Polynyas in the Southern Ocean They are vast gaps in the sea ice around Antarctica. By exposing enormous areas of seawater to the frigid air, they help to drive the global heat engine that couples the ocean and the atmosphere

by ArnoldL. Gordon and Josefino C. (omiso

uring the austral winter(the In these regions, called polynyas, the largest and longest-lived polynyas. It months between June and Sep­ surface waters of the Southern Ocean is not yet fully understood what forces Dtember) as much as 20 million (the ocean surrounding Antarctica) are create and sustain open-ocean polyn­ square kilometers of ocean surround­ bared to the frigid polar atmosphere. yas, but ship and satellite data gath­ ing Antarctica-an area about twice Polynyas and their effects are only ered by us and by other investigators the size of the continental U.S.-is cov­ incompletely understood, but it now are enabling oceanographers to devise ered by ice. For more than two cen­ appears they are both a result of the reasonable hypotheses. turies, beginning with the voyages of dramatic interaction of ocean and at­ Captain James Cook in the late 18th mosphere that takes place in the Ant­ n order to understand the specific century, explorers, whalers and scien­ arctic and a major participant in it. forces that create and sustain po­ tists charted the outer fringes of the The exchanges of energy, water and Ilynyas and the effects polynyas can ice pack from on board ship. Never­ gases between the ocean and the at­ have, one must first understand the theless, except for reports from the mosphere around Antarctica have a role the Southern Ocean plays in the few ships that survived after being major role in determining the large­ general circulation of the world ocean trapped in the pack ice, not much was scale motion, temperature and chemi­ and in the global climate as a whole. known about the ice cover itself. cal composition of the ocean and at­ Because the Southern Ocean has ex­ Since 1973, however, satellites bear­ mosphere throughout the globe. Po­ tensive open connections to the rest ing passive microwave sensors have lynyas accelerate these processes by of the world ocean, and because it is made it possible to survey the ice exposing the surface of the Southern much larger than the polar ocean in cover routinely from space, and these Ocean directly to the atmosphere. the Arctic region, the Antarctic plays a observations have brought about a There are two general kinds of po­ larger role in the global ocean circu­ marked change in investigators' view lynyas: coastal polynyas and open­ lation than the Arctic does. of the Antarctic ice. One of the most ocean polynyas. Coastal polynyas de­ When solar radiation warms the surprising findings of the satellite era velop when strong local winds origi­ tropical and subtropical oceans, the is that the ice cover is not at all a nating on the Antarctic continent blow resulting heat diffuses downward into continuous blanket. Within its borders ice away from the shoreline as it freez­ deeper water. Heat from the sun also there are many small breaks, on the es, leaving an ice-free area between causes relatively high rates of evapo­ scale of from one to 10 kilometers, the coast and the ice pack. Even before ration, which raises the salinity of the called leads. More surprisingly, there satellite data were available, investiga­ surface waters; the excess salt, like the are sometimes vast regions-up to tors had known about the existence of heat, diffuses into the depths. The 350,000 square kilometers in area­ coastal polynyas. They were very sur­ warmth and salinity of the deep water that are completely free of ice. prised, however, to learn from satellite are also enhanced by deep-reaching observations how widespread coastal mixing events in the North Atlantic, polynyas are and how large they can in which relatively warm, high-salinity ARNOLD L. GORDON and JOSEFINO C. COMISO combine sea- and satellite­ become. (In some coastal polynyas the surface water is mixed downward. The based research in their study of polyn­ distance from the shoreline to the resulting mass of warm, salty deep yas. Gordon is professor of physical border of the ice pack is as much as 50 water gradually spreads toward the oceanography at Columbia University to 100 kilometers.) Southern Ocean. In the Antarctic this and is on the senior staff of Columbia's Open-ocean polynyas form within water, now called the Circumpolar Lamont-Doherty Geological Observato­ the body of the ice cover, far from the Deep Water, rises toward the surface. ry. His research is focused on large-scale coast. Some of these are by far the There it gives up some heat to the circulation and mixing in the ocean and their relation to the global climate. Co­ miso is a physical scientist at the Na­ tional Aeronautics and Space Adminis­ WEDDELL POLYNYA, an enormous ice-free region amid the ice cover near the Weddell tration's Goddard Space Flight Center. Sea, formed during three consecutive Southern Hemisphere winters. In these satel­ He went to Goddard in 1979 after doing lite images, which were made in September of 1974 (top), 1975 (middle) and 1976 research in experimental particle phys­ (bottom), the white region represents the Antarctic landmass and colored regions ics at various institutions. At Goddard he has helped to demonstrate the utility represent ocean areas covered by various concentrations of sea ice. Pink and purple of passive-microwave and infrared data regions are almost completely covered by ice and light blue regions are ice-free. In from satellites in the study of oceanic summer the polynya disappeared with the melting of the ice cover. At its largest and atmospheric processes. the polynya measured about 350 by 1,000 kilometers.It had measurable effects on the temperature of the underlying ocean at depths as great as 2,500 meters.

