Spring Wfo Ocean Eddies , • • r ..*.-•• ••.- ' V m mm1 • Anticyclonic Gulf Stream /~i 1 1 1 1 1 1 1 1 1 1 1 1 \u 1 1 ij 1 1 1 „!>•-». MK-H HI. I.IMC-K It- I. !<••>< Figure 5. The shape of the temperature distribution for the experiment shown in Figure 6. The Gulf Stream separates cold water, to the north, from warm water, to the south. As it meanders, a cold eddy breaks off and wanders to the south. cold water in the north to displace lighter, warm water at the same level in the south. This same desire of the fluid to come back to equilibrium is active in any array of baroclinic eddies. The theory for this process quite accurately predicts that eddies of 100-200 kilometers in diameter should appear, and this is a partial rationalization of the observed size of eddies. A fifth, and my final, example of eddy motion is the most difficult to describe mathematically. This is the simple tendency for a rough sea floor to break up currents into little eddies. The effect is amplified by the earth's rotation, which tends to pull the flow at different depths into vertically coherent alignment. A typical representation of the hills and valleys of the sea floor, used in my simulations, is shown in Figure 7. Free currents can flow easily along contours of constant height, but waves result wherever fluid is forced to move across these same contours. Figure 8 shows the flow in a computer simulation of the deep ocean, over a typical pattern of seamounts and ridges. The increasing complexity and small size of the deep eddies are evident. When all of the important features of the ocean are put into the theories or simulations of ocean eddies (density stratification, beta effect, mean currents, wind stress, rough bottom topography, for a start), one finds emerging various hybrids of the above special examples. But it is comforting that, despite a great deal of variation from place to place in the ocean, the predicted eddies resemble the 200-kilometer-wide cells seen at sea. And, what is most amusing, none of those features can be omitted without making the simulated patterns grossly in error. Interaction Between Eddies and Circulation Waves on the Gulf Stream front represent one genuine interaction between eddies and the time- averaged flow (otherwise known as the Eulerian circulation); here the circulation produces eddies. But the events can also work in the opposite direction, for eddies, if fast enough, can gang up to drive a systematic flow. It is just possible that the ocean works as a sort of Rube Goldberg device, with the wind driving a strong circulation that breaks down into eddies; the eddies then drift off and radiate into the far reaches of the ocean, where they recombine to drive new elements of circulation. Models of this sort are now being explored. Some rigorous theory suggests that, indeed, the chaotic stirring ability of eddies (which would quickly diffuse a huge blob of ink) induces, by necessity, a circulation of the sea at the next larger scale. 35 Figure 6. Streamline patterns in a computer model of the Gulf Stream. The upper sequence shows the upper-ocean flow, the lower sequence shows the flow in the deeper water. The current, initially narrow and intense in the upper water, and nonexistent in the deep water, begins to meander, and spontaneously breaks into a pattern of eddies, in both deep and shallow water. Time is in months. 1111 1 1 nil I Illl miiiiiiiiiiiiiii i*» •.• Sfv ».» INC* ».« iiHMii iiiimmmiii •"'*' '!!«!! S 1! I !,' t . i ,i lit,!: t; 'IIIIII 36 iiiiimliiiinnrmiJHfiiiiiiil'iiiniiiifliiiHH1 LO*-I.-«.M«»II Ml* 1.MK«II INC- S.«7t-H Figure 7. Perspective view of the sea- floor topography used in simulations of ocean eddy fields. The region is 2000 kilometers square; consequently, much small-scale detail had to be deleted from the model. 31 STREAMFUNCTION FIELDS/10 FOR T = 250 DAYS LAYER 1 LAYER 2 LAYER 4 LAYER 6 CONTOUR INTERVAL = 20 KM? /OAT FOR LAYERS 1 AND 2 = 10 KM1 /DAT FOR LATERS 4 ANO • 500 KM Figure 8. Flow patterns at four different depths in a fast-ocean model. Layer one is nearest the surface, layer six is nearest the bottom. Notice the smaller scale of the deep eddies, which respond to the sea-floor topography. Here the domain is 1 000 kilometers square. Tlie motions at the shallower levels are what I have called thermodine eddies. (From W. B. Owens, 1975, A numerical study of mid-ocean mesoscale eddies, Ph.D. dissertation, Johns Hopkins University, Department of Earth and Planetary Sciences.) 