REPORTS ON RESEARCH AT THE WOODS HOLE OCEANOGRAPHIC INSTITUTION

Volume 37, Number 1 Spring 1994

Atlantic Ocean

* .,' I Circulation Oceanus REPORTS ON RESEARCH AT THE WOODS HOLE OCEANOGRAPHIC INSTITUTION

Vol. 37, No. 1 Spring 1994 ISSN 0029-8182

Atlantic Ocean Circulation

The Atlantic Ocean 1

Progress in Describing and Interpreting Its Circulation By Michael S. McCartney A Primer on Ocean Currents 3

Mi'i'i^iiri'ii/fi/tK

Towards a Model of Atlantic Ocean Circulation 5 The Plumbing of the Cl/ii/nlr's Rnil/u/m-

: By MichaelS. Mri''nrli/i //

Dynamics and Modeling of Marginal Sea Outflows 9 Out They Go, Doini, and Around tin' \\brlil By James F. Price

Meddies, Eddies, Floats, and Boats 12

1 How Do Mediterranean ni/d Atlantic Waters .l/ii .-' By Amy Bower Where Currents Cross 16

Interarlnni the and the Western Current t CiilfSln'am >i i nuii'< is published semi-annually by the Woods Hole of Deep Boundary Rolinl S. P/,i;iil Oceanographic Institution. Woods Hole. MA (C~i4:i By 508-457-2 i.cxi 271H Giant Eddies of South Atlantic Water Invade the North 19 iii-ciniiix and its logo are Registered Trademarks ol the Woods Hole Oceanographio Institution, All Rights Reserved Disrupted Flou- andSwiriing H'ati'n By Philip L Richardann In North America, an Oceanns Package. including

1 ) .mil 1,; nu i ( neusletiei l-:.i/ilnr, an 8-page junior-high-selnm! How Does the \\'/i in tin' l>,r/> Si

America, the annual fee fur fii-fniiin, uK is S:;ii. su inagaz Wave-Induced Recirculations 26 for single copies. To receive the publications, iili'ascfall (loll Abyssal free) 1-M! s^.Vnni.l m HIate: WHOI Oceanus Package Snrvice Turnini'i This Wti/i and That Way P.O. Box NY 1:^17 Center, Wlii. Syracuse, By Michael A. Spall riirrksslinulil hcilraui a I'S hank in US dollars and mailr p,i\ahlr I" \\uinls Hole Oceaiioguplnc Institution. When .sending rhange of address, please inrludr mailing label

( 'la mis lor missing numbers from ihe I'S will be honored within

Ihree months Ol publir'.il ovei -IMS, fl\e innntlis

I'eriin.ssioii lo phoidciipv for mlrriial or personal use or tin- internal in use is In <>,,'uiiu\ personal of specific clients granted Editor: Vicky Cullen Designer: Jim Canavan

Illi' . I In hlil. and othel llsns M-Hl,li'lril Hltll I lie '"pM'mlll Woods Hole Institution I'lralanrr Cfiiln (('('(' I, prn\ idcd that tile base fee nf S^ p,-i rMp\ Oceanographic ul I be arlicle is to CCC, 21 St i rcl, Salem, paid directly Congress Robert B. Gagosian, Director Guy W. Nichols, Chairman of the Board of Trustees M \ iilii^n Special requests should be addressed to Ckwiv//s James M. President of the Charles A. Dana, III, President of the Associates - Clark, Corporation ISSN 0029 3182 !$2.00* n.iGSTR l^s :!72 IHi7 Printed in

SAor in \'iU-d paper Woods Hole Oceanographic Institution is an Equal Employment Opportunity and Affirmative Action Employer Athanasius Kircher's seventeenth The Atlantic Ocean century view of Atlantic Ocean circulation. Progress in Describing and Interpreting Its Circulation

was the first physical oceanographer, world oceans' vast subtropical gyres. However, the the first to observe and ponder the cause ocean circulation was deduced from extremely sparse of ocean movements? The answer is lost data (nonexistent data in many regions). The voids

in an Whoin history, of course, but the first writings were filled with considerable imagination, on the subject date to early Greek civilization. By mid imagination almost completely unconstrained by seventeenth century, expanding exploration of the New physical principals, as physicists had not yet focused on World had yielded enough geographic information to the ocean. For example, Kircher was unaware of the allow the Jesuit priest Athanasius Kircher to construct physical impossibility of two currents, like those a chart of the world as well as detailed renditions of diagrammed east of Brazil, passing through one

smaller regions, such as the Atlantic sector from his another, and he was particularly fond of the notion that

hand on this page. Much on this chart is recognizable subterranean passages linked maelstroms and currents,

today, both in regard to principal land masses and dispersing their entrances all over the world's oceans. aspects of surface currents. The classic era of physical oceanography began These charts include the first interpretations of the about 200 years later with recognition of the central

OCEANUS on role Earth's rotation plays in the We initially planned this collection of reports to focus on Atlantic circulation. physics of ocean motion and the current research deep of we ended with is an fielding of expeditions for the The diversity topics up of of the basic of explicit purpose of exploring ocean indication two aspects nonlinearity circulation. In a 1954 pamplet physical oceanography. First is the high degree of entitled "Why do our ideas about the linkage among physical phenomena: As some particu- ocean circulation have such a lar phemenon is examined, its physics often turns out linked to that of other peculiarly dream like quality," Henry to be directly phenomena. Stommel described this classic era of Second is the significant linkage among the scientists ocean is nonlinear, it is physical oceanography : themselves: Because physics a 1 a cruise, or to be a specialist using narrowly "( ) grand expedition, nearly impossible on a defined brings back many hydrographic defined technique or working narrowly but stations' worth of data; topic. We don't have to be experts in everything, need a of a broad (2) extensive plots, graphs, and we do good understanding spectrum Henry Stammel, Atlantis IL 1963 tabulations of the data are made and of phenomena that modify the physics of the ones we published for the benefit of future want to understand. has a generations; The WHOI Physical Oceanography Department with 6 (3) certain of the more striking 32-member scientific staff working postdoctoral and 34 features of the data plots are noted; associates, a 25-member technical staff, are students on Ph.D.s, all ably supported (4 ) some plausible hypotheses graduate working oceano- advanced to explain them. by 36 assistants. The staff works on physical This procedure usually exhausted graphic problems throughout the world's oceans, from millimeters to the energy of those involved, and spanning the full range of scales almost always the funds, and the megameters, and utilizing a wide range of techniques to study usually stopped at this stage." from intensive field-measurement programs equally However, the real focus of intensive theoretical and numerical modeling studies. Stommel's pamphlet was a call to the It is the largest concentration of physical oceanogra- of these scientists have research community for a fifth step: phers in the world. Seven of their indi- "All of these steps are absolutely contributed reports on one each many John Erika Dan, 1962 Swallow, a of how two necessary, of course; otherwise we vidual projects. They include study Kircher's would know nothing at all about the currents cross not through each other, as in other. deals ocean as it really is. But for the full drawing, but one beneath the One report development of the science there with the very real complexity of the currents east of the actual of the must be one additional step: Brazil, and the rest with "plumbing" Kircher's subterranean (5) the plausible hypotheses must deep circulation, not passages, of circulation be tested by specially designed but instead an intricate system abyssal observations. In this way theories pathways by which waters sink from discrete source the Atlantic Ocean. can be rejected or accepted, or may regions and move through Michael S. be modified to become acceptable." McCartney Senior Scientist Shortly thereafter Stommel provided a prototype for this modern Physical Oceanography Department era of physical oceanography by Editor's note: There are references in this volume to hypothesizing on theoretical grounds many the work of Stommel, a oceanographer the nature of the basic structure of Henry physical Val Worth ingtoii, Atlantis, 1948 extraordinaire, who died in January 1992 at 71, a loss deeply the abyssal circulation. The felt by his colleagues both at WHOI and throughout the world. validated the hypothesis was almost immediately by in A special issue oWceanus entitled A Tribi/tr H, ///// direct measurement of John Swallow (the program Sinniiiii'l was published by the Institution in 1992. The inside recipient early in 1994 of WHOl's first Henry Stommel back cover of this issue features the awarding of the first Medal in Oceanography) and Val Worthington (WHOI medal in his name. Physical Oceanography Department Chairman from 1974tol981), It is now about 37 years since Stommel published his pamphlet, and we can happily report much progress

toward fulfilling his model for modern physical occanographic research. We still feel very much like explorers when we cany out our measurement programs, but we conduct and interpret them within a framework of increasingly sophisticated physical understanding. Atmospheric Pressure Oceanic Dynamic Height

70 60= 50

Note the similarity of these imagined circulation fieldsfor the atmosphere and ocean. Atmospheric pressure contours show a low over Canada and a high over southeastern United States. Arrows indicate direction offlow deducedfrom the geostrophic relation (that between horizontal pressure variations and horizontal currents), with higher speeds where the contours are closer together. Contour lines in the western North Atlantic illustrate an analogous field, the dynamic height at the sea surface. The compressed bundle emergingfrom the south in mid off the coast ofFlorida is the Florida Current, which initiates the GulfStream system. T)>e ocean surface is about a meter higher basin than it is inshore of the GulfStream, and this drives theflow, indicated by arrows. Contours that peel off to the south indicate a partial clockwise circulation pattern around the high-elevation region, analogous to the clockwiseflow around high pressures in the atmosphere. A Printer on Ocean Currents