90 SCIENTIFIC AMERICAN June 1988

© 1988 SCIENTIFIC AMERICAN, INC 1974

45'W 45"£ I I

SCIENTIFIC AMERICAN June 1988 91

© 1988 SCIENTIFIC AMERICAN, INC -

MERIDIONAL CIRCULATION PATTERN of the Southern Ocean termediate Water. Another fraction of the surface water moves (the ocean surrounding Antarctica) is dominated by the up­ toward the pole; it eventually sinks and flows northward as welling of a warm, salty water mass called the Circumpolar Antarctic Bottom Water. The circulation is influenced by polyn­ Deep Water and its transformation into Antarctic Surface Wa­ yas. In so-called open-ocean polynyas (center) heat is vented to ter, which ultimately sinks to become Antarctic Intermediate the atmosphere rapidly because there is no insulating layer of Water and Antarctic Bottom Water. The circulation is driven by ice. Convection currents, in which warm water rises as cold wind and the exchange of heat and fresh water between the water sinks, accelerate the movement of warm water toward ocean and the atmosphere. The upwelled water is converted the surface. So-called coastal polynyas (right) form when ice is into cold, relatively fresh surface water as it loses heat to the blown away from the continent as fast as it freezes. The re­ atmosphere and gains fresh water from precipitation and the sulting open area vents heat rapidly, and the salt that is re­ melting of ice. Some of the surface water moves northward and jected when sea ice forms increases the density of the surface gradually sinks, contributing to the formation of Antarctic In- water and accelerates the formation of Antarctic Bottom Water.

atmosphere and becomes a cold, ocean, by drawing water to the sur­ heat is transferred from the Equator dense mass of water, which sinks face, allowing it to approach equilibri­ to the pole helps to determine the toward the sea floor. This so-called um with the temperature and compo­ degree to which temperature varies Antarctic Bottom Water then moves sition of the atmosphere and forcing it with latitude on a global scale. This northward along the ocean floor, trav­ downward again. The process cools in turn has a part in driving many of eling well beyond the Equator; it is the deep ocean, lowers its salinity and the processes that govern the earth's the major mass of bottom water in restores oxygen levels that have been weather and climate. the world. As it moves north it grad­ depleted by organisms. The deep­ ually mixes upward, and the cycle of ocean overturning may also be signifi­ erhaps the most important for­ warming and cooling begins again. cant in establishing a rough balance ces driving overturning in the The Southern Ocean is thus one part between the oceanic concentrations of PSouthern Ocean are buoyancy of an enormous heat engine that such dissolved gases as carbon diox­ and the wind. Buoyancy gives the drives the motions of much of the ide and the levels of those gases in the ocean a layered structure, in which world ocean. In addition, the massive atmosphere. Hence the overturning is masses of water are stratified accord­ overturning of water that takes place an important factor in such events as ing to their density (which in turn is in the Antarctic provides a mechanism the "greenhouse warming" that may determined by the water's tempera­ by which the chemistry of the atmos­ take place as carbon dioxide gas is ture and salinity). The wind, on the phere can affect the chemistry of the added to the atmosphere. other hand, tends to disturb this strat­ deep ocean and vice versa. The over­ The heat engine's effects do not ification. In addition, the exchange of turning in a sense ventilates the deep stop there. The overall rate at which heat and fresh water between the sur-