38 Topography '- • Remember, too, that the ocean behaves differently at different depths. The eddies in fast oceans are able to transmit energy from the shallow water to the deep; there, remote from the force of the wind, they can again recombine to drive a large-scale circulation. This action is particularly important to the fate of water masses in the depths that are known to travel horizontally across entire oceans. A particular example of the mean flow produced by a random pattern of eddies is shown in Figure 9. Here, as time progresses, the flow pattern more and more resembles the contours of ocean depth. Again, this induction effect appears to be a necessary consequence of the ability of eddies to diffuse a patch of marked fluid. Conclusion I have suggested that an ocean with a flat bottom, driven by very gentle winds blowing across its surface, would respond in a predictable fashion, with the eddy and circulation components formally independent. But the actual winds are far too strong for this to be a complete picture. Instead, a chaotic state prevails in which the circulation and eddies may in turn feed upon one another. Whereas eddies in slow oceans behave predictably as waves, those in fast oceans interact violently with one another, and react to the sea-floor topography. An empirical understanding of, and a firm statistical theory for, these interactions are beginning to emerge. The behavior of eddies will one day be as well grounded as, say, the classical mechanics of a vibrating string or a gas composed of colliding molecules. Peter Rhines is a senior scientist in the Department of Physical Oceanography, Woods Hole Oceanographic Institution. He visits NCAR (National Center for Atmospheric Research) in Boulder, Colorado, for computer modeling. Streamlines Figure 9. T/ie development of strong eddies over rough bottom topography, in a single-layer (no pycnocline) ocean model. Tfre eddies spontaneously develop into flow along the height contours of the topography. (Courtesy of G. Holloway and M. Hendershott) Figures 2-7 from P. Rhines. "The dynamics of unsteady currents,' in The Sea, vol. 6. Marine Modeling, edited by E. D. Goldberg, I. N. McCave, J. J. O'Brien, and J. H. Steele, to be published in 1976 by John Wiley & Sons, New York. 39 POLYGON-70: A Soviet Oceanographic Experiment by Nicholas Fofonof f The word nOJIVlFOH implies a test range or proving ground. For Soviet oceanographers, the series of POLYGON experiments conducted during the past two decades represents a focused effort to obtain oceanographic data to test theoretical concepts of ocean circulation. The POLYGON is the proving ground for ideas. As observed by Academician L. M. Brekhovskikh and his colleagues (Proc. R. S. E., vol. 72, p. 34, 1972): ... a collection of observations cannot replace a purposefully planned experiment. And as it has been figuratively put down by H. Sverdrup, at its early stage of development the situation in physical oceanography has been such that observations were carried out by many while few pondered over them. A sort of crisis of the 1940's and 1950's led to a new situation when, as V. B. Stockmann said, too many scientists were engaged in theory and calculations and only very few conducted purposeful measurements in the open ocean. These scientists also point out that recent developments in theoretical oceanography, meteorology, and related sciences, combined with advances in technology, have converted descriptive physical oceanography into a physical science of the ocean. A similar realization of the gap between theory and experiment among Western oceanographers was expressed, for example, by H. Stommel in 1954 in a privately circulated note whimsically titled "Why do our ideas about ocean circulation have such a peculiarly dream-like quality?" in which he pleaded for oceanographic experiments designed specifically to test theoretical ideas. POLYGON Experiments Soviet oceanographers credit V. B. Stockmann with developing the concept of methodological research that evolved into the present series of POLYGON experiments. In 1935 Stockmann and I. I.Ivanovsky studied the structure of turbulence in the Caspian Sea. Using rudimentary equipment, they made a series of measurements over a three-week period from two anchored ships and found that current fluctuations were not closely linked to local winds and persisted even during periods of no winds. Their measurements enabled them to estimate momentum transfer by evaluating the correlations between current components— the Reynolds stresses extensively studied in turbulent flows. At Stockmann's initiative, a long-range plan was developed during the postwar years to examine the structure of ocean currents and the relationship between the currents and the density field. It was not until 1956 that an experiment of comparable scope was carried out. A POLYGON in the Black Sea was instrumented with a moored buoy (42° 50'N, 40° 25' W) carrying current meters at 20 and 100 meters depth. Currents were recorded at 20-minute intervals for 18 days. During the