water generally of low salinity are formed in the Southern Ocean. Such water masses spread into the interior of the global ocean and to a large extent are responsible for the anomalous cold, low-salinity state of the modern oceans. The circumstances leading to this role for the Southern Ocean are related to the existence of a deep-ocean circumpolar belt around Antarctica that was established some 25 million years ago by the shifting lithospheric plates which make up the Earth's surface (see below Ocean basins (ocean)). This belt establishes the Antarctic Circumpolar Current, which isolates Antarctica from the warm surface waters of the subtropics. The Antarctic Circumpolar Current does not completely sever contact with the lower latitudes. The Southern Ocean does have access to the waters of the north, but through deep- and bottom-water pathways. The basic dynamics of the Antarctic Circumpolar Current lifts dense deep water occurring north of the current to the ocean surface south of it. Once exposed to the cold Antarctic air masses, the upwelling deep water is converted to the cold Antarctic Bottom Water and Antarctic Intermediate Water. The southward and upwelling deep water, which carries heat injected into the deep ocean by processes farther north, is balanced by the northward spread of cooler, fresher, oxygenated water masses of the Southern Ocean. It is estimated that the overturning rate of water south of the Antarctic Circumpolar Current amounts to 35 to 45 million cubic metres per second, most of which becomes Antarctic Bottom Water.

The primary site of Antarctic Bottom Water formation is within the continental margins of the Weddell Sea, though some is produced in other coastal regions, such as the Ross Sea. Also, there is evidence of deep convective overturning farther offshore. Antarctic Bottom Water, formed at a rate of 30 million cubic metres per second, slips below the Antarctic Circumpolar Current and spreads to regions well north of the equator. Slowly upwelling and modified by mixing with less dense water, it returns to the Southern Ocean as deep water.

The remaining upwelling of deep water spreads near the surface to the north, where it forms Antarctic Intermediate Water within the Antarctic Circumpolar Current zone and spreads along the base of the thermoclines farther north. This water mass forms a sheet of low-salinity water that demarcates the lower boundary of the subtropical thermocline. It upwells into the thermocline, partly compensating for the sinking of North Atlantic Deep Water.

There are many types of ocean waves. Waves differ from each other in size and in terms of the forces that drive them. Waves represent an oscillatory motion of seawater at regular time intervals or periods. Some may be running, or progressive, waves in which the crests propagate, while others are stationary, or standing, waves. Two of the more common types of waves, gravity waves and tides, are considered here. For gravity waves, the stabilizing force—i.e., the force that attempts to restore the crests and troughs of the waves to the average sea level—is the Earth's gravity. The distance between the crests, or wavelength, of gravity waves range from a few centimetres to many kilometres. Tiny waves at the ocean surface with a wavelength of less than 1.7 centimetres are called capillary waves (capillary wave). Their restoring force is the surface tension of seawater. Capillary waves are direct products of the wind stress exerted on the sea surface and tend to feed wind energy into gravity waves, which characteristically have longer wavelengths.

Of the nontidal kinds of running surface waves, three types may be distinguished: wind waves and swell, wind surges, and sea waves of seismic origin (tsunamis).

Wind waves are the wind-generated gravity waves. After the wind has abated or shifted or the waves have migrated away from the wind field, such waves continue to propagate as swell.

The theory of waves starts with the concept of simple waves, those forming a strictly periodic pattern with one wavelength and one wave period and propagating in one direction. Real waves, however, always have a more irregular appearance. They may be described as composite waves, in which a whole spectrum of wavelengths, or periods, is present and which have more or less diverging directions of propagation. In reporting observed wave heights and periods (or lengths) or in forecasting them, one height or one period is mentioned as the height or period, however, and some agreement is needed in order to guarantee uniformity of meaning. The height of simple waves means the elevation difference between the top of a crest and the bottom of a trough. The significant height, a characteristic height of irregular waves, is by convention the average of the highest one-third of the observed wave heights. Period, or wavelength, can be determined from the average of a number of observed time intervals between the passing of successive well-developed wave crests over a certain point, or of observed distances between them.

Wave period and wavelength are coupled by a simple relationship: wavelength equals (wave velocity) wave period times wave speed, or L = TC, when L is wavelength, T is wave period, and C is wave speed.

The wave speed of surface gravity waves depends on the depth of water and on the wavelength, or period; the speed increases with increasing depth and increasing wavelength, or period. If the water is sufficiently deep, the wave speed is independent of water depth. This relationship of wave speed to wavelength and water depth (d) is given by the equations below. With g being the gravity acceleration (9.8 metres per second squared), C² = gd, when the wavelength is 20 times greater than the water depth (waves of this kind are called long gravity waves or shallow-water waves); and C² = gL/2π, when the wavelength is less than two times the water depth (such waves are called short waves or deep water waves). For waves with lengths between 2 and 20 times the water depth, the wave speed is governed by a more complicated equation combining these effects:

where tanh is the hyperbolic tangent.

A few examples are listed below for short waves, giving the period in seconds, the wavelength in metres, and wave speed in metres per second:

Waves often appear in groups as the result of interference of wave trains of slightly differing wavelengths. A wave group as a whole has a group speed that generally is less than the speed of propagation of the individual waves; the two speeds are equal only for groups composed of long waves. For deepwater waves, the group velocity (V ) is half the wave speed (C). In the physical sense, group velocity is the velocity of propagation of wave energy. From the dynamics of the waves, it follows that the wave energy per unit area of the sea surface is proportional to the square of the wave height, except for the very last stage of waves running into shallow water, shortly before they become breakers.

The height of wind waves increases with increasing wind speed and with increasing duration and fetch of the wind (i.e., the distance over which the wind blows). Together with height, the dominant wavelength also increases. Finally, however, the waves reach a state of saturation because they attain the maximum significant height to which the wind can raise them, even if duration and fetch are unlimited. For instance, winds of 5 metres per second, 15 metres per second, and 25 metres per second may raise waves with significant heights up to 0.5 metre, 4.5 metres, and 12.5 metres, respectively, with corresponding wavelengths of 16 metres, 140 metres, and 400 metres, respectively.

After becoming swell, the waves may travel thousands of kilometres over the ocean, particularly if the swell is from the large storms of moderate and high latitudes, whence it easily may travel into the subtropical and equatorial zones, and the swell of the trade winds, which runs into the equatorial calms. In traveling, the swell waves gradually become lower; energy is lost by internal friction and air resistance and by energy dissipation because of some divergence of the directions of propagation (fanning out). With respect to the energy loss, there is a selective damping of the composite waves, the shorter waves of the wave mixture suffering a stronger damping over a given distance than the longer ones. As a consequence, the dominant wavelength of the spectrum shifts toward the greater wavelengths. Therefore, an old swell must always be a long swell.

When waves run into shallow water, their speed of propagation and wavelength decrease, but the period remains the same. Eventually, the group velocity, the velocity of energy propagation, also decreases, and this decrease causes the height to increase. The latter effect may, however, be affected by refraction of the waves, a swerving of the wave crests toward the depth lines and a corresponding deviation of the direction of propagation. Refraction may cause a convergence or divergence of the energy stream and result in a raising or lowering of the waves, especially over nearshore elevations or depressions of the sea bottom.

In the final stage, the shape of the waves changes, and the crests become narrower and steeper until, finally, the waves become breakers (surf). Generally, this occurs where the depth is 1.3 times the wave height.

Running wind surges are long waves caused by a piling up of the water over a large area through the action of a traveling wind or pressure field. Examples include the surge in front of a traveling storm cyclone, particularly the devastating hurricane surge caused by a tropical cyclone, and the surge occasionally caused by a wind convergence line, such as a traveling front with a sharp wind shift.

A tsunami (Japanese: tsu, “harbour,” and nami, “sea”) is a very long wave of seismic origin that is caused by a submarine or coastal earthquake, landslide, or volcanic eruption. Such a wave may have a length of hundreds of kilometres and a period on the order of a quarter of an hour. It travels across the ocean at a tremendous speed. (Tsunamis are long waves traveling at the wave speed given by C² = gd.) To a depth of 4,000 metres, for instance, the corresponding wave speed is about 200 metres per second, or 720 kilometres per hour. In the open ocean the height of tsunamis may be less than one metre, and they pass unnoticed. As they approach a continental shelf, however, their speed is reduced and their height increases dramatically. Tsunamis have caused enormous destruction of life and property, piling up in coastal waters at places thousands of kilometres away from their point of origin, particularly in the Pacific Ocean.

A freestanding wave may arise in an enclosed or nearly enclosed basin as a free swinging or sloshing of the whole water mass. Such a standing wave is also called a seiche, after the name given to the oscillating movements of the water of Lake Geneva, Switz., where this phenomenon first was studied seriously. The period of oscillation is independent of the force that first brought the water mass out of equilibrium (and that is supposed to have ceased thereafter), but depends only on the dimensions of the enclosing basin and on the direction in which the water mass is swinging. Assuming a simple rectangular basin of constant depth and the most simple lengthwise oscillation, the period of oscillation (T) is equal to two times the length of the basin divided by the wave speed computed from the shallow-water formula above. This relationship may be written: T = L/C, in which L equals two times the length of the basin and C is the wave speed found from the formula, using the known depth of the basin. Besides this fundamental tone (or response to stimuli), the water mass also may swing according to an overtone, showing one or more nodal lines across the basin.

The water in an open bay or marginal sea also may perform such a free oscillation as a standing wave, the difference being that in an open bay the greatest horizontal displacements are not in the middle of the bay but at the mouth. For the fundamental period of oscillation, the formula given above is used with a wavelength equal to four times the length (from the mouth to the closed end) of the bay. In practice, of course, it is more difficult than that, because the form of a bay or marginal sea is irregular and the depth differs from place to place. The North Sea has a period of lengthwise swinging of about 36 hours. The cause of such free oscillations may be a temporary wind or pressure field, which brought the sea surface out of its horizontal position and which afterward ceased to act more or less abruptly, leaving the water mass out of equilibrium.

Gravity waves also occur on internal “surfaces” within oceans. These surfaces represent strata of rapidly changing water density with increasing depth, and the associated waves are called internal waves. Internal waves manifest themselves by a regular rising and sinking of the water layers around which they centre, whereas the height of the sea surface is hardly affected at all. Because the restoring force, excited by the internal deformation of the water layers of equal density, is much smaller than in the case of surface waves, internal waves are much slower than the latter. Given the same wavelength, the period is much longer (the movements of the water particles being much more sluggish), and the speed of propagation is much smaller; the formulas for the speed of surface waves include the acceleration of gravity, g, but those for internal waves include the gravity factor times the difference between the densities of the upper and the lower water layer divided by their sum.

The cause of internal waves may lie in the action of tidal forces (the period then equaling the tidal period) or in the action of a wind or pressure fluctuation. Sometimes, a ship may cause internal waves (dead water) if there is a shallow, brackish upper layer.

The tides may be regarded as forced waves, partially running waves and partially standing waves. They are manifested by vertical movements of the sea surface (the height maximum and minimum are called high water 【HW】 and low water 【LW】) and in alternating horizontal movements of the water, the tidal currents. The words ebb and flow are used to designate the falling tide and the rising tide, respectively.

The forces that cause the tides are called the tide-generating forces. A tide-generating force is the resultant force of the attracting force of the Moon or the Sun and the force of inertia (centrifugal force) that results from the orbital movement of the Earth around the common centre of gravity of the Earth- Moon or Earth-Sun system.

Considering the Earth-Moon system, at any time the tide-generating force is directed vertically upward at the two places on the Earth where the Moon is in the vertical (on the same and on the opposite side of the Earth); it is directed vertically downward at all places (forming a circle) where the Moon is in the horizon at that moment. At all other places, the tide-generating force also has a horizontal component. Because this pattern of forces is coupled to the position of the Moon with respect to the Earth and because for any place on the Earth's surface the relative position of the Moon with respect to that place has, on the average, a periodicity of 24 hours 50 minutes, the tide-generating force felt at any place has that same periodicity. When the Moon is in the plane of the equator, the force runs through two identical cycles within this time interval because of the symmetry of the global pattern of forces described above. Consequently, the tidal period is 12 hours 25 minutes in this case; it is the period of the semidiurnal lunar tide. The fact that the Moon is alternately to the north and to the south of the equator causes an inequality of the two successive cycles within the time interval of 24 hours 50 minutes. The effect of this inequality is formally described as the superposition of a partial tide called the diurnal lunar tide, with the period of 24 hours 50 minutes, on the semidiurnal lunar tide.

In the same manner, the Sun causes a semidiurnal solar tide, with a 12-hour period, and a diurnal solar tide, with a 24-hour period. In a complete description of the local variations of the tidal forces, still other partial tides play a role because of further inequalities in the orbital motions of the Moon and the Earth.

The interference of the solar-tidal forces with the lunar-tidal forces (the lunar forces are about 2.2 times as strong) causes the regular variation of the tidal range between spring tide, when it has its maximum, and neap tide, when it has its minimum.

Although the tide-generating forces are very small in comparison with the Earth's force of gravity (the lunar tidal force at its maximum being only 1.14 × 10^-7 times the force of gravity), their effects upon the sea are considerable because of their horizontal component. Since the Earth is not surrounded by an uninterrupted envelope of water but rather shows a very irregular alternation of sea and land, the mechanism of the response of the oceans and seas to the tidal forces is extremely complex. A further complication is caused by the deflecting force of the Earth's rotation (the Coriolis force; see above).

In enclosures formed by gulfs (gulf) and bays (bay), the local tide is generated by interaction with the tides of the adjacent open ocean. Such a tide often takes the form of a running tidal wave that rotates within the confines of the enclosure. In some semi-enclosed seas, such as the Mediterranean, Black, and Baltic seas, a standing wave, or tidal seiche, may be generated by the local tide-raising forces.

In these seas, the tidal range of sea level is only on the order of centimetres. In the open ocean, it generally is on the order of tens of centimetres. In bays and adjacent seas, however, the tidal range may be much greater, because the shape of a bay or adjacent sea may favour the enhancement of the tide inside; in particular, there may be a resonance of the basin concerned with the tide. The largest known tides occur in the Bay of Fundy (Fundy, Bay of), where spring tidal ranges up to 15 metres have been measured.

Tidal bores form on rivers (river) and estuaries near a coast where there is a large tidal range and the incoming tide is confined to a narrow channel. They consist of a surge of water moving swiftly upstream headed by a wave or series of waves. Such bores are quite common. There is a large one, known as the mascaret on the Seine (Seine River), which forms on spring tides and reaches as far upriver as Rouen. There is a well-known bore on the Severn, in England, and another forms on the Petitcodiac River, which empties into the Bay of Fundy in New Brunswick. The classic example is the bore on the Ch'ien-t'ang described by Commander W. Usborne Moore of the British navy in 1888 and 1892. He reported heights of 2.5 to 3.5 metres.

When a tidal bore forms in a river, the direction of flow of the water changes abruptly as the bore passes. Before it arrives, the water may be still or, more usually, a small freshwater current flows outward toward the sea. The tide comes in as a “wall of water” that passes up the river. Behind the bore, the current flows upriver. At the division between the moving water behind the bore and the still water in front, there is a wave, the water surface behind being higher than it is in front. This wave must travel more quickly than the water particles behind it, because, as the advancing water travels upriver, it collects the still water in front and sets it in motion. Upriver, the advancing tide will consist not of salt water from the sea but rather of fresh water that has passed farther down and been collected and returned in front of the incoming tide. It is therefore necessary to distinguish between the velocity of the advancing wave and that of the water particles just behind it.

Density currents are currents that are kept in motion by the force of gravity acting on a relatively small density difference caused by variations in salinity, temperature, or sediment concentration. As noted above, salinity and temperature variations produce stratification in oceans. Below the surface layer, which is disturbed by waves and is lighter than the deeper waters because it is warmer or less saline, the oceans are composed of layers of water that have distinctive chemical and physical characteristics, which move more or less independently of each other and which do not lose their individuality by mixing even after they have flowed for hundreds of kilometres from their point of origin.

An example of this type of density current, or stratified flow, is provided by the water of the Mediterranean Sea as it flows through the Strait of Gibraltar (Gibraltar, Strait of) out into the Atlantic (Atlantic Ocean). Because the Mediterranean is enclosed in a basin that is relatively small compared with the ocean basins and because it is located in a relatively arid climate, evaporation exceeds the supply of fresh water from rivers. The result is that the Mediterranean contains water that is both warmer and more saline than normal deep-sea water, the temperature ranging from 12.7° to 14.5° C and the salinity from 38.4 to 39.0 parts per thousand. Because of these characteristics, the Mediterranean water is considerably denser than the water in the upper parts of the North Atlantic, which has a salinity of about 36 parts per thousand and a temperature of about 13° C. The density contrast causes the lighter Atlantic water to flow into the Mediterranean in the upper part of the Strait of Gibraltar (down to a depth of about 200 metres) and the denser Mediterranean water to flow out into the Atlantic in the lower part of the strait (from about 200 metres to the top of the sill separating the Mediterranean from the Atlantic at a depth of 320 metres). Because the strait is only about 20 kilometres wide, both inflow and outflow achieve relatively high speeds. Near the surface the inflow may have speeds as high as two metres per second, and the outflow reaches speeds of more than one metre per second at a depth of about 275 metres. One result of the high current speeds in the strait is that there is a considerable amount of mixing, which reduces the salinity of the outflowing Mediterranean water to about 37 parts per thousand. The outflowing water sinks to a depth of about 1,500 metres or more, where it encounters colder, denser Atlantic water. It then spreads out as a layer of more saline water between two Atlantic water masses.

Density currents caused by suspended sediment concentrations in the oceans are called turbidity currents. They appear to be relatively short-lived, transient phenomena that occur at great depths. Turbidity currents are thought to be caused by the slumping of sediment that has piled up at the top of the continental slope, particularly at the heads of submarine canyons (see below Continental margins: Submarine canyons (ocean)). Slumping of large masses of sediment creates a dense sediment-water mixture, or slurry, which then flows down the canyon to spread out over the ocean floor and deposit a layer of sand in deep water. Repeated deposition forms submarine fans, which are analogous to the alluvial fans found at the mouths of many river canyons. Sedimentary rocks that are thought to have originated from ancient turbidity currents are called turbidites.

Although large-scale turbidity underflows have never been directly observed in the oceans, there is much evidence supporting their occurrence. This evidence may be briefly summarized: (1) Telegraph cables have been broken in the deep ocean in a sequence that indicates some disturbance at the bottom (bottom water) moving from shallow to deep water at speeds on the order of 20 to 75 kilometres per hour, or 10 to 40 knots. The trigger for this phenomenon is commonly, though not exclusively, an earthquake near the edge of the continental slope. The only disturbance that seems capable of being transmitted downslope at the required speed is a large turbidity current. The best-known example of such a series of cable breaks took place in the North Atlantic following the 1929 earthquake under the Grand Banks of Newfoundland, but other examples have been described from the Magdalena River delta (Colombia), the Congo delta, the Mediterranean Sea north of Orléansville and south of the Straits of Messina, and Kandavu Passage, Fiji. (2) Cores (core sampling) taken from the ocean bottom in the area downslope from cable breaks reveal layers of sand interbedded with normal deep-sea pelagic or hemipelagic oozes (sediments formed in the deep sea by quiet settling of fine particles). In the case of the cable breaks south of the Grand Banks, a large-diameter core taken from the axis of a submarine canyon in the continental slope contained 1 centimetre of gray clay underlain by at least 20 centimetres of gray pebble and cobble gravel. Cores farther south showed a graded layer about one metre thick of coarse silt and fine sand. The presence of these gravel and sand layers is consistent with the hypothesis that they were deposited by the turbidity current that broke the cables. (3) Coring has revealed layers of fine-grained sand or coarse silt at many other localities in the abyssal plains of the oceans. These layers are generally moderately well sorted and contain microfossils characteristic of shallow water that are also size-sorted. In some cases the layers are laminated and arranged in a definite sequence. It is clear that the sand forming these layers has been moved down from shallow water, and in many cases the only plausible mechanism appears to be a turbidity current. (4) At the base of many submarine canyons (submarine canyon) there occur very large submarine fans (submarine fan). Deep-sea channels on the fan surfaces extend for many tens of kilometres and have depths of more than 100 metres and widths of one kilometre or more. Submarine levees are a prominent feature, and these project above the surrounding fan surface to elevations of 50 metres or higher. The gross characteristics of such channels suggest that they were formed by a combination of erosion and deposition by turbidity currents. (5) Thick deposits of interbedded graded sandstones (sandstone) and fine-grained shales (shale) are common in the geologic record. In some cases there is good fossil evidence that the shales were deposited in relatively deep water, perhaps as much as several thousand metres deep. Relatively deepwater deposition is also suggested by the absence of sedimentary structures characteristic of shallow water. The interbedded sandstones, however, contain shallow-water fossils that are sorted by size, have a sharp basal contact with the shale below and a transitional contact with the shale above, and display a characteristic sequence of sedimentary structures. The structures include erosional marks made originally on the mud surface but now preserved as casts on the base of the sandstone bed (sole marks) and internal structures including some or all of the following: massive graded unit, parallel lamination, ripple cross-lamination or convolute lamination, and an upper unit of parallel lamination. This combination of textural and structural features can be explained by deposition from a current that slightly erodes the bottom and then deposits sand that becomes finer grained as the velocity gradually wanes. The properties inferred from these ancient sandstone deposits are consistent with the properties of turbidity currents inferred from laboratory experiments.

In spite of the convincing nature of the evidence, there are still some objections to the turbidity current hypothesis. Most geologists and oceanographers accept that such currents exist and that the currents are important agents of erosion and sediment deposition, in both modern and ancient seas, but researchers believe that the turbidity current hypothesis has been overworked. There is evidence, for example, which suggests that currents flowing parallel to submarine contours exist in many ocean basins. These bottom currents have been observed in a few cases, and velocities as high as 20 to 50 centimetres per second have been recorded. These currents can produce some of the features that previously had been attributed to turbidity current action. Moreover, nearly all features of sands that are produced by turbidity currents can be formed by shallow-water action, such as fluvial processes. Hence the problem of discriminating between deposits formed by turbidity currents and deposits formed by other current types is quite complex and requires a careful assessment of all lines of evidence in each case. Some ancient sandstones have been interpreted as “fluxoturbidites” because the sedimentary structures and other properties suggest a transporting agent intermediate between turbidity currents and large-scale slumping and sliding of sediment.

The notion of a connection (air–sea interface) between the temperature of the surface layers of the oceans and the circulation of the lowest layer of the atmosphere, the troposphere, is a familiar one. The surface mixed layer of the ocean is a huge reservoir of heat when compared to the overlying atmosphere. The heat capacity of an atmospheric column of unit area cross-section extending from the ocean surface to the outermost layers of the atmosphere is equivalent to the heat capacity of a column of seawater of 2.6-metre depth. The surface layer of the oceans is continuously being stirred by the overlying winds and waves, and thus a surface mixed layer is formed that has vertically uniform properties in temperature and salinity. This mixed layer, which is in direct contact with the atmosphere, has a minimum depth of 20 metres in summer and a maximum depth exceeding 100 metres in late winter in the mid-latitudes. In lower latitudes the seasonal variation in the mixed layer is less marked than at higher latitudes, except in regions such as the Arabian Sea where the onset of the southwestern Indian monsoon may produce large changes in the depth of the mixed layer. Temperature anomalies (i.e., deviations from the normal seasonal temperature) in the surface mixed layer have a long residence time compared with those of the overlying turbulent atmosphere. Hence they may persist for a number of consecutive seasons and even for years.