92 SCIENTIFIC AMERICAN June 1988

© 1988 SCIENTIFIC AMERICAN, INC face water and the atmosphere alters water continues to be pushed either the density of the surface water, and toward or away from the pole by the so it can stimulate convection cur­ prevailing winds. The poleward-flow­ rents (in which buoyant water moves ing portion continues to increase in upward and denser water moves density as it mixes with the brine re­ downward) that lead to a rearrange­ jected by newly forming sea ice. When ment of the stratification. this surface water reaches the Antarc­ The top layer of water in the South­ tic continental shelf, it sinks, flowing ern Ocean, extending to a depth of off the continental shelf to the ocean about 100 meters, is called the Antarc· bottom, where it forms the Antarctic tic Surface Water. The Antarctic Sur­ Bottom Water. In this way the effects face Water is cold because of its prox­ of buoyancy and wind have combined imity to the extremely cold Antarctic to draw warm deep water from north­ atmosphere, and it is relatively fresh ern latitudes, transfer its heat to the because in the Southern Ocean precip· Antarctic atmosphere and send it itation (and the melting of Antarctic Sinking toward the ocean floor. glacial ice) exceeds evaporation. Although the formation of the ice North of about 65 degrees south cover acts to accelerate this heat en­ latitude, prevailing winds push much gine, the presence of the ice itself of the surface water farther to the paradoxically tends to impede the north-away from the pole-while deep-ocean overturning. The ice cover south of that latitude other winds insulates the ocean surface, inhibiting push surface water to the south, to­ the venting of oceanic heat. Some heat ward the pole. The resulting diver­ is still conducted to the atmosphere gence at 65 degrees south causes through the ice, but conduction is not an upwelling: deeper water is lifted nearly as efficient as the direct expo­ to replace the water that has been sure of open water to the atmosphere pushed north or south. The effect would be; a layer of ice one meter thick is to bring the warmer, saltier layer (somewhat thicker than most of the below the Antarctic Surface Water up Antarctic ice cover) can reduce the toward the surface, where it is con­ heat loss of the winter ocean by a verted into Antarctic Surface Water. factor of as much as 10 to 100. The source of the warmer water is the Circumpolar Deep Water, which oW do polynyas, in which the is drawn to the Southern Ocean in ice cover is breached, affect part by the wind-driven upwelling. Al­ Hthese processes? This question though the Circumpolar Deep Water, would be impossible to address with­ which extends to a depth of roughly out data from satellites. Satellite ob­ 2,000 meters, is warmer than the sur­ servations provide consistent, repeti­ face water, it is still slightly denser tive and long-term coverage, and they because it has relatively high salinity. also make it possible for investigators For that reason it tends to remain to follow changes that take place over below the surface water. time frames as short as a week or even Nevertheless, the surface water and a single day. the deep water south of about 60 Satellite-borne sensors can operate degrees south latitude are so similar in a variety of wavelengths and ob­ in density that their stratificationpat­ serving modes. The sensor that has tern is only marginally stable. Very shown itself to be the most versatile small forces can disturb it. Those for­ and useful for observations in polar ces are provided by the wind and by regions is the passive microwave radi­ changes in the freshwater balance of ometer, which detects the radiation the surface layer. Strong winds can having wavelengths between one mil­ greatly enhance the mixing of the two limeter and one meter that is emitted layers, particularly if ice floes are pres­ naturally by various materials. This ent to help create turbulence. The for· radiation is emitted during dark peri­ mation of sea ice in the winter also has ods as well as during daylight hours a role in this mixing: when seawater and passes through most cloud cover, freezes, the ice crystals reject much of making it possible to observe a given the seawater's salt, expelling it in a region at any time of day and through concentrated brine that increases the almost any kind of weather. The mi­ EXTENT of the Antarctic ice cover varies salinity of the surface layer and there­ crowave radiometer is ideal for moni­ from season to season. These satellite by makes it more susceptible to mix­ toring sea· ice cover, because at some images were made in (from top to bottom) ing with the underlying deep water. microwave wavelengths the emissivity March, June, September and December When the Circumpolar Deep Water of ice and water is very different. of 1984-respectively the southern sum· mixes upward as a result of these The amount of radiation emitted in mer, fall, winter and spring. In the south­ processes, it vents heat to the atmos­ a given wavelength by an area of sea ern summer ice covers about four mil­ phere and cools, becoming denser. At ice depends on a number of variables, lion square kilometers of ocean, and in the same time much of the surface including the ice's temperature, thick- the winter it covers about 20 million.