same period, temperature and salinity were measured from two ships. The data from this experiment permitted a more detailed look at current fluctuations over time scales of a few hours to a few days. From the results of this work, R. V. Ozmidov concluded that the current fluctuations had characteristics of locally isotropic turbulence in that no preferred direction could be found. The 1956 experiment is noteworthy in that a number of young Soviet oceanographers were brought together in a cooperative venture. It is this group that evolved and extended the POLYGON concept and that forms the core of present-day POLYGON oceanographers in the U.S.S.R. The first venture to carry out similar research in the deep ocean occurred in summer 1958. Three buoys were anchored in the northeastern Atlantic (53° 44'N, 17° 31' W) in a right triangle with separations of 70 and 90 nautical miles. Measurements with current meters were made at depths of 50, 100, 200, and 400 meters recording at 30-minute intervals for about 14 days. Hydrographic stations were taken periodically at each location to determine the temperature, salinity, and density fields. This experiment revealed major 40 spatial variations of the currents over the short distances between the moorings, thus casting doubt on the ability to predict motion and density fields from a few point observations and indicating the need for a statistical approach. In the decade that followed, new techniques of current measurement were developed in the U.S. and U.S.S.R. In 1965 the Woods Hole Oceanographic Institution Buoy Group (more formally called the Moored Array Project) established the 10-year time-series station (Site D, 39° 20'N, 70°W) at which many of the techniques and equipment for successful long-term deep-sea moorings were developed and tested. The next POLYGON experiment was deployed in 1967 in the Arabian Sea by the Institute of Oceanology. This program called for 7 moorings instrumented at 1 1 depths from 25 to 1200 meters, with additional short-term moorings to measure to 4000 meters depth and to sample high frequencies (sampling at 5-minute intervals). A hydrographic survey was made in a 5-degree square enclosing the L-shaped moored array. The experiment lasted two months in order to examine spatial-scale fluctuations of 20 to 1 50 kilometers and time scales ranging from a few minutes to two weeks and more. A site with flat bottom topography was selected to minimize complications of flow interaction with the bottom. Program objectives included the following: -To define the statistical characteristics of the currents, in particular their temporal and spatial spectra. -To estimate how the magnitudes of terms in the equations of motion are affected by the area chosen for averaging. -To test the "frozen turbulence" hypothesis (an assumption made to estimate spatial dimensions from a one-point time-series measurement). -To determine location of energy sources in the frequency spectrum and to find direction of energy flow at the largest scale. -To test the balance between pressure gradients and Coriolis forces (geostrophy). How well is the geostrophic balance satisfied as a function of scale or averaging? What are the adjustment time constants? -To investigate properties of the vertical and horizontal structure of the velocity and other hydrological fields. The Arabian Sea experiment represented considerable progress in the evolution of the POLYGON concept. Except for duration and level of effort, it had all the major components of present-day experiments. The results indicated that a period of two months was inadequate to resolve the full-time variability of the currents and that larger and longer POLYGONS would be necessary to encompass the full range of ocean variability. Such programs would call for further development of resources and organization. Plans for a long-term open-ocean experiment were already being formulated at the time of the Arabian Sea POLYGON. Also, initial discussions with U.S. oceanographers in 1969 led to consideration of possibilities for future joint efforts. These discussions continued through the early 1970s and established the foundations for POLYMODE, a major joint experiment scheduled for 1977-78, continuing at least one year and combining the oceanographic resources of both countries. POLYGON-70 Several oceanographic laboratories of the Soviet Academy of Sciences, under the scientific leadership of L. M. Brekhovskikh of the Acoustics Institute and V. B. Kort of the P. P. Shirshov Institute of Oceanology in Moscow, joined forces to deploy POLYGON-70 in the Cape Verde Basin of the tropical northeastern Atlantic (16° 30' N, 33° 30' W) from February to September 1970. The basic design of POLYGON-70 grew from objectives and considerations