Observational studies to investigate the relationship between anomalies in ocean surface temperature and the tropospheric circulation have been undertaken primarily in the Pacific and Atlantic. They have identified large-scale ocean surface temperature anomalies that have similar spatial scales to monthly and seasonal anomalies in atmospheric circulation. The longevity of the ocean surface temperature anomalies, as compared with the shorter dynamical and thermodynamical “memory” of the atmosphere, has suggested that they may be an important predictor for seasonal and interannual climate anomalies.

First, it is useful to consider some examples of the association between anomalies in ocean surface temperature and irregular changes in climate. The Sahel, a region that borders the southern fringe of the Sahara in Africa, experienced a number of devastating droughts during the 1970s and '80s, which can be compared with a much wetter period during the 1950s. Data was obtained that showed the difference in ocean surface temperature during the period from July to September between the “driest” and “wettest” rainfall seasons in the Sahel after 1950. Of particular note were the higher-than-normal surface temperatures in the tropical South Atlantic, Indian, and Southeast Pacific oceans and the lower-than-normal temperatures in the North Atlantic and Pacific oceans. This example illustrates that climate anomalies in one region of the world may be linked to ocean surface temperature changes on a global scale. Global atmospheric modeling studies undertaken during the mid-1980s have indicated that the positions of the main rainfall zones in the tropics are sensitive to anomalies in ocean surface temperature.

Shorter-lived climate anomalies, on time scales of months to one or two years, also have been related to ocean surface temperature anomalies. The equatorial oceans have the largest influence on these climate anomalies because of the evaporation of water. A relatively small change in ocean surface temperature, say, of 1° C, may result in a large change in the evaporation of water into the atmosphere. The increased water vapour in the lower atmosphere is condensed in regions of upward motion known as convergence zones. This process liberates latent heat of condensation, which in turn provides a major fraction of the energy to drive tropical circulation and is one of the mechanisms responsible for the El Niño/Southern Oscillation phenomenon discussed later in this article.

Given the sensitivity of the tropical atmosphere to variations in tropical sea surface temperature, there also has been considerable interest in their influence on extratropical circulation. The sensitivity of the tropospheric circulation to surface temperature in both the tropical Pacific and Atlantic oceans has been shown in theoretical and observational studies alike. Figures were prepared to demonstrate the correlation between the equatorial ocean surface temperature in the east Pacific (the location of El Niño) and the atmospheric circulation in the middle troposphere during winter. The atmospheric pattern was a characteristic circulation type known as the Pacific-North American (PNA) mode. Such patterns are intrinsic modes of the atmosphere, which may be forced by thermal anomalies in the tropical atmosphere and which in their turn are forced by tropical ocean surface temperature anomalies. As noted earlier, enhanced tropical sea surface temperatures increase evaporation into the atmosphere. In the 1982–83 El Niño event a pattern of circulation anomalies occurred throughout the Northern Hemisphere during winter. These modes of the atmosphere, however, account for much less than 50 percent of the variability of the circulation in mid-latitudes, though in certain regions (northern Japan, southern Canada, and the southern United States), they may have sufficient amplitude for them to be used for predicting seasonal surface temperature perhaps up to two seasons in advance.

The response of the atmosphere to mid-latitude ocean surface anomalies has been difficult to detect unambiguously because of the complexity of the turbulent westerly flow between 20° and 60° latitude in both hemispheres. This flow has many properties of nonlinear chaotic systems and thus exhibits behaviour that is difficult to predict beyond a couple of weeks. The atmosphere alone can exhibit large fluctuations on seasonal and longer time scales without any change in external forcing conditions, such as ocean surface temperature. Notwithstanding this inherent problem, some effects of ocean surface temperature anomalies on the atmosphere have been observed and modeled.

The influence of the oceans on the atmosphere in the mid-latitudes is greatest during autumn and early winter when the ocean mixed layer releases to the atmosphere the large quantities of heat that it has stored up over the previous summer. Anomalies in ocean surface temperature are indicative of either a surplus or a deficiency of heat available to the atmosphere. The response of the atmosphere to ocean surface temperature, however, is not random geographically. The circulation over the North Atlantic and northern Europe during early winter has been found to be sensitive to large ocean surface temperature anomalies south of Newfoundland. When a warm positive anomaly exists in this region, an anomalous surface anticyclone occurs in the central Atlantic at a similar latitude to the temperature anomaly, and an anomalous cyclonic circulation is located over the North Sea, Scandinavia, and central Europe. With colder than normal water south of Newfoundland, the circulation patterns are reversed, producing cyclonic circulation over the central Atlantic and anticyclonic circulation over Europe. The sensitivity of the atmosphere to ocean surface temperature anomalies in this particular region is thought to be related to the position of the overlying storm tracks and jet stream. The region is the most active in the Northern Hemisphere for the growth of storms associated with very large heat fluxes from the surface layer of the ocean.

Another example of a similar type of air-sea interaction event has been documented over the North Pacific Ocean. A statistical seasonal relationship exists between the summer ocean temperature anomaly in the Gulf of Alaska (Alaska, Gulf of) and the atmospheric circulation over the Pacific and North America during the following autumn and winter. The presence of warmer-than-normal ocean surface temperature in the Gulf of Alaska results in increased cyclone development during the subsequent autumn and winter. The relationship has been established by means of monthly sea surface temperature and atmospheric pressure data collected over 30 years in the North Pacific Ocean.

The air-sea interaction events in both the North Pacific and North Atlantic oceans discussed above raise questions as to how the anomalies in ocean surface temperature in these areas are initiated, how they are maintained, and whether they yield useful information for atmospheric prediction beyond the normal time scales of weather forecasting (i.e., one to two weeks). Statistical analysis of previous case studies have shown that ocean surface temperature anomalies initially develop in response to anomalous atmospheric forcing. Once developed, however, the temperature anomaly of the ocean surface tends to reinforce and thereby maintain the anomalous atmospheric circulation. The mechanisms thought to be responsible for this behaviour in the ocean are the surface wind drift, wind mixing, and the interchange of heat between the ocean and atmosphere. The question of prediction is therefore difficult to answer, as these events depend on a synchronous and interconnected behaviour between the atmosphere and the surface layer of the ocean, which allows for positive feedback between the two systems.

Tropical cyclones represent still another example of sea-air interactions. These storm systems are known as hurricanes (hurricane) in the North Atlantic and eastern North Pacific and as typhoons (typhoon) in the western North Pacific. The winds of such systems revolve around a centre of low pressure in an anticlockwise direction in the Northern Hemisphere and in a clockwise direction in the Southern Hemisphere. The winds attain velocities in excess of 115 kilometres per hour, or 65 knots, in most cases. Tropical cyclones may last from a few hours to as long as two weeks, the average lifetime being six days.

The oceans provide the source of energy for tropical cyclones both by direct heat transfer from their surface (known as sensible heat) and by the evaporation (vaporization) of water. This water is subsequently condensed within a storm system, thereby releasing latent heat energy. When a tropical cyclone moves over land, this energy is severely depleted and the circulation of the winds is consequently weakened.

Such storms are truly phenomena of the tropical oceans. They originate in two distinct latitude zones, between 4° and 22° S and between 4° and 35° N. They are absent in the equatorial zone between 4° S and 4° N. Most tropical cyclones are spawned on the poleward side of the region known as the intertropical convergence zone (ITCZ).

More than two-thirds of observed tropical cyclones originate in the Northern Hemisphere, and roughly the same proportion occur in the Eastern Hemisphere. The North Pacific has more than one-third of all such storms, while the southeast Pacific and South Atlantic are normally devoid of them. Most northern hemispheric tropical cyclones occur between May and November, with peak periods in August and September. The majority of southern hemispheric cyclones occur between December and April, with peaks in January and February.

The formation of tropical cyclones is strongly influenced by the temperature of the underlying ocean or, more specifically, by the thermal energy available in the upper 60 metres of ocean waters. Typically, the underlying ocean should have a temperature in excess of 26° C in this layer. This temperature requirement, however, is only one of five that need to be met for a tropical cyclone to form and develop. The other preconditions relate to the state of the tropical atmosphere between the sea surface and a height of 16 kilometres, the boundary of the tropical troposphere. They can be summarized as follows:

All these conditions may be met but still not lead to cyclone formation. It is thought that the most important factor is the presence of a large-scale cyclonic circulation in the lower troposphere. The above conditions occur for a period of 5 to 15 days and are followed by less favourable conditions for a duration of 10 to 20 days.

Once a tropical cyclone has formed, it usually follows certain distinct stages during its lifetime. In its formative stage the winds are below hurricane force and the central pressure (atmospheric pressure) is about 1,000 millibars. The formative period is extremely variable in length, ranging from 12 hours to a few days. This stage is followed by a period of intensification, when the central pressure drops rapidly below 1,000 millibars. The winds increase rapidly, and they may achieve hurricane force within a radius of 30 to 50 kilometres of the storm centre. At this stage the cloud and rainfall patterns become well organized into narrow bands that spiral inward toward the centre. In the mature phase the central pressure stops falling and, as a consequence, the winds no longer increase. The region of hurricane force winds, however, expands to occupy a radius of 300 kilometres or more. This expansion is not symmetrical around the storm centre; the strongest winds occur toward the right-hand side of the centre in the direction of the cyclone's path. The period of maturity may last for one to three days. The terminal stage of a tropical cyclone is usually reached when the storm strikes land, followed by a resultant increase in energy dissipation by surface friction and a reduction in its energy supply of moisture. A reduction in moisture input into the storm system may also take place when it moves over a colder segment of the ocean. A tropical cyclone may regenerate in higher latitudes as an extratropical depression, but it loses its identity as a tropical storm in the process. The typical lifetime of a tropical cyclone from its birth to death is about six days.

The paths of tropical cyclones show a wide variation. In both the North Atlantic and the North Pacific, the paths tend to be initially northwestward and then recurve toward the northeast at higher latitudes. It is now known that the tracks of tropical cyclones are largely determined by the large-scale tropospheric flow. This fact opens up the possibility that, with the aid of high-resolution numerical models, accurate predictions of their tracks may become feasible. The development of polar-orbiting and geostationary satellites has made it possible to accurately track cyclones over the remotest areas of the tropical oceans.

A tropical cyclone can affect the thermal structure and currents in the surface layer of the ocean waters in its path. Cooling of the surface layer occurs in the wake of such a storm. Maximum cooling occurs on the right of a hurricane's path in the Northern Hemisphere. In the wake of Hurricane Hilda's passage through the Gulf of Mexico in 1964 at a translational speed of only five knots, the surface waters were cooled by as much as 6° C. Tropical cyclones that have higher translational velocities cause less cooling of the surface. The surface cooling is caused primarily by wind-induced upwelling of cooler water from below the surface layer. The warm surface water is simultaneously transported toward the periphery of the cyclone, where it downwells into the deeper ocean layers. Heat loss across the air-sea interface and the wind-induced mixing of the surface water with those of the cooler subsurface layers make a significant but smaller contribution to surface cooling.

In addition to surface cooling, tropical cyclones may induce large horizontal surge currents and vertical displacements of the thermocline. The surge currents have their largest amplitude at the surface, where they may reach velocities approaching one metre per second. The horizontal currents and the vertical displacement of the thermocline observed in the wake of a tropical cyclone oscillate close to the inertial period. These oscillations remain for a few days after the passage of the storm and spread outward from the rear of the system as an internal wake on the thermocline. The vertical motion may transport nutrients from the deeper layers into the sunlit surface waters, which in turn promotes phytoplankton blooms (i.e., the rapid growth of diatoms and other minute one-celled organisms). The ocean surface temperature normally recovers to its precyclone value within 10 days of a storm's passage.

Tropical cyclones play an important role in the general circulation of the atmosphere, accounting for 2 percent of the global annual rainfall (precipitation) and between 4 and 5 percent of the global rainfall in August and September at the height of the Northern Hemispheric cyclone season. For a local area, the occurrence of a single tropical cyclone can have a major impact on the region's annual rainfall. Furthermore, tropical cyclones contribute approximately 2 percent of the kinetic energy of the general circulation of the atmosphere, some of which is exported from the tropics to higher latitudes.

This major current system, as described earlier, is a western boundary current that flows poleward along a boundary separating the warm and more saline waters of the Sargasso Sea to the east from the colder, slightly fresher continental slope waters to the north and west. The warm, saline Sargasso Sea, composed of a water mass known as North Atlantic Central Water, has a temperature that ranges from 8° to 19° C and a salinity between 35.10 and 36.70 parts per thousand. This is one of the two dominant water masses of the North Atlantic Ocean, the other being the North Atlantic Deep Water, which has a temperature of 2.2° to 3.5° C and a salinity between 34.90 and 34.97 parts per thousand, and which occupies the deepest layers of the ocean (generally below 1,000 metres). The North Atlantic Central Water occupies the upper layer of the North Atlantic Ocean between roughly 20° and 40° N. The “lens” of this water is at its lowest depth of 1,000 metres in the northwest Atlantic and becomes progressively shallower to the east and south. To the north it shallows abruptly and outcrops at the surface in winter, and it is at this point that the Gulf Stream is most intense.

The Gulf Stream flows along the rim of the warm North Atlantic Central Water northward from the Florida Straits along the continental slope of North America to Cape Hatteras. There, it leaves the continental slope and turns northeastward as an intense meandering current that extends toward the Grand Banks of Newfoundland. Its maximum velocity is typically between one and two metres per second. At this stage, a part of the current loops back onto itself, flowing south and east. Another part flows eastward toward Spain and Portugal, while the remaining water flows northeastward as the North Atlantic Drift (also called the North Atlantic Current) into the northernmost regions of the North Atlantic Ocean between Scotland and Iceland.

The southward-flowing currents are generally weaker than the Gulf Stream and occur in the eastern lens of the North Atlantic Central Water or the subtropical gyre (see above Circulation of the ocean waters: Wind-driven circulation: The subtropical gyres (ocean)). The circulation to the south on the southern rim of the subtropical gyre is completed by the westward-flowing North Equatorial Current, part of which flows into the Gulf of Mexico; the remaining part flows northward as the Antilles Current. This subtropical gyre of warm North Atlantic Central Water is the hub of the energy that drives the North Atlantic circulation. It is principally forced by the overlying atmospheric circulation, which at these latitudes is dominated by the clockwise circulation of a subtropical anticyclone. This circulation is not steady and fluctuates in particular on its poleward side where extratropical cyclones in the westerlies periodically make incursions into the region. On the western side, hurricanes (during the period from May to November) occasionally disturb the atmospheric circulation. Because of the energy of the subtropical gyre and its associated currents, these short-term fluctuations have little influence on it, however. The gyre obtains most of its energy from the climatological wind distribution over periods of one or two decades. This wind distribution drives a system of surface currents in the uppermost 100 metres of the ocean. Nonetheless, these currents are not simply a reflection of the surface wind circulation as they are influenced by the Coriolis force (see above Circulation of the ocean waters: Wind-driven circulation: Coriolis effect (ocean)). The wind-driven current decays with depth, becoming negligible below 100 metres. The water in this surface layer is transported to the right and perpendicular to the surface wind stress because of the Coriolis force. Hence an eastward-directed wind on the poleward side of the subtropical anticyclone would transport the surface layer of the ocean to the south. On the equatorward side of the anticyclone the trade winds would cause a contrary drift of the surface layer to the north and west. Thus surface waters under the subtropical anticyclone are driven toward the mid-latitudes at about 30° N. These surface waters, which are warmed by solar heating and have a high salinity by virtue of the predominance of evaporation over precipitation at these latitudes, then converge and are forced downward into the deeper ocean.

Over many decades this process forms a deep lens of warm, saline North Atlantic Central Water. The shape of the lens of water is distorted by other dynamical effects, the principal one being the change in the vertical component of the Coriolis force with latitude known as the beta effect. This effect involves the displacement of the warm water lens toward the west, so that the deepest part of the lens is situated to the north of the island of Bermuda rather than in the central Atlantic Ocean. This warm lens of water plays an important role, establishing as it does a horizontal pressure gradient force in and below the wind-drift current. The sea level over the deepest part of the lens is about one metre higher than outside the lens. The Coriolis force in balance with this horizontal pressure gradient force gives rise to a dynamically induced geostrophic current, which occurs throughout the upper layer of warm water. The strength of this geostrophic current is determined by the horizontal pressure gradient through the slope in sea level. The slope in sea level across the Gulf Stream has been measured by satellite radar altimeter to be one metre over a horizontal distance of 100 kilometres, which is sufficient to cause a surface geostrophic current of one metre per second at 43° N.

The large-scale circulation of the Gulf Stream system is, however, only one aspect of a far more complex and richer structure of circulation. Embedded within the mean flow is a variety of eddy structures that not only put kinetic energy into circulation but also carry heat and other important properties, such as nutrients for biological systems. The best known of these eddies are the Gulf Stream rings, which develop in meanders of the current east of Cape Hatteras. Though the eddies were mentioned as early as 1793 by Jonathan Williams, a grandnephew of Benjamin Franklin, they were not systematically studied until the early 1930s by the oceanographer Phil E. Church. Intensive research programs were finally undertaken during the 1970s. Gulf Stream rings have either warm or cold cores. The warm rings are typically 100 to 300 kilometres in diameter and have a clockwise rotation. They consist of waters from the Gulf Stream and Sargasso Sea and form when the meanders in the Gulf Stream pinch off on its continental slope side. They move generally westward, flowing at the speed of the slope waters, and are reabsorbed into the Gulf Stream at Cape Hatteras after a typical lifetime of about six months. The cold core rings, composed of a mixture of Gulf Stream and continental slope waters, are formed when the meanders pinch off to the south of the Gulf Stream. They are a little larger than their warm-core counterparts, characteristically having diameters of 200 to 300 kilometres and an anticlockwise rotation. They move generally southwestward into the Sargasso Sea and have lifetimes of one to two years. The cold-core rings are usually more numerous than warm-core rings, typically 10 each year as compared with five warm-core rings annually.

This western boundary current is similar to the Gulf Stream in that it produces both warm and cold rings. The warm rings are generally 150 kilometres in diameter and have a lifetime similar to their Gulf Stream counterparts. The cold rings form at preferential sites and in most cases drift southwestward into the Western Pacific Ocean. Occasionally a cold ring has been observed to move northwestward and eventually be reabsorbed into the Kuroshio.

A significant characteristic of the large-scale North Atlantic circulation is the poleward transport of heat. Heat is transferred in a northward direction throughout the North Atlantic. This heat is absorbed by the tropical waters of the Pacific and Indian oceans, as well as of the Atlantic, and is then transferred to the high latitudes, where it is finally given up to the atmosphere.

The mechanism for the heat transfer is principally by thermohaline circulation rather than by wind-driven circulation (see above Circulation of the ocean waters: Thermohaline circulation (ocean)). Circulation of the thermohaline type involves a large-scale overturning of the ocean, with warm and saline water in the upper 1,000 metres moving northward and being cooled in the Labrador, Greenland, and Norwegian seas. The density of the water in contact with the atmosphere is increased by surface cooling, and the water subsequently sinks below the surface layer to the lowest depths of the ocean. This water is mixed with the surrounding water masses by a variety of processes to form North Atlantic Deep Water. The water moves slowly southward as the lower limb of the thermohaline circulation. It is this overturning circulation that is responsible for the warm winter climate of northwestern Europe (notably the British Isles and Norway) rather than the horizontal wind-driven circulation discussed above. The North Atlantic Drift (North Atlantic Current), which is an extension of the Gulf Stream system to the south, provides this northward flow of warm and saline waters into the polar seas. This feature makes the circulation of the North Atlantic Ocean uniquely different from that of the Pacific Ocean, which has a less effective thermohaline circulation. Although there is a northward transfer of heat in the North Pacific, the subtropical wind-driven gyre in the upper ocean is mainly responsible for it. Thus the Kuroshio on the western boundary of the North Pacific gyre is principally driven by the surface wind circulation of the North Pacific.

Studies of the sediment cores obtained from the ocean floor have indicated that the ocean surface temperature was as much as 10° C cooler than today in the northernmost region of the North Atlantic Ocean during the last glacial maximum some 18,000 years ago. This difference in surface temperature would indicate that the warm North Atlantic Drift was much reduced compared to what it is at present, and hence the thermohaline circulation was considerably weaker. In contrast, the Gulf Stream was probably more intense than it is today and exhibited a large shift from its present path to an eastward flow at 40° N.

As was explained earlier, the oceans can moderate the climate of certain regions. Not only do they affect such geographic variations, but they also influence temporal changes in climate. The time scales of climate variability range from a few years to millions of years and include the so-called ice age cycles that repeat every 20,000 to 40,000 years, interrupted by interglacial periods of “optimum” climate, such as the present. The climatic modulations that occur at shorter scales include such periods as the Little Ice Age from the early 16th to the mid-19th centuries, when the global average temperature was approximately 1° C lower than it is today. Several climate fluctuations on the scale of decades have occurred in the 20th century, such as warming from 1910 to 1940, cooling from 1940 to 1970, and the warming trend since 1970.

Although many of the mechanisms of climate change are understood, it is usually difficult to pinpoint the specific causes. Scientists acknowledge that climate can be affected by factors external to the land-ocean-atmosphere climate system, such as variations in solar brightness, the shading effect of aerosols injected into the atmosphere by volcanic activity, or the increased atmospheric concentration of “greenhouse” gases (e.g., carbon dioxide, nitrous oxide, methane, and chlorofluorocarbons) produced by human activities. However, none of these factors explain the periodic variations observed during the 20th century, which may simply be manifestations of the natural variability of climate. The existence of natural variability at many time scales makes the identification of causative factors such as human-induced warming more difficult. Whether change is natural or caused, the oceans play a key role and have a moderating effect on influencing factors.

The shortest, or interannual, time scale relates to natural variations that are perceived as years of unusual weather—e.g., excessive heat, drought, or storminess. Such changes are so common in many regions that any given year is about as likely to be considered as exceptional as typical. The best example of the influence of the oceans on interannual climate anomalies is the occurrence of El Niño conditions in the eastern Pacific Ocean at irregular intervals of about 3–10 years. The stronger El Niño episodes of enhanced ocean temperatures (2°–8° C above normal) are typically accompanied by altered weather patterns around the globe, such as droughts in Australia, northeastern Brazil, and the highlands of southern Peru, excessive summer rainfall along the coast of Ecuador and northern Peru, severe winter storminess along the coast of central Chile, and unusual winter weather along the west coast of North America.

The effects of El Niño have been documented in Peru since the Spanish conquest in 1525. The Spanish term “la corriente de El Niño” was introduced by fishermen of the Peruvian port of Paita in the 19th century; it refers to a warm, southward ocean current that temporarily displaces the normally cool, northward-flowing Humboldt, or Peru, Current. (The name is a pious reference to the Christ child, chosen because of the typical appearance of the countercurrent during the Christmas season.) By the end of the 19th century Peruvian geographers recognized that every few years this countercurrent is more intense than normal, extends farther south, and is associated with torrential rainfall over the otherwise dry northern desert. The abnormal countercurrent also was observed to bring tropical debris, as well as such flora and fauna as bananas and aquatic reptiles, from the coastal region of Ecuador farther north. Increasingly during the 20th century, El Niño has come to connote an exceptional year rather than the original annual event.