SCIENTIFIC AMERICAN June 1988 93

© 1988 SCIENTIFIC AMERICAN, INC ness, salinity and snow cover. By mak­ irregularities in the satellite's orienta­ steady temperature in the process of ing observations in a number of wave­ tion. In 1987 a new sensor, called the freezing. The heat flux to the atmos­ lengths simultaneously, it is possible Special Sensor Microwave Imager, was phere due to latent-heat polynyas is to eliminate the complications intro­ launched as part of the Defense Mete­ thought to be in excess of 300 watts duced by these factors and to derive a orological Satellite Program, ensuring per square meter, enough to add a good estimate of the total ice cover in continuous satellite coverage. lO-centimeter-thick layer of new ice a given region. Furthermore, by com­ (which is subsequently removed by bining observations in different wave­ he existence and extent of po­ the wind) to the coastal region every lengths, one can derive many other lynyas are among the most in­ day. The ice generated in latent-heat geophysical parameters, such as the Tteresting finilings to come from polynyas may supply much of the sea temperature of the ice and the sea satellite observation of the Antarc­ ice in the adjacent ocean. surface, the amount of water vapor in tic. As we have mentioned, there are Meanwhile, as the ice forms within the atmosphere, the position of the ice two kinds of polynyas: coastal polyn­ the polynya, the brine it rejects raises edge and even the speed of the wind yas and open-ocean polynyas. Satellite the salinity of the water overlying the over the ocean. data, shipboard observations and the­ continental shelf and helps to drive The first spaceborne passive micro­ oretical analysis have now shown that the formation of the Antarctic Bottom wave sensor was the Electrically Scan­ the two kinds of polynyas are the re­ Water. According to rough calcula­ ning Microwave Radiometer on board sult of very different causes and have tions based on satellite data, the for­ the Nimbus 5 satellite, which was quite different effectson the deep­ mation of ice in coastal polynyas launched in December, 1972. This ra­ ocean overturning. leaves behind enough salt to form diometer, which measured microwave The coastal polynyas are essentially bottom water from Antarctic Surface radiation in only one frequency, trans­ sea-ice factories. They appear when Water at the rate of about 10 million mitted good data for about four years. local winds push newly forming sea cubic meters per second. For later observations investigators ice away from the continent as quickly The enormous scale at which coast­ relied on the Scanning Multichannel as it freezes. This exposes an area of al polynyas create ice, cool the ocean Microwave Radiometer on board the open ocean on which more ice can and add salt to the Antarctic Bottom satellite Nimbus1. This instrument, form, continuing the process. Coastal Water took oceanographers complete­ which recorded the intensity and po­ polynyas are also called latent-heat ly by surprise when it was first de­ larization of radiation in fivedistinct polynyas, because the heat they re­ duced from satellite observations. The frequencies, was in operation from lease to the atmosphere is primarily in knowledge we have gained about 1978 until 1987, when the microwave the form of the "latent heat" that liq­ coastal polynyas has helped to explain scanner was turned off because of uid water gives off while it hovers at a several previously puzzling features

SATElliTE IMAGES of the Weddell Polynya made in August, identify such important features as the border between the 1975, illustrate the advantages of passive-microwave imagery Antarctic ice cover and the open ocean (boundary between blue (right) over conventional infrared imagery (left). In the infrared and gray at top left) and the edge of the continental landmass image the polynya is visible as the dark area at the center, but (orange border at bottom right). In this image the various col­ cloud cover makes it impossible to determine the polynya's ors represent the "brightness" of the microwave radiation emit­ extent. In the passive-microwave image it is possible not only ted by materials on the earth's surface. White and gray indio to determine the precise shape of the polynya but also to cate the faintest emission, which signifies areas of open water,

94 SCIENTIFIC AMERICAN June 1988

© 1988 SCIENTIFIC AMERICAN, INC of the Antarctic ocean, such as the high salinity of the waters overlying the continental shelf (which had been measured during the austral sum­ mers) and the large quantities of Ant­ arctic Bottom Water that flow from the Antarctic into the world ocean_