similar to those of earlier experiments, but the range and resolution of spatial scales studied were increased over those of the Arabian Sea POLYGON. A total of 1 7 surface-float moorings were set within a 2-degree Figure 1. Tfiree-day averages of current at 300 meters from POL YGON- 70. Current vectors are shown at each of the 1 7 mooring locations. (Adapted from L. M. Brekhovskikh eta!., 1971, Deep-Sea Res. 18:1189-1206, and L. M. Brekhovskikh etal., 1972, Proc. R.S.E. (B) 72:557-56.; 41 Photographs taken aboard the Soviet research vessel Akademik Kurchatov in the Cape Verde Basin (off West Africa) in 1970 during the POL YGON experiment. (Photos by Charles Ross of Bedford Institute, Halifax, Nova Scotia. Text by Robert Heinmiller, POL YMODE Executive Manager, MIT.) Launching of the Soviet CTD (conductivity /temperature /depth) profiler "AMCT" ("Stork"). The instrument measured conductivity and temperature as a function of depth down to 200 meters, with the data transmitted up the conducting wire to a recording and processing unit aboard the ship. \ A cylindrical surface mooring float on the foredeck after recovery. The float was made of low-density foam, with a central pipe for rigidity. The mast, with a sheet-metal radar reflector at the top, had broken off while the mooring was on station. It swung alongside the float, held by a wire-rope lift line, and gradually cut the gouge visible in the side of the float. 42 Anchors used on the POL YGON moorings. Made of cast iron, they were strung together in groups of three with heavy chain, for a total weight of about 700 kilograms. A swivel was inserted at the point where the anchors were attached to the mooring line. Racks of Nansen bottles (for water sampling) in the passageway on the starboard side of the Kurchatov. Additional stocks of bottles were carried in the hold for the seven- month trip. 43 square in the form of a cross with north-south and east-west arms as shown in Figure 1 . K. V. Konaev of the Acoustics Institute suggested the design based on antenna theory, to measure the selected band of wavelengths or spatial scales. This twofold increase in scale and threefold increase in duration represented a major expansion of the level of effort. Furthermore, the Arabian Sea experiment involved one ship (R/V Vitiaz), whereas POLYGON-70 required six ships— three to maintain the buoy array and three more to carry out the entire experiment. R/V Dimitri Mendeleev (command ship of POLYGON-70) and R/V Akademik Kurchatov of the Institute of Oceanology maintained seven mooring positions each. The remaining three were tended by R/V Andrei Vilkitsky of the Navy Hydrographic Service. The moorings were replaced at 25-day intervals throughout the experiment— an operation requiring 7-10 days to complete. Visual inspections of each mooring were made about every 3 days, and any that drifted from their grid locations were repositioned. As a result, 90 percent of the moorings were recovered. However, such close inspection and maintenance consumed a large amount of ship time that could have been devoted to other research, and hydrographic coverage was therefore not as intensive as that in the U.S. MODE-1 (Mid-Ocean Dynamics Experiment, 1973). Six hydrographic surveys were made approximately once a month during the experiment. Of these, two were large-scale surveys in a 2-by-2- degree square in August and a 3-by-3-degree square in September. Three other surveys examined the central region containing the moored array. In addition, north-south sections were occupied to investigate the surrounding ocean structure. In all, about 500 hydrographic stations were completed. The program was supplemented by a variety of other measurements. R/V Akademik Vernadsky of the Marine Hydrophysical Institute of the Ukranian Academy of Sciences conducted microstructure studies using continuously profiling instruments. The two research ships of the Acoustics Institute, R/V Sergei Vavilov and R/V Peter Lebedev, carried out special acoustic experiments in addition to the hydrographic observations. Separate programs to study meteorology, air-sea interaction, geophysics, geochemistry, and biology were included in the overall project. The major result of POLYGON-70 was the documentation of an energetic anticyclonic (clockwise) eddy that dominated the current field over the major portion of the experiment, as indicated by the current vectors in Figure 1. The eddy was elliptical, oriented NW-SE, with a length of 400 kilometers and a width of 200 kilometers. It drifted slowly westward at speeds of 4-6 centimeters per second and displayed orbital speeds ranging from 10 centimeters per second at 1500 meters depth, to 20-25 centimeters per second at 200-500 meters. The hydrographic surveys showed a density