Measurements and Lingo ofPhysical Oceanographers

Volumes of Water Moving in Currents rampages plagued the midwest last summer, accounts for only about 0.02 million cubic meters per second, express current flow in "millions of one ten-thousandth of the Gulf Stream's Oceanographerscubic meters per second," a term difficult for most roughly transport! people to comprehend. A large current, such as the Gulf Suvam south of Nova Scotia, transports more than 150 of Seawater million cubic meters per second, and typical transports Properties for the smaller deep western boundary currents are 10 Though seawater's salinity is expressed in to 20 million cubic meters per second. Various dense Salinity:grams of dissolved solids per kilogram of seawater, overflows from marginal seas such as the Mediterra- today's methods measure seawater's electrical conduc-

nean are even smaller, 1 to 3 million cubic meters per tivity and then employ a well-known conversion

second. For comparison, the sum of all the rivers algorithm to determine salinity. While coastal waters flowing into the Atlantic is about 0.6 million cubic- can exhibit a wide range of salinity as a result of meters per second. The Amazon contributes about a freshwater runoff, most of the world ocean lies in the

third of that total, while the Mississippi River, whose narrow salinity range 33.8 to 36.8.

OCEANUS Temperature: Oceanographers commonly refer to (Greenland and Norwegian seas), the major source for The Nordic the "potential" temperature of a parcel of water. This the rest of the North Atlantic Deep Water. recognizes that a parcel of water sinking from a surface Seas' sources are confined by the ridge system that

source will, if it does not mix or exchange heat with connects Iceland to Greenland and Europe, but spill

surrounding waters, become slightly warmer as the over the ridge to form dense overflows. The last source pressure on it increases with depth. For example, if a for the North Atlantic Deep Water is the Antarctic //parcel of surface seawater starts with a temperature of Bottom Water: Some of this water flows from the Brazil into the 0C and salinity 35 and descends to 3,000 meters as part Basin in the South Atlantic across the equator

of an overflow, it may warm as much as 0.3C. An North Atlantic (see articles by Hogg and Spall), where instrnrtmnt lowered into this overflow would sense this its ultimate fate is to upwell into the lower part of the

armer which is called the "in situ" North Atlantic Water Pickart's article). Major temperature, Deep (see currents like the Gulf Stream and eddies like the temperature,/ However, for most calculations and , analyses ooeanographers use the potential temperature, North Brazil Current retroflection eddies (see this of and thus Richardson's the water which co/rects for effect pressure article), may penetrate deep remain/0.0C. and mix or disrupt the deep-current flow.

Tha North Atlantic is the warmest and most saline of

the world's oceans, having a mean potential tempera- Current Velocity Measurements to ture of i.08C and mean salinity of 35.09, compared measure current speed in two ways. of 34.72. Most the globbl average of 3.51C and salinity OceanographersFor direct measurements, moored current meters of the waVmer andoKffe saline waters of the world are record the speed of water flowing past them at intervals concentratecTnTme kilometer of the upper subtropical of hours or days, and surface or subsurface drifters and circulation in what is called the tropical regimes reveal the pathways of the parcels of water in which main thermocline of decrease in (a region rapid they were launched. Indirect estimates use a relation with in the North Atlantic temperature depth, typically between horizontal pressure variations and horizontal About 77 of world-ocean Ot^uppfen kilometer). percent currents (called the geostrophic relation). This is 'volume is colder than with salinities in the 4C, analogous to a meteorologist's use of atmospheric narrow 34.1 to 35.1. At the sea relatively range surface, pressure charts to deduce the wind field. In both cases about 26 of the surface area is colder than only percent the geostrophic velocity (of the wind or the ocean and it is within this area that the volume of 4C, large current) is parallel to the pressure contours on the cold water its characteristics before acquires sinking chart and inversely proportional to the contour spacing. and like those discussed in this traveling along paths For the ocean, the pressure is determined from the field publication. of density as deduced from an equation of state that links density to the actual measurements of tempera- Water Masses ture and salinity. However, this calculation in the ocean the of the large volume of cold water described above yields the velocity difference between depth Thecomprises the "deep" and "bottom" waters. In the calculation and a "reference level." To convert this Atlantic there are several sources for these waters in velocity difference to the absolute velocity, tin-

both hemispheres. Each hemisphere's sources are reference level velocity must be determined by direct blended by circulation and mixing. The net effect of measurement, estimation, or some other means. Since northern sources dominates the deep water, while the the number of direct velocity measurements is generally net effect of southern sources dominates the bottom very small relative to the hydrographic database

water. In regions where the two water masses coexist, available for application of the geostrophic relation, the bottom water lies beneath the deep water, although more often than not the "other means" are needed. because of their mixing there is no sharp demarcation These include: between the two water masses away from their sources. "Budgets," where a reference-level velocity is deduced Oceanographers often recognize these dominant by the requirement that the the sum of all the flows into sources through the proper names North Atlantic Deep and out of a region must be zero (since sea level isn't Water and Antarctic Bottom Water. The Mediterranean changing). outflow descends through the thermocline to form a "Water-mass analysis," where water-mass distribution a mid-depth saline "tongue" of Mediterranean Water, suggests flow pattern velocities. (For example, given which mixes downward through the uppermost part of source region may cause a property "tongue" along the North Atlantic Deep Water.(Authors Bower and which the source's characteristic properties propagate Price discuss the Mediterranean outflow.) Other and then mix with the surrounding water. This may sources for this uppermost part of the North Atlantic show the flow direction and allow estimation of a Deep Water are found in the Labrador Sea, where speed.) a of zero winter cooling produces a large volume of water at "Level of no motion inference," where zone zones of water potential temperatures between 3.0C and 4.5C. Even speed is assumed to lie between two colder waters are produced in the Nordic Seas flowing in opposite directions.

SPRING 1994 artney conduc-

tivfty/terkgeraifyre/ dejitlt water Towards a Model of mpling rosette. Atlantic Ocean Circulation

The Plumbing of the Climate's Radiator Michael S. McCartney Senior Scientist, Physical Oceanography Department

sense of theme of physical oceanographers' While in some sheer numerical thKe_arej^ot research is the exploration and explanation of ocean circulation measurements, tbj^j/culatior ocean circulation as a natural physical remarkably complex, so the measurllmeru^flfe actiy Thephenomenon with scales from millimeters to rather sparse relative to what is needea megameters. In this context, the word "model" usually phenomena unambiguously. Thus we also become of ocean brings to mind a theoretician analytically or numeri- modelers, for even a basic description to in words cally solving some system of equations that might circulation is a model: an attempt provide of capture the essence of the physics of some particular and pictures a synthesis, or simplified interpretation, part of the ocean circulation. In other words, the the measurements. theoretician's model attempts to explain the causes and One such model, the popular "conveyor belt" mechanisms of observed phenomena in the ocean, or interpretation (overleaf), has emerged from the the last on the perhaps predict unobserved ones. increasing focus (over roughly decade) Those of us who make ocean circulation observa- ocean's role in global climate. Its underpinnings are, to of the oldest observations and ! ions often get involved in a different sort of model. however, traced some The conveyor belt 'ollow the footsteps of image of the An Oceanic Conveyer Belt centuries of ocean Atlantic Ocean's explorers who have tried role in exchanging make maps of Atlantic warm and cold Ocean surface circula- waters with the tion. One of the oldest of rest of the world's oceans. Developed these, by the Jesuit by W. S. Broecker of Athanasius Kircher, dates the Columbia from 1678 (seepage 1). University's It is important to Lamont-Doherty recognize two things Earth Observatory, about circulation it represents the schematics. Atlantic as a First, they radiator that are almost instantly converts imported 2 rendered obsolete by new warm water to < measurements, for our cold exported science is still in a basic water. The their early interpretations as expressed by Benjamin exploration phase. Second, they represent a subjec- intensity of the Count of Rumford, in 1797 and elaborated tive synthesis: Other oceanographers might empha- conversion rate is Thompson, William and Prestwich in the size different that is, draw the picture and generally esti- by Carpenter Joseph aspects, mated at between 1870s. This very simple conveyor-belt model is based on estimate the transport amplitudes somewhat 10 and 20 m ill/on a few fundamental circulation observations: differently the product is very much "in the eye of cubic meters per Cold water with identifiable North Atlantic character- the beholder!" second. istics can be traced in diluted form through most of the The schematic opposite points out a problem world ocean. inherent in the process of sorting out ocean To conserve mass, an equal volume of warm water circulation's role in climate. We show 13 million cubic

must replenish the cold water draining out of the North meters per second of warm water entering the North Atlantic. Atlantic across the equator; this is our estimate of the

There is a large liberation of heat from the ocean to magnitude of cold-water production in the North the atmosphere at middle and high North Atlantic Atlantic and thus of the intensity of the conveyor belt. latitudes. However, the Gulf Stream south of Nova Scotia