As Peruvians began to exploit the guano of marine birds for fertilizer in the early 20th century, they noticed El Niño-related deteriorations in the normally high marine productivity of the coast of Peru as manifested by large reductions in the bird populations that depend on anchovies (anchovy) and sardines for sustenance. The preoccupation with El Niño increased after mid-century, as the Peruvian fishing industry rapidly expanded to exploit the anchovies directly. (Fish meal produced from the anchovies was exported to industrialized nations as a feed supplement for livestock.) By 1971 the Peruvian fishing fleet had become the largest in its history; it had extracted very nearly 13 million metric tons of anchovies in that year alone. Peru was catapulted into first place among fishing nations, and scientists expressed serious concern that fish stocks were being depleted beyond self-sustaining levels, even for the extremely productive marine ecosystem of Peru. The strong El Niño of 1972–73 captured world attention because of the drastic reduction in anchovy catches to a small fraction of prior levels. The anchovy catch did not return to previous levels, and the effects of plummeting fish meal exports reverberated throughout the world commodity markets.

El Niño was only a curiosity to the scientific community in the first half of the 20th century, thought to be geographically limited to the west coast of South America. There was little data, mainly gathered coincidentally from foreign oceanographic cruises, and it was generally believed that El Niño occurred when the normally northward coastal winds off Peru, which cause the upwelling of cool, nutrient-rich water along the coast, decreased, ceased, or reversed in direction. When systematic and extensive oceanographic measurements were made in the Pacific in 1957–58 as part of the International Geophysical Year, it was found that El Niño had occurred during the same period and was also associated with extensive warming over most of the Pacific equatorial zone. Eventually tide-gauge and other measurements made throughout the tropical Pacific showed that the coastal El Niño was but one manifestation of basinwide ocean circulation changes that occur in response to a massive weakening of the westward-blowing trade winds in the western and central equatorial Pacific and not to localized wind anomalies along the Peru coast.

The wind anomalies are a manifestation of an atmospheric counterpart to the oceanic El Niño. At the turn of the century, the British climatologist Gilbert Walker set out to determine the connections between the Asian monsoon and other climatic fluctuations around the globe in an effort to predict unusual monsoon years that bring drought and famine to the Asian sector. Unaware of any connection to El Niño, he discovered a coherent interannual fluctuation of atmospheric pressure over the tropical Indo-Pacific region, which he termed the Southern Oscillation (SO). During years of reduced rainfall over northern Australia and Indonesia, the pressure in that region (e.g., at what are now Darwin and Jakarta) was anomalously high and wind patterns were altered. Simultaneously, in the eastern South Pacific pressures were unusually low, negatively correlated with those at Darwin and Jakarta. A Southern Oscillation Index (SOI), based on pressure differences between the two regions (east minus west), showed low, negative values at such times, which were termed the “low phase” of the SO. During more normal “high-phase” years, the pressures were low over Indonesia and high in the eastern Pacific, with high, positive values of the SOI. In papers published during the 1920s and '30s, Walker gave statistical evidence for widespread climatic anomalies around the globe being associated with the SO pressure “seesaw.”

In the 1950s, years after Walker's investigations, it was noted that the low-phase years of the SOI corresponded with periods of high ocean temperatures along the Peruvian coast, but no physical connection between the SO and El Niño was recognized until Jacob Bjerknes (Bjerknes, Jacob), in the early 1960s, tried to understand the large geographic scale of the anomalies observed during the 1957–58 El Niño event. Bjerknes, a meteorologist, formulated the first conceptual model of the large-scale ocean-atmosphere interactions that occur during El Niño episodes. His model has been refined through intensive research since the early 1970s.

During a year or two prior to an El Niño event (high-phase years of the SO), the westward trade winds typically blow more intensely along the equator in the equatorial Pacific, causing warm upper-ocean water to accumulate in a thickened surface layer in the western Pacific where sea level rises. Meanwhile, the stronger, upwelling-favourable winds in the eastern Pacific induce colder surface water and lowered sea levels off South America. Toward the end of the year preceding an El Niño, the area of intense tropical storm activity over Indonesia migrates eastward toward the equatorial Pacific west of the International Date Line (which corresponds in general to the 180th meridian of longitude), bringing episodes of eastward wind reversals to that region of the ocean. These wind bursts excite extremely long ocean waves, known as Kelvin waves (imperceptible to an observer), that propagate eastward toward the coast of South America, where they cause the upper ocean layer of relatively warm water to thicken and sea level to rise.

The tropical storms of the western Pacific also occur in other years, though less frequently, and produce similar Kelvin waves, but an El Niño event does not result and the waves continue poleward along the coast toward Chile and California, detectable only in tide-gauge measurements. Something else occurs prior to an El Niño that is not fully understood: as the Kelvin waves travel eastward along the equator, an anomalous eastward current carries warm western Pacific water farther east, and the warm surface layer deepens in the central equatorial Pacific (east of the international dateline). Additional surface warming takes place as the upwelling-favourable winds bring warmer subsurface water to the surface. (The subsurface water is warmer now, rather than cooler, because the overlying layer of warmer water is now significantly deeper than before.) The anomalous warming creates conditions favourable for the further migration of the tropical storm centre toward the east, giving renewed vigour to eastward winds, more Kelvin waves, and additional warming. Each increment of anomalies in one medium (e.g., the ocean) induces further anomalies in the other (the atmosphere) and vice versa, giving rise to an unstable growth of anomalies through a process of positive feedbacks. During this time, the SO is found in its low phase.

The current focus of oceanographic research is on understanding the circumstances leading to the demise of the El Niño event and the onset of another such event several years later. The most widely held hypothesis is that a second class of long equatorial ocean waves—Rossby waves (Rossby wave) with a shallow surface layer—is generated by the El Niño and that they propagate westward to the landmasses of Asia. There, the Rossby waves reflect off the Asian coast eastward along the equator in the form of upwelling Kelvin waves, resulting in a thinning of the upper ocean warm layer and a cooling of the ocean as the winds bring deeper, cooler water to the surface. This process is thought to initiate one to two years of colder-than-average conditions until Rossby waves of a contrary sense (i.e., with a thickened surface layer) are again generated, functioning as a switching mechanism, this time to start another El Niño sequence.

Another goal of scientists is to understand climate change on the scale of centuries or longer and to make projections about the changes that will occur within the next few generations. Yet, determinations of current climatic trends from recent data are made difficult by natural variability at shorter time scales, such as the El Niño phenomenon. Many scientists are attempting to understand the mechanisms of change during an El Niño event from improved global measurements so as to determine how the ocean-atmosphere engine operates at longer time scales. Others are studying prehistoric records preserved in trees, sediments, and fossil corals in an effort to reconstruct past variations, including those like the El Niño. Their aim is to remove such short-term variations so as to be able to make more accurate estimates of long-term trends.

While the ocean basins lie much lower than sea level, the continents stand high—about one kilometre above sea level. The physical explanation for this condition is that the continental crust is light and thick, whereas the oceanic crust is dense and thin. Both the continental and oceanic crust lie over a more uniform layer called the mantle. As an analogy, one can think of a thick piece of styrofoam and a thin piece of wood floating in a tub of water. The styrofoam rises higher out of the water than the wood.

The ocean basins are transient features over geologic time, changing shape and depth while the process of plate tectonics proceeds. The surface layer of the Earth, the lithosphere, consists of a number of rigid plates that are in continual motion. The boundaries between the lithospheric plates form the principal relief features of the ocean basins: the crests of oceanic ridges are spreading centres where two plates move apart from each other at a rate of several centimetres per year. Molten rock material wells up from the underlying mantle into the gap between the diverging plates and solidifies into oceanic crust, thereby creating new ocean floor. At the deep-sea trenches, two plates converge, with one plate sliding down under the other into the mantle where it is melted. Thus, for each segment of new ocean floor created at the ridges, an equal amount of old oceanic crust is destroyed at the trenches, or so-called subduction zones (see below Deep-sea trenches (ocean) and also the article plate tectonics). It is for this reason that the oldest segment of ocean floor, found in the far western Pacific, is apparently only about 200 million years old, even though the age of the Earth is estimated to be at least 4.6 billion years.

The dominant factors that govern seafloor relief and topography are the thermal properties of the oceanic plates, tensional forces in the plates, volcanic activity, and sedimentation. In brief, the oceanic ridges rise about two kilometres above the seafloor because the plates near these spreading centres are warm and thermally expanded. In contrast, plates in the subduction zones are generally cooler. Tensional forces resulting in plate divergence at the spreading centres also create block-faulted mountains and abyssal hills, which trend parallel to the oceanic ridges. Seamounts and guyots, as well as abyssal hills and most aseismic ridges, are produced by volcanism. Continuing sedimentation throughout the ocean basin serves to blanket and bury many of the faulted mountains and abyssal hills with time. Erosion plays a relatively minor role in shaping the face of the deep seafloor, in contrast to the continents. This is because deep ocean currents are generally slow (they flow at less than 50 centimetres per second) and lack sufficient power.

Mapping the characteristics of the ocean basin has been difficult for several reasons. First, the oceans are not easy to travel over; second, until recent times navigation has been extremely crude, so that individual observations have been only loosely correlated with one another; and, finally, the oceans are opaque to light—i.e., the deep seafloor cannot be seen from the ocean surface. Modern technology has given rise to customized research vessels, satellite and electronic navigation, and sophisticated acoustic instruments that have mitigated some of these problems.

The Challenger Expedition, mounted by the British in 1872–76, provided the first systematic view of a few of the major features of the seafloor. Scientists aboard the HMS Challenger determined ocean depths by means of wire-line soundings and discovered the Mid-Atlantic Ridge. Dredges brought up samples of rocks and sediments off the seafloor. The main advance in mapping, however, did not occur until sonar was developed in the early 20th century. This system for detecting the presence of objects underwater by acoustic echo provided marine researchers with a highly useful tool, since sound can be detected over several thousands of kilometres in the ocean (visible light, by comparison, can only penetrate 100 metres or so of water).

Modern sonar systems include the Seabeam multibeam echo sounder and the GLORIA scanning sonar (see undersea exploration: Methodology and instrumentation: Exploration of the seafloor and the Earth's crust (undersea exploration)). They operate on the principle that the depth (or distance) of the seafloor can be determined by multiplying one-half the elapsed time between a downgoing acoustic pulse and its echo by the speed of sound in seawater (about 1,500 metres per second). Such multifrequency sonar systems permit the use of different pulse frequencies to meet different scientific objectives. Acoustic pulses of 12 kilohertz (kHz), for example, are normally employed to measure ocean depth, while lower frequencies—3.5 kHz to less than 100 hertz (Hz)—are used to map the thickness of sediments in the ocean basins. Very high frequencies of 100 kHz or more are employed in side-scanning sonar to measure the texture of the seafloor. The acoustic pulses are normally generated by piezoelectric transducers. For determining subbottom structure, low-frequency acoustic pulses are produced by explosives, compressed air, or water-jet implosion. Near-bottom sonar systems, such as the Deep Tow of the Scripps Institution of Oceanography (in La Jolla, Calif., U.S.), produce even more detailed images of the seafloor and subbottom structure. The Deep Tow package contains both echo sounders and side-scanning sonars, along with associated geophysical instruments, and is towed behind a ship at slow speed 10 to 100 metres above the seafloor. It yields very precise measurements of even finer-scale features than are resolvable with Seabeam and other comparable systems.

Another notable instrument system is ANGUS, a deep-towed camera sled that can take thousands of high-resolution photographs of the seafloor during a single day. It has been successfully used in the detection of hydrothermal vents at spreading centres (see below Oceanic ridges (ocean)). Overlapping photographic images make it possible to construct photomosaic strips about 10–20 metres wide that reveal details on the order of centimetres.

Three major navigation systems are in use in modern marine geology. These include electromagnetic systems such as loran and Earth-orbiting satellites (see undersea exploration: Basic elements of undersea exploration: Navigation (undersea exploration)). Acoustic transponder arrays of two or more stations placed on the seafloor a few kilometres apart are used to navigate deeply towed instruments, submersibles, and occasionally surface research vessels when detailed mapping is conducted in small areas. These systems measure the distance between the instrument package and the transponder sites and, using simple geometry, compute fixes accurate to a few metres. Although the individual transponders can be used to determine positions relative to the array with great accuracy, the preciseness of the position of the array itself depends on which system is employed to locate it.

The oceanic crust differs from the continental crust in several ways: it is thinner, denser, younger, of different chemical composition, and created in a different plate-tectonic setting. The oceanic crust is formed at spreading centres on the oceanic ridges, whereas continental crust is formed above the subduction zones. The oceanic crust is about six kilometres thick. It is composed of several layers, not including the overlying sediment. The topmost layer, about 500 metres thick, includes lavas (lava) of basaltic composition (i.e., rock material consisting largely of plagioclase 【feldspar】 and pyroxene). The lavas are generally of two types: pillow lavas and sheet flows. Pillow lavas appear to be shaped exactly as the name implies—like large overstuffed pillows about one metre in cross section and one to several metres long. They commonly form small hills tens of metres high at the spreading centres. Sheet flows have the appearance of wrinkled bed sheets. They commonly are thin (only about 10 centimetres thick) and cover a broader area than pillow lavas. There is evidence that sheet flows are erupted at higher temperatures than those of the pillow variety. On the East Pacific Rise at 8° S latitude, a series of sheet flow eruptions (possibly since the mid-1960s) have covered more than 220 square kilometres of seafloor to an average depth of 70 metres.

Below the lava is a layer composed of feeder, or sheeted, dikes that measures more than one kilometre thick. Dikes are fractures that serve as the plumbing system for transporting magmas (molten rock material) to the seafloor to produce lavas. They are about one metre wide, subvertical, and elongate along the trend of the spreading centre where they formed, and they abut one another's sides—hence the term sheeted. These dikes are also of basaltic composition. There are two layers below the dikes totaling about 4.5 kilometres in thickness. Both of these include gabbros (gabbro), which are essentially basalts with coarser mineral grains. These gabbro layers are thought to represent the magma chambers, or pockets of lava, that ultimately erupt on the seafloor. The upper gabbro layer is isotropic (uniform) in structure. In some places, this layer includes pods of plagiogranite, a differentiated rock richer in silica than gabbro. The lower gabbro layer has a stratified structure and evidently represents the floor or sides of the magma chamber. This layered structure is called cumulate, meaning that the layers (which measure up to several metres thick) result from the sedimentation of minerals out of the liquid magma. The layers in the cumulate gabbro have less silica but are richer in iron and magnesium than the upper portions of the crust. olivine, an iron-magnesium silicate, is a common mineral in the lower gabbro layer.

The oceanic crust lies atop the Earth's mantle, as does the continental crust. Mantle rock is composed mostly of peridotite, which consists primarily of the mineral olivine with small amounts of pyroxene and amphibole.

Knowledge of the structure and composition of the oceanic crust comes from several sources. Bottom sampling during early exploration brought up all varieties of the above-mentioned rocks, but the structure of the crust and the abundance of the constituent rocks were unclear. Simultaneously, seismic refraction experiments enabled researchers to determine the layered nature of the oceanic crust. These experiments involve measuring the travel times of seismic waves (seismic wave) generated by explosions (e.g., dynamite blasts) set off over distances of several tens of kilometres. The results of early refraction experiments revealed the existence of two layers beneath the sediment cover. More sophisticated experiments and analyses led to dividing these layers into two parts, each with a different seismic wave velocity, which increases with depth. The seismic velocity is a kind of fingerprint that can be attributed to a limited number of rock types. Sampled rock data and seismic results were combined to yield a model for the structure and composition of the crust.

Great strides in understanding the oceanic crust were made by the study of ophiolites. These are slices of the ocean floor that have been thrust above sea level by the action of plate tectonics. In various places in the world, the entire sequence of oceanic crust and upper mantle is exposed. These areas include, among others, Newfoundland and the Pacific Coast Ranges of California, the island of Cyprus in the Mediterranean Sea, and the mountains in Oman on the southeastern tip of the Arabian Peninsula. Ophiolites reveal the structure and composition of the oceanic crust in astonishing detail. Also, the process of crustal formation and hydrothermal circulation, as well as the origin of marine magnetic anomalies (see below), can be studied with comparative clarity. Although it is clear that ophiolites are of marine origin, there is some controversy as to whether they represent typical oceanic crust or crust formed in settings other than an oceanic spreading centre—behind island arcs, for example.

The age of the oceanic crust does not go back farther than about 200 million years. Such crust is being formed today at oceanic spreading centres. Many ophiolites are much older than the oldest oceanic crust, demonstrating continuity of the formation processes over hundreds of millions of years. Methods that may be used to determine the age of the crustal material include direct dating of rock samples by radiometric dating (measuring the relative abundances of a particular radioactive isotope and its daughter isotopes in the samples) or by the analyses of fossil evidence, marine magnetic anomalies, or ocean depth. Of these, magnetic anomalies deserve special attention.

A marine magnetic anomaly is a variation in strength of the Earth's magnetic field caused by magnetism in rocks of the ocean floor. Marine magnetic anomalies typically represent 1 percent of the total geomagnetic field strength. They can be stronger (“positive”) or weaker (“negative”) than the average total field. Also, the magnetic anomalies occur in long bands that run parallel to spreading centres for hundreds of kilometres and may reach up to a few tens of kilometres in width.

Marine magnetic anomalies were first discovered off the coast of the western United States in the late 1950s and completely baffled scientists. The anomalies were charted from southern California to northern Washington and out several hundred kilometres. Victor Vacquier, a geophysicist, noticed that these linear anomalies ended at the fracture zones mapped in this area. In addition, he noticed that they had unique shapes, occurred in a predictable sequence across their trends, and could be correlated across the fracture zones. Soon thereafter, linear magnetic anomalies were mapped over the Reykjanes Ridge south of Iceland. They were found to occur on both sides of the ridge crest and parallel to it. Simultaneously, Alan Cox and several other American geophysicists documented evidence that the Earth's magnetic field had reversed in the past: the north magnetic pole had been the south magnetic pole about 700,000 years ago, and there were reasons to believe older reversals existed. Also at this time, Robert S. Dietz and Harry H. Hess were formulating the theory of seafloor spreading—the hypothesis (seafloor spreading hypothesis) that oceanic crust is created at the crests of the oceanic ridges and consumed in the deep-sea trenches.

It remained for Frederick J. Vine and Drummond H. Matthews of Great Britain and Lawrence W. Morley of Canada to put these observations together in a theory that explained marine magnetic anomalies. The theory rests on three assumptions: (1) that the Earth's magnetic field periodically reverses (geomagnetic reversal) polarity; (2) that seafloor spreading occurs; and (3) that the oceanic crust is permanently magnetized as it forms and cools at spreading centres. The theory expresses the assumptions—namely, that the oceanic crust records reversals of the Earth's field as it is formed during seafloor spreading. Positive anomalies result when the crust is magnetized in a “normal” polarity parallel to the ambient field of the Earth, and negative anomalies result when the crust is “reversely” magnetized in an opposite sense. As the magnetized crust moves down the flanks of a ridge away from the spreading centre, it remains permanently magnetized and “carries” the magnetic anomalies along with it. (For further details about paleomagnetism and seafloor spreading, see plate tectonics: Historical overview: Renewed interest in continental drift (plate tectonics).)

A brilliant leap in understanding was now possible. If the age of the field reversals were known, the age of the ocean crust could be predicted by mapping the corresponding anomaly. By the mid-1960s, Cox and his colleagues had put together a schedule of reversals for the last four or five million years by studying the ages and magnetic polarities of lava flows found on land. Vine and the Canadian geologist J. Tuzo Wilson applied the time scale to marine magnetic anomalies mapped over the Juan de Fuca Ridge, a spreading centre off the northwest United States. They thus dated the crust there and also computed the first seafloor spreading rate of about 30 millimetres per year. The rate is computed by dividing the distance of an anomaly from the ridge crest by the age of the anomaly twice. Thus the oceanic crust at the Juan de Fuca Ridge is moving at about 15 millimetres per year away from the ridge crest and at about 60 millimetres per year away from the crustal segment on the opposite side of the crest.

During the 1960s and '70s marine magnetic anomalies were mapped over wide areas of the ocean basins. By using estimates of the ages of oceanic crust obtained from core samples by deep-sea drilling, a magnetic anomaly time scale was constructed, and at the same time the spreading history for the ocean basins covering the last 200 million years or so was proposed.

It is thought that the most important contributor to marine magnetic anomalies is the layer of lavas in the upper oceanic crust. A secondary contribution originates in the upper layer of gabbros. The dike layer is essentially demagnetized by the action of hydrothermal waters at the spreading centres. The dominant mechanism of permanent magnetization is the thermoremanent magnetization (or TRM) of iron-titanium oxide minerals. These minerals lock in a TRM as they cool below 200° to 300° C in the presence of the Earth's magnetic field. Although several processes are capable of altering the TRM, including reheating and oxidation at the seafloor, it is remarkably robust, as is evidenced by magnetic anomalies as old as 165 million years in the far western equatorial Pacific.

The largest features of the ocean basin are the oceanic ridges. In the past these features were referred to as mid-ocean ridges, but, as will be seen, the largest oceanic ridge, the East Pacific Rise, is far from a mid-ocean location, and the nomenclature is thus inaccurate. Oceanic ridges are not to be confused with aseismic ridges, which have an entirely different origin (see below).

Oceanic ridges are linear mountain chains comprising the largest features on Earth. They are found in every ocean basin and appear to girdle the Earth. The ridges rise from depths near 5 kilometres to an essentially uniform depth of about 2.6 kilometres and are roughly symmetrical in cross section. They can be thousands of kilometres wide. In places, the crests of the ridges are offset across transform faults, or fracture zones, which can be followed down the flanks of the ridges. (Transform faults are those along which lateral movement occurs.) The flanks are marked by sets of mountains and hills that are elongate and parallel to the ridge trend.

New oceanic crust (and part of the upper mantle, which, together with the crust, makes up the lithosphere) is formed at seafloor spreading centres at the crests of the oceanic ridges. Because of this, certain unique geologic features are found there. Fresh basaltic lavas are exposed on the seafloor at the ridge crests. These lavas are progressively buried by sediments as the seafloor spreads away from the site. The flow of heat out of the crust is many times greater at the crests than elsewhere in the world. Earthquakes are common along the crests and in the transform faults that join the offset ridge segments. Analysis of earthquakes occurring at the ridge crests indicates that the oceanic crust is under tension there. A high-amplitude magnetic anomaly is centred over the crests because fresh lavas at the crests are being magnetized in the direction of the present geomagnetic field.