structure consistent with the currents but yielded lower speeds than the direct current measurements. Preliminary results of POLYGON-70 were made available in 1972 to U.S. oceanographers for planning MODE-1 (see page 45). U.S.S.R. oceanographers K. Chekotillo, L. Fomin, and M. Koshliakov participated in a number of planning sessions in 1972 and were observers (together with Y. Grachev) during the experiment itself. Their experience was useful because it encouraged use of a mapping array and a more intensive density program in attempting a more detailed evaluation of the geostrophic balance than had been tried in POLYGON-70. Analyses of both POLYGON-70 and MODE-1 data are still underway. Preliminary findings are being communicated among the scientists through meetings, newsletters, and formal publications. Since MODE-1, the pace of international meetings has escalated in preparation for POLYMODE. Scientists from both countries have pooled their knowledge and resources for a determined attack on the description and understanding of the eddy structure in the North Atlantic. POLYMODE represents the next major step in increasing our knowledge of eddies and their role in ocean circulation. Nicholas Fofonoffis a senior scientist in the Department of Physical Oceanography, Woods Hole Oceano graphic Institution. 44 The MicUOcean Dynamics Experiment by Carl Wunsch The Mid-Ocean Dynamics Experiment (MODE-1) was a large, intensive, and logistically complicated program conducted by physical oceanographers in the North Atlantic in mid- 1973. From the participants' point of view, it marked an important step in changing fundamental ideas about the physics of the ocean. But, almost as important, it altered the way oceanographers go about their business. MODE-1 was an exercise in the organizational aspects of oceanography almost as much as it was a scientific experiment, and its organization has tended to become a model for many other large oceanographic projects. For several years, beginning in 1970, MODE-1 absorbed the energies and talents of a substantial number of physical oceanographers from the United States and the United Kingdom and also consumed a significant fraction of U.S. and U.K. ship and equipment resources. Some scientists not involved in the experiment have occasionally claimed that this enormous focusing of effort was not entirely salutary— and not always when the individuals concerned were competing for the same resources. Origins of the Program As with many scientific ideas, the roots of MODE-1 cannot be traced precisely. Most participants would undoubtedly cite as the key progenitor the measurements made by J. Swallow and J. Crease during the Aries expedition of 1959-60 (see page 20). That work can, in turn, be traced to a suggestion published by H. Stommel in 1955, and even before that to the general ideas of the time. The Aries measurements are indeed an obvious predecessor of MODE-1 because they suggested that something fundamental might be wrong with ideas about how the circulation of the ocean worked. Theoretical models of ocean dynamics extant in the 1950s postulated a direct coupling between the ocean and the forces acting on it (large-scale winds and solar heating). The hypothesized response was thus a smooth, sluggish mean flow. An analogy can be made by supposing that in the atmosphere there were only climate- hot tropics, cold poles, large regions of desert, and so on, with a slow drift of air between them. There would be no weather in the commonly accepted sense, for example, no northeasters, hurricanes, and squalls. Of course, no one would leave out the weather in trying to realistically model the atmosphere, for by observation alone it is obviously a dominant feature. But in the ocean, only observations of climate were available, recorded in the atlases of the temperature and salinity and mean currents compiled so painstakingly over the years, beginning with the Challenger expedition (1872-76) and before. Thus when new technology made available direct measurements of deep-ocean currents, one could see for the first time that there might be something analogous to weather in the ocean. If tliis were indeed the case, then by analogy both to meteorology and to laboratory studies of turbulence, most details of theories of how the ocean worked might then be incorrect. For example, meteorologists believed for many years that "weather" simply represented the means by which the overall climatic circulation dissipated energy. That is, energy was put into the system by the overall heating at the equator and cooling at the poles, and storms in general represented a mechanism by which the system generated smaller scales on the road to ultimate frictional dissipation. But beginning in the early 1900s and culminating with observations made possible in the 1950s, meteorologists found that the system was much more complicated, and elegant, than they had thought. In many parts of the atmosphere, the smaller-scale storms (often called eddies) were feeding energy into the climatic circulation— driving it, instead of the reverse. Thus the notion of a climatic circulation that could be considered in any way independent of the storms or eddies also present was quite wrong. 