he forces that create and sustain open-ocean polynyas are some­ Twhat more complex than those underlying coastal polynyas. Open­ ocean polynyas probably form be­ cause of convection cells: vertical cir­ culation patterns in which, over an COSMONAUT POLYNYA, in the Cosmonaut Sea at 45 degrees east longitude and 65 area a few kilometers across, warm degrees south latitude, marks a region where polynyas often appear. The fact that water (in this case water from the certain geographic locations are the sites of frequent polynyasand areas of poor Circumpolar Deep Water) rises to the ice cover lends support to the theory that the topography of the ocean bottom surface, cools and then sinks, to be can influence the formation of polynyas. This image was made in September, 1986_ replaced by more rising warm water. In open-ocean polynyas the continu­ ously rising warm water would melt Ocean cannot have a diameter great­ for several reasons. One reason was its any ice on the surface and prevent er than about 10 to 30 kilometers. A sheer size: at its largest it measured further ice from forming. Open-ocean large, self-sustaining polynya would some 350 by 1,000 kilometers. polynyas are also called sensible-heat therefore probably have to consist of a Another remarkable feature was its polynyas, because the process in number of convection cells pressed persistence. It was firstobserved in which they give up heat is directly together "shoulder to shoulder." the austral winter of 1974. During the associated with a change in tempera­ Sensible-heat polynyas have a num­ following summer the ice cover in that ture-it can be sensed. ber of important effects. For one, their region of the Southern Ocean melted Because the layered structure of the convection cells draw up warm water completely, as it normally does, but Southern Ocean is only marginally sta­ from the Circumpolar Deep Water and when the ice cover formed again in the ble, isolated convection cells can form accelerate the process in which heat winter of 1975, the Weddell Polynya under the sea ice in many places. Why is transferred to the atmosphere. It re-formed as well. It had moved slight­ do only some of them lead to polyn­ is not yet known how important this ly to the west (in keeping with the yas? The answer may have to do with effect is in the overall action of the prevailing surface current in that re­ spatial scale. When convection is trig­ global heat engine, because it is not gion), but its shape and size were gered under a cover of sea ice, the known whether large polynyas form much the same as they were in 1974. initial burst of heat rising to the sur­ frequently enough to be climatically The following year saw the same se­ face melts most if not all of the ice significant. The effect could be sub­ quence of events: when the ice cov­ immediately above the cell. This cre­ stantial, however, because the ocean er re-formed in the winter of 1976, ates a filmof fresh water at the sur­ sheds heat from open water so much the polynya reappeared, slightly to the face. The fresh water is less dense more efficiently than through ice. west of its 1975 position and some­ than the warm salty water, and so it Sensible-heat polynyas also pro­ what smaller. is stable: it damps out the convection, mote chemical interaction between In 1977 the Weddell Polynya did not because the warm water cannot rise the ocean and the atmosphere. Unlike appear, but its persistence over the past it to the surface. Even if that water under the ice pack, water ex­ three previous winters indicates the initial sheet of fresh water were not posed in a polynyacan exchange gases existence during the summer of some sufficient to damp out the convection, relatively freely with the atmosphere, "memory" mechanism, some large­ other sections of ice would flow into absorbing some and releasing others, scale residual effectthat could re­ the polynya and melt, perhaps eventu­ before it sinks again. Although the create the polynya after the summer ally providing enough fresh water to rapid rate of overturning within a po­ melt. It is likely that the key to this halt the convection. lynya would prevent any specific wa­ memory mechanism is found in salini­ If the polynya initially had a large ter parcel from attaining full equilibri­ ty. When the ice cover melted in the area, however, it could be self-sus­ um with the atmosphere, it would still austral spring and summer, areas out­ taining. The surface area of a polyn­ allow significant exchange of gases. side the polynya would have been cov­ ya, within which convection occurs, is The existence of large sensible-heat ered by a thin layer of relatively fresh proportional to the square of its radi­ polynyas could thus alter the nature water. Because of its buoyancy, the us. The amount of ice that can float of the ocean-atmosphere interaction fresh water would have stayed on the into a polynya is proportional to its and thereby have a great effect on the surface and not mixed with the water perimeter, which is directly propor­ chemistry and climate of the atmos­ below. In the area where the polynya tional to the radius. Thus if a polynya phere and the deep ocean worldwide. had been, however, the surface would had a radius above some minimum have been salty, because there would threshold, ice could not flow in quickly uch of our knowledge con­ have been no sea ice to melt there. enough to stop the convection and the cerning open-ocean polynyas The following winter, when the cold polynya would survive. Mhas come from studying one atmosphere started to remove heat For reasons that have to do with the spectacular polynya that formed in from the surface water, the layered fluid dynamics of the rotating earth, the mid- 1970's near the Weddell Sea. structure of the waters in the region the convection cells in the Southern The Weddell Polynya was remarkable of the polynya would have become