The model, as far as it goes, is consistent with these transports about six times this amount! The "recirculat-

observations. In particular, the estimated replenish- ing gyre" components of measured North Atlantic warm-

ment rate, the estimated intensity of heat loss, and the water transport thus greatly exceed the amount that is observed temperature change of the warm-to-cold estimated to convert from warm to cold. (The recircu- conversions are in harmony. lating gyre can be seen in the figure opposite as the Gulf An observationist builds on a simple model like the Stream water that moves across and down the middle of conveyor belt by checking its consistency with observa- the Atlantic to rejoin the northward flowing currents.)

tions other than the ones on which it was based. The Thus, the part of the circulation field most important

broader the base of such consistency, the more "real" for climate issues is a small fraction of the field actually

the model is perceived to be, and the more valuable it being measured. The physical phenomenon responsible

becomes as a benchmark for testing a theoretician's for the northward transport of warm water through the model or as a guide for theoretical models for North Atlantic is a linked set of gyre flows rather than example, what is the circulation that the model should the simple image of the upper limb of the conveyor belt. duplicate? But when consistency with observations is The physics of the western intensification of these gyres

lacking, the observationist must augment the simple was first deduced by Henry Stommel in 1948, and their model by building in missing pieces or perhaps relationship to the wind and thermal and hydrological

discard it as a flawed starting point. forcing remains a very active research topic to this day. and The figure opposite is an example of an augmented The synthesis of observations, their interpretation, in version of part of the conveyor-belt model. WHO I their physical modeling proceed parallel. from Senior Scientist Bill Schmitz and I developed it over Similar degrees of complexity are emerging the last few years to describe the field of horizontal studies of the cold-water part of the system, studies warm-water circulation in the North Atlantic (see that illustrate the model-building process. The

publication footnote). Such circulation schematics dominant components of the meridional flow of cold attempt to integrate all relevant measurements and water are western intensified both within the Atlantic

interpretations, in this case both our own and those of and within its subbasins, which are defined by abyssal our many colleagues around the world, present and ridges. In Kircher's old circulation-system map on page into past. Such circulation schematics have an even longer 1, water was imagined to descend subterranean history than the conveyor-belt interpretation, for we passages and to reemerge in faraway places. The real

SPRING 1994 plumbing of the system is internal lo the ocean itself, than directly towards their destinations. lint it achieves a similar result: Waters sink in a few The latest work on the Atlantic circulation model places to inid-depth or the seafloor (but not with the involves the cold-water system near the equator. For

\ lolcncc of maelstroms!) and are earned by current this, MIT/WHOI Joint Program graduate student Marjy systems to the far reaches of the world ocean by deep Friedrichs, Associate Scientist Mindy Hall, and I currents. The width of these currents is restricted In examined the southward cold-water transport in the the internal physics of the ocean and steered by the tropics of the North and South Atlantic. We found the alnssal topography (not subterranean passages!) North Atlantic's transport dominated by a colder water The physics of this deep western intensification was mass than the South Atlantic's. The shift does not occur first elucidated by Henry Stommel in the mid 1950s as a smooth north-to-south transition, for the contrast

(see page 22). Such intensified flows are "easier" to in the western basin persists even quite close to the measure because they are geographically small. equator. Measurements indicate that a zonal upwelling

Measurements of the intensity of these southward flows cell in the cold water along the equator achieves the show a magnitude often twice or more than that transition (bottom figure overleaf). expected from the simple conveyor-belt intepretations. One of the difficulties we face in interpreting our Flow measurements in the basins' interior regions have data is visualization of the three-dimensional structure revealed that a gyre component is responsible for the of the circulation. In the top figure overleaf, the vertical larger-than-expected boundary current flows. This structure of the deep circulation was averaged to focus system of gyres is more broken up by abyssal bathymetry on the overall horizontal flow. In the bottom figure, part in a (top figure overleaf) than are the large warm-water of the vertical structure of that deep flow is shown gyres. This model preserves the intensity of the perspective view, with the ocean above 2,000 meters net meridional flows of cold removed, and with the view northwest towards cnmeyor-belt conception's looking A simplified of a in the \\ ater. but it is also consistent with measurements of the continental slope Brazil from point rendition of the the intensity of deep western boundary currents and middle South Atlantic south of the equator. The result system of linked interior flows. Combining this cold-water schematic looks a bit like an expressway cloverleaf, but with no recirculation gyres andflow pathways with the warm-water circulation of the figure below connections between the two levels. We believe that the of imported warm- gives a representation of the full three-dimensional indicated seven million cubic meters per second water as it makes ocean circulation that the quasi-two-dimensional eastward flow of the colder level upwells in the eastern its wayfrom the belt it is more Atlantic to the indicated seven million conveyor simplifies. Perhaps aptly tropical supply equator to the water described as a baggage carousel than a conveyor belt, cubic meters per second of westward flowing on subpolar basins of with water parcels mostly going round and round rather the warmer level of the cloverleaf. the North Atlantic, with flow rates in millions of cubic meters per second, indicated by circled numbers. Partition of the 13 million cubic meters per second estimated net northward warm-

waterfloiv into distinct cold-water production regions is given by the numbers in squares: 7 million cubic meters per second sinking in the Labrador Sea, 4 million cubic meters per second (pink squares) entering the Nordic Seas to return as dense overflows into the region south of Iceland, and 2 million cubic meters per second entrained into 100 these overflows.

OCEANUS Principal path- The temperature change involved in this equatorial tnii/.v offlowfor the upwelling cell, about 0.5C, is small, but trie recognition combined deep and of the circuitous route that cold water from the North bottom water Atlantic takes in order to move southward across the (all water equator has implications for the mode of the system's colder than to climate A in the 3C) response change. perturbation circulation warm- to cold-water conversion process that provides hi the North and the "northern source" water to the cold-water limb of tropical Atlantic, the conveyor belt does not simply move through the with a perspective Atlantic relatively undiluted in a deep western view of the seafloor boundary current "filament." The recirculation gyres bathymetry that dilute the source water's characteristics with those of contains and the basin's interior water mass. The diluted shapes the water that reaches the is pathways. Red deep equator mostly pathways indicate diverted into the interior and further altered southward before returning to the western boundary to trendingflows, continue southward through the South and blue north- Atlantic where we are finding additional ward. The recirculation gyres. These new observa- parallel tions and a new black interpretations present to the theoreticians, strips modeling challenge are the and, for we observationists, it points to axes of new measurements that will improve the intense our interpretive circulation model. deep recirculai- This work was the ing gyres to either supported by Office of \ni nl the National Science Foundation side of the Gulf h'lwarch, mi,l lln \nlionat Oceanography unit Mn/n- Stream (counter- Ail in i miration. One the most recent clockwise to its x/i/in-n- of ir/ri'iml "North Atlantic north and /i/ih/icationsonitis Circulation" (W.J. Schmitz and M.S. McCartney, clockwise to its Reviews of Geophysics 31, paiji's 29 In 49). south). Additional counterclockwise Mike McCartney came to WHOI 20 years recirculating gyres ago with degrees in theoretical fluid dynamics. are in shown Within two weeks of his arrival Fritz Fuglister,

green. Tfiese gyres one of the real pioneers of Atlantic physical mix the character- oceanography, took him to sea on the istics of the institution's research vessel Chain. It changed

northward and his outlook 1 Recently Mike added up his southwardflowing research cruise time and found that he has waters, and averaged a month per year, including eight of the Atlantic He that the main intensify the crossings says over the is that the cruises come southwardflou< change years in for none in 1993 and a along the western clumps, example scheduled 90 in 1994 for of the boundary. days crossings South Pacific and of the Antarctic Circumpolar 2.4Cto3.2C Current Between the cruises and the fun of (near 2500 meters) the data's about the 1.8Cto2.4C untangling message (near 3500 meters) circulation and the physics, he further relaxes by taking long solo cruises on his old gaff cutter and untangling all her rigging He

doesn't understand why people think it odd that he spends so much time on the water

An interpretation of tropical cold-water circulation in the Atlantic, looking northward/ram beloiv the equator. Numbers indicateflow rates in millions of cubic meters per second. In the North Atlantic, a deep western boundary current is dominated by deep water near 2C. Tliis southward current mostly turns eastward, awayfrom the continental slope, after crossing the equator, and is warmed by downward diffusion of heat in the equatorial zone, causing an "upwelling" into the wanner layer. Westward/low of this layer in the equatorial zone (dominated by deep tvater near2.7C) diverges near the western boundary info northward and southwardflows, with the latter dominating the deep South Atlantic iceatern boundary current.