45 80 40e 35C 30< 25< 20 WOODS HOLE •;$* P?s^ \ -- -..^ . K -. - *•„:- 'A.* *. BEAUFORT g3& *-'^ — $^ - • BERMUDA -^^j FLOAT LISTENING STATION MODE REGION MIAMI^ FLOAT LISTENING STATION FLOAT LISTENING STATION FLOAT LISTENING STATION W A," M o . 75 70C 65* 46 The possibility that this might also be the case in the ocean was obvious to most of those who thought about it. The importance of the question of how the general circulation of the ocean is controlled should not be underestimated. Apart from the purely intellectual challenge of knowing how this remarkably complex and interesting enormous fluid dynamical machine works, there is the practical aspect of understanding the effects of ocean currents on global climate, fisheries, and pollutants. Between 1960 and 1970, almost nothing specific was done about the problem. For, while the Aries measurements were the result of a new technology making available new measurements and changing ideas, there seemed no way to follow up all the disturbing possibilities suggested. The technology was not in fact adequate to deal with the problem. For as Swallow has pointed out many times, the vessel available could not keep up with the floats, and it appeared they would have to be followed for many months or more for one to learn much. Altogether the instruments seemed mismatched to the magnitude of the problem they suggested. During the 1960s oceanographers had occasional discussions about whether anything might be done, but the consensus was always that appropriate instrumentation was not available to operate in the deep sea for the very long periods of time that seemed necessary. However, technical developments were underway that kept provoking these intermittent discussions: for example, neutrally buoyant floats that could be tracked from land stations (rather than from ships) at 1000-kilometer ranges for up to a year; deep-sea moored buoys capable of staying in place for several months at a time; deep-sea pressure instruments that could function almost like barometers; and instruments capable of measuring the vertical profile of horizontal currents in a few hours (see page 59). Planning Stage In early 1970, scientists involved in the new instrumentation began to meet informally; a consensus emerged that, finally, something could and should be done. The physical oceanographic Location and topography of the MODE-1 area are indicated on a portion of the Heezen and Tlrarp physiographic diagra m of th e North A tlan tic. (Portion of the physiographic diagram of the North Atlantic Ocean, published by the Geological Society of America, Boulder, Colorado. Copyright 1968, Bruce C. Heezen and Marie Tliarp. Reproduced with permission.) community as a whole, including theorists, was brought into communication, and serious planning for the experiment began. Because of the developmental nature of many of the available instruments, as well as general ignorance about the phenomenon to be studied, the planning of MODE-1 went through several stages. The ignorance factor was in many ways the biggest problem. If one were looking for ocean weather for the first time, where should one go, what should one measure and for how long? How big were the "storms"? For this reason, MODE-1 was originally called Pre-MODE and envisioned as the pilot experiment for a proper mid-ocean dynamics experiment to come later (the name was later changed when the organizers were told that governmental funding would not be available for "pre-anything"-it had to be a "real" experiment). The new and not-as-yet thoroughly tested nature of much of the instrumentation led the initial planning to be based upon the cautious principle of children playing in a sandbox: nominally playing together, but each in fact building his own sandcastle. MODE-1 was to consist of individual experiments conducted in the same area; but because of fears of instrumentation failures, no one element was to become so crucial that its failure or loss would jeopardize the others. As it turned out, this quasi- independence was vital. In retrospect, the planning of MODE-1 was comparatively simple. The experiment had a high degree of inevitability about it— a consensus really did exist that a program of this sort was necessary and feasible at the time. The planning was made fairly simple for