SCIENTIFIC AMERICAN June 1988 95

© 1988 SCIENTIFIC AMERICAN, INC microwave measurements reveal rela­ tively small polynyas or areas of very thin and patchy ice cover in the vicini­ ty of the Maud Rise. Further evidence is found in an area called the Cosmo­ naut Sea near 65 degrees south lati­ tude and 45 degrees east longitude, where the ocean circulation, perhaps in response to the topography of the bottom, also seems to be precondi­ tioned for the formation of convection cells. In the Cosmonaut Sea, as over the Maud Rise, small polynyas and areas of diminished ice cover often occur. These polynyas in general do not survive an entire winter. To test this hypothesis, one of us (Gordon) helped organize efforts to push through the ice cover and reach the Maud Rise region in winter. Our first attempt was in 198 1 on the Soviet ship Mikhail Somov. On this trip we were not able to penetrilte all the way to the Maud Rise,although we got close enough to be able to gather ex­ tremely useful data on the interac­ tions between the deep water and the surface layer below the ice cover. We found an unexpectedly large entrain­ ment of deep water into the winter surface layer, suggesting an active ver­ tical flux of deep water. POlAR VIEW reveals several open-ocean polynyas (1-3) and coastal polynyas (A-f). We succeeded in reaching the Maud The image shows the average ice cover during a three-day span in the winter of 1980. Rise on our next attempt, in the aus­ tral winter of 1986, as part of a multi­ national expedition on the West Ger­ unstable: the salty surface water, on overturning necessary to transfer that man vessel Polarstern of the Alfred becoming colder and therefore dens­ much heat could have been as high as Wegener Institute for Polar and Ma­ er, would have sunk and warmer wa­ six million cubic meters per second rine Research in Bremerhaven. Al­ ter from lower layers would have ris­ during the winter, when the polynya though the analysis of our data from en, establishing new convection cells was active. In the process the polyn­ that expedition is incomplete, we and re-creating the polynya. In other ya could have produced as much as found there was indeed an upwelling areas the relatively fresh surface wa­ half of the Antarctic Bottom Water of deep water over the Maud Rise. The ter would simply have frozen, rees­ that flowed out of the Weddell Sea. fact that there was an upwelling in tablishing the Antarctic ice cover. 1986, even in the absence of a polynya, The effects of the Weddell Polynya hat caused the Weddell Polyn­ indicated that the topography of the were in keeping with its size. The po­ ya to form exactly where it ocean bottom was probably deflect­ lynya left a clear imprint on waters far Wdid? The most plausible an­ ing the flow of Circumpolar Deep Wa­ below the surface. For example, meas­ swer holds that the topography of the ter upward. In addition to elucidating urements of temperature at various ocean bottom played a role. The po­ mechanisms associated with the for­ depths made in the area of the polyn­ lynya's first appearance was directly mation of the Weddell Polynya, the ya during the summers of 1973 and above a feature called the Maud Rise, Polarstern voyage represents a rela­ 1977 (that is, before and after the oc­ which is an underwater seamount. tively new and exciting way to study currence of the polynya) reveal dra­ Here the bottom is some 3,500 meters the ocean: combining large-scale satel­ matic changes in the temperature of closer to the surface than it is in other lite observations of the surface with the deep water. The changes seem regions. Perhaps the warm Circumpo­ local, ship-based research. particularly dramatic when one bears lar Deep Water is deflected toward the in mind how very stable the deep wa­ surface as it passes over the Maud he study of polynyas is leading ters of the ocean are over short time Rise. This would "precondition" the oceanographers to a new under­ scales. The temperatures measured in region so that other events, such as Tstanding of the mechanisms at the region in 1977 were colder than greater than average mixing of the work in the Southern Ocean. As a re­ those measured in 1973 by as much as water by windstorms, would be more sult it is bringing about new ways of .8 degree, in depths all the way down likely to bring warm, salty water into conSidering the processes that govern to 2,500 meters. the surface layer, triggering the forma­ much of the world's climate, and it The missing heat had probably been tion of convection cells. raises important new questions. For carried to the surface by convection Satellite data from the years since example, does the overturning in large in the polynya. Reasonable estimates 1976 yield some evidence for the hy­ polynyas, such as the Weddell Polynya, suggest that the rate of convective pothesis. In many of those years the alter the rate at which carbon dioxide