SPRING 1994 Jim Price discusses the Mediterranean

over/loin Dynamics and Modeling of Marginal Sea Outflows

Out They Go, Down, and Around the World

James F. Price

Associate Scientist, Physical Oceanography Department

waters of the deep ocean carry a huge the deep water leads to the highest productivity found

volume of heat and dissolved gases along in the oceans. This global transport aspect of deep- paths that cross ocean basins and may ocean circulation has long been recognized as an Theencircle the globe. These paths all originate important element of the ocean's geochemistry and has in one of a handful of polar marginal seas where cooling been studied intensively by oceanographers since at or evaporation makes water sufficiently dense that it least the 1920s. can sink to great depths in the open ocean. The paths It is a sign of the times that we now also think of the can be thought of as terminating where the deep water deep ocean as a storage site, inadvertent or planned, for returns to the sea surface, usually after many centuries, anthropogenic substances such as radioactive wastes and after being slowly modified by mixing with warmer and greenhouse gases including carbon dioxide and waters and by a rain of organic material from the upper fluorocarbons. Most of these are harmful to the ocean. In special regions where deep water comes to atmospheric environment, and some of these sub- the surface in greater volume, along the equator and in stances may also harm sea life. To evaluate the risks, we coastal upwelling zones, the resulting nutrient load of need to know with some confidence where and in what Oer - these materials of the Mediterranean outflow in the Gulf of in 9y OK,,-5s quantity Cadiz, 'Pat/ 0n are injected into the collaboration with Tom Sanford of the University of deep sea, and where and Washington and Rolf Lueck of the University of Victoria.

when they will reappear We measured current profiles through the outflow and

at the sea surface. Our some properties of the turbulence as well (see figure at

present knowledge of left). These data were analyzed by Molly O'Neil Baringer, deep-ocean currents is then a Ph.D. student in the MIT/WHOI Joint Program, elementary, and reliable and by Greg Johnson, a recent WHOI graduate then at the estimates of deep-ocean University of Washington as a postdoctoral researcher. transport and storage The data showed that the Mediterranean outflow begins are very difficult to with highly saline and very dense water and yet ends up

make. Research aimed making intermediate (rather than deep) water because it

at with oceanic as it first to \ improved understand- mixes intensely water begins in t ing of deep-ocean descend the continental slope the Gulf of Cadiz (see

'- circulation is thus likely figure below left). The large bottom slope combined with

to be one of the focal the large initial density causes a very strong buoyancy

points for oceanography in (gravitational) acceleration that increases the current the coming decade. This speed to more than a meter per second. This strong

Profiles of research effort will go forward at many current causes instability at the interface between the currents, salinity, institutions using ideas and resources from North Atlantic water and the Mediterranean water, which and turbulent many disciplines (see, for example, the 1992 in turn causes mixing that dramatically reduces the energy dissipation Paleoceznography Reports on Research, page 11). salinity and density of this outflow. through the core of One of circulation is the The end result is that mixed Mediterranean Water the Mediterranean intriguing aspect deep of the dense outflows that settles into the North Atlantic thermocline at a of outflow (which is dynamics marginal-sea depth the saline bottom initiate deep circulation. It is tempting to say that these about 1,000 meters. Mixing (or entrainment) causes layer approxi- outflows "drive" the deep circulation, but the processes the transport to increase from about .7 million cubic 100 meters mately that cause the deep water to return to the surface are meters per second at the Strait of Gibraltar to about 2.5 thick). T)ie Strait probably of equal importance in the long run. What we million cubic meters per second in the North Atlantic. of Gibraltar is to can say with some confidence is that the transport, The mixed Mediterranean water makes an important the right, and the temperature, and salinity of the outflow waters are salinity anomaly that spreads across the North Atlantic outflow current is in Bower's article on moving southwest crucial factors setting the properties of the deep basin (see Amy page 12). into the Gulf of water, and they are clearly also crucial in the overall A study of historic data on the subpolar outflows Cadis;. problem of understanding deep-water circulation. (two flow from the Norwegian and Greenland seas and The first important step in my research program to one from the Weddell Sea) showed a similar pattern of understand outflows was to make a detailed field survey strong mixing as these outflows began to descend the

continental slope. However, these subpolar outflows all produce bottom water despite beginning with a lesser Deep Water Formation density than does the Mediterranean outflow. Why the in Marginal Seas difference? To understand outflow mixing in a quantitative and OPEN OCEAN ultimately predictive way, Baringer and I developed a numerical model that simulates a single outflow

descending into a resting ocean. The elements of the model dynamics were known from the field study: We knew that topographic and Coriolis accelerations (an effect of Earth's rotation that accelerates moving ocean

water to the right in the Northern Hemisphere and to the left in the Southern Hemisphere) were essential, and we had evidence that mixing occurred primarily in

response to strong currents. These and other insights, when combined with the conservation laws for mass

and momentum, led to a simple but somewhat realistic outflow model that we could use to interpret observa- A schematic three in the water a showing stages formation of deep by marginal tions and explore outflow dynamics. The bottom figure sea. A shallow and confined marginal sea acts as a concentration basin that opposite shows observed and simulated density and produces a comparatively small volume ofvery dense water. Before this dense speed along the path of the Mediterranean outflow. The water can settle into the open ocean it must descend the continental shelf and model has some skill at simulating this outflow, and it slope, and may mix significantly with oceanic waters along the way. The has useful for the differences and water thatfinally reaches the deep sea may be quite differentfrom the water proven understanding thatfirst came out of the marginal sea. similarities between this and the other major outflows.

10 SPRING 1994 1 '/'// imrk Ha Experiments with the model gave some surprising truxf/ii/ili'il !>// Marginal Sea Outflows Office of.\'iinil Ht'ncnrrli. The Gut/ results. The most important is the apparent high of Cadis results were descr/h*

necessarily give up some resolution of time or space that we can easily achieve in our single- purpose outflow models. In the near term we will work on ways to include outflow models within the framework of

the general circulation models. Our goal is to develop outflow models that are compatible with 200 1 50 the lower resolution of the general Distance Downstream (kilometers) circulation models, and yet remain Tlie observed and simulated responsive to changing bottom topography, sea level, or (data points) (solid lines) and the of the Mediterranean climate. Once we have done this, we can analyze the speed density along path outflow, ne simulated density shows a very rapid general circulation models to learn how and where decrease where the outflow her/ins to descend the materials are absorbed and transported by deep-ocean continental slope (about 30 kilometers downstream circulation, and learn where they may come to perhaps from the start). Tlie profiles in the topfigure opposite, the surface again. made in this region, indicate very strong outflow currents and intense turbulence due to miring. Amy Bower works with RAFOS floats. ^ VT^^^^^___ _^^^^^_^^^ ,^^^^^ _ _ ^ ,^_ _ _ _ Meddies, Eddies, floats ^ and Boats

Mediterranean and Atlantic Waters Mix?

Amy S. Bower Assistant Scientist, Physical Oceanography Department

of the most prominent features of the slope of Spain and Portugal in a relatively narrow jet

North Atlantic Ocean is the tongue of salty known as the Mediterranean Undercurrent. Eventually, water that extends from Portugal westward at the Mediterranean Water leaves the coast and spreads Onemid depth. This tongue results from an out into the North Atlantic to form the tongue of salty exchange flow between the Mediterranean Sea and the water shown in the top figure opposite. Atlantic through the Strait of Gibraltar. Atlantic Water Traditionally, physical oceanographers have thought flows through the strait into the Mediterranean, where it Mediterranean Water to be carried out into the North is converted into saltier Mediterranean Water by an Atlantic by steady currents, and further dispersed by

excess of evaporation over precipitation. Mediterranean some random mixing. While these processes certainly Water, which is heavier than Atlantic Water clue to its contribute to the spreading of Mediterranean Water, the higher salt content, escapes into the North Atlantic discoveiy of a new type of eddy in the Atlantic has forced through the strait under the incoming Atlantic Water. us to revise our thinking about how Mediterranean

After leaving the strait, the dense Mediterranean Water Water is carried from its source. These eddies, found at

sinks and turns to the right, following the continental about 1,000 meters, are rapidly rotating lenses that

SPRING 1994 contain a con' of warm,

highly saline Mediterra- nean Water. One of the

earliest observations of

such an eddy occurred in lH7i; when researchers making temperature and

salinity measurements near the Bahamas

encountered a patch of unexpectedly warm and

salty water. The water property characteristics of this patch or lens

suggested it was of

Mediterranean origin. An

acoustically tracked float quickly deployed in the 20 - lens showed a persistent clockwise rotation.

This discovery prompted a more rigorous

search for salty lenses in

the eastern Atlantic, - closer to the Mediterra- 100 L 80 60 40 20 20- nean. In the last decade.

many new lenses have been found and studied using

hydrographic measurements, velocity profilers, and meddies were probably already at least one year old, and Salinity on the

floats. Eastern Atlantic salt lenses, now known as had traveled over 1,000 kilometers from their formation 10C surface (about 1,000 "meddies," for Mediterranean eddies, are typically 100 site. Meddy 1 was tracked with floats for two years, and meters deep)from kilometers in diameter, centered at 1,000 meters, and surveyed hydrographically four times. It traveled about a 1976 monograph extend over about 800 meters vertically. The water in 2,000 kilometers and decayed slowly over that time. by Valentine the cores is as meddy Worthington. The 35 much as 4C warmer T T T high-salinity water MEDDY 2 and 1 thousand Mediterranean part per NOV85 of re saline than the origin is shown in Madeira shades of red. ackgro u nd fluid,

indicating that form I'mii 2000 Mediterranean ^ JULY 86 MEDDY 1 Hyeres Undercurrent water Seamounf 30 somewhere along the 2000 1 ' coast of or 4000 Spain JUN 85 Portugal. Due to their Second Crui: rapid rotation, mixing ~' Canary ls.~ between the meddy Float cores and the surround-

ing fluid is limited. Thus Trajectories Trajectories of meddies are capable of in Meddies 25 three acoustically carrying the salty trackedfloats Mediterranean Water far OCT85 deployed in from its source. This is MEDDY 3 meddies (Mediter-

illustrated in the figure ranean eddies). Tfie at right, which shows JULY 86 rapid clockwise looping the trajectories of three of thefloats is floats launched in characteristic of eastern Atlantic 20 meddies. (Courtesy meddies. When first ofPhilip identified, these 30 25 20 Richardson)

OCEANUS John Kemp how long does it take a meddy to come to life? Since to prepares deploy meddles play a significant role in carrying salty water a sound-source from the Mediterranean into the Atlantic, we must learn mooring aboard more about their life histories if we are to understand R/V Oceanus in how the distributions of temperature and salinity are May 1993. Tim and in the North other sound maintained Atlantic. sources are used to A new program, called A Mediterranean Undercur- track RAFOSfloats. rent Seeding Experiment (AMUSE), is currently underway to address these questions. This study is being

conducted jointly with Laurence Armi of Scripps Institution of Oceanography, and Isabel Ambar of the University of Lisbon. In this experiment, 40 subsurface RAFOS floats are being launched in the Mediterranean

Undercurrent south of Portugal over a six-month period, at the rate of two per week. They are being tracked acoustically using an array of German, French, and US sound beacons moored in the eastern North Atlantic.