several reasons: the "mid-ocean" location (a 600-kilometer-wide area between Bermuda and Florida) was dictated largely by the need for proximity to U.S. East Coast ports, and the need to be within the tracking range of the large neutrally buoyant floats; the duration was constrained by the endurance of the instrumentation; and, most important, almost nothing was known about open-ocean variability. For this last reason, no matter what the detailed design of the experiment, whether or not mistakes were made, success was virtually assured. Almost any new measurements were bound to be enormously enlightening. Very early on, the scientists responsible for bringing together disparate investigators, instruments, institutions, and ships to collaborate in one small area at a single time, found the need for an administrative structure. Problems to be discussed ranged from the question of whether there was any relevant theory of the circulation of 47 the ocean, to such specific issues as making sure that acoustic signals from different investigators did not interfere with each other, and that the six ships and two aircraft involved could communicate with each other and with the scientists ashore. Consequently an elaborate committee structure evolved, as shown in a simplified schematic in Figure 1 , and nearly three years of meetings and discussions ensued. The sandbox principle was generally accepted, but of course everyone wanted to make the best use of the simultaneous presence of diverse instruments: to achieve redundancy and intercomparisons; to take advantage of anything the theorists could agree on; and, generally, to make the best possible experiment. Of necessity, steering groups emerged, as did a MODE-1 bureacracy as an inevitable adjunct. Most of the MODE-1 scientists were unfamiliar with this kind of organization in their professional lives, and many of them found the whole business of dealing with a centralized, highly planned experiment, increasingly traumatic. Traditionally, sea-going oceanographers have been an individualistic lot, with a strong tendency to be scientific-loners. A not untypical working life meant going to sea for a few weeks each year, with the scientist in complete charge of the use of the ship, and while at sea responsible in his science to no one but himself. Upon completion of the work at sea, several months would be devoted to working up the data for publication; then plans made for the next set of observations at sea, and the whole process would begin again. So, despite the fact that working at sea required a crew and scientific watchstanders, the scientist in charge was more or less captain of his own fate and able to do effective "small science." There can be few more tangible, and deeply satisfying, experimental scientific experiences than having nearly complete control of the movements of a large vessel and perhaps 50 supporting crew and scientists. With the advent of MODE-1 many oceanographers saw all this changing. Almost everything had to be negotiated with a scientific bureaucracy, endless meetings and negotiations had to be endured, and even the traditional autonomy of the chief scientist while at sea seemed to be disappearing— he was now answerable to a communications center in Bermuda and to a willful executive committee. In general, the specter of high-energy physics emerged-it looked as though the resulting scientific papers would need 25 co- authors. Big science really seemed to have arrived in oceanography, and what many people saw of it, they disliked. Much of the experimental nature of CC Co-Chairman EO Executive Officer P Principal Investigator IDOE The National Science Foundation's International Decade of Ocean Exploration ONR Office of Naval Research (together with IDOE, major source of funding of MODE-1 ) Figure 1. MODE-1 scientific management diagram. (From "MODE-1: The Program and the Plan," Appendix 2. MODE-1 Executive Office, 54-141 7, MIT, Cambridge, Massachusetts.) 48 Cameraman films the launching of a current meter from the fantail ofR/V Chain for a documentary on MODE-1 entitled The Turbulent Ocean. (Harold Armstrong) MODE-1 thus consisted of finding out if a lot of individualistic oceanographers could in fact bring themselves to work together. Another semisociological problem lay in the relation between theorists and sea-going scientists. Of course, there is and always has been an intimate and continuing interaction between theory and experiment in physical oceanography, but this has tended to exist between individuals, and then mostly through the reading of each other's papers. In MODE-1, theorists and experimentalists were trying to work directly together in planning a joint experiment on a very large scale. Theory was meant to advance hand-in-glove