96 SCIENTIFIC AMERICAN June 1988

© 1988 SCIENTIFIC AMERICAN, INC is exchanged between the ocean and FOR A MOMENT, NO CROWDS. NO CHEERS, NOTHING BETWEEN YOU the atmosphere? The deep water that AND THE PURE JOY OF VICTORY, OMEGA. FOR ALL YOUR SIGNIFICANT MOMENTS. wells up south of 60 degrees south latitude is high in carbon dioxide and releases significant quantities of it to the atmosphere. Because polynyas ac­ ceierate the upward flux of Circumpo­ lar Deep Water, one might expect they would enhance the rate of carbon di­ oxide release, perhaps adding to the projected global greenhouse warm­ ing. On the other hand, as we men­ tioned above, the rate of overturning in polynyas is so great that no sin­ gle water parcel spends enough time near the surface to exchange much carbon dioxide with the atmosphere. The net effectof polynyas on atmos­ pheric carbon dioxide levels is there­ fore uncertain. Conversely, it is also not known what effect the projected warming might have on the frequency or extent of future polynyas. How important are polynyas gen­ erally in the deep-ocean overturning that takes place in the Antarctic? We know the Weddell Polynya had quite a pronounced effect, cooling a great vol­ ume of ocean and giving surface water access to the atmosphere over a large area. Are such large polynyas rare events, or do they occur relatively of­ ten? And how much of the overturn­ ing that takes place near the conti­ nental shelf is mediated directly by OMEGA ALWAYS MARKS SIGNIFICANT MOMENTS the coastal polynyas? Oceanographers IN THE OLYMPICS, IN THE SPACE PROGRAIvl. IN SIGNIFICANT LIVES LIKE 'rOURS. now have the tools to answer these THE OMEGA SEAMASTER. ADVANCED SWISS TECHNOLOGY. WATER RESISTANT 10 30 METERS important questions and many oth­ IN STAINLESS STEEL AND 18K GOLD. ers. The combination of comprehen­ sive, long-term observations by satel­ o lite and detailed observations from the field in winter is making it possible OMEGA to construct ever more precise models of these vital and far-reaching proces­ OMEGA. OFFICIAL TIMEKEEPER OF THE OLYMPIC GAMES, CALGARY AND SEOUL 1988. ses in the Southern Ocean. .MAYOD·S Fine Je\\'elerssince ��. FURTHER READING WEDDELL DEEP WATER VARIABILITY. Ar­ nold L. Gordon in Journal of Marine Research, Vol. 40, Supplement, pages 119-217; 1982. ANTARCTIC SEA ICE, 1973-1976: SAT­ ELLITE PASSIVE-MICROWAVE OBSERVA­ TIONS. H. Jay Zwally, Josefino C. Comi­ so, Claire L. Parkinson, William J Camp­ FREE SCIENCE BOOKS CATALOG bell, Frank D. Carsey and Per Gloersen. Send for your free catalog of new and recent publications from W. H. Freeman and NASA Special Publication 459; 1983. ANTARCTIC OFFSHORE LEADS AND Po­ Company, the book publishing arm of Scientific American for more than 20 years. LYNYAS AND OCEANOGRAPHIC EFFECTS. Discounts and bonus books offered. Jay Zwally, Josefino C. Comiso and H. o Yes, I'd like a free copy of your latest science book catalog. Arnold L. Gordon in Oceanology of the Antarctic Continental Shelf,edited o Please send literature about the Scientific American Library, too. by Stanley S. Jacobs. American Geo­ Name ______physical Union, 1985. RECURRlNG POLYNYAS OVER THE COSMO­ Address ______NAUT SEA AND THE MAUD RISE. J C. Comiso and A L. Gordon in Journal of City, State, Zip ______92, Geophysical Research, Vol. No. C3, , Mail to: Gail Harleston pages 2819-2833; March 15, 1987. - W. H. Freeman and Company, 41 Madison Avenue, New York, NY 10010

SCIENTIFIC AMERICAN June 1988 97

© 1988 SCIENTIFIC AMERICAN, INC