(RAFOS is the reverse spelling of the acronym SOFAR, for SOund Fixing And Ranging, and refers to the fact

that RAFOS floats listen for sound signals transmitted by moored sound sources, rather than themselves transmit-

ting to moored listening stations, as SOFAR floats do.)

slowly over that time. Meekly 2, on the other hand, The floats will remain underwater for about one year,

dispersed its salty core catastrophically when it collided following the movement of Mediterranean Water as it with a seamount. leaves the coast and mixes into the Atlantic. Some floats

Although the general characteristics of meddles have will most likely be trapped in newly formed meddies

been well-documented, fundamental questions about (see schematic below) . At the end of their mission, each their formation remain. For example, where do meddles float will automatically drop a ballast weight, rise tn the

form? What physical process is responsible for their surface, and transmit its stored acoustic tracking data,

formation? How many meddles are born each year, and as well as temperature and pressure data, by radio signals to receivers on two satellites maintained

by the French-run Service ARGOS. The data will then be sent automati-

cally over the Internet network to computers at WHOI. The floats Cape St Vincent themselves are consid-

Vilamoura ered expendable, though we have recovered a few

test floats that were set to

surface after a month.

It would be impracti-

cal, costly, and unneces-

sary to retain a large oceanographic research vessel for the float 36 - Hypothetical deployments. Just 2 Meddy meters long and weighing Formation only about 10 kilograms, which look like Site the floats, giant test tubes, can

easily be lifted and dropped over the side by one or two people. 35 Vertical temperature 11 10 9 8 7 are to A schematic diagram showing the Mediterranean Undercurrent (arrows) and the expendable profiles necessary locate the float launch bathythermograph (XBT) line/float-launch site south ofPortugal,

SPRING 1994 position within the Undercurrent, but these can be obtained from a small vessel using a portable expend 39 able bathythermograph (XBT) system. A charter arrangement was deemed the best alternative, and we PORTUGAL have the engaged 73-foot motor-sailing vessel, Kialoa II, for the six months of float deployments. (A former ocean racer, Kiulon II is owned and operated by Frank Robben, a retired University of California, Berkeley, mechanical

engineering professor.) Each deployment trip takes 38 about 24 hours tn complete. The vessel leaves 8 Vilumoura, a small town on the south coast of Portugal, o o -fc. at dawn, ami reaches the first of the 20 XBT station sites

by late morning. XBT stations are made approximately 1 e\en :!i> minutes while the vessel motors or sails along a

transect crossing the Undercurrent . When the center of the current is located, two RAFOS floats are lowered by

hand over the side, about 5 kilometers apart. The XBT 37 Cape stations are then continued to the offshore edge of the St. Vincent Undercurrent, and the vessel returns to Vilamoura in

the early morning hours of the next day. Some cruises are being extended to three or four days to recover test t'loats or to make more extensive XBT surveys of the Undercurrent.

Most of our results will not be available until the

middle of 1994, but some tantalizing observations have been made with a few floats deployed for 30-day test Trajectory of a RAFOSfloat deployedfrom Kialoa II in July 1993. Tlie dots indicate the The missions. The trajectory of one test float is shown at float's daily position. persistent clockwise looping of thefloat, starting at the southwest corner ofPortugal, represents the documented right superimposed on the bottom topography. This float first evidence of a neu' meddy 's birth. drifted downstream in the Undercurrent for four days until it reached Cape St. Vincent, at the southwest corner of Portugal. It then started looping clockwise and Thi.\ irnrk ixjiimli'd by the National Science Foundation. moved away from the coast into deeper water. The float Amy Bower grew up in Rockport, Massachusetts, where her made before the its eight complete loops end of mission. interest in science was nurtured by countless visits to Boston's This looping action, and the temperature record along Museum of Science, weather-watching, and high school math club. She discovered her sea (and the field of the float trajectory, indicate that this float was caught in legs physical oceanography) as an undergraduate on FW Westward, Sea Education Association's a new meekly the first time the birth of a meddy has research schooner. Lessons learned on board Westward have been documented. helped her conduct science under sail on Kialoa II \Vhen this experiment is complete, we will have many more trajectories illustrating how Mediterranean Water makes its way into the deep North Atlantic. We have already seen that Cape St. Vincent is a site of meddy formation. Other locations have been suggested, and our experiment will help to confirm or reject these sites of meddy formation. New sites may also be discovered. Once the major formation sites have been identified, the next step in understanding the medily story will be to Amy Bower and Joao Martins make detailed hydro- (University of graphic surveys during the Lisbon) recover a formation process to test RAFOSfloat on determine the mecha- the charter vessel nisms responsible for this Kialoa II hi July remarkable phenomenon. 199.1 Where Currents Cross

Intersection of the GulfStream and the Deep Western Boundary Current

Robert S. Pickart

Associate Scientist, Physical Oceanography Department

Gulf Stream is perhaps the most widely "separation" from the boundary is enormous, yet the known ocean current because its impacts precise reasons for the Gulf Stream separation (and range from effects on trans-Atlantic shipping that of other western boundary currents) remain Theto influences on the European climate. Near unclear. The physics of the separation process is a topic its origin off the southeast coast of the US, the Gulf of active research. Stream (or Florida Current, as it is known there) flows As the Gulf Stream turns offshore, into deeper water,

just off the continental boundary in several hundred it crosses the Deep Western Boundary Current

meters of water. After paralleling the boundary for (DWBC), which flows southwest toward the equator. roughly 1,000 kilometers the current abruptly turns The DWBC, carrying water recently formed at high offshore near Cape Hatteras, NC, and flows northeast latitudes, is the primary means for replenishing the toward Europe. Beyond Cape Hatteras the Gulf Stream deep waters of the vast ocean basins. Historically this extends to several thousand meters depth and mean- crossing was thought to be of little consequence, with above the ders significantly, in marked contrast to the shallow, the upper-layer Gulf Stream simply passing

nearly straight Florida Current. The impact of this deeper boundary current. However, new modeling

SPRING 1994 Lateral distribu- 40' N tions ofchloro- - Upper fluorocarbons Layer '(CFCs) help 38' 500-800 reveal thefate

Western Boundary 36 Current as it encounters the Gulf Stream. 34 CFC maps at three different lK depths in Slr the Deep Western Bound- 30 ary Current are shown in relation to the water parcel trajectories calculated from the density field. Study region of the GulfStream -Deep Western Crosses Boundary Current (DWBC) crossover, offshore of Cape denote Hatteras, NC. Tfie historic paths of the GulfStream shipboard- (from satellite observations) and Dtt'BC (from water sun>ey property measurements) are shoum, along with measure- locations of moored instruments. ment sites, and dotted lines results and observations reveal that the crossover has a trace the profound impact on both currents, and may in fact underly- contribute to the Gulf Stream mechanism separation ing itself. My fieldwork to address the Gulf Stream-DWBC bathym- interaction has included a shipboard survey of the etry. crossover region in collaboration with a three-year moored instrument array (see figure above). Since the upper portion of the DWBC extends to shallower depths than the separating Gulf Stream, one question concerns how this upper DWBC can survive its crossing and progress past the Gulf Stream. The shipboard survey measured density and various chemical properties of the water including chlorofluo- rocarbons (CFCs), which are used as refrigerants and aerosol propellants. The newly formed DWBC water is recognizable by its high CFG content, absorbed from the atmosphere while at the surface in high latitudes before sinking to form the DWBC. The lateral CFC distribu- tions from the shipboard survey reveal that the upper part of the DWBC is actually sheared in half by the Gulf Stream. At shallow depths a tongue of high-CFC water is pulled offshore, implying that this portion of the

DWBC turns with the Gulf Stream (top figure at right).