The depths over the oceanic ridges are rather precisely correlated with the age of the ocean crust; specifically, it has been demonstrated that the ocean depth is proportional to the square root of crustal age. The theory explaining this relationship holds that the increase in depth with age is due to the thermal contraction of the oceanic crust and upper mantle as they are carried away from the seafloor spreading centre in an oceanic plate. Because such a plate is ultimately about 100 kilometres thick, contraction of only a few percent predicts the entire relief of an oceanic ridge. It then follows that the width of a ridge can be defined as twice the distance from the crest to the point where the plate has cooled to a steady thermal state. Most of the cooling takes place within 70 or 80 million years, by which time the ocean depth is about 5 to 5.5 kilometres. Because this cooling is a function of age, slow-spreading ridges, such as the Mid-Atlantic Ridge, are narrower than faster-spreading ridges, like the East Pacific Rise (see below). Further, a correlation has been found between global spreading rates and the transgression and regression of ocean waters onto the continents. During the Early Cretaceous period about 100 million years ago, when global spreading rates were uniformly high, oceanic ridges occupied comparatively more of the ocean basins, causing the ocean waters to transgress (spill over) onto the continents, leaving marine sediments in areas now well away from coastlines.

Besides ridge width, other features appear to be a function of spreading rate. Global spreading rates range from 10 millimetres per year (mm/yr total rate) or less up to 160 mm/yr. Oceanic ridges can be classified as slow (up to 50 mm/yr), intermediate (up to 90 mm/yr), and fast (up to 160 mm/yr). Slow-spreading ridges are characterized by a rift valley at the crest. Such a valley is fault-controlled. It is typically 1.4 kilometres deep and 20 to 40 kilometres wide. Faster-spreading ridges lack rift valleys. At intermediate rates, the crest regions are broad highs with occasional fault-bounded valleys no deeper than 200 metres. At fast rates, an axial high is present at the crest. The slow-spreading rifted ridges have rough faulted topography on their flanks, while the faster-spreading ridges have much smoother flanks.

Oceanic spreading centres are found in all the ocean basins. In the Arctic Ocean a slow-rate spreading centre is located near the eastern side in the Eurasian basin. It can be followed south, offset by transform faults, to Iceland. Iceland has been created by a hot spot (see below) located directly below an oceanic spreading centre. The ridge leading south from Iceland is named the Reykjanes Ridge, and, although it spreads at 20 mm/yr or less, it lacks a rift valley. This is thought to be the result of the influence of the hot spot.

The Mid-Atlantic Ridge extends from south of Iceland to the extreme South Atlantic Ocean near 60° S latitude. It bisects the Atlantic Ocean basin, which led to the earlier designation of mid-ocean ridge for features of this type. The Mid-Atlantic Ridge became known in a rudimentary fashion during the 19th century. In 1855 Matthew Fontaine Maury (Maury, Matthew Fontaine) of the U.S. Navy prepared a chart of the Atlantic in which he identified it as a shallow “middle ground.” During the 1950s the American oceanographers Bruce Heezen and Maurice Ewing proposed that it was a continuous mountain range.

In the North Atlantic the ridge spreads slowly and displays a rift valley and mountainous flanks. In the South Atlantic spreading rates are between slow and intermediate, and rift valleys are generally absent, as they occur only near transform faults.

A very slow oceanic ridge, the Southwest Indian Ridge, bisects the ocean between Africa and Antarctica. It joins the Mid-Indian and Southeast Indian ridges east of Madagascar. The Carlsberg Ridge is found at the north end of the Mid-Indian Ridge. It continues north to join spreading centres in the Gulf of Aden and Red Sea. Spreading is very slow at this point but approaches intermediate rates on the Carlsberg and Mid-Indian ridges. The Southeast Indian Ridge spreads at intermediate rates. This ridge continues from the western Indian Ocean in a southeasterly direction, bisecting the ocean between Australia and Antarctica. Rifted crests and rugged mountainous flanks are characteristic of the Southwest Indian Ridge. The Mid-Indian Ridge has fewer features of this kind, and the Southeast Indian Ridge has generally smoother topography. The latter also displays distinct asymmetric seafloor spreading south of Australia. Analysis of magnetic anomalies shows that rates on opposite sides of the spreading centre have been unequal at many times over the past 50 or 60 million years.

The Pacific-Antarctic Ridge can be followed from a point midway between New Zealand and Antarctica northeast to where it joins the East Pacific Rise off the margin of South America. The former spreads at intermediate to fast rates.

The East Pacific Rise extends from this site northward to the Gulf of California, where it joins the transform zone of the Pacific-North American plate boundary. Offshore from Chile and Peru, the East Pacific Rise is currently spreading at fast rates of 159 mm/yr or more. Rates decrease to about 60 mm/yr at the mouth of the Gulf of California. The crest of the ridge displays a low topographic rise along its length rather than a rift valley. The East Pacific Rise was first detected during the Challenger Expedition of the 1870s. It was described in its gross form during the 1950s and '60s by oceanographers, including Heezen, Ewing, and Henry W. Menard. During the 1980s, Kenneth C. Macdonald, Paul J. Fox, and Peter F. Lonsdale discovered that the main spreading centre appears to be interrupted and offset a few kilometres to one side at various places along the crest of the East Pacific Rise. However, the ends of the offset spreading centres overlap each other by several kilometres. These were identified as a new type of geologic feature of oceanic spreading centres and designated overlapping spreading centres. Such centres are thought to result from interruptions of the magma supply to the crest along its length and define a fundamental segmentation of the ridge on a scale of tens to hundreds of kilometres.

Many smaller spreading centres branch off the major ones or are found behind island arcs. In the western Pacific, spreading centres occur on the Fiji Plateau between the New Hebrides and Fiji Islands and in the Woodlark Basin between New Guinea and the Solomon Islands. A series of spreading centres and transform faults lie between the East Pacific Rise and South America near 40° to 50° S latitude. The Scotia Sea between South America and the Antarctic Peninsula contains a spreading centre. The Galápagos spreading centre trends east-west between the East Pacific Rise and South America near the Equator. Three short spreading centres are found a few hundred kilometres off the shore of the Pacific Northwest. These are the Gorda Ridges off northern California, the Juan de Fuca Ridge off Oregon and Washington, and the Explorer Ridge off Vancouver Island. In a careful study of the seafloor spreading history of the Galápagos and the Juan de Fuca spreading centres, the American geophysicist Richard N. Hey developed the idea of the propagating rift. In this phenomenon, one branch of a spreading centre ending in a transform fault lengthens at the expense of the spreading centre across the fault. The rift and fault propagate at one to five times the spreading rate and create chevron patterns in magnetic anomalies and the grain of the seafloor topography resembling the wake of a boat.

From the 1970s highly detailed studies of spreading centres using deeply towed instruments, photography, and manned submersibles have resulted in new revelations about the processes of seafloor spreading. The most profound discoveries have been of deep-sea hydrothermal vents (see below) and previously unknown biological communities.

Spreading centres are divided into several geologic zones. The neovolcanic zone is at the very axis. It is 1 to 2 kilometres wide and is the site of recent and active volcanism and of the hydrothermal vents. It is marked by chains of small volcanoes or volcanic ridges. Adjacent to the neovolcanic zone is one marked by fissures in the seafloor. This may be 1 to 2 kilometres wide. Beyond this point occurs a zone of active faulting (fault). Here, fissures develop into normal faults with vertical offsets. This zone may be 10 or more kilometres wide. At slow spreading rates the faults have offsets of hundreds of metres, creating rift valleys and rift mountains. At faster rates the vertical offsets are 50 metres or less. A deep rift valley is not formed because the vertical uplifts are cancelled out by faults that downdrop uplifted blocks. This results in linear, fault-bounded abyssal hills and valleys trending parallel to the spreading centre.

Warm springs emanating from the seafloor in the neovolcanic zone were first found on the Galápagos spreading centre. These waters were measured to have temperatures about 20° C above the ambient temperature. In 1979 hydrothermal vents with temperatures near 350° C were discovered on the East Pacific Rise off Mexico. Since then, similar vents have been found on the spreading centres off the Pacific Northwest coast of the United States, on the south end of the northern Mid-Atlantic Ridge, and at many locations on the East Pacific Rise.

Hydrothermal vents are localized discharges of heated seawater. They result from cold seawater percolating down into the hot oceanic crust through the zone of fissures and returning to the seafloor in a pipelike flow at the axis of the neovolcanic zone. The heated waters often carry sulfide minerals of zinc, iron, and copper leached from the crust. Outflow of these heated waters probably accounts for 20 percent of the Earth's heat loss. Exotic biological communities exist around the hydrothermal vents. These ecosystems are totally independent of energy from the Sun. They are not dependent on photosynthesis but rather on chemosynthesis by sulfur-fixing bacteria. The sulfide minerals precipitated in the neovolcanic zone can accumulate in substantial amounts and are sometimes buried by lava flows at a later time. Such deposits are mined as commercial ores in ophiolites on Cyprus and in Oman.

Magma chambers have been detected beneath the crest of the East Pacific Rise by seismic experiments. (The principle underlying the experiments is that partially molten or molten rock slows the travel of seismic waves and also strongly reflects them.) The depth to the top of the chambers is about two kilometres below the seafloor. The width is more difficult to ascertain, but is probably one to four kilometres. Their thickness seems to be about two to six kilometres based on studies of ophiolites. The chambers have been mapped along the trend of the crest between 9° and 13° N latitude. The top is relatively continuous, but is apparently interrupted by offsets of transform faults and overlapping spreading centres.

Fracture zones occur less frequently on the East Pacific Rise, but they offset the ridge by a greater amount. More than a dozen can be found between 20° N and 30° S. Typical offsets are roughly 100 kilometres. Several fracture zones more than 3,000 kilometres long are found off the shore of western North America. These include the Mendocino, Murray, Molokai, and Clarion fracture zones. They are not associated with a ridge crest. Rather, they occur on the west flank of the defunct Pacific-Farallon oceanic ridge. The Farallon Plate has all but disappeared down a subduction zone that extended along the entire coast of California and Baja California until about 25 to 30 million years ago. Subduction now occurs north of the Mendocino Fracture Zone. These fracture zones off western North America were among the first mapped. Menard has traced them almost 10,000 kilometres westward across the Pacific. The continental margin of northern California is displaced to the right where the Mendocino Fracture Zone and its transform portion, the Gorda Escarpment, intersect it.

The portion of a fracture zone between different offset spreading centres constitutes a transform fault. Transform faults also connect spreading centres to subduction zones (deep-sea trenches). Faults of this kind are the only segments of fracture zones that are seismically active. J. Tuzo Wilson (Wilson, J. Tuzo) recognized this and other features and explained the phenomenon as a transfer of motion from one spreading centre to another. The American geologist W. Jason Morgan, one of the several outstanding pioneers in plate tectonics, recognized that transform faults are zones where opposing lithospheric plates slip past one another. Morgan proposed that opposing plates along an oceanic ridge crest offset by fracture zones are divided by the spreading centres and transform faults. The inactive portions of the fracture zone on the ridge flanks are scars on the ocean floor created in the transform faults.

Not everywhere in the ocean basins are plate motions exactly parallel to transform faults. In places where a component of opening motion occurs across the transform, volcanic activity results, and the fracture zone is termed a leaky transform fault. South of New Zealand, between it and the Pacific-Antarctic Ridge, a component of shortening is occurring across a transform called the Macquarie Ridge. Here, subduction may be taking place at a slow rate.

The Pacific basin is rimmed by trenches of both marginal and island arc varieties. Marginal trenches bound the west side of Central and South America from the Gulf of California to southern Chile. Although they are deeply buried in sediment, trenches are found along the western North American continental margin from Cape Mendocino (in northern California) to the Canadian border. The Aleutian Trench extends from the northernmost point in the Gulf of Alaska west to the Kamchatka Peninsula in the Soviet Union. It can be classified as a marginal trench in the east but is more properly termed an island arc west of Alaska.

In the western Pacific, the trenches are associated with island arcs. These include the Kuril, Japan, Bonin, Mariana, Ryukyu, and Philippine trenches that extend from Kamchatka to near the Equator. A complex pattern of island arcs is found in Indonesia. The major island arc here is the Java Trench extending from northern Australia to the northwestern end of Sumatra in the northeast Indian Ocean. The region of New Guinea and the Solomon Islands includes the New Britain and Solomon trenches, the latter of which joins the New Hebrides Trench directly to the south. East of this area the Tonga and Kermadec trenches extend south from the Fiji Islands to New Zealand.

Two island arcs occur in the Atlantic Ocean. The South Sandwich Trench is located west of the Mid-Atlantic Ridge between South America and Antarctica. The Puerto Rico Trench joins the Lesser Antilles Island arc in the eastern Caribbean.

Some seafloor features bear the name trench and are deep linear troughs but are not subduction zones. The Vema Trench on the Mid-Indian Ridge is a fracture zone. The Vityaz Trench northwest of Fiji is an aseismic (inactive) feature of unknown origin. The Diamantina trench (Diamantina Fracture Zone) extends westward from the southwest coast of Australia. It is a rift valley that was formed when Australia separated from Antarctica between 60 and 50 million years ago.

The deepest water on Earth (11,034 metres) is located in the southern end of the Mariana Trench near Guam. A few trenches are partially filled with sediments derived from the bordering continents. The Aleutian Trench is effectively buried east of Kodiak Island in the Gulf of Alaska. Here, the ocean floor is smooth and flat. To the west farther from the sediment supply on Alaska, the trench reaches depths of more than seven kilometres. The Lesser Antilles trench in the eastern Caribbean also is buried by sediments originating from South America.

Oceanward of trenches the seafloor is usually bulged upward in an outer ridge or rise of up to 1,000 metres relief. This condition is thought to be the elastic response of the oceanic plate bending down into a subduction zone. The landward or island-arc slope of the trench is often interrupted by a submarine ridge, which sometimes breaks the ocean surface, as in the case of the Java Trench. Such a ridge is constructed from deformed sediments scraped off the top of the descending oceanic plate and is termed an accretionary prism. A line of explosive volcanoes (volcano), extruding (erupting) a lava that forms the volcanic rock andesite, is found on the overriding plate usually 100 kilometres or so from the trench. In marginal trenches these volcanoes form mountain chains, such as the Cascades in the Pacific Northwest or the great volcanoes of the Andes. In island arcs they form active volcanic island chains, such as the Mariana Islands.

Behind the volcanic line of island arcs are sometimes found young, narrow ocean basins. These basins are bounded on the opposite side by submarine ridges. Such interarc, or backarc, basins are sites of seafloor spreading directly caused by the dynamics of subduction. They originate at the volcanic line, so that the outer bounding submarine ridge, or third arc, represents an older portion of the volcanic line that has spread away. These backarc basins bear many of the features characteristic of oceanic spreading centres. Well-studied examples of these features are found in the Lau Basin of the Tonga arc and also west of the Mariana Islands. The Sea of Japan originated from backarc spreading behind the Japanese arc that began some 30 million years ago. At least two backarc basins have opened behind the Mariana arc, creating seafloor in two phases from about 30 to 17 million years ago in the western Parece Vela Basin and from 5 million years ago in the Mariana Trough next to the islands.

In some oceans the basin floors are crossed by long, linear and mountainous aseismic ridges. The term aseismic distinguishes these ridges from oceanic spreading centres because the former lack earthquakes. Most aseismic ridges are constructed by volcanism from a hot spot and are composed of coalescing volcanoes of various sizes. A hot spot is a magma-generating centre fixed in the Earth's deep mantle and leaves a trail of volcanic outpourings on the seafloor as an oceanic plate travels over it. This form of volcanism is not associated with the volcanism at spreading centres and is distinct from it chemically in that the magma extruded onto the surface has a higher alkali composition. (For additional information on hot spots, seevolcano: Volcanism and tectonic activity: Intraplate volcanism (volcano).)

The Hawaiian-Emperor chain is the best displayed aseismic ridge. Earthquakes do occur here, but only at the end of the ridge where volcanism is current—in this case, on the island of Hawaii (commonly known as the Big Island) to the southeast end of the island chain. Taking into account the relief of the island of Hawaii above the seafloor, it is the largest volcanic edifice on Earth. The Hawaiian-Emperor chain stretches from the Big Island to the intersection of the Kuril and Aleutian trenches in the northwest Pacific. There are roughly 18 volcanoes or seamounts (see below) per 1,000 kilometres along the Hawaiian segment and 13 per 1,000 kilometres on the Emperor portion beyond the bend. The Hawaiian Islands are a part of the chain—the young part—that rises above sea level. The Hawaiian-Emperor chain has two main trends: (1) from the Hawaiian Islands west to the Kammu and Yūryaku seamounts (near 32° N, 168° W), the trend of the Hawaiian portion is just west of northwest; and (2) from this point to the Aleutian Trench, the trend of the Emperor segment is north-northwest. The hot spot interpretation infers that this change in trend is due to a change in the direction of Pacific Plate motion, from north-northwest prior to 38 million years ago (the age of the ridge at the change in trend) to west of northwest until the present day. Radiometric dating of rocks from the ridge indicates that it is 70 million years old at its extreme north end.

Other prominent aseismic ridges include the Ninetyeast Ridge and the Chagos-Laccadive Plateau in the Indian Ocean and the Walvis Ridge and Rio Grande Rise in the South Atlantic. The Ninetyeast Ridge is thought to have originated from hot spot volcanic activity now located at the Kerguelen Islands near Antarctica. These islands lie atop the Kerguelen Plateau, which also originated from volcanism at this hot spot. The Ninetyeast Ridge stretches parallel to 90° E longitude in a long, linear chain of seamounts and volcanic ridges from the Andaman Islands in the Bay of Bengal more than 4,500 kilometres to the south where it intersects Broken Ridge at 30° S latitude. Broken Ridge is an aseismic ridge and was once part of the Kerguelen Plateau. It was split away from the plateau as Australia separated from Antarctica.

Core samples of the seafloor along the Ninetyeast Ridge have been retrieved through deep-sea drilling. Analyses of the samples show that the ridge is slightly less than 30 million years old in the south and about 80 million years old in the north. Additionally, sediments on the ridge indicate that parts of it were above sea level while it was being built near a spreading centre. The ridge then subsided as it rode north on the Indian Plate.

The Walvis Ridge and Rio Grande Rise originated from hot spot volcanism now occurring at the islands of Tristan da Cunha 300 kilometres east of the crest of the Mid-Atlantic Ridge. The Walvis Ridge trends northeast from this location to the African margin. The Rio Grande Rise trends roughly southeast from the South American margin toward the Mid-Atlantic Ridge. Both the Walvis Ridge and Rio Grande Rise began forming from the same hot spot near the spreading centre as the South Atlantic was in its initial opening stages 100 to 80 million years ago. The spreading centre shifted west of the hot spot about 80 million years ago, ending construction of the Rio Grande Rise but continuing to build the Walvis Ridge. Volcanic activity has since diminished, resulting in the younger part of the latter ridge being smaller. The findings of ocean drilling on the Rio Grande Rise show that it was once a volcanic island some two kilometres high.

Seamounts are submarine volcanoes with more than 1,000 metres of relief. Aseismic ridges are built by chains of overlapping seamounts. A seamount is akin to a subaerial shield volcano in that it also has gently sloping sides (5° to 15°) and is constructed by nonexplosive eruptions of alkaline basalt lavas that are thought to originate from depths of roughly 150 kilometres. About 2,000 seamounts are known; they are most common in the Pacific and on fast-spreading ridges. Like the Hawaiian-Emperor chain, the lines of seamounts and islands trending northwest-southeast in the central and south Pacific (Marshall Islands, Line Islands, Tuamotu Archipelago, and Cook and Austral Islands) may be due to hot spot volcanism. Isolated seamounts also occur, and many of these are located in the far western Pacific. Another group of smaller seamounts is found in the northeastern Pacific.

Flat-topped seamounts are called guyots (guyot). They are particularly abundant in the western Pacific and along the Emperor seamount chain. Bottom samples and drill cores of shallow-water sediments and fossils capping guyots have been retrieved. The presence of such geologic materials suggest that guyots are seamounts that have had their peaks planed off by wave action and have since subsided below sea level. The western Pacific guyots are capped by drowned coral atolls and reefs (coral reef). These reefs are generally of Late Cretaceous age (about 95 million years old). The cause of the subsidence is attributed to the sinking of the seafloor as it moves down the flanks of an oceanic ridge. However, the reason for the demise of the coral reefs on the Cretaceous guyots is less clear. Under normal conditions, coral growth can easily keep up with sinking due to seafloor spreading. The Cretaceous guyots may have resulted from the northward drift of seamounts and reefs on the Pacific Plate away from the tropical zone of favourable growth. Another hypothesis is that the reefs were killed by unusually anoxic (oxygen-depleted) conditions that developed suddenly, a situation possibly related to intense seafloor volcanism in the Pacific at this time.

Abyssal hills (abyssal hill) are low-relief (less than 1 kilometre) features usually 1 to 10 kilometres wide and elongate parallel to spreading centres or to marine magnetic anomalies located in the vicinity of the latter. The tops of the hills are often flat, in which case they have steep sides. Gently sloping sides, however, are equally common. Abyssal hills are extremely numerous, so much so that Menard declared them “the most widespread physiographic forms of the face of the earth.” Abyssal hills are most common in the Pacific basin, where they cover 80 to 85 percent of the seafloor. Because they cover the entire flanks and crests of the oceanic ridges, such hills are thought to form during crustal accretion at spreading centres. They are commonly associated with intermediate- and fast-spreading ridges. On slow-spreading ridges, such as the Mid-Atlantic, the topographic features are much larger and have steeper sides. Bottom-sampling and seismic reflection studies reveal that abyssal hills are relief features on the top of the oceanic crust; they are not constructed from ocean-bottom sediments. In areas such as the abyssal plains (see below), abyssal hills are buried by sediments.

Apparently the hills are constructed by two processes: volcanism and block faulting. The relative contribution of each may depend on the spreading rate. At slower rates, faulting of the oceanic crust is a dominant factor in forming the relief, and the relief of the hills is greater as the rate is slower. At the crest of a spreading centre, volcanism in the neovolcanic zone initiates the construction of volcanic hills. The zone of active faulting is where they form or are modified by block faulting. The existence of discrete and separate volcanic hills indicates that volcanism at a spreading centre is episodic.

The ocean basin floor is everywhere covered by sediments of different types and origins. The only exception are the crests of the spreading centres where new ocean floor has not existed long enough to accumulate a sediment cover. Sediment thickness in the oceans averages about 450 metres. The sediment cover in the Pacific basin ranges from 300 to 600 metres thick, and that in the Atlantic is about 1,000 metres. Generally, the thickness of sediment on the oceanic crust increases with the age of the crust. Oceanic crust adjacent to the continents can be deeply buried by several kilometres of sediment. Deep-sea sediments can reveal much about the last 200 million years of Earth history, including seafloor spreading, the history of ocean life, the behaviour of the Earth's magnetic field, and the changes in the ocean currents and climate.

The study of ocean sediments has been accomplished by several means. Bottom samplers, such as dredges and cores up to 30 metres long, have been lowered from ships by wire to retrieve samples of the upper sediment layers. Deep-sea drilling has retrieved core samples from the entire sediment layer in several hundred locations in the ocean basins. The seismic reflection method has been used to map the thickness of sediments in many parts of the oceans. Besides thickness, seismic reflection data can often reveal sediment type and the processes of sedimentation. (For more information on the equipment and techniques used by investigators to study deep-sea sediments, see undersea exploration.)

Deep-sea sediments can be classified as terrigenous, originating from land; biogenic, consisting largely of the skeletal debris of microorganisms; and authigenic, formed in place on the seafloor. Pelagic sediments, either terrigenous or biogenic, are those that are deposited very slowly in the open ocean either by settling through the volume of oceanic water or by precipitation. The sinking rates of pelagic sediment grains are extremely slow because they ordinarily are no larger than several micrometres. However, fine particles are normally bundled into fecal pellets by zooplankton, which allows sinking at a rate of 40 to 400 metres per day.