with new observations in a way that had not happened before. At Sea At the end of three years of planning, argument, crisis, and much sweat, the experiment went to sea east of Florida in early March 1973. Its final complement involved an international group of more than 50 oceanographers representing 15 institutions, several hundred scientific participants, 6 ships, and 2 aircraft. A list of all institutions and principal investigators and most of their projects is shown in Table 1 . The field experiment was very complex. Moorings and floats were being set, the aircraft were dropping profilers, the ships were all measuring the density field, and continual changes in strategy and tactics were being made as data came in and problems occurred. To tie together the disparate elements and to provide communications, a "hot line" was set up linking the ships and aircraft by radio to a "hot-line center" in Bermuda and by leased cable to five U.S. institutions on the mainland. Innumerable conference calls ensued in which scientists on several ships talked with the center chief scientist on Bermuda and, simultaneously, with members of the executive committee and others at the various U.S. institutions. If having control of a single ship can provide a sense of power, the notion of controlling a fleet is an even more tempting prospect. Not surprisingly, a number of would-be admirals emerged. But despite an occasional tremendous clash of wills, common sense usually prevailed. Participants generally followed the principle that the man-on- the-spot, the chief scientist on the ship at sea, is in the best position to judge his problems and possibilities. And most of the scientists at sea seemed happy to have the advice of those ashore on how best to use the ships for the overall good. At the end of the intensive four-month field period, a vast amount of new data emerged and the long process of making sense of it began. 49 Table 1. Projects, principal investigators and institutions in MODE-1 Moored current meter arrays 16 moorings with 4-8 current meters each 5 moorings with 4 current meters each 8 moorings with 1 or 2 current meters each Bottom-mounted instruments 2 IGPP capsules, 1 -month lifetime; 1 IGPP capsule, 1 -year lifetime (temperature, current, pressure bottom kilometer) 6 inverted echo-sounders 3 electric-field recorders and 3 bottom-mounted magnetometers 5 bottom-pressure recorders (fused silica bourdon type) Float tracking 20 long-range SOFAR-type floats using MILS listening stations 36 intermediate-range acoustic floats tracked by shipborne hydrophones Hydrophone arrays for locating SOFAR floats Density measurements Shipboard STD and CTD casts Moored thermal array 60 temperature-pressure recorders (on WHOI moorings) Towed instruments STD tows to map isopycnal surfaces N. Fofonoff, W. Schmitz, and F. Webster J. Swallow J. Knauss and W. St urges W. Munk, F. Snodgrass, and W. Brown H. T. Rossby C. Cox, V. Vacquier, J. Filloux, and R. Parker D. J. Baker, Jr. A. Voorhis, D. C. Webb, and H. T. Rossby J. Swallow R. Walden and H. Bertaux D. Hansen J. Crease A. Leetmaa R. Scarlet C. Wunsch E. Katz and R. Nowak Woods Hole Oceanographic Institution National Institute of Oceanography, England University of Rhode Island Institute of Geophysics and Planetary Physics, Univ. of Calif., San Diego Yale University Scripps Institution of Oceanography Harvard University Woods Hole Oceanographic Institution and Yale University National Institute of Oceanography, England Woods Hole Oceanographic Institution Atlantic Oceanographic and Meteorological Laboratory National Institute of Oceanography, England Atlantic Oceanographic and Meteorological Laboratory Massachusetts Institute of Technology Massachusetts Institute of Technology and Draper Laboratory Woods Hole Oceanographic Institution 50 Free-fall instruments Velocity profilers acoustically tracked by using bottom-mounted transponders Electric-field free-falling probe and bottom recorders Displacement-type current probe, airborne expendable (2000) Numerical modeling and theoretical studies Synoptic maps for MODE-1 Interactions between short internal gravity waves and larger- scale motions in the ocean MODE-1 array design as an inverse problem Theory and computer experiments on oceanic eddies and waves Analytic and numerical studies of mesoscale motions A theoretical-numerical study of geostrophic eddy motions in the oceans Administrative Funds for travel, executive officer, meetings, etc. 5 Filloux-type bottom- mounted tide gauges and 1 Hewlett Packard pressure gauge Monitoring earth's magnetic field at island stations Bottom-mounted vertical electric-field measurements Bottom-mounted magnetometers T. Pochapsky T. Sanford W. S. Richardson Columbia University Woods Hole Oceanographic