This is confirmed by the trajectories of water parcels calculated from the density field. Below the separating

Gulf Stream, the CFC distribution and trajectories are much different. Here a significant portion of the DWBC does progress past the Gulf Stream (although it bends Average Crossover / Maximum Crossover / Minimum Crossover / - - ^ of incidence = 7 (angle of incidence 4) / (angle 37) / (angle of incidence ) /

270 90 270 90 270

180

DWBC Gulf Stream Topography

Hi ree worth - years seaward, apparently due to the Gulf Stream above it layer Gulf Stream affect the deeper portions of the moored of middle figure, previous page). This "splitting" of the DWBC? The importance of these two boundary currents measurements lias DWBC obviously impacts the transport of properties by to the general circulation of the North Atlantic and the revealed that the the current to lower latitudes. The Gulf Stream's significant consequences of this crossing motivate angle at which the entrainment of newly formed water from the DWBC is continued of their interaction. This will GulfStream and investigation Deep Western an effective means of replenishing the interior basin. also shed light on similar crossovers in other areas. Boundary Current This entrainment also alters the dynamical structure of 77m- ii'urk mix supportedjointly by (he National Science cross varies the Gulf Stream itself. Whether this contributes to the Fan ndnlii in unit tlir OJJ/ri'ufXni nl /,'./,, // // issiinniinri-i'it in continually in actual separation as recent models suggest remains mi (ii'l/i'lf fi/ntled "Hoiv Does the Deep Western Boundary Current time. Tlie Deep to be sorted out. Cross the GuffStream?" (Journal of Physical Oceanography, Western Boundary December 1993, pages 2602-2616). The situation at depth is different still. The CFC Current responds and water near the 3,000-meter Robert Pickart was introduced to to changes in Gulf map parcel trajectories oceanography (and Stream orientation level show that the DWBC passes beneath the Gulf sea sickness) in college as a participant in WHOI's Summer Student Fellow program. Despite recent experiences, such with approxi- Stream with minimal difficulty and, as the bottom as knocked off his feet he a one- being by large waves, enjoys mately figure overleaf shows, there is also a significant inflow to sea and the excitement of new month going collecting delay. of water from offshore. This is not to however, that say, observations. His primary research interests are deep the Gulf Stream has no impact on the deepest portion of circulation and ventilation. His most challenging en- the DWBC. The three-year moored measurements deavor, however, is raising his four young children with his wife Anne simultaneously tracked the movement and orientation of the Gulf Stream and measured the flow of the near- bottom DWBC. These observations revealed that the

path and transport of the DWBC are continually altered by fluctuations of the upper-layer Gulf Stream. In particular, a change in the Gulf Stream's angle of separation causes a change in the trajectory of the

DWBC (see the figure above), while an increase in Gulf Stream transport leads to a weakening of the DWBC. The effect of the Gulf Stream angle change is in accord with recent theory, though the reason for the DWBC weakening with increased Gulf Stream transport is still unclear. While these recent measurements have improved our understanding of the complex Gulf Stream-DWBC

crossover process, they have also led to more intriguing questions. For example, by what physics does the upper-

SPRING 1994 Phil Richardson

hefts a RAFOSfloat aboard tf/FKnorr. Giant Eddies of South Atlantic Water Invade the North

Disrupted Flow and Swirling Waters

Philip L. Richardson Senior Scientist, Physical Oceanography Department

the equatorial region of the Atlantic, the North in the Atlantic and raise many questions about their Brazil Current follows the Brazilian coast numbers, life histories, and importance in the general

northwestward before turning sharply to the right circulation. Do these eddies provide a path for Inbetween 5N and 10N to cross the Atlantic as the significant amounts of upper-layer South Atlantic water North Equatorial Countercurrent. In 1990, satellite to travel up the coast? Do they transport water extend ocean color images were used to identify large, 400- primarily in the near-surface layer or do they kilometer-diameter, clockwise-rotating eddies that deeply into and below the thermocline? Do they drive appeared to be separating from the North Brazil recirculating flows in the underlying deep water as

Current as it turned sharply to the east, much as warm- eddies are thought to do in the Gulf Stream system? six core Gulf Stream rings form from northward meanders. From 1989 to 1992 we were able to track

Because these eddies originate from retroflections or retroflection eddies for the first time using surface in the sharp changes in current direction, they are called drifters and subsurface floats that were trapped North Brazil Current retroflection eddies. Satellite eddies' closed circulation and looped in them for as observations show them to be some of the largest eddies long as five months (see lower figure on page 20). The

OCEANUS Retro/lection Tlte inset schematic map eddies (orange) North Brazil shows the major tropical are water parcels Current currents between July and when the that pinch offfrom Retroflection September, the North Brazil North Brazil current Eddies Current and retroflects andfeeds into continue up the the eastwardflowing Smith American North Equatorial coast, instead of Countercurrent. In remaining with con trast, from Jan uary the North Equato- through July the North Equatorial rial Countercur- Countercurrent disap- Countercurrent rent. Approxi- pears in the western, (NECC) mately three of tropics, westward these eddiesform velocities are observed in each year starting this area, and the North in July, when the Brazil current continues North Brazil up the coast as the Current takes a Guyana Current. North Brazil sharp right turn, Current or retroflection. (NBC) After an eddy pinches off near

8N, the retrofl.ec- tion forms again 40 near farther south, looping trajectories were used to a latitude 5N to of describe the number, movement, and Float and Drifter 6N. Retroflection characteristics of these eddies. eddies are about Trajectories surface had WO kilometers hi Looping trajectories overall diameter diameters up to 250 kilometers with near the surface swirl speeds as fast as 80 centimeters and drift per second, dropping to diameters of northwestward at 140 kilometers at 900 meters with around 10 swirl speeds of 35 centimeters per centimeters per second. The deepest looper was at second. Tliey are 1,200 meters with a maximum thought to be diameter of 100 kilometers and a swirl A responsiblefor signifi- speed of 20 centimeters per second. carrying .. i I i cant amounts The from tin- of eddy shape, determined 40 Stm tli Atlantic loop diameters, appears to be an ~\ r water northward inverted cone. The data suggest that into the Caribbean Eddy Trajectories at least three such eddies form each Current. year from July to March. They move northwestward along the South

American coast with a mean velocity of 10 centimeters per second, and seem to disintegrate when they

encounter a 1,000-meter ridge between Barbados and Tobago. The water advected northward by the eddies probably then enters the B Caribbean Current. o

Retroflection eddies appear to a volume of South Atlantic water carry significant A) Composite of looping trajectories measured by northward into the North Atlantic, short-circuiting the surface drifters (42,45,57, 71) and subsurface SOFAR eddies 1989 longer route around the gyre formed by the North floats (22B, 22C, 28B) in retroflection from to 1992. were recorded Equatorial Countercurrent and the North Equatorial Surface drifters' positions by orbiting satellites several times per day. Acoustic Current. Each eddy transports about a million cubic signals transmitted daily by thefloats were recorded meters of water per second; three eddies per year by a moored array of listening stations. account for as much as a quarter of the total northward B) Inferred trajectories ofsi.r retroflection eddies. in the limb of the thermohaline transport upper Three different eddies were observed during the period (temperature and salinity driven) circulation cell. August '1989 to April 1990(288, 22B, 45/57).

SPRING 1994 Estimates place the total Surface-drifting uppcr-layr current transport buoy trajectories in the North of water into the North Brazil Current Atlantic at about 13 million retroflection cubic meters per second. during 1983 to This upper-layer northward 1985. All,five transport is balanced In ihr available drifters southward flow of cold North retroflected during the months Atlantic Deep Water beneath ofJuly to Decem- the eddies. (See figures on ber, implying that pages 6 and 7.) there was a The discovery of numer- complete break in ous eddies translating up the the north- westward coast helps explain a flow. discrepancy between two earlier data sets, from drifting buoys and historical ship drifts, that show continuous flow up the coast to the Caribbean from January through June. From July their rapid movement. Our knowledge about many through December, however, all available surface subsurface currents is even more rudimentary, which drifters (during 1983 to 1985) in the North Brazil suggests that we will find many more surprises in our

Current retroflected into the countercurrent (see figure ocean-circulation studies. above). In these same months, ocean-color images also Fundingfor the work described was provided by the National showed Amazon Water around the retroflection flowing Science Foundation. Recent scientific journal articles on I In' into the countercurrent, implying a complete surface subject of this article include "Tracking Ocean Eddies" (Philip L. American 261-27 disruption of flow up the coast. These drifters and Richardson, Scientist, May^lune 1993, pages 1) and "North Bras/I Ciinr/il Eddies" ('PL Richardson, G.E. images contradict historical ship drifts, which show a Hafford, R. Limeburner, and W.S. Brown, Journal of Geophysical continuous northwestward current there during July Research, 1994, Vol. 99, pages 5081-5093). through December, with a branch feeding into the Phil Richardson grew up on a cattle ranch in California countercurrent. An explanation for this discrepancy where he spent long hours chasing cows. He ran away to sea involves retroflection eddies. The continuous current and eventually earned a Ph.D in oceanography from the seen in is an artifact of ship-drift maps probably University of Rhode Island. His early experience on the ranch has averaging many years of velocity measurements on the proven valuable in his recent efforts to follow floats and drifters inshore side of the mean path of eddies, where eddy in the oceans. swirl velocity is 60 W northwestward. This is demonstrated by a map of surface-drifter velocity, including newer trajectories in retroflection eddies (see figure to right), that agrees with the earlier ship-drift data. The Guyana Current, if it exists during the months of July through December, is not a smoothly flowing current but instead consists primarily of a train of retroflection eddies.