Terrigenous sediments are transported to the oceans by rivers and wind. The sediments that reach the continental shelf are often stored in submarine canyons on the continental slope. Turbidity currents carry these sediments down into the deep sea (see above Density currents in the oceans: Turbidity currents (ocean)). These currents create sedimentary deposits called turbidites, which are layers up to several metres thick composed of sediment particles that grade upward from coarser to finer sizes. The turbidites build sedimentary deep-sea fans (submarine fan) adjacent to the base of the continental slope. Turbidites also are found below the major river deltas of the world where they build features called abyssal cones. The largest of these is the Ganges Fan (also called the Ganges Cone or Bengal Cone) in the Bay of Bengal east of the Indian subcontinent. It measures 3,000 kilometres long (north-south) by 1,000 kilometres wide (east-west) and is up to 12 kilometres thick. The Bengal Cone is forming from rock material eroded from the Himalayas and transported by the Ganges and Brahmaputra rivers.

Abyssal plains (abyssal plain) are formed by the accumulation of turbidites beyond the limits of deep-sea fans and abyssal cones in locations where there is a very large sediment supply. In contrast to fans and cones, abyssal plains are flat and featureless. They are prominent near both margins of the Atlantic and in the northeast Pacific. Tectonic and climate controls have influenced the formation of abyssal plains. The last major glaciation near the end of the Pleistocene epoch about 10,000 years ago greatly increased erosion and sediment supply to the deep sea, but deep-sea trenches interrupted the flow of turbidity currents to the ocean floor. Off the Pacific Northwest coast of the United States, however, the trenches were filled by turbidites, and subsequent turbidity currents passed beyond them to form the Alaska and Tufts abyssal plains.

Brown clays are a variety of pelagic sediment, mostly of terrigenous origin, which are composed largely of four different clay minerals: chlorite, illite, kaolinite, and montmorillonite. By definition, clays have less than 30 percent biogenic components. Quartz, volcanic ash, and micrometeorites are common as minor constituents. Brown clays are widespread in the deeper areas of the oceans below four kilometres. They dominate the floor of the central North Pacific. Clays accumulate very slowly, averaging about one millimetre per 1,000 years. The type of clay found in a given area is a function of the source region on land and the climate. For example, chlorite is dominant in polar regions and kaolinite in the tropics. Clays are introduced into the oceans by river transport, although kaolinite is also carried by the wind from the arid regions of Africa and Australia. montmorillonite is an alteration product of volcanic material and can form from either wind-blown volcanic ash or basaltic glass on the seafloor.

Sediments composed mostly or entirely of volcanic ash are commonly found adjacent to the island arcs and marginal trenches. These are normally deposited as turbidites. Volcanic ash that has been ejected higher than five kilometres during an eruption can be carried by wind and settle out through the atmosphere and oceans as pelagic sediment. The ocean floor encircling Antarctica is covered by glacial marine sediments. These sediments are carried by icebergs from the continent as far north as the Antarctic Convergence at 45° to 55° latitude.

Biogenic oozes are pelagic sediments that have more than 30 percent skeletal material. They can be either carbonate (or calcareous) ooze or siliceous ooze. The skeletal material in carbonate oozes is calcium carbonate usually in the form of the mineral calcite but sometimes aragonite. The most common contributors to the skeletal debris are such microorganisms as foraminiferans and coccoliths, microscopic carbonate plates that coat certain species of marine algae and protozoa. Siliceous oozes are composed of opal (amorphous, hydrated silica) that forms the skeleton of various microorganisms, including diatoms, radiolarians, siliceous sponges, and silicoflagellates. The distribution of biogenic oozes depends mainly on the supply of skeletal material, dissolution of the skeletons, and dilution by other sediment types, such as turbidites or clays.

Primary productivity in the ocean surface waters controls supply to a large extent. Productivity is high at the Equator and in zones of coastal upwelling and also where oceanic divergences occur near Antarctica. Productivity is lowest in the central areas of the oceans (the gyres) in both hemispheres. Siliceous oozes are more reliable indicators of high productivity than carbonate oozes. This is because silica dissolves quickly in surface waters and carbonate dissolves in deep water; hence, high surface productivity is required to supply siliceous skeletons to the ocean floor. Carbonate oozes dominate the deep Atlantic seafloor, while siliceous oozes are most common in the Pacific; the floor of the Indian Ocean is covered by a combination of the two.

Carbonate oozes cover about half of the world's seafloor. They are present chiefly above a depth of 4,500 metres; below that they dissolve quickly. This depth is named the Calcite Compensation Depth (or CCD). It represents the level at which the rate of carbonate accumulation equals the rate of carbonate dissolution. In the Atlantic basin the CCD is 500 metres deeper than in the Pacific basin, reflecting both a high rate of supply and low rate of dissolution in comparison to the Pacific. The input of carbonate to the ocean is through rivers and deep-sea hydrothermal vents. Variation in input, productivity, and dissolution rates in the geologic past have caused the CCD to vary over 2,000 metres. The CCD intersects the flanks of the world's oceanic ridges, and as a result these are mostly blanketed by carbonate oozes.

Siliceous oozes predominate in two places in the oceans: around Antarctica and a few degrees of latitude north and south of the Equator. At high latitudes the oozes include mostly the shells of diatoms. South of the Antarctic Convergence diatom oozes dominate the seafloor sediment cover and mix with glacial marine sediments closer to the continent. Seventy-five percent of all the oceans' silica supply is being deposited in the area surrounding Antarctica. Radiolarian oozes are more common near the Equator in the Pacific. Here, both siliceous oozes and calcareous oozes occur, but carbonate deposition dominates the region immediately near the Equator. Siliceous oozes bracket the carbonate belt and blend with pelagic clays farther north and south. Because siliceous skeletons dissolve so quickly in seawater, only the more robust skeletal remains are found in the siliceous oozes. Thus, fossils of this kind are not completely representative of the organisms living in the waters above.

The most significant authigenic sediments in the ocean basins today are metal-rich sediments and manganese nodules. Metal-rich sediments include those enriched by iron, manganese, copper, chromium, and lead. These sediments are common at spreading centres, indicating that processes at the centres are responsible for their formation—specifically, hydrothermal circulation is the controlling factor. Deep-sea drill cores have revealed the presence of metal-rich sediments on top of ancient oceanic crust away from ridge crests. It can be inferred from this that the processes controlling their formation existed in the past, but with variations. Which type of enriched sediment is deposited depends on the degree of mixing between the hydrothermal water deep in the crust at a spreading centre and the cold seawater percolating down into the crust. Little mixing produces sulfides, liberal mixing yields manganese-rich crustal material, and intermediate conditions give rise to sediments enriched in iron and manganese.

Manganese nodules are pebbles or stones about the size of walnuts that are built of onionlike layers of manganese and iron oxides. Minor constituents include copper, nickel, and cobalt, making the nodules a potential ore of these valuable elements. Mining of manganese nodules has been the subject of study and experimentation since the 1950s (see below Economic aspects of the oceans: Sources of minerals and other raw materials (ocean)). The nodules grow very slowly, about one to four millimetres per million years. They are found in areas of slow sedimentation, usually five millimetres per thousand years or less. The North and South Pacific (Pacific Ocean) hold the greatest concentration of manganese nodules; in some places, the nodules cover 90 percent of the surface of the ocean floor. Coverages this high also are found in the southernmost South Atlantic. The Indian Ocean floor is largely devoid of manganese nodules. Because seawater is supersaturated in manganese, the direct precipitation of the element onto an available surface is the most likely mode of nodule formation.

Two significant mysteries surround manganese nodules. Drilling and coring in the sediment column has shown that nodules are vastly more abundant at the seafloor than below it and that the rate of growth of nodules is 10 times slower than the lowest known sedimentation rates. If such is the case, the nodules should be quickly buried and should be common in the sediment below the seafloor. Current theories for explaining these observations propose that bottom currents keep areas of nodule growth free of sediment deposition and that burrowing organisms nudge and roll the nodules in the process of feeding, thereby keeping them at the surface of the seafloor. Observations in the deep sea support both explanations.

The patterns of sedimentation in the ocean basins have not been static over geologic time. The existing basins, no more than 200 million years old, contain a highly variable sedimentary record. The major factor behind the variations is plate movements and related changes in climate and ocean water circulation. Since about 200 million years ago, a single vast ocean basin has given way to five or six smaller ones. The Pacific Ocean basin has shrunk, while the North and South Atlantic basins have been created. The climate has changed from warm and mild to cool, stormy, and glacial. Plate movements have altered the course of surface and deep ocean currents and changed the patterns of upwelling, productivity, and biogenic sedimentation. Seaways have opened and closed. The Strait of Gibraltar, for example, was closed off about 6 million years ago, allowing the entire Mediterranean Sea to evaporate and leave thick salt deposits on its floor. Changes in seafloor spreading rates and glaciations have caused sea level to rise and fall, greatly altering the deep-sea sedimentation pattern of both terrigenous and biogenic sediments. The CCD has fluctuated more than 2,000 metres in response to changes in carbonate supply and the corrosive nature of ocean bottom waters. Bottom currents have changed, becoming erosive or nondepositional in some regions to produce unconformities and redistributing enormous volumes of sediment to other locations. The Pacific Plate has been steadily moving northward, so that biogenic sediments of the equatorial regions are found in drill cores taken in the barren North Pacific.

Through most of geologic time, probably extending back 2 billion years, the ocean basins have both grown and been consumed as plate tectonics continued on Earth. The latest phase of ocean basin growth began just less than 200 million years ago with the breakup of the supercontinent Pangaea, the enormous landmass composed of nearly all the present-day continents. Since that time, the major developments have included a shrinking of the Pacific basin at the expense of the growing Atlantic and Arctic basins, the opening of the Tethys seaway circling the globe in tropical latitudes and its subsequent closing, and the opening of the Southern Ocean (see above General considerations (ocean)) as the southern continents moved north away from Antarctica.

As was noted earlier, the oldest known oceanic crust (estimated to be about 200 million years old) is located in the far western equatorial Pacific, east of the Mariana Island arc. The Pacific Ocean floor at this site was generated during seafloor spreading from a pattern of ridges and plates that had existed for some unknown period of time. At least five different seafloor spreading centres were involved. In the Indian Ocean the oldest segment of seafloor was formed about 165 to 145 million years ago by the rifting away of Africa and South America from Gondwana, a supercontinent consisting largely of the present-day continents of the Southern Hemisphere. At this time, Africa was joined to South America, Eurasia, and North America. Today, this old seafloor is found along the east coast of Africa from the Somali Basin to the east coast of South Africa and adjacent to Queen Maud Land and Enderby Land in East Antarctica.

Close to 180 million years ago (but before 165 million years ago), North America and Eurasia, which together made up most of the large northern continent of Laurasia, began drifting away from Africa and South America, creating the first seafloor in the central region of the North Atlantic and opening the Gulf of Mexico. The Tethys seaway also opened during this rifting phase as Europe pulled away from Africa. Shortly after this time continental fragments, including possibly Tibet, Myanmar (Burma), and Malaya, rifted away from the northwest coast of Australia and moved northward, thereby creating the oldest seafloor in the Timor Sea. During this period spreading continued in the Pacific basin with the growth of the Pacific Plate and the consumption by subduction of its bordering plates, including the Izanagi, Farallon, and Phoenix. The Pacific Plate moved northward during this phase and continues to do so today.

India and Madagascar, as a unit, rifted away from Australia and Antarctica prior to 130 million years ago and began drifting northward, creating seafloor adjacent to Western Australia and East Antarctica. Possibly simultaneously or shortly after this rifting began, South America started to separate from Africa, initiating the formation of seafloor in the South Atlantic Ocean.

Between 90 and 80 million years ago, Madagascar and India separated, and the spreading ridges in the Indian Ocean were reorganized. India began drifting northward directly toward Asia. During this same period Europe, joined to Greenland, began drifting away from North America, which resulted in the emergence of the seafloor in the Labrador Sea and the northernmost Atlantic Ocean. This spreading phase affected the passages in the Tethys seaway between Europe (Iberia) and northwest Africa, intermittently opening and closing it. In the southwest Pacific, New Zealand, along with the Lord Howe Rise and the Norfolk Ridge, rifted away from Australia and Antarctica between 80 and 60 million years ago, opening the Tasman Sea.

About 60 million years ago a new rift and oceanic ridge formed between Greenland and Europe, separating them and initiating the formation of oceanic crust in the Norwegian Sea and the Eurasian basin in the eastern Arctic Ocean. The Amerasian basin in the western Arctic Ocean had formed during an earlier spreading phase from about 130 to 110 million years ago. Between 60 and 50 million years ago, significant events occurred in the Indian Ocean and southwest Pacific. Australia began drifting northward, away from East Antarctica, creating seafloor there. The northward movement of Australia resulted in the emergence of several subduction zones and island arcs in the southwest and equatorial Pacific. The Indian subcontinent first touched against the Asian continent about 53 million years ago, developing structures that preceded the main Himalayan orogeny (mountain-building event), which began in earnest some 40 million years ago.

Less than 30 million years ago, seafloor spreading ceased in the Labrador Sea. Along the west coast of North America, the Pacific Plate and the North American Plate converged along what is now California shortly after 30 million years ago. This resulted in the cessation of a long history of subduction in the area and the gradual conversion of this continental margin to a transform fault zone. Continued closure between Africa and Europe, which began about 100 million years ago, caused the isolation of the Mediterranean Sea, so that by 6 million years ago it had completely evaporated.

The present-day Mediterranean seafloor was formed during a complex sequence of rifting between small plates in this region, beginning with the separation of North America and Europe from Africa about 200 million years ago. In the eastern Mediterranean, the seafloor is no older than about 100 million years. West of Italy it was created during subsequent spreading between 30 and 20 million years ago.

The Caribbean Sea and the Gulf of Mexico (Mexico, Gulf of) formed as a result of the relative movement between North America and South America. The seafloor of the Gulf of Mexico began forming some 160 to 150 million years ago. A proto- or ancient Caribbean seafloor also was formed during this period but was later subducted. The present Caribbean seafloor consists of a captured piece of the Farallon Plate (from the Pacific basin) and is estimated to be for the most part of Cretaceous age (i.e., about 120 to 85 million years old).

The seafloor in the western portion of the Philippine Sea developed between 60 and 35 million years ago. In the east, it was formed by backarc spreading from 30 million years ago (see above). The origin of the older crust is not completely clear. It either was created by spreading in the Pacific basin and subsequent capture by the formation of the Bonin and Mariana arcs, or it resulted from backarc spreading behind trenches to the south.

Through knowledge of the ocean sedimentary record, the history of plate motions, glacial changes, and established relations between present sedimentation patterns and environmental factors, it is possible to reconstruct an oceanographic history for approximately the past 200 million years. This is the emerging field of paleoceanography.

Prior to the breakup of Pangaea, one enormous ocean, Panthalassa, existed on Earth. Currents in this ocean would have been simple and slow, and the Earth's climate was, in all likelihood, warmer than today. The Tethys seaway formed as Pangaea broke into Gondwana and Laurasia (see above). In the narrow ocean basins of the central North Atlantic, restricted ocean circulation favoured deposition of evaporites (halite, gypsum, anhydrite, and other less abundant salts). Evaporites also were deposited some 100 million years ago in the equatorial regions of the South Atlantic during the early opening of this ocean.

Sequences of organic-rich, black shales (black shale) were deposited during the early phases of spreading in the North and South Atlantic. These sediments indicate anoxic conditions in the deep ocean waters. The oceans must have been well stratified into dense layers to prevent the overturning and mixing required to replace depleted oxygen. Black shales also were deposited in the older areas of the eastern Indian Ocean.

During the time interval between 200 and 65 million years ago, but especially from 100 to 65 million years ago, microplankton abundance and diversity increased enormously in the oceans. This resulted in increased deposition of biogenic sediments in the ocean basin. During Cretaceous time (from 144 to 66.4 million years ago), sea level was often high, and shallow seas lapped onto the continents. This may have provided an environment favourable to the explosion in the numbers of species of foraminiferans, diatoms, and calcareous nannoplankton. Increased abundance of calcareous nannoplankton shifted the locus of carbonate sedimentation from shallow seas to the deep ocean. The end of Cretaceous time is marked by a sudden extinction of many life-forms on Earth, and marine organisms were no exception (see Cretaceous Period). Coccolithophores (calcareous nannoplankton) and planktonic foraminiferans were particularly affected, and only a few species survived. Ocean sediments were suddenly less biogenic, and clays became widespread.

After Cretaceous time the Earth underwent a gradual cooling, especially at high latitudes. Deep-sea sedimentation changed as thermohaline bottom water circulation became fully developed (see above Circulation of the ocean waters: Thermohaline circulation (ocean)). The CCD rose in the Pacific and dropped in the Atlantic as a result of changes in thermohaline circulation. An event of major significance was the spreading away of Australia from Antarctica beginning about 53 million years ago. This separation initiated limited circum-Antarctic circulation, which isolated Antarctica from the warmer oceans to the north, and led to cooling, which set the stage for later major glaciation.

At the Eocene-Oligocene boundary (36.6 million years ago), Antarctic Bottom Water began to form, resulting in greatly decreased bottom-water (bottom water) temperatures in both the Pacific and Atlantic oceans. Bottom-living organisms were strongly affected, and the CCD suddenly dropped from about 3,500 metres to approximately 4,000 to 5,000 metres in the Pacific. Bottom-water temperatures were generally warm, 12° to 15° C, during the time preceding this event. In a study of deep-sea sediment core material from near Antarctica, J.P. Kennett and Lowell D. Stott of the United States discovered that there was a period between roughly 50 and 35 million years ago when deep waters were very warm (20° C) and salty. The origin of these ocean waters was most likely in the low latitudes and resulted from high evaporation rates there.

The modern oceans are distinguished by very cold bottom water. The gradual changes toward this condition began 10 million years after the origination of Antarctic Bottom Water. Particularly significant among these changes was the closing of the Tethys seaway as Australia and several microcontinents moved north into the Indonesian region. Also, Australia moved far enough north that circum-Antarctic surface circulation became fully established.

The modern ocean circulation patterns and basin shapes were mostly in place by the beginning of Miocene (Miocene Epoch) time (nearly 24 million years ago). An exception was an ocean connection between the Pacific and Caribbean Sea in Central America that persisted until about 3 million years ago. Major and probably permanent ice sheets on Antarctica formed during Miocene time, and glacial sediments began to dominate the seafloor surrounding the continent shortly thereafter. Siliceous oozes also became widespread around Antarctica. Siliceous sedimentation increased in this area at the expense of siliceous sedimentation in equatorial regions. Ocean circulation became more vigorous, global climate became cooler, and sedimentation rates in the ocean basins increased. Planktonic microorganisms were segregated into latitudinal belts. Bottom-water flow north through the Drake Passage between South America and Antarctica began in Miocene time, resulting in erosion and nondeposition of sediments in the southwest Atlantic and southeast Pacific oceans. Also during Miocene time rifting between Greenland and Europe had progressed to a point where a connection was established between the North Atlantic and the Norwegian Sea. This resulted in the formation of North Atlantic Deep Water (see above Circulation of the ocean waters: Thermohaline circulation (ocean)), which began flowing south along the continental rise of North America at this time. Sediments redistributed and deposited by this deep current are called contourites and have been extensively studied by Bruce Heezen, Charles D. Hollister, and Brian E. Tucholke, among others.

Sudden global cooling set in near the end of the Miocene some 6 million years ago. The strength of ocean circulation must have increased, as evidence of increased upwelling and biological productivity is present in ocean sediments. Diatomaceous sediments were deposited in abundance around the rim of the Pacific. This cooling event is synchronous with a drop in sea level, thought to be about 40 or 50 metres by various authorities, and probably corresponds to the further growth of the Antarctic ice sheet. This lowered sea level, coupled with the closure of narrow seaways probably due to plate movements, isolated the Mediterranean Sea. Subsequently, the sea dried up, leaving evaporite deposits on its floor. The Swiss geologist Kenneth J. Hsü and the American oceanographer William B.F. Ryan have concluded that the Mediterranean probably dried up about 40 times as seaways opened and closed between 6 and 5 million years ago. This evaporation removed about 6 percent of the salt from the world ocean, which raised the freezing point of seawater and promoted further growth of the sea ice surrounding Antarctica.

Enormous ice sheets emerged in the Northern Hemisphere between 3 and 2 million years ago, and the succession of Quaternary glaciations began at 1.6 million years ago. The exact cause of the glacial period is unclear, but it is most likely related to the variability in solar isolation, increased mountain building, and an intensification of the Gulf Stream at 3 million years ago due to the closing off of the Pacific-Caribbean ocean connection in Central America. The Quaternary glaciations, of which there were probably 30 episodes, left the most dramatic record in ocean sediments of any event in the previous 200 million years. Terrigenous sedimentation rates greatly increased in response to fluctuations in sea level of up to 100 metres and a more extreme climate. Biogenic sedimentation also increased and fluctuated with the glacial episodes. Deep-sea erosion began in many places as a result of intensified bottom-water circulation.

Since continental margins are the shallowest parts of the world's oceans, they are most affected by changes in sea level. Worldwide changes in sea level, called eustatic sea-level changes, have occurred throughout geologic history. The most common causes of such sea-level changes are global climatic fluctuations that lead to major glacial advances and retreats—i.e., ice ages and interglacial periods. Other causes that are not as well understood may include major mountain-building events and isostatic changes in crustal plates. When continental glaciers advance, as they did several times during the Pleistocene Epoch (which extended from about 1,600,000 to 10,000 years ago), water that would normally be in the oceans is locked up as ice on land, resulting in a drop in sea level. As the glaciers retreat, more water is fed to the ocean basins and the sea level rises. Fluctuations from highstand to lowstand have totaled 250 metres or more during Cenozoic time (roughly the last 66.4 million years), with concomitant fluctuations in exposure and flooding of the continental margins. (During a highstand the sea level is above the edge of the continental shelf, while during a lowstand it is below the shelf edge.)

Rivers (river) bring a variety of sediments to the coast. These are classified by their mineralogy and by particle size and include sand, silt, and clay. Sand to sedimentologists is a grain of any composition from 63 to 2,000 micrometres (0.0063 to 0.2 centimetres) in its largest diameter. Silt is 4 to 62 micrometres, and clay is any particle less than 4 micrometres. Most of the detrital minerals brought to the continental margins by rivers in sand and silt sizes are quartz, feldspars, and mica; those of clay size are a suite of clay minerals that most commonly include smectite, kaolinite, and illite. (Clay can, in other words, refer either to particle size or to a group of minerals.) These, then, are the mineral constituents, which, together with calcium carbonates produced in the oceans by biogenic activity as shells and the hard parts of plants and animals, go to make up the sedimentary packages that are deposited on and constitute a fundamental part of continental margins.