It may seem surprising that we are just beginning to Mean vectors calculated nil available understand a major current velocity by grouping surface-drifter velocity measurementsfrom 1983 to 1993 into l-by-l bins. The newer data, system like this. Our which are similar to historical ship drifts, include trajectories in several ignorance has been caused retroflection eddies and show that on average the North Brazil Current runs partially by a lack of good in continuously up the coast into the Guyana Current, with someflaw peeling off situ data sets and partially by andfeeding into the countercurrent. The mean northwestward current is the intermittent character of partly an artifact of averaging the clockui.ie rotating swirl velocity ofseveral these powerful eddies and eddies as they drifted northwestward along the coast.

OCEANUS Nelson Hofig with RAFqSfloats that are bei

hi the la/xii

The Deep Basin Experiment How Does The Water Flow in the Deep South Atlantic? Nelson Hogg Senior Scientist, Physical Oceanography Department

from his success at explaining why Almost immediately, using conventional hydrography currents like the Gulf Stream are found along and a novel neutrally buoyant float that John Swallow the western boundaries of all ocean basins, (UK Institute of Ocean Sciences) had recently invented, FreshHenry Stommel proposed almost 40 years ago Val Worthington (WHOI) and Swallow discovered the that similar features in the Western Current" of the scheme on thjjlwere abyss. According "Deep Boundary part to his scheme, these would be fed by water made dense the Continental Rise south of Cape Cod. Then, with help in winter. This convective from Stommel and others and these same IjXpolar regions during using floats, dilation would be completed by a slow bleed of water Swallow set out to confirm the existence of the slow

from the deep boundary currents into the ocean interior circulation that was an integral part of the

interior, and a broad rising through the ocean ther- scheme. Instead, they found that the deep ocean, away mocline (a region of rapid decrease in temperature from the boundary, was dominated by a vigorous eddy

with into it like the that depth ) the upper ocean where would then field, much weather systems dominate our I" 1 carried poleward (in. for example, the Gulf Stream) day-to-day existence on land. With the ship-based to complete the circuit. technology then available, there was no hope of being

SPRING 1994 able to extract the very weak mean flow in the presence circulation. We have centered our program on the Brazil Upper (left) and lower of so much noise from the eddies. Basin for several reasons: French and German oceanog- (right) Atlantic Ocean Advances in technology over the last four decades raphers had already planned extensive hydrographic circulation scheme nun make it possible to revisit this issue, and a group of work in this region as their contributions to the World proposed by Henry Brazilian scientists are Ocean Circulation work in the US, German, French, and Experiment, previous Stommel in 1957. engaged in an attempt to reveal the secrets of the deep area suggests that the return circuit through the Water is made thermocline is surprisingly dense by winter strong in this area, and the cooling, evapora- of the Brazil tion, and ice - geometry formation in the Basin is relatively polar regions and simple there are just then moves three major connecting equatorward as 10 to passages neighboring narrow, relatively basins and the bottom and swiftflows along boundaries are smooth, the western

for the Mid- margins. Water 20 except slowly leaks away Atlantic ridge portion on into the interior the basin's eastern side. and then upward The observational across the 30 attack is multipronged. thermocline to First we will complete an eventually return extensive network of to the poles.

"hydrographic lines," 40 making temperature,

salinity, and other measurements. The

sampling strategy is

designed to split the basin BOS - into a number of boxes. Using various conserva- tion principles (such as

60W 50' 40 30 20 10 10E for mass, momentum,

heat, and salt), we hope Tlie bathymetry of the South Atlantic showing pathwaysfor theflow of the two to determine whether the main deep water masses, North Atlantic Deep Water (NADW) and Antarctic Bottom Water (AABW). rising process occurs

OCEANUS 23 A schematic illustration ofhow the deepest water entering the Brazil Basin front the south through the Verna Channel can conserve both mass and heat by rising and changing density. Tlie water entering the basin below a certain temperature does not leave at that temperature, but instead rises and warms through downward diffusion ofheat. Tfiis balance allows estimation of thermal

diffiisivity, a measure of the rate of mixing in the ocean.

uniformly across the basin, as in Stommel's visionary electronics and mooring hardware now allows us to set scheme, or perhaps is localized along the boundaries, as these moorings for at least two years and anticipate a

others have suggested. Current-meter moorings will 90 percent data return. A similar array, moored across

also be deployed, mostly along the boundaries and the Vema Channel some 15 years ago, was part of the

within the deep connecting passages. Progress in inspiration for the Deep Basin Experiment: It showed

that the deepest, or

"bottom," water, was pouring into the Brazil Basin but had no exit other

than to warm and rise.

Swallow's neutrally uoyant floats have evolved considerably over

the last 40 years. We will

be setting more than 200 floats, mostly of a variety known as RAFOS floats

developed originally by Tlie Deep Basin Tom Rossby at the Experimentfield University of Rhode Island program. Filled and now commercially circles locate in North acoustic sound produced MA. These sources that will Falmouth, listen for moored- trackflocks of floats ni'iilrallt/ buoyant acoustic-source signals floats, rectangles that can travel great indicate moored distances in the ocean, as I'linrtit-ineter Moored Arrays much as 1,500 kilometers arrays, and lini's Sound Sources for our instruments and showplanned Q we shipboard depths. Consequently, need nine In/drographic 40 only moorings surveys. 20 10 to ensure that a given

SPRING 1994 1 float will he able to hear at least two sound sources I In ocean interior. Small but measurable amounts of

Accurate clocks and knowledge of the speed of sound this material are injected into the ocean over a small in water then allows triangulation on the float to area at a specified depth. Over the next few months determine its position to within about a kilometer. The eddy processes stir the injected material both horizon- miniaturization of electronics allows these instruments tally and vertically into a smooth distribution, much to be housed within what amount to large test tubes like stirring cream in a coffee cup. The rate at which about 2 meters long. At the end of their two-to-three- the material spreads vertically is a direct measure of

\ear missions, they drop a weight and rise to the the process invoked by Stommel so long ago. Following surface where they report their information to home very successful use of this technique in the thermocline base through a satellite link. With these flocks of long- of the eastern North Atlantic in 1992, Ledwell is now lived floats, we anticipate now being able to fulfill John planning to perform a similar experiment in the depths Swallow's dream of directly measuring the weak of the Brazil Basin to answer the question of how interior flow. rapidly processes away from the boundaries are able to At least two observational techniques are available carry the deep water back up to the thermocline on to us now that were unknown 40 years ago. Over that their return poleward journey. time, industry has been producing ever-increasing The Deep Basin Experiment is in the early stages of amounts of chemicals called chlorofluorocarbons for field work, with a great deal of instrumentation use as refrigerants and aerosol propellants, among installed in the Brazil Basin. Data collection will other things. Chlorofluorocarbons disperse through the continue for several years, and then we can begin to atmosphere and from there dissolve in the surface attack the scientific issues that have been unanswered layers of the ocean and are incorporated in the sinking for the past 40 years. polar waters that are the sources for deep boundary I'S (null/ mi fur I lif fleet i Hiix/i/ K.i-/n'i-/nii'/il K firm nlnl by the currents. In the 1950s tests addition, hydrogen bomb Xiil/n/iai Science Foundation. introduced radioactive tracers into the amosphere and Nelson Hogg dreamt of being on warm, sunny beaches by the ultimately into the oceans, tritium being one measur- sea while growing up in northern Manitoba. It wasn't until he had able tracer that is taken the ocean. These become up by a summer job during college at the Defense Research Establishment markers of the source water and allow us to estimate in Dartmouth, Nova Scotia, that he had that opportunity during an the time since that water was last in contact with the instrument-testing cruise to Bermuda That was enough to convince him that oceanography had advantages over high-energy physics sea surface (figure below). For most of the time since then, he has been at WHOI We have had no control over the input of these industrial and military pollutants to the oceans. Recently, however, WHOI Associate Scientist Jim Lcdwell pioneered the use of a deliberately released, nontoxic chemical compound (known as sulfur hrxafluoride) to directly infer vertical mixing rates in

South Atlantic Ventilation Experiment, Leg 1

A section showing tritium concentra- tion along a line starting near 10S on the Brazilian Coast and angling northeastward to cross the equator near20Watthe Mid-Atlantic 250 500 750 1000 Ridge. T)ie section Distance (km) ends at Africa.