A constant battle is being waged between the rivers that bring sediments eroded from the land to the sea and the waves and currents (ocean current) of the receiving body of water. This dynamic struggle goes on year after year, century after century, sometimes for millions of years. Take, for example, the north coast of the Gulf of Mexico into which the Mississippi River flows. The continental margin at this site is subject to relatively low wave and current energy, and so the river has filled up most of the adjacent continental shelf with a delta and dumps roughly 256 million tons of sediment each year directly at the top of the continental slope. By contrast, the Columbia River in the Pacific Northwest of the United States carries 131 million tons to the coast, where the sediments are attacked by the large waves and currents normal for that margin. As a result, sediments are widely dispersed, and the shelf is not filled with a large subaerial delta.

The effects of this battle are easily seen where human activities have interfered with the transport of sediments to the sea by major rivers. For example, the Nile River delta is retreating rapidly, widening the submerged portion of the continental margin, because the Aswan High Dam has trapped much of the sediment normally fed to the delta front. The lower Mississippi River has been artificially maintained in a channel by high man-made levees. These have stopped the floods that fed much of the western delta margin. Because of this, coupled with a slow rise in sea level and the effects of canals dug in the delta wetlands, the coast has begun to retreat significantly.

When rivers carrying sediment from the interiors of continents reach the sea, several things happen. Velocity in the river jet decreases rapidly, and the sand particles drop out to be picked up by the waves and currents along the coast, where they feed beaches or barrier island systems (see structural landform: Landforms produced by coastal processes: Landforms of depositional coasts (coastal landforms)). If the river has a large enough discharge, the finer-than-sand-sized materials may be carried for kilometres onto the margin in a fresh- or brackish-water plume. The surf system then acts as a wave filter, trapping the sand in the coastal zone but allowing the finer materials to be carried out onto the margin. When estuaries are the receiving bodies of water on the coastal boundaries of continental margins, as in the case of the east coast of North America, virtually all the sediments brought down by the rivers are trapped within the confines of the estuaries.

In addition to the two primary types of continental margin, there also are special types that do not readily fit either category. One of the most intensely studied margins of the world is the Borderland, the continental margin of southern California and northern Baja California. It consists of a series of offshore basins and ridges, some of which are exposed as islands. This system of basins and ridges formed as the result of faulting (fault) associated with the movement of the Pacific Plate past the North American Plate. It remains tectonically active today and is related to the San Andreas Fault system of California. A second special type is the marginal plateau. The Blake Plateau off the east coast of Florida is a good example. Such a plateau constitutes a portion of a continental margin that has many of the features of a normal system but is found at much greater depth—1,000 metres in the case of the Blake Plateau.

Continental margins can be either constructional or erosional (erosion) over varying periods of geologic time, depending on the combination of factors discussed above. When deposition exceeds erosion, the margin grows seaward, a process of progradation that builds out as well as up. When the erosive forces are predominant, the margin remains static or actually retreats over time. Some geologists think that the continental margin of the eastern United States has retreated as much as 5–30 kilometres since Cretaceous time—i.e., during the last 66 million years or so.

Continental margins are very significant economically. Most of the major fisheries of the world are located on them. Of these, sport fisheries and related tourist industries are becoming increasingly important to the economies of developed nations. Paradoxically, continental margins also are one of the world's biggest dump sites. All kinds of wastes are disposed of along the margins, and the effects of pollution on their environment and ecology have become a major global concern (see also below Economic aspects of the oceans: Waste disposal and other related actions (ocean)).

Continental margins are the only parts of the world's oceans to be effectively exploited for mineral resources. Millions of tons of sand are mined by dredges each year off the U.S. coasts alone for beach renourishment projects. From time to time placer deposits also have been worked. Examples include tin off Indonesia, gold off Alaska, and diamonds off Namibia. By far and away the largest mineral resources to be exploited from continental margins are oil and natural gas. Exploration of the continental margins by major oil companies has intensified and is expected to continue for the foreseeable future because the margins are the most likely sites of giant undiscovered petroleum deposits. Continental margins are made of thick accumulations of sedimentary rock, the type of rock in which oil and gas generally occur. In fact, most of the sedimentary rocks exposed on the continents were originally deposited on continental margins; thus, even the hydrocarbon deposits found on land were formed for the most part on ancient continental margins.

The American oceanographer Donald J.P. Swift has called continental shelves palimpsests, parchment writing tablets upon which stories are written after each previous writing has been erased. Each new stand of sea level “writes” a new story of sedimentation on the shelf after the previous episode has been erased by the rise or fall that preceded it but with traces of the previous environment of deposition or last erosional event remaining. The eraser is the surf, a high-energy force that erodes and reworks everything as it passes over, winnowing out the finer-than-sand-sized sediment and leaving the coarser material behind. An interpreted seismic line shows the complicated array of channels (eroded and then filled), old deltaic deposits, ancient erosional surfaces, and winnowed sand bodies that make up the continental shelf southwest of Cape San Blas on the panhandle of Florida.

How the above processes affect any particular margin depends on its tectonic setting and the size of the rivers that drain into it. On continental shelves backed by high mountain ranges, such as the Pacific coast of North and South America, the difference between high and low sea-level stands may be difficult to detect, being one of degree perhaps noticeable only by marginally increased sedimentation rates during lowstands. In many ways, continental shelves on tectonically active margins at present sea levels approximate lowstands on trailing-edge, or passive, margins.

When sea level is lowered on a trailing-edge shelf that has no adjacent high mountains, such as the Atlantic coast of North America, rivers (river) are rejuvenated—i.e., their base level is lowered and they begin to erode their beds, carrying sediment from the continent across the former continental shelf that is now exposed and depositing it at the new coast. When sea level falls below the shelf break, the coast lies on the continental slope. As sea level rises again on tectonically stable or sinking shelves, medium- and small-sized river mouths drown and estuaries form, trapping the sediment (sedimentation) within them and starving the shelves. In these cases, sediment for the shelf is primarily produced by erosion of the coastline as the surf zone advances landward with rising sea level. Fine-grained material is winnowed out, to be either deposited back in the estuaries or carried in steps by advective processes across the shelf to the deeper water beyond. As a result, continental shelf surfaces on trailing-edge margins into which no large rivers flow are veneered with a sand sheet lying over a complex of older deposits, some of which peek through the surface as outcrops—vestiges of an earlier story written on the palimpsest. Large rivers that drain a large, high continent, such as the Mississippi, are able to keep pace with rising sea level and deliver enough sediment to keep an estuary from forming, and, at a high stillstand like that of the present, even fill their entire shelf area. (For a description of modern deposits of this type, see river: Rivers as agents of landscape evolution: Deltas (river).)

For many years after World War II, the period when many of the world's continental shelves were first described in detail, it was thought that the sand deposits on continental shelves were “relict,” deposits left stranded by a higher sea level from the higher-energy regime of the surf zone that passed over them perhaps as much as a few thousand years before. Geophysical investigations of the shelf area since the mid-1970s have revealed the presence of many types of sand waves and ripple marks in seafloor sediments that show submerged continental shelf sediments to be constantly undergoing reworking and erosion. As scientific understanding of the physical processes that affect continental shelves has increased, it has been found that currents set up by large winter storms, monsoons, hurricanes, and typhoons are reworking the bottom by winnowing out the fine-grained materials and carrying them either back into the estuaries or beyond the shelf break, where they are lost from the system.

In short, the kind of sediment that covers the surface of a continental shelf is determined by the interplay among the tectonic setting, the size of the rivers that empty into it (size based on how much sediment they carry), and the wave energy that affects it, just as is the case with continental margins in general. Shelves such as that of western Florida that have been cut off from clastic input (i.e., sediments composed chiefly of quartz and clay minerals derived from erosion of the continent) may be covered with carbonate sediments. In some cases, as in the islands of The Bahamas, the carbonate shelf, called a bank, is cut off from a continental source by deep water. Continental shelves with rivers that carry sediments from the continents to the shelf and beyond only at lowstands of sea level and those that drain mountainous areas on high-energy coasts are dominated by quartz sands, and shelves with rivers that drain large continental areas and carry enough sediment to keep abreast of sea-level rise or dominate ambient wave-energy conditions will accumulate muddy sediment deposits out across their surfaces.

About 8.5 percent of the ocean floor is covered by the continental slope-rise system. This system is an expression of the edge of the continental crustal block. Beyond the shelf-slope break, the continental crust thins quickly, and the rise lies partly on continental crust and partly on the oceanic crust of the deep sea. Inclination of the continental slope averages about 4°, although it can approach vertical on carbonate margins, on faulted margins, or on leading-edge, tectonically active margins such as that off the above-mentioned Borderland. Steep slopes usually have either a very poorly developed continental rise or none at all and are called escarpments. Partway down the continental slope there is often a second inflection point where the gradient drops to 1:100 to 1:700. This lower gradient area defines the continental rise, which may extend seaward several hundred kilometres to the deep seafloor at an average depth of about 4,000 metres. Although the slope-rise was formerly often divided into separate provinces, the trend among sedimentologists is to consider it as part of one continuum. The sedimentary processes that form the slope and rise, however, are quite distinct.

The continental slopes are temporary depositional sites for sediments. During lowstands of sea level, rivers may dump their sedimentary burden directly on them. Sediments build up until the mass becomes unstable and sloughs off to the lower slope and the continental rise. During highstands of sea level, these processes slow down as the coastline retreats landward across the continental shelf, and more of the sediments delivered to the coast are trapped in estuaries and lagoons. Still the process continues, albeit slowly, as sediments are brought across the shelf break by winnowing of the shelf surface and by advection. Slopes are sometimes scoured by such major ocean currents as the Florida Current that work to erode their surfaces. Off active major deposition centres, such as the Mississippi delta, slope sequences may accumulate through progradation, while the active slope front is continuously shedding sediments downslope by gravity processes.

Continental rises are major depositional regimes. They consist of thick sequences of continental material that build up as the result of three sedimentary processes.

The first such process is a downslope movement of sediments by mass wasting (mass movement), a set of gravity-deposition events, including submarine landslides, slumps, debris flows, and high-velocity, sediment-laden density flows known as turbidity currents (see above). Several phenomena may initiate gravity events. In tectonically active areas earthquakes are important triggering mechanisms. Even in the Atlantic they play a significant role. One of the few documented major gravity events took place on the Grand Banks of Newfoundland in 1929, when an earthquake triggered a gravity flow that possibly attained velocities of more than 90 kilometres per hour and was traced for hundreds of kilometres as it successively broke transatlantic cables. Other triggering events may be oversteepening of deposits on the sharply inclined portions of the continental slope, breaking internal waves that have been shown to affect the upper slope, and storm waves and storm-induced currents.

A second process that may be equally important, although its overall significance is subject to considerable scientific debate, is deposition from bottom currents that flow parallel to the slope of the continental rise—namely, contour currents. Resulting sediment accumulations are called contourites. The major points of contention concerning the efficacy of contour currents are (1) whether or not they are strong enough—they flow at a speed of about 20 centimetres per second—to build the huge thicknesses of sediment that make up the rises, and (2) how the sediments get into the contour currents in the first place. It is probable that most of the mass of rise material is originally brought downslope by gravity events and then redistributed by contour currents.

Vertical settling through the water column of both clastic and biogenic particles is the third contributor of slope and rise sediments. These pelagic sediments are composed of clay minerals and fine-grained particles (chiefly quartz, mica, and carbonate) swept off the continental shelf, wind-blown dust, organic detritus, and the tests of planktonic plants and animals. Chief among the last group are the tests of foraminiferans, pteropods, and coccolithophores that are composed of calcium carbonate and those of diatoms and radiolarians that are made of silicon dioxide.

Originating on the continental shelf and cutting into the continental slope-rise system are some of the most spectacular physiographic features on the Earth—the submarine canyons. They have been the object of investigation for many years. Much of the definitive work was done by Francis P. Shepard of the United States, widely regarded as the father of modern marine geology, and Robert F. Dill, one of his students.

Submarine canyons come in various shapes and sizes. Many of those that have been extensively studied resemble terrestrial river-cut canyons. Submarine canyons are often V-shaped and have steep walls, sinuous courses, and tributaries that feed into them. Many end in a series of distributary channels through the continental rise (see river: Rivers as agents of landscape evolution: Deltas: Morphology of deltas (river)). Distributary channels often emerge in the deep sea on abyssal plains that are found adjacent to the continents, especially in the Atlantic basin. The scale of many canyons is truly herculean. For example, the Hudson (Hudson Canyon) submarine canyon, which serves as a type canyon for those of the east coast of the United States, originates (or heads) on the outer edge of the continental shelf off New York harbour at a depth of about 90 metres. Hudson Canyon wends its way down the continental slope and enters the continental rise at a depth of about 2,100 metres. Higher up, where the canyon floor reaches a depth of 1,800 metres, the walls are nearly 1,220 metres high. The valley and its distributaries can be traced seaward across the continental rise for at least 300 kilometres to a depth of more than 4,500 metres.

Submarine canyons act as conduits to bring sand-sized sediments (sedimentation) from the continental margins to the deep sea. During lowstands of sea level, rivers empty directly into the heads of many Atlantic canyons. Sand and mud are carried down these systems, many times bypassing the slope-rise system to be carried directly out onto the abyssal plains on the deep seafloor. Cores taken in abyssal plains in the Atlantic Ocean off North Carolina show distinctive shelf molluscan fragments in gravity-flow deposits that have traveled hundreds of kilometres and that can be traced back through the Hatteras submarine canyon system. During highstands of sea level, the submarine canyons off the east coast of North America are many tens of kilometres from the coastline, and the downslope movement through them is slowed down dramatically or perhaps even ceases. On the west coast of the continent where active tectonism results in a narrow margin backed by high mountains, a situation exists today that is roughly analogous to the one that prevails on a passive margin during lowstands of sea level. Submarine canyons originate close to the coast, intersecting the longshore current system and siphoning off sand to basin floors in the Borderland (see above).

The origin of submarine canyons is still not completely understood. Probably several processes contribute to their formation. The underlying question is how erosion occurs, often in solid rock, thousands of metres below the lowest known stand of sea level. Some canyons may be located on fault lines or other structural zones of weakness. They certainly are scoured by gravity flows of various types, and the collapsing of the outer portions of canyon walls plays an important part in widening them.

Submarine canyons are not present everywhere. They are rare on margins that have extremely steep continental slopes or escarpments. The reasons for this are unclear. It simply is one of many aspects of the mechanisms of submarine canyon formation that remain to be discovered.

Coral reefs are masses of carbonate (carbonate mineral) of lime built up from the seafloor by the accumulation of the skeletons of a profusion of animals and algae; (algae) eventually they rise to the surface of the water.

Reef-building corals, chiefly the stony corals or Scleractinia, grow best in shallow, sunlit water, between the low-water mark and a depth of 11 metres, but they can still construct reefs in water as deep as 40 metres, and they may have a sparse existence between 40 and 55 metres. These corals prefer water of normal salinity and with an annual maximum temperature above 22° C but below 28° C. Their reef-building activities, however, may be carried on in waters whose minimum temperature in winter is not less than 15° C.

A second group of corals in present-day seas grows in thickets and coppices that develop banks rather than reefs on the outer, deeper, colder, and darker parts of continental shelves and platforms. These organisms flourish in water with a winter minimum temperature ranging between about 4° and 15° C at depths of about 60 to 200 metres. In any one thicket there are commonly only two genera of delicately branching corals involved. Such coral banks are known along the eastern Atlantic shelf edge (or continental slope) from Norway to the Cape Verde islands and again off the Niger River delta and in the west Atlantic around the Gulf of Mexico, The Bahamas, and the Orinoco River delta. Off New Zealand such banks have been recognized on the Campbell Plateau and the Chatham Rise; they also occur in the northwest Pacific near Japan.

The third coral assemblage of the modern seas is associated with even colder or deeper seas; it consists of small, solitary corals of relatively few genera, known from the abyssal (abyssal zone) floors of the oceans and from the shelves around Antarctica, Patagonia, and the Falkland Islands in waters 2° to 6° C in temperature.

Buried fossil reefs on ancient continental shelves are targets for petroleum exploration. The porosity of reefs and the characteristic curvature of nonporous enclosing sediments cause them to be prospective reservoirs for oil and gas (natural gas). The rich oil fields of Alberta, for example, are associated with Devonian reefs (about 408 to 360 million years old). Fossil reefs recently have become targets for metal prospecting because some corals contain small percentages of metals, such as zinc and copper, selectively incorporated from seawater. A living coral reef may also have economic potential in that it constitutes a major tourist attraction.

Water conditions favourable to the growth of reefs exist in tropical or near-tropical surface waters. Regional differences may result from the presence or absence of upwelling currents of colder waters or from the varying relation of precipitation to evaporation.

Tropical seas are well lit, the hours of daylight varying with the latitude. Light intensity (luminous intensity) and radiant energy also vary with the depth. Thus, at latitude 32°44′ N (the Madeira Islands) the “day” in March has a length of 11 hours at a depth of 20 metres, 5 hours at 30 metres, and only about a quarter of an hour at 40 metres. Nearer the pole these figures decrease further. Light intensity has a profound effect on the growth of the individual reef-coral skeleton because of the symbiont zooxanthellae of reef corals (see below Biological factors (ocean)). The number of species present on a reef also may be related to light intensity and radiant energy.

Turbidity may be high in lagoons (lagoon), where shallow water lies over a silt-covered seafloor and where storms and windy periods cause considerable disturbance of the bottom silt. The average transparency may be low (about 12 metres), and light penetration is reduced.

Inside the Great Barrier Reef, on the shallow continental shelf of Queensland, the oxygen content of the water is high, exceeding 90 percent saturation most of the time; in deeper water, during the calm periods of the rainy season, the saturation may fall to about 80 percent. Plant nutrients such as phosphate and nitrate show no seasonal change in quantity; both are present in very small quantities throughout the year. Constant mixing of the shallow sea prevents any stratification of the nutrients. As a result, growth of phytoplankton is possible and almost uniform throughout the year, providing a constant supply of food (food chain) for the zooplankton, which in turn form the chief food supply of the corals. Some nutrients enter the lagoonal waters with the oceanic water that flows through the reef openings, but the dissolved phosphates in the lagoons are probably derived chiefly from bacterial decomposition of the organic matter on the sea bottom, as well as from detritus swept in from the reef surfaces. This environmental pattern is typical of many atoll lagoons.

Minute quantities of metallic elements are present in solution in seawater and also occur in marine invertebrate skeletons, though not in the same proportions as in the surrounding water. magnesium and strontium are the most frequently occurring trace elements (trace element) in reef skeletons and are measured in parts per thousand, but barium, manganese, and iron are also present and can be measured in parts per million. In Pacific corals, 2.17 parts per million of uranium have been found, and in Florida coral, 2.36–2.95 parts per million. Strontium is concentrated in aragonitic skeletons, and magnesium in calcitic skeletons; coral aragonite has a higher strontium content than (some) molluscan aragonite; the magnesium content in the calcite of coralline algae is high; that of barnacle shells is low (11.5 parts per thousand). By identifying these trace elements and their degree of assimilation in different organisms, sediments formed predominantly of coral-skeletal detritus can be distinguished from sediment derived chiefly from mollusks or coralline algae.

Atomic-absorption spectrophotometry has shown that ultratraces of metals are present in the aragonite skeleton of the hydrozoan coral Millepora from a reef flat on the Coral Sea Plateau off Queensland. These are, in parts per billion: lead (100), copper (71), cadmium (23), cobalt (17), nickel (1,480), iron (507), and zinc (507).

Another aspect of reef geochemistry is the carbon and oxygen isotopic composition of coral skeletons and shells. Determination of the number of carbon isotopes present provides a method of assessing the age of a sample, and determinations of oxygen isotopes present are useful in indicating water-temperature changes that occurred during the period of growth of the reef.

Winds and currents are important in shaping individual reefs and in determining the orientation, shape, and position of the coral sand cays (cay), or “low islands,” that develop on reefs. Currents are primarily those generated by the prevailing winds, but, in areas where the tidal range is great, tidal effects may become paramount.

Cays may be round, oval (or boat-shaped), or irregular in outline. They originate when sediment is lifted from the reef surface and carried leeward by waves or tidal currents and then deposited where the water velocity is reduced abruptly. Thus, they commonly form on the more protected leeward end of the reef. Wind action at low tide on these deposits may build dunes above the high-water mark. Beach rock may form by carbonate cementation of grains in deposits lying between tide levels. It then acts as a stabilizing factor. Storm waves may drive forward coral fragments derived from “stag-horn” corals growing on the windward slopes of the reef, forming shingle banks; successive, superposed banks may thus be formed. The shingle on the banks may become cemented and thus add considerable stability to the cay, as does the growth of vegetation. Hurricanes, however, may carve back the shorelines of even stabilized cays. Huge, isolated boulders of coral or coral limestone are fairly common along reef margins. Some may be remnants of a once-emergent reef platform; others are hurricane or storm jetsam.

Coral reefs are best developed where the mean annual surface-water temperatures are approximately 23° to 25° C. No significant reefs occur where such temperatures fall below about 18° C, although a few reef-coral species can exist in temperatures considerably below this. Seasonal temperature differences on any one reef are usually slight, as are differences due to depths of water or situation on the reef.

Seawater of normal oceanic salinity (between 30 and 40 parts per thousand), to which corals are restricted, is normally supersaturated in calcium carbonate (CaCO₃), so that adequate ionized calcium (Ca²⁺) is available for the skeleton-forming process. Floods of fresh water may destroy life on inshore fringing reefs. A luxuriant reef on Stone Island, near Bowen, Queens., Australia, was killed to a depth of three metres below mean tide level by a week of cyclonic rains, in which 90.7 centimetres of rain coincided with full-moon spring tides.

The most significant biological determinant of reef accumulation is the presence of zooxanthellae (zooxanthella) in the living tissues of all reef corals and of many massive-shelled mollusks (Tridacnidae) and other shelled invertebrates, as well as in the soft-bodied hydrozoans, scyphozoans, and anthozoans. Zooxanthellae are now known to represent the vegetative stages of dinoflagellate algae, and their association with reef corals is symbiotic (mutualism)—i.e., mutually helpful. In temperate seas they occur only occasionally. Their profusion in reef animals is no doubt connected with the greater light intensity and radiant energy of reef waters, for, like other plants, zooxanthellae require sunlight for photosynthesis. They remove at the source part of the carbon dioxide, together with nitrogen, phosphorus, and (doubtless) sulfur, produced by metabolic breakdown within the coral and which would otherwise be excreted by the corals. They greatly aid in the formation of the coral skeleton by increasing the speed with which the carbon dioxide produced in coral metabolism is removed and the speed with which the skeletal calcium carbonate is formed. Corals also may gain some nutrient from their zooxanthellae, but they probably do not need the oxygen produced during photosynthesis.

The productivity of reefs is a current focus of interest. A constant supply of food in the form of zooplankton is essential to reef corals, which are carnivorous. The zooplankton supply is dependent on an adequate phytoplankton supply, and the phytoplankton, in turn, require an adequate supply of plant nutrients dissolved in the water. An atoll in the open ocean may be compared to an oasis in a desert, as a localized centre of high productivity. In the warm, well-lit, and well-mixed lagoon waters there is a rapid turnover of the endemic planktonic (floating or swimming) and benthic (bottom-dwelling) population, perhaps 12.5 times per year. Carbonate skeletal matter is accumulated perhaps 1,000 times faster on the summit of an atoll than in the surrounding deeps.

Certain biological factors may contribute to the destruction of coral reefs: the fish and invertebrates that feed on the soft tissues of reef builders and the organisms that bore into coral rock. Of the former, the most destructive such enemy yet known is Acanthaster planci, the crown-of-thorns starfish, which, during the 1960s multiplied spectacularly and removed the soft tissues from large areas of many reefs in the southwest Pacific. A. planci feeds by everting its stomach and liquifying and absorbing the tissues of the corals. By the late 1970s it had become apparent, however, that the sudden spread of A. planci was part of the organism's natural life cycle and that the coral reefs could regenerate rapidly after such an infestation. Coral-rock borers include boring algae, boring sponges (of great significance), various polychaete and sipunculid worms, and many bivalves and a few gastropods. These organisms usually penetrate the rock mechanically but in some cases do so chemically. Extensive damage is caused both by their own activities and by the assistance they give to the erosive action of the sea. A phenomenon known as “bleaching” caused extensive devastation among coral reefs in the east Pacific during the early 1980s and in the Caribbean during the mid- to late 1980s. It is called bleaching because zooxanthellae (principally the so-called brown algae) are expelled, leaving the white coral exposed. While the cause of bleaching is not yet fully known and extensive research is under way, it is believed that the most likely factor is unusually high seawater temperatures, approximately 30° C.

Charles Darwin (Darwin, Charles) concluded in 1842 that barrier reefs began as reefs fringing the land around which they now form a barrier and that oceanic atoll reefs began as reefs fringing a volcanic island. Subsidence of the land fringed was thought to allow the reef to grow upward (and outward over its own forereef debris). Maximum growth would occur at the seaward edge, and lagoons would develop between the ascending barrier, or atoll, reef and the land or volcanic cone. When the volcanic cone became completely submerged, the atoll lagoon would contain only coral islands. Fundamentally, Darwin's concept is still valid, although many consider submergence by the rise of sea level, following melting of Pleistocene ice sheets, to be a better explanation of the latest upward growth of many reefs, particularly on continental shelves. Mid-ocean stages of coral reef development are explained by plate tectonic theory (see above Ocean basins (ocean)), according to which the ocean floor subsides as it spreads outward from oceanic ridges. The Hawaiian Islands, with barrier reefs in the southeast grading to atolls in the northwest, is a good example of this.

A reef whose surface lies above high-tide mark, either by uplift or by eustatic regression of the sea (which is determined by ice sheet-sea level relations), is subject to planing by marine erosion. If planing off is complete, a flat-topped submerged platform results; if subsidence or eustatic submergence intervenes, a wave-cut terrace is left around the reef. Terraces that may have formed in this way are known around many reefs. Some annular reefs may develop without relation to subsiding volcanic cones. When reef platforms have been uplifted above sea level, they are subjected to subaerial erosion. Surface slope, or gradient, determines the amount of runoff and is a prime factor in this erosion. Two secondary processes also are involved: (1) case hardening of steep, bare limestone surfaces by recrystallization caused by alternate wetting and drying, so that walls or knife edges result from weathering; and (2) continuous subsoil solution, if surfaces are nearly horizontal and runoff is diminished. These processes combine to produce a prominent rim and a saucer-shaped interior in emerged limestone islands. With submergence, algal and coral growth resumes, the fastest growth being on the rim and on any pinnacles that may be left. Thus an atoll or annular reef may develop along the rim around the low-lying central region, which becomes a lagoon, and coral knolls grow on former pinnacles in the lagoonal area.

Fringing reefs form a veneer in the shallow water near or at the shore of the mainland or of islands. On shorelines where bays receive large quantities of terrestrial mud, sand, and fresh water, fringing reefs are intermittent and are restricted to promontories. Along limestone coasts, however, coastal erosion is by solution, little mud or sand is supplied, and coral growth may be almost continuous along the shore. Fringing reefs may extend as far as 1,500 metres from shore. They show ecological zonation parallel to the shore. Along mainland shores with easily eroded rocks, mudbanks may be washed into the reef flat and colonized by mangroves. Inner parts of reef flats may show low mesas of middle Holocene (perhaps 5,000 years old) and even Pleistocene emergent reefs that have not been quite planed off by marine abrasion. Geomorphologic features peculiar to raised fringing reefs have been described for the Solomon Islands, and there the complex problem of dead reefs and dead patches on reefs arises. The surface of the reef flat is mostly of dead coral, but pools occur in which live coral colonies flourish. An algal rim formed by red calcareous algae may develop on a fringing-reef margin if the reef faces strong waves and swells the year around. This rim is commonly less spectacular than the algal ridge of the windward edge of Pacific oceanic atolls and may be developed merely as an algal platform, or pavement, on which algal encrustation over corals is thin. The seaward slope of a fringing reef, like the seaward slope of an atoll reef, is characterized by a zone of grooves and spurs to a depth of perhaps 15 metres, and it is in this zone that vigorous growth occurs. Under stable shelf conditions, the fringing reef will extend outward as the spurs elongate and the grooves fill or roof over with coral and algal growth. In tectonic belts (zones of uplift and deformation), fringing reefs may be uplifted intermittently, resulting in parallel, stepped subaerial terraces such as those of the Finsch Coast of Papua New Guinea.

Platform and patch reefs are characteristic of continental shelves; they may or may not lie behind a barrier reef. Reefs grow actively outward as well as upward, especially in the stable conditions of a continental shelf. Any given reef, having depth and temperature fixed by its location, will have its shape determined by the direction and force of the water currents that bring food and by the shape of the base on which it grows. Where the forces of growth are equal in all directions, radial expansion results in platformlike reefs. With further radial growth, lagoonal platform reefs develop. If the reef grows on a sand bank, elongation may result. The shape of an elongated platform reef may be determined by the orientation of rising and falling tidal currents. These may be directly opposed to each other. The boat-shaped reefs of Torres Strait (Torres Strait Islands), between Australia and New Guinea, apparently developed in such a pattern. Where wave-generated currents are asymmetric, horseshoe reefs develop, with convexity facing the current and the leeward ends curving round to partly surround a lagoon. Low Islets, made famous by the Great Barrier Reef Expedition of 1928–29, is the best-known example of this type. A sand cay (or cays) commonly develops on one or both of the leeward wings. Those parts of Pacific platform reefs that face strong and persistent currents characteristically have a low algal rim from which radiate grooves and spurs.

Barrier reefs (barrier reef) commonly present to the ocean and the trade winds a steep wall, some dropping abruptly 1,000 to 5,000 metres. On the lagoon side they grade off gently with a wedge of sediment dotted by small patch reefs, coral knolls, and coral heads. Depths in the lagoon may reach 50 to 80 metres. According to the strength of surf and swell, an algal ridge, rim, platform, or pavement develops and is commonly emergent at low water. The seaward slope has radial grooves and spurs, the grooves forming surf channels and imparting great wave resistance. Ecological zonation on the seaward slope is difficult to study because of the dangers of surf and swell. Zonation on the reef flat, parallel to the wall, is dependent mainly on depth.

A characteristic form in a barrier reef system is the wall, or ribbon, reef, emergent at low tide, such as Yonge Reef, Queens., Australia. A ribbon reef flat is commonly only 300 to 450 metres wide from the seaward wall to its lagoonward edge. Its ends may curve leeward and border the passages between it and the next reefs in line. Rarely, there may be an unvegetated sand cay. A wall reef may develop irregular leeward prongs normal (perpendicular) to its axis by vigorous coral growth favoured by augmented turbulence associated with a high tidal range. The open leeward zone may be very wide (three kilometres) and support large, scattered reef clumps that are always submerged.

The oceanic atoll reefs of the Pacific Ocean rise from volcanic cones that have subsided, probably intermittently, in areas of oceanic deeps. According to the Darwinian subsidence theory, the annular atoll reefs extend and grade downward into barrier reefs. The theory suggests that these barrier reefs originated as reefs that fringed a volcanic cone. On the other hand, the compound atoll reefs of the Indian Ocean, such as the Maldives and the Laccadive Islands, and of the Queensland Plateau of the Pacific Ocean are believed to have grown above foundered continental (rather than oceanic) crustal segments. The Nicaraguan atolls rise from a seafloor of 1,000 metres or more in a volcanic province.

Lagoons are, as indicated above, areas of relatively shallow water situated in a coastal environment and having access to the sea but separated from the open marine conditions by a barrier. The barrier may be either a sandy or shingly wave-built feature, or it may be a coral reef. Thus, there are two main types of lagoons: (1) elongated or irregular stretches of water that lie between coastal barrier islands and the shoreline and (2) circular or irregular stretches of water surrounded by coral atoll reefs or protected by barrier coral reefs from direct wave action. Lagoons of the first type are characterized by quiet water conditions, fine-grained sedimentation, and, in many cases, brackish marshes. Water movements are related to discharge of river flow through the lagoon and to the regular influx and egress of tidal waters through the inlets that normally separate the barrier islands. Lagoons of the second type are best exemplified by the roughly circular quiet waters that are surrounded by coral atoll reefs. Lagoon depths are maintained at a moderate level by sedimentation, and this compensates for the subsidence that commonly attends reef formation (see above Coral reefs, coral islands, and atolls (ocean)). Because the reef is an organic structure, the lagoonal sediments contain much calcareous material. The sheltered waters support a distinctive flora and fauna.

Coastal lagoons are widely distributed throughout the world and have been estimated to constitute about 13 percent of the total world coastline. Lagoons are more common on coasts with moderate to low tidal ranges; for example, they occur widely on low coasts of the southern Baltic, southeast North Sea, Black Sea, Caspian Sea, and Mediterranean Sea and of the southeastern United States and the Gulf of Mexico. Lagoon coasts also occur along southern Brazil; the east coast of Madagascar; northeastern Russia; Japan; and isolated parts of Africa, India, Australia, and New Zealand.

Lagoons are generally characteristic of coasts of low or moderate energy, occurring especially on the east coasts of continents where the swells are less violent and in high latitudes where offshore ice provides some protection. They also are associated with low coasts and rarely occur where high cliffs form the coast. They can form only where there is abundant sediment for construction of the protective barrier islands. Too much sediment from the mainland, however, can lead to delta formation rather than lagoons, although lagoons frequently occur along the outer delta margin and between delta distributaries (see also river: Rivers as agents of landscape evolution: Deltas (river)).

Coral lagoons are restricted to tropical open seas that provide the conditions necessary for coral growth. They occur widely in the western Pacific, in parts of the Indian Ocean, and in isolated places in the Caribbean, mainly within 25° latitude of the equator. Coral lagoons are of great importance to many island communities in the Pacific, particularly where they provide the only quiet water for use as harbours, although the passage through the reef into the lagoon is often perilous.

The clarity of the water of the coral lagoon contrasts with the considerable amount of fine sediment in barrier island lagoon water. The extensive growth of salt marsh in such lagoons often detracts from or precludes their use as harbours, but where the water is deep enough these lagoons provide sheltered anchorage and good conditions for small-boat sailing in quiet water.

Coral lagoon dimensions range from small atolls to those so wide that the coral reefs on the far side cannot be seen across the lagoon. Atoll widths range from about 2.5 to nearly 100 kilometres, but the mean value is about 20 kilometres. Depths rarely exceed 60 metres and many are shallower, usually less than 20 metres deep. The lagoon of Mayotte island in the Comoro archipelago in the Indian Ocean attains a maximum depth of about 92 metres but generally is shallower, and it is about 16 kilometres in width at its widest point; this lagoon lies behind a barrier reef that encircles the island, forming an atoll about 55 kilometres in diameter.

Barrier island lagoons are usually elongated, though irregular ones may occur where river estuaries flood behind barriers. This occurs on the east coast of the United States, where lagoons extend intermittently for nearly 1,500 kilometres along the coast. The Gippsland Lakes in Victoria, Australia, exemplify a complex lagoon system formed behind a 149-kilometre beach. Elongated lagoons up to 64 kilometres in length lie behind the beach barrier, and larger lagoons, such as Lake Wellington, lie behind the southwestern end. Postglacial subsidence has flooded the lowland in this area. The lagoons are shallow: Lake Wellington is less than 3.5 metres in depth, and much of Lake King is less than 6 metres deep. Scour holes as deep as 16.5 metres do occur, however. The elongated lagoon behind the barrier is only 1 to 1.5 metres deep, typical of barrier island lagoons.

The degree of water circulation depends on the width of the tidal inlets, the tidal range, and the amount of runoff from adjacent land areas. Maximum velocities are attained at the points where the water passes through the barriers. In the entrance to the Gippsland lagoons, for example, tidal currents reach 5.6 kilometres per hour, but river floods that escape to the ocean can raise the velocity to 13 kilometres per hour. Water may be blown into the lagoon by strong winds; the increased level results in an outflowing current when the wind drops. Seiches can be created in this way. Small waves can be generated within lagoons when the wind blows along their maximum dimension. These may reach 1.25 metres in height and 1.5 to 9 metres in length in the Gippsland lagoons. In coral atoll lagoons there is little or no runoff, and seawater moves in and out through the passes, where tidal currents reach their maxima. Velocities of 19 to 22 kilometres per hour have been recorded in the Hao Channel of the Tuamotu Archipelago.

In the Mayotte Lagoon the water has the same properties as the upper layers of the open ocean. The salinity is close to 35 parts per thousand, and the temperature varies between 27° and 24° C. This is typical of coral lagoons, but the temperature and salinity of barrier island lagoons are more variable because of their wider climatic range. Because they are shallow, lagoon waters approximate the air temperature: colder than the open ocean in winter, warmer in summer.

Salinities decrease as a function of the amount of runoff entering the lagoon in relation to the tidal influx. Lagoons may be considered brackish, marine, or hypersaline. Brackish lagoons receive much runoff, and salinity increases toward the tidal inlets. The Gippsland lagoons exemplify this type. The salinity at the inner end varies from 0.5 to 5 parts per thousand according to season, and central values vary between 5 and 20 parts per thousand. Hypersaline lagoons occur where evaporation exceeds inflow. Laguna Madre in Texas and Sivash Sound in the Black Sea have salinities of 65 and 132 parts per thousand, respectively. Salt deposits may form in these conditions. The denser saline water tends to lie beneath the fresher water where mixing is not severe.

Lagoons behind coastal barriers normally are zones of fine sedimentation. Their bottom profiles, therefore, build up gradually with time. Typical depths of the Texas lagoons vary between 1.25 and 3.5 metres, and their floors are flat. Early theories that attempted to relate the form of the offshore and lagoon profile are no longer held; and, because the lagoon profile changes with continued deposition, it cannot be used to establish the process of lagoon formation. The profile is usually gently undulating, but it may contain deeper channels, especially near the tidal inlets. Profiles across coral lagoons either are smooth and flat from calcareous sedimentation or contain knolls of growing or dead coral. There are 2,300 such knobs in the Enewetak Lagoon in the Marshall Islands.

The essential feature that causes the lagoon to exist is the barrier that separates it from the ocean. In the coral lagoon the formation of the reef depends on the existence of suitable conditions for reef growth, which have already been mentioned briefly.

The barrier island lagoons, on the other hand, depend not on organic processes but on waves, which act in a suitable environment on an adequate supply of bottom material, most commonly sand. Barrier islands are formed in those areas where the coastal slope is flatter than the equilibrium slope required by the long constructive swells—i.e., the waves that build up the foreshore in front of their breakpoint. They are, therefore, found on low coasts. They may occur in areas of subsidence, stability, or emergence, wherever sufficient sand exists.

The barrier islands that form the lagoons are made of sand, but the sediments of the lagoon are usually finer, as conditions are quieter. The lagoonal muds differ from shelf muds. Glauconite is rare in lagoon muds, but oyster reefs may occur as along the Texas coast. The muds are found away from the channels, in which only coarse sediment can rest, owing to strong currents. Flocculation in the saline lagoon water expedites clay deposition. The source of the fine sediment is from inland areas, and transport is by rivers. The details of lagoon sedimentation vary with the nature of the river load. Sedimentation rates are much greater in the lagoon than the adjacent open ocean, because deposition is continuous over much of the lagoon and is often aided by flocculation and vegetation. In the Texas lagoons from 1875 to 1936, 20 centimetres of deposition occurred in spite of 30 centimetres of subsidence; the sedimentation rate, therefore, was about one metre in 100 years.

The barrier islands are formed by the waves, which build up their equilibrium profile by deposition on a gradient that is too flat. The level of the growing accumulation may be raised by the wind, forming dunes. Where the land behind the growing barrier is low, it will become flooded to form a lagoon if sea level rises slowly. Such a rise of sea level has occurred during the past 20,000 years. As long as the barrier island can maintain its level above the sea, as sea level rises, the lagoon will exist until it is filled with sediment. Not all lagoons and barrier island complexes have formed during rising sea-level conditions, but, where there is evidence that no open-sea foreshore sediments lie on the land side of the barrier, this hypothesis seems likely. In some barriers, however, outbuilding of material by glacial outwash streams or rivers may provide a suitably low gradient and enough sediment to form a barrier, as along the south coast of Iceland. In other areas material carried alongshore to form a spit may develop into a bay-mouth barrier, enclosing a lagoon. Such features can be of sand or shingle. The Fleet, a brackish body of water behind Chesil Beach in southern England, is an example of the latter type.

Waves within the lagoon may develop cuspate spits along the land side of the barrier and the inner edge of the lagoon. These features may eventually break the lagoon into almost circular or oval water bodies. Examples occur in the Chukchi Sea lagoons in northeastern Russia and elsewhere where vegetation does not form marshland.

Storms or tsunamis exert an effect on lagoons when they breach or overtop the barriers around a lagoon by raising the water level temporarily. Major changes in configuration can occur in a short time. Hurricanes, for instance, can cause serious effects on the coast of Texas, along which lagoons are common. Padre Island was lowered to below mean low tide at its southern end by one storm in 1919, and several washover channels were cut. The mainland shore also suffered erosion. Deposition may also occur; saline marls have been laid down on freshwater marsh, and small beach ridges may be built inside the lagoon where the high water level drives sand inland over mud. Coral reefs are more resistant to storms than are mobile sandy barriers.

Lagoons of both types change with time. In both, a relative rise of sea level with time is important in the development of the lagoon. In coral atolls there is evidence from deep boring that Darwin's (Darwin, Charles) original subsidence hypothesis of atoll formation, via barrier reefs from fringing reefs around a subsiding volcanic peak, is substantially correct in many cases. As long as the coral can maintain its growth at a suitable level as its foundation subsides, the atoll will continue to enclose a lagoon, which is floored by coral or calcareous sediment derived from the reef and which maintains its depth by growth or deposition.

The postglacial rise of sea level also has influenced barrier island formation in many instances. When sea level rises too fast, a barrier may be drowned and its lagoon will cease to exist.

Estuaries are partially enclosed bodies of coastal water that have access to the sea and that receive a freshwater contribution from the land. Physically this would include river mouths (as in the case of the Mississippi River) or coastal plain drainage areas of low relief (e.g., Chesapeake Bay) that are flooded by a rise in sea level. Other examples would be submerged fjords (Scoresby Sound, Greenland) and structural basins (San Francisco Bay) and the bodies of water behind spits (Hurst Castle spit, Eng.) or barrier beaches (Ninety Mile Beach, Australia). In the case of spits and barrier beaches, the definitions of lagoons and estuaries overlap (see above Lagoons (ocean) and especially the article river: Rivers as agents of landscape evolution: Estuaries (river)).

Classification of estuaries is made on the basis of salinity distribution. Though the terminology differs slightly among different specialists on the subject, there are several basic types. They are as follows: (1) vertically mixed, wherein salinity, while constant from top to bottom at any site, increases from land to sea; (2) slightly stratified, in which saline water circulates in at the bottom, mixes with fresh water, and then flows out at the top (salinity thus increases with depth and out toward the sea); (3) highly stratified, which is similar to the slightly stratified type, but is limited to the upper layer of water above the outer sill of a fjord; and (4) salt wedge, where saline seawater intrudes in as a wedge at the bottom, while fresh water flows out and over it at the top.

Any concavity (bay) of a coastline or reentrant of the sea, regardless of size, depth, configuration, and geologic structure, may be called a gulf or bay. The nomenclature for features of this type is far from uniform; names that may refer to sizable gulfs and bays in various places include bight, firth, sound, and fjord. A number of pronounced concavities of oceanic margins have no proper name at all.

The problem of terminology extends to the difference between gulfs and seas. There are many small seas, such as the Sea of Marmara (11,000 square kilometres) and the Sea of Azov (38,000 square kilometres), which, strictly speaking, are really gulfs of the ocean or other seas (the Sea of Azov is a gulf of the Black Sea). The Gulf of Aden (about 270,000 square kilometres), another example, is part of the Arabian Sea, and these water bodies have a common regime (similar tides, precipitation, evaporation, and so forth). The narrow sound of Bab el-Mandeb connects the gulf with the vast Red Sea (438,100 square kilometres) and exhibits a number of specific geomorphic features. The Red Sea, in turn, has two small gulfs to the north—namely, those of Suez (Suez, Gulf of) and Aqaba. (Aqaba, Gulf of)

The shape and bottom topography of gulfs and bays are amazingly diverse. They are determined by the geologic structure and development of the region. Homogeneous bedrock of low strength or resistance results in simple shapes and shallow depths. The Gulf of Riga (Riga, Gulf of) (at the Baltic Sea) is a possible example of the type. Long narrow arms with approximately parallel shores of the south Kara Sea extend inland for about 800 kilometres. They occupy troughs that originated by erosion during a period of lower sea level (Baidaratskaya Bay, Obskaya (Ob, Gulf of) Bay with Tazovskaya Bay tributary, Yenisey Bay, Gydanskaya Bay). Deep, angular gulfs, on the other hand, are created along fractures, faults, and rifts (e.g., Varanger Fjord); they usually have irregular bottom topography. Parallel fractures form extremely deep, narrow gulfs with parallel shores, such as the Gulf of California. Genuine fjord-gulfs are notable for their very high length to width ratios (up to 50:1). In regions that have undergone nonuniform deformation and uplift, gulfs and bays of complicated and irregular shape and bottom topography are consequently formed; the Gulf of St. Lawrence is an example.

Gulfs are connected with the sea by means of one or more straits. Sometimes there may be an archipelago in the mouth of the gulf, as in the Gulf of Bothnia. There are some gulfs that open into the sea or into another gulf on opposite sides (Baffin Bay, Gulf of Aden, and the Gulf of Oman).

Single gulfs usually are formed along linear shores of the continent. If the shoreline is irregular and has a complex geologic structure, groups of gulfs of a similar nature may occur. Most shorelines have small reentrants of various size that are called bays. These features are strongly influenced by local conditions, and they are not described or classified within the context of this section, which treats major water bodies of the world. For additional information on the dynamics of water within gulfs and bays, see above Waves of the sea (ocean).

These bodies of water may differ from the adjacent ocean (or sea) by virtue of water properties and dynamics and processes of sedimentation. Such differences are determined by the size and the shape of a given gulf, by the depth and bottom topography, and, to a considerable extent, by the degree of isolation from the ocean. Climatic conditions also are important. Isolation from an adjacent ocean depends on the ratio of width of mouth to total surface area of a gulf or on the cross section of the mouth to total water volume. If there is a sill (a submarine ridge or rise), the ratio of depth above the sill to the depth of the gulf is of great importance. No extensive comparisons of these ratios have been made to date; hence any analysis of controlling variables must remain somewhat qualitative.