OCEANUS Mike Spall rtnixan ocean circulation model. Wave-Induced Abyssal Recirculations

Turning This Way and That Way

Michael A. Spall Associate Scientist, Physical Oceanography Department

is generally believed that deep-ocean circulation small regions at high latitudes and uniformly upwell

plays a fundamental role in the global climate throughout the lower latitudes into the upper ocean. system. The abyssal circulation transports cold, This simple model predicted that a series of deep Itfresh water formed at high latitudes toward the western boundary currents in the world's ocean basins mid and low latitudes, where it upwells to maintain the would carry the waters away from their regions of main thermocline (a region of rapid decrease in formation, and that the basin interiors would be temperature with depth) in the presence of downward characterized by very weak poleward flow. heat diffusion. This is thought to be accomplished However, recent basin-scale hydrographic measure- through a complex pattern of boundary currents and ments in the North Atlantic and in the Brazil Basin of interior the location of the World flows, although the dynamics, and even the the South Atlantic Ocean ( paths, of these flows are not well understood. The Ocean Circulation Deep Basin Experiment) indicate

traditional view of abyssal circulation stems from Henry that strong flows are not confined to the western

Stommel's late 1950s and early 1960s theoretical work, boundary regions. Instead, the abyssal ocean appears to which assumes that deep waters are formed in very contain significant large-scale recirculation gyres,

26 SPRING 1994 Brazil Basin, with bottom topography approximating the deep bowl shape of the Brazil Basin between the South

American coast and the Mid-Atlantic Ridge. The basin's southern and northern limits are partially blocked by

ridges, representing the Rio Grande Rise in the south and the Ceara Rise to the north. The basin extends

approximately 3,000 kilometers in the north-south direction and 2,000 kilometers east-west. We know that

Antarctic Bottom Water is formed at very high latitudes in the southern oceans and spreads northward, passing through the Brazil Basin into the North Atlantic. We approximate this flow into the Brazil Basin by introduc- ing a deep western boundary current of Antarctic Bottom Water through a channel at the southern boundary of the domain. The flow exits through a channel adjacent to the western boundary at the domain's northern limit. Transport within each outflow

layer may be different from that at the inflow, allowing for overall warming of bottom waters. We have conducted two sets of experiments, one Schematic of the model domain representative of the with a steady and one with an unsteady deep western Brazil Basin in the South Atlantic Ocean. Tlie basin is boundary current. We keep the physical configura- bounded to the west by South America, to the east by the tion stratification, topography, and flow strength Mid-Atlantic Ridge, to the south by the Rio Grande Rise, constant, and render the flow steady by increasing the and to the north by the Ceara Rise. Narrow channels current's the bottom. This allows allow Antarctic Bottom Water to flow into the basin drag against approach us to the influence of from the south and out to the north, where it crosses the investigate time-dependent equator and enters the North Atlantic Ocean. motion on the mean state under otherwise similar flow

conditions and model physics. The schematic below cross-basin flows, eastern boundary currents, and shows the mean flow of Antarctic Bottom Water in the waves Mike article on basin with a western topographic (see McCartney's deep steady deep boundary Schematic of the page 5 for a discussion of the deep mean flows). The current. Consistent with traditional models of deep mean circulation presence of these currents implies that the deep ocean circulation driven by large-scale upwelling, the ofAntarctic is influenced more Bottom by complex physics than were dominant circulation features are the northward . Waterfrom the model with a included in the simple early models. Understanding the flowing deep western boundary current and a large-scale steady deep structure and forcing mechanisms of abyssal recircula- cyclonic (clockwise in the Southern Hemisphere) western boundary tion and their relation to the western gyres, deep current. Colored boundary' currents and upwelling, is essential if we are lines indicate to fully understand deep-ocean circulation and its bottom water role in the global climate system. depth (yellowfor red I have been using a numerical model to shallow, for deep) while arrows investigate some of the consequences of more indicateflow complex physics on abyssal circulations. The direction. The bold model the approximates continuously blue arrows show stratified ocean as three constant-density where the water layers. The two deeper layers represent the flow's into and out lower and upper Antarctic Bottom Water of the basin. Most the water is (Antarctic Bottom Water is a cold, fresh water of carried to the mass that originates near Antarctica and flows north over the into the Brazil Basin from the south); the sloping bottom shallowest layer represents the upper ocean. This along the western is well system suited for studying abyssal circula- boundary in the tion because it allows for steep and tall bottom deep western topography, time-dependent motions, such as waves boundary current. There is a and eddies, and vertical mixing between water cyclonic (clockwise) masses. The effect of mixing between the layers is recirculation with paramaterized here as a uniform upwelling of water a depressed center between the two deeper layers. in the basin The figure above shows the model configuration. interior. The domain is roughly patterned on the geometry of the

OCEANUS 27 aspects of the observations agree with the model circulation, including an eastward flow of lower and westward flow of upper Antarctic Bottom Water. A key to determining if the observed large-scale recirculation gyre is a result of wave propagation over the western

slope, as predicted by the model, would be the presence of topographic waves with energy propagation into the basin interior. RAFOS floats and current meters

recently deployed as part of the World Ocean Circula- tion Deep Basin Experiment should be helpful in

testing this hypothesis. (See Nelson Hogg's article on of float page 22 for a discussion the experiment. )

Tliix ii'ork ix mijiportril hi/ the National Scienc.e Foundation

a ml I In- ivTii/iin. Bundesnrinisterfur Forschung und Technologic. out I'll, n/iiin imrl of the resfinr// f/,'vr//W l/c/r mix I'uiricil

///< -i 'sk/inde while Michael Spall was rixilinii li/xli/iiljiii- Mi n in Kiel. Germany. A manuscript describing this work has been submittedfor publication in the Journal of Marine Research.

while Michael Spall became interested in oceanography working on his Ph.D in Applied Mathematics at Harvard in a University. He came to Woods Hole 1990 after spending postdoc at the National Center for Atmospheric Research in Boulder, Colorado He enjoys spending time with his family,

sailing, fishing, playing volleyball, and studying oceanography.

Schematic of the recirculation gyre in the basin interior. The flow indicated mean circulation here is essentially void of any time-dependent motion. ofAntarctic bottom To investigate the influence of time-dependent Waterfrom the motion, we repeat the calculation above with reduced model with an bottom drag, which allows small perturbations in the unsteady deep waves anil western boundary mean flow to grow into large amplitude current. Tliedeep eddies. The schematic above indicates that the mean westem boundary deep western boundary current continues to flow to the current still flows north, but now the mean interior circulation is to the north along anticyclonic (counterclockwise), opposite to that found the western in the absence of time-dependent motions. boundary, but, now The nature of the variability near the western the direction of c 400- is indicated in the at a kinetic- in the basin boundary figure right by ; flow CJl (U near the interior is energy time series over the sloping bottom 1i <~ c o.Si ant/cyclonic boundary. We find the passage of high-frequency events m a- "8 (counter-clock- with velocities as fast as 15 to 20 centimeters per second wise) with a raised 200 - about once each year. These currents are much faster center. Hie than the mean speed of the deep western boundary unsteady western current, which is approximately 3 to 4 centimeters per boundary current second. These events are the of entirely changes energetic signature the waves that are by mean-flow topographic Rossby generated 4 6 direction in the meandering of the deep western boundary current and Time (years)

basin interior. 1 that propagate downslope into the deep basin interim . These waves carry energy from the deep western Kinetic over the bottom near the boundary current into the basin interior and entirely energy sloping western boundary. Tlie high-frequency variability change the direction of flow around the deep basin. marks the passage of topographic waves that are Recent observations of Antarctic Bottom Water generated in packets by the meandering deep western circulation within the Brazil Basin indicate the boundary current about once per year. Energy these of a basin-scale recirculation presence anticyclonic gyre waves carry into the basin interior causes the large with strength and distribution similar to that found in scale recirculation gyres seen in the two previous the model when waves ;nv allowed to develop. Other figures to change directions.

SPRING 1994 MBL WHO! LIBRARY

rich or within the refuge of a monastery. We can count ourselvesfortunate

Institution Director Robert Gagosian reads the citation for the Henry Stommel Medal in Oceanography before presenting the certificate to British oceanogra- pher.John Su'alloiv. Elizabeth Stommel later presented the medal to Swallow. Tlie medal likeness ofHenry Stommel was sculpted by Judith Munk, a La Mia, California artist and architect, and wife ofScripps oceanographer Walter Munk.

The Henry Stommel Medal in Oceanography Established by the Trustees of the Woods Hole Oceanographic Institution Awarded To John Crossely Swallow, <&.$. For fundamental and enduring contributions to observing and understanding ocean processes, and in recognition of his exemplary seagoing oceanographic studies of North Atlantic, Mediterranean, and Indian Ocean circulation

February 9, 1994

Stommel Medal Award Committee Chair Nelson The first Henry Stommel Medal in Oceanography Hogg described Swallow's capacity for hard work at was presented early this year to British oceanogra- sea as "legendary." In her remarks before presenting ffi pher John Crossley Swallow, a long-time Stommel CD the medal named for her husband, Elizabeth Stommel friend and colleague who is best known for his said of Henry: "It was important, he said, to live with invention of the neutrally buoyant float (otherwise it GO the ocean: to watch, to feel, to hear it, to confront known as the "Swallow float") and its pioneering use c-t-O mindfully each day, to let it permeate one's life." The in uncovering important elements of the ocean's 3 legendary combined work of Henry Stommel and John general circulation. Stommel and Swallow shared 3 ct> Swallow, at sea and ashore, have contributed mightily many interests, but especially their desire to under- to the world's of ocean currents. stand ocean circulation. knowledge

Henry Stommel, 1920-1992 Woods Hole Oceanographlc Institution Woods Hole, MA 02543 508457-2000