A high sill hampers the water exchange between an ocean and gulf and may lead to stagnation (oxygen deficiency) as is found in some fjords of Norway, in the Red Sea, and, particularly, in the Black Sea. Also, the presence of a sill causes independent circulation of gulf waters, generated by local winds and the runoff of rivers. Sills are not indispensable for the formation of an independent circulation, however. A narrow mouth, as in the Gulf of Bothnia (Bothnia, Gulf of), leads to the same result.

In humid climates, the waters of gulfs are freshened by river runoff. Salinity is particularly low in the gulfs of the Baltic Sea and along the southern coast of the Kara Sea. Water becomes almost fresh in their heads, especially in the spring when snow begins to thaw. Gulfs of the arid zone suffer from intensive evaporation and receive little river runoff. Thus, salinity increases markedly in this climatic regime—up to 60 parts per thousand in the Persian Gulf and up to 350 parts per thousand in the Kara-Bogaz-Gol (a gulf of the Caspian Sea). In addition to its effect on salinity, river runoff delivers organic matter and nutrient salts that may determine the specific features of life in the gulfs. The number of genera and species of organisms is small, but the organisms present tend to develop in quantities. That is why shrimp, oyster, and other fisheries are concentrated in many gulfs.

Funnel-shaped gulfs, in which the depth gradually decreases headward, usually have resonant tides (tide). The tidal range at the head of such gulfs is several times greater than that in the open ocean (e.g., Bristol Channel, Río de la Plata, Mezenskaya Bay, Shelikhova Gulf). The world maximum tidal range has been registered in the Bay of Fundy (18 metres). The regularity (magnitude and frequency) of the flood tide may be distorted in such instances, and the duration of the flood tide may become much shorter than that of the ebb tide. This may cause the phenomenon of tidal bore, in which a steep wave will move rapidly upstream for dozens of kilometres.

Gulfs of simple shape with a narrow mouth and a high degree of isolation from the ocean are often subject to seiches (seiche). These free oscillations can result from rapid changes of atmospheric pressure and, of course, from tectonic movements such as earthquakes. Seiches gradually decrease, but some oscillation continues long after their cause disappears. A high rise of the water (storm surge) occurs in long and shallow gulfs if winds from the sea are protracted. Such phenomena are difficult to predict, and the high water levels may cause floods. Seiches commonly occur at the heads of Helgoländer Bay in the North Sea and in the Gulf of Finland (Finland, Gulf of).

Certain aspects of sedimentation are affected by the isolation of gulfs from the ocean and river runoff. The rate of sediment accumulation in gulfs of limited area may be very high. This, of course, is a function of river discharge; sediment composition is usually similar to that of the load transported by entering rivers. Deposition of calcium carbonate often occurs in shallow gulfs in the arid zones where few if any perennial streams exist. The bottoms of long gulfs (or gulfs having sills) are usually covered with silt even at the shallowest depths (e.g., Hudson Bay, the Po Hai, the inlets or gubas of the Kara Sea, the Gulf of Riga). Only strong tidal currents (ocean current) can prevent this siltation and, in some cases, cause the opposite phenomenon of bottom erosion. Currents maintain the existence of or actively deepen bottom troughs in narrow-mouthed gulfs whose depths are over 200 metres, whereas depths of adjacent parts of the open ocean are only on the order of some dozens of metres.

Waves (wave) of the open ocean either do not penetrate into comparatively isolated gulfs or, if they do, they become greatly reduced after entry. Small local waves that are related to gulf size prevail there. This tends to make gulfs quite navigable, and seaports and harbours have generally been situated on them.

Areas situated in open concavities of the continental coast (Gulf of Alaska (Alaska, Gulf of), Bay of Biscay (Biscay, Bay of), Gulf of Guinea (Guinea, Gulf of), Great Australian Bight, Bay of Bengal, Gulf of Tehuantepec, for example) are classified as the A₁ group. The depth of these gulfs in the region of the mouth usually is on the order of kilometres. The continental shelf and slope are generally pronounced. The general shape of such gulfs is simple; width of mouth usually exceeds its length. Water circulation and its physical properties are similar to those of the oceans. The character of the marine faunas does not differ from that of oceanic areas.

Ocean gulfs, such as the gulfs of Oman (Oman, Gulf of), California (California, Gulf of), Aden (Aden, Gulf of), and some others, have smaller areas and are isolated to a lesser degree. These gulfs and bays, in group A₃, have shapes that are determined by young faults and fractures. Depths in these gulfs generally exceed one kilometre. Unlike the previous group, in which gulfs might be of composite geologic structure, these occupy areas that have undergone only a single episode of deformation.

Gulfs of intercontinental and marginal seas are considered to be a third category. These may be divided into group C₁, which consists of gulfs of basin seas, including the deepwater part only (Aqaba) or both the deepwater and the shelf parts (the Gulf of Honduras), and group C₂, the shelf gulfs of the same seas (e.g., the Persian Gulf, Gulf of Suez, Anadyrsky Gulf (Anadyr, Gulf of), Bristol and Norton channels, Shelikhova Gulf). (Shelikhov, Gulf of)

The sea is generally accepted by scientists as the place where life began on Earth. Without the sea, life as it is known today could not exist. Among other functions, it acts as a great heat reservoir, leveling the temperature extremes that would otherwise prevail over the Earth and expand the desert areas. The oceans provide the least expensive form of transportation known, and the coasts serve as a major recreational site. More importantly, the sea is a valuable source of food and a potentially important source of energy and minerals, all of which are required in ever-increasing quantities by industrialized and developing nations alike.

From the beginning of recorded history, people have used the sea as a means of transporting themselves and their goods. The bulk of the tonnage of products transported throughout the world today continues to be moved in ocean vessels. The size of these vessels ranges from small boats capable of carrying a few tons to bulk carriers (e.g., supertankers) capable of transporting more than 500,000 tons of oil. The cost of transporting goods on the ocean depends on the product, the form of shipment, and the type of vessel. As the per capita consumption of materials increases, the outlook for marine transportation is one of ever-increasing tonnages and size of carrying vessels.

Since the laying of the transatlantic cable in the 19th century, the oceans have served as a major means of communication between continents and islands. Hundreds of seafloor cables connect many large centres of world population. With the development of satellite communications, seafloor cables as a means of communication have decreased somewhat in importance, but they will continue to carry information for many decades to come.

In addition to communications, cable and pipes laid on the seafloor carry electrical energy, oil, and other commodities in many parts of the world.

Many millions of tons of edible fish and shellfish are taken from the oceans each year. Yet, the food-producing potential of the sea is considerably greater than this. The prevailing methods by which fish are taken from the oceans are inefficient. This problem is compounded by the fact that only a handful of varieties are targeted by commercial fishermen (anchovy, sardine, herring, cod, mackerel, and pollack constitute more than half of the total annual catch). Overfishing of certain varieties and areas along the continental shelves also has resulted in declining numbers and therefore catches. The decimation of the whale (many species are near extinction) and also of the anchovy off the waters of Peru exemplifies the consequences of overfishing.

Several alternatives have been proposed to enhance productivity. One possibility is to extract protein concentrates from all types of fish, promoting the use of varieties formerly ignored for foodstuff. It has been estimated that this procedure could produce a sustained yield of two billion tons of food annually for the world's populace. Another alternative would be to use protein derived from algae cultivated on the continental shelf. A more viable course of action would be to develop the continental shelf for fish- and kelp-farming (aquaculture). The Japanese have instituted a substantial research program in this area. They have farmed oysters in their oceanic bays for many years and more recently raised sea bream and shrimp in protected environments. Commercial shrimp and oyster farms also have been developed in the United States in the shallow waters of estuaries and bays. Both commercial and research aquaculture projects have been conducted to raise abalone, Maine lobster, salmon, and edible forms of seaweed under controlled conditions. Limited sea farming has been practiced in France and Italy, both of which raise mussels and certain other shellfish.

Various marine organisms have been artificially cultivated as sources for therapeutic drugs. For example, during the late 1980s American researchers set out to grow microalgae that have a high concentration of a substance called beta-carotene, which is believed to prevent or retard the spread of certain forms of cancer.

The need for water for domestic and agricultural purposes has grown steadily throughout the latter half of the 20th century. As a result, increasing attention has been given to desalting ocean waters and brackish waters in inland seas. Throughout the world, more than 3,500 land-based desalination plants, producing a total of more than eight billion litres (in excess of two billion gallons) per day, were in operation by the early 1990s. In general, the desalination plants are located in areas where the population has outstripped the onshore water supply and where high-cost desalinated water can be afforded. This situation tends to arise in coastal-desert areas or on densely populated islands, because the cost of pumping water through pipelines to interior areas would add prohibitively to the basic cost at the sites of desalination. The United States operates a little more than 30 percent of the world's desalination facilities. Another 20 percent or so are found in the Middle East, chiefly in Israel, Saudi Arabia, Kuwait, and the United Arab Emirates. Together, these Middle Eastern plants produce roughly two-thirds of the world's desalinated water.

A population usually can afford to pay about 10 times as much for water for domestic purposes as it does for agricultural water. Large-scale nuclear desalination facilities promise to lower the cost of desalted water at the desalination sites to a level that most industries and some agricultural enterprises can afford. As yet, however, no such plant has been constructed.

At present desalination is accomplished primarily by distillation and membrane processes. Distillation processes involve some form of evaporation and subsequent condensation. Many of the largest commercial desalination plants use multiple-effect distillation (i.e., multistage-flash distillation). Roughly half of the world's desalting facilities employ distillation processes and account for approximately 75 percent of all desalinated water produced annually.

Membrane processes for desalting include reverse osmosis and electrodialysis. Of the two, reverse osmosis is the more widely used, particularly for desalting brackish waters from inland seas. In this method the natural process of osmosis is reversed by applying pressure to brine that is in contact with an osmotic membrane. The membrane impedes the passage of salt ions while allowing fresh water to move through. In electrodialysis an electric potential is used to drive positive and negative ions of the dissolved salts through membranous filters, thereby sharply reducing the salt content of the water between the filters.

In the future, it can be expected that the ocean will become an increasingly important source of fresh water. If production and transportation costs can be lowered sufficiently, it may be possible to produce fresh water to irrigate large areas that border the oceans in many parts of the world.

There are several recognized techniques by which energy can be extracted from the sea. The major problem in taking energy resources from the sea is that they tend to be diffused over a large lateral area. A point-concentration energy source is necessary if it is to be exploited economically.

Hydraulic turbine-generator units are presently used to extract energy from ocean tides, although on a very limited scale. As of the late 1980s, operating tidal units included a 240-megawatt plant in France on the estuary of the Rance River and several smaller installations, as, for example, a 40,000-kilowatt pilot plant in Russia on the Barents Sea. There are few sites throughout the world that are suitable for harnessing tidal energy without constructing prohibitively expensive damlike structures. (These are required to trap water at high tide and release it at low tide to turn the hydraulic turbine; see turbine: Water turbines (turbine).) Thus, it seems unlikely that tidal power will play a more significant role in the coming years.

Another more promising technology, known as ocean thermal energy conversion (OTEC), makes use of the temperature differential between the warm surface waters of the oceans, heated by solar radiation, and the deeper cold waters to generate power in a conventional heat engine. The difference in temperature between the surface and lower water layer can be as large as 50° C over vertical distances of as little as 90 metres in some ocean areas. To be economically practical, the temperature differential should be at least 20° C in the first 1,000 metres below the surface.

The OTEC concept was first proposed in the early 1880s by the French engineer Jacques-Arsìne d'Arsonval. His idea called for a closed-cycle system, a design that has been adapted for most present-day OTEC pilot plants. Such a system employs a secondary working fluid (a refrigerant) such as ammonia. Heat transferred from the warm surface ocean water causes the working fluid to vaporize through a heat exchanger. The vapour then expands under moderate pressures, turning a turbine connected to a generator and thereby producing electricity. Cold seawater pumped up from the ocean depths to a second heat exchanger provides a surface cool enough to cause the vapour to condense. The working fluid remains within the closed system, vaporizing and reliquefying continuously.

Some researchers have centred their attention on an open-cycle OTEC system that employs water vapour as the working fluid and dispenses with the use of a refrigerant. In this kind of system warm surface seawater is partially vaporized as it is injected into a near vacuum. The resultant steam is expanded through a low-pressure steam turbogenerator to produce electric power. Cold seawater is used to condense the steam, and a vacuum pump maintains the proper system pressure.

During the 1970s and '80s the United States, Japan, and several other countries began experimenting with OTEC systems in an effort to develop a viable renewable energy source. In 1979 American researchers put into operation the first OTEC plant able to generate usable amounts of electric power—about 15 kilowatts of net power. This unit, called Mini-OTEC, was a closed-cycle system mounted on a U.S. Navy barge a few kilometres off the coast of Hawaii. In 1981–82 Japanese companies tested another experimental closed-cycle OTEC plant. Located on the Pacific Island republic of Nauru, this facility produced 35 kilowatts of net power. Since that time researchers have continued developmental work to improve heat exchangers and to devise ways of reducing corrosion of system hardware by seawater.

The prospects for the commercial application of OTEC technology seem bright, particularly on islands and in developing nations in the tropical regions where conditions are most favourable for OTEC plant operation. It has been estimated that the tropical ocean waters absorb solar radiation equivalent in heat content to that of about 170 billion barrels of oil each day. Removal of this much heat from the ocean would not significantly alter its temperature, but would permit the generation of about 10 million megawatts of electricity on a continuous basis.

In the mid-1950s, the production of oil and gas from oceanic areas was negligible. By the early 1980s, about 14 million barrels per day, or about 25 percent of the world's production, came from offshore wells, and the amount continues to grow. More than 500 offshore drilling and production rigs were at work by the late 1980s at more than 200 offshore locations throughout the world drilling, completing, and maintaining offshore oil wells (see oil shale (petroleum)). Estimates have placed the potential offshore oil resources at about 2 trillion barrels, or about half of the presently known onshore potential oil sources.

It was once thought that only the continental-shelf areas contained potential petroleum resources, but discoveries of oil deposits in deeper waters of the Gulf of Mexico (about 3,000 to 4,000 metres) have changed that view. It is now believed that the continental slopes and neighbouring seafloor areas contain large oil deposits, thus enhancing potential petroleum reserves of the ocean bottom.

Offshore drilling is not without its drawbacks, however. Not only is it difficult and expensive to drill on the continental shelf and in deeper water, but there is also the risk of accidental discharges of oil that can cause serious damage to the environment and marine organisms. In spite of technological advances, inadvertent leaks, sometimes on a large scale, continue to occur.

The rivers of the world dump billions of tons of material into the oceans each year. Seafloor springs and volcanic eruptions also add many millions of tons of elements. Even the winds contribute solid materials to the oceans in appreciable quantities. Most of these sediments rapidly settle to the seafloor in nearshore areas, in some cases forming potentially valuable placer mineral deposits. The dissolved load of the rivers, however, mixes with seawater and is gradually dispersed over the total oceanic envelope of the Earth. Because of the nature of the minerals and their mode of formation, it is convenient to consider the occurrence of ocean deposits in several environments—marine beaches, seawater, continental shelves, sub-seafloor consolidated rocks, and the marine sediments of the deep-sea floor. Minerals are mined from all of these environments except for the deep-sea floor, which was only recently recognized as a repository for mineral deposits of unbelievable extent and significant economic value.

Minerals that resist the chemical and mechanical processes of erosion in nature and that possess a density greater than that of the Earth's common minerals have a tendency to concentrate in gravity deposits known as placers (placer deposit). During the Pleistocene glaciations, sea level was appreciably lowered as the ocean water was transferred to the continental glaciers. Because of the cyclical nature of the ice ages and the intervening warm periods, a series of beaches were formed in nearshore areas both above and below present sea level. Also, when sea level was lowered in past ages, the streams that today flow into the sea coursed much further seaward, carrying placer minerals to be deposited in channels that are now submerged. With geophysical exploration techniques, these channels and beaches can be easily delineated, even though these features are totally covered by Holocene sediments—i.e., those deposited during the past 10,000 years.

Sand and gravel are mined from a number of offshore locations around the world, generally with hydraulic dredges. They are used primarily for construction purposes or for beach replenishment or nearshore fills.

Of considerable interest are the seafloor phosphorite deposits on the coastal shelves of many nations. The phosphorite off California occurs as nodules that vary in shape from flat slabs a few metres across to small spherical forms termed oolites. The nodules commonly are found as a single layer at the surface of coarse-grained sediments. Phosphorite composition from the California offshore area is surprisingly uniform and contains potentially economically attractive amounts of phosphorus. Another type of phosphate deposit has been discovered off the west coast of Mexico. It contains as much as 40 percent apatite (common phosphate mineral), and some experts have speculated that up to 20 billion tons of recoverable phosphate rock exist in the deposit.

Mineral deposits of enormous size and potential economic significance have been discovered on the deep ocean floor. Minerals formed in the deep sea are frequently found in high concentrations because there is relatively little clastic material generated in these areas to dilute the chemical precipitates.

An estimated 10¹⁶ tons of calcareous oozes, formed by the deposition of calcareous shells and skeletons of planktonic organisms, cover some 130 million square kilometres of the ocean floor. In a few instances, these oozes, which occur within a few hundred kilometres of most nations bordering the sea, are almost pure calcium carbonate; however, they often show a composition similar to that of the limestones used in the manufacture of portland cement.

Covering about 39 million square kilometres of the ocean floor in great bands across the northern and southern ends of the Pacific Ocean and across the southern ends of the Indian and Atlantic oceans are other oozes, consisting of the siliceous shells and skeletons of plankton animals and plants. Normally these oozes could serve in most of the applications for which diatomaceous earth is used, for fire and sound insulation, for lightweight concrete formulations, as filters, and as soil conditioners.

An estimated 10¹⁶ tons of red clay covers about 104 million square kilometres of the ocean floor. Although compositional analyses are not particularly exciting, red clay may possess some value as a raw material in the clay-products industries, or it may serve as a source of metals in the future. The average assay for alumina is about 15 percent, but red clays from specific locations have assayed as high as 25 percent alumina; copper contents as high as 0.20 percent also have been found. A few hundredths of a percent of such metals as nickel and cobalt and a percent or so of manganese also are generally present in a micronodular fraction of the clays and in all likelihood can be separated and concentrated from the other materials by a screening process or by some other physical method.

Underlying the hot brines in the Red Sea are basins containing metal-rich sediments that potentially may prove to be of considerable significance. It has been estimated that the largest of several such pools, the Atlantis II Deep, contains several billion dollars worth of copper, zinc, silver, and gold in relatively high grades. These pools lie in about 2,000 metres of water midway between The Sudan and the Arabian Peninsula. Because of their gellike nature, pumping these sediments to the surface may prove relatively uncomplicated. These deposits are forming today under present geochemical conditions and are similar in character to certain major ore deposits on land.

The discovery, in 1978, of polymetallic sulfides at the mid-ocean spreading centres has aroused much interest. As noted elsewhere in the article, these sulfides include sediments enriched in iron and manganese. Sites of rich deposits have been located at the Galápagos spreading centre in the Gulf of California and at the East Pacific Rise.

From an economic standpoint, the most interesting oceanic sediments are manganese nodules—small, black to brown, friable lumps found to be widely distributed throughout the major oceans in the late 19th century by the Challenger and Albatross expeditions.

Many theories have been proposed to account for the formation of manganese nodules, the best probably being that the ocean is saturated at its present state of acidity-alkalinity in iron and manganese. For this reason, these elements precipitate as colloidal particles that gradually increase in size and filter down to the seafloor. Colloids of manganese and iron oxides collect many metals and tend to agglomerate as nodules at the seafloor rather than settle as particles in the general sediments. An estimated 1.5 trillion tons of manganese nodules lie on the Pacific Ocean floor alone. Averaging about four centimetres in diameter and found in concentrations as high as 38,600 tons per square kilometre, these manganese nodules contain as much as 2.5 percent copper, 2.0 percent nickel, 0.2 percent cobalt, and 35 percent manganese. In some deposits, the content of cobalt and manganese is as high as 2.5 percent and 50 percent, respectively. Such concentrations would be considered high-grade ores if found on land, and, because of the large horizontal extent of the deposit, they are a potential source of many important industrial metals.

Relatively simple mechanical cable bucket or hydraulic dredges with submerged motors and pumps can effect the mining of the nodules at rates as high as 10,000 to 15,000 tons per day, from depths as great as 6,000 metres. The estimated costs of mining and processing the nodules indicate that copper, nickel, cobalt, and other metals can be economically produced from this source.

The prospects of mining the manganese nodules and metal-rich sediments have brought home the need to resolve long-standing legal problems relating to the ownership of marine resources. During the 18th century the extent of the territorial sea (and therefore rights) was established as 3 nautical miles (5.6 kilometres) from a nation's shoreline. The area beyond the territorial sea, the so-called high seas, was regarded as open to all nations. By the mid-1940s technological advances had extended offshore oil drilling beyond the territorial limit. This situation, together with the desire of various coastal nations to protect their fishing grounds, eventually resulted in an attempt to codify international law concerning territorial waters, ocean resources, and sea lanes. A 1982 treaty calling for the enactment of the United Nations Law of the Sea (Sea, Law of the) Convention was signed by 138 nations; some 30 other states, including the United States, the United Kingdom, and West Germany, refused to sign, however. The treaty extended the territorial limit of each coastal nation to a distance of 12 nautical miles and granted it sovereign rights over natural resources—living and nonliving—within an exclusive economic zone (or EEZ) of 200 nautical miles. The nations that refused to sign the treaty objected to its provisions governing seabed mining. The treaty declared the minerals on the seafloor beneath the high seas the “common heritage of mankind” and stipulated that their exploitation be directed by a global authority. While private and national mining concerns are allowed to conduct exploration and set up extraction operations, the question of seabed mineral ownership and mining rights remains largely unresolved. This situation is viewed in some quarters as the primary obstacle to full and effective utilization of seabed resources.

One of the least known but most significant uses of the sea is as an enormous dump site. In the past, the oceans were able to assimilate the wastes of society without noticeable adverse effects. However, industrialization and other concomitant developments, along with sharp increases in global population, have given rise to quantities and forms of waste that are now taxing the capacity of the oceans to absorb them. Extensive marginal areas of the oceans have been heavily polluted by human wastes ranging from the raw sewage of urban centres to junked appliances and automobiles. Less apparent but more insidious forms of pollution are toxic chemicals, nuclear wastes, and oily bilges pumped by practically all vessels using petroleum for power (see also above Chemical and physical properties of seawater: Composition of seawater: Effects of human activities (ocean)).

Some other human activities are equally harmful to the marine environment. Massive oil spills from tanker accidents, such as the 1989 mishap involving the Exxon Valdez in Prince William Sound, Alaska, have not only disfigured innumerable beaches and estuaries but caused widespread damage to wildlife as well. Large power plants are generally located along coastlines to reduce the costs involved in cooling their condensers by water-circulation systems. Although the whole of the ocean never will be affected by the waste heat dissipated by these plants, detrimental environmental effects can be caused in the immediate area of the power-plant outfall. Herbicides and pesticides (especially the organochlorides still used by some countries) reach the oceans via the wind and rivers and contaminate marine organisms.

The fringes of the oceans—the beaches, lagoons, and bays—are the most sensitive to human action, but the continued dumping of wastes, attended by other abuses, will eventually affect the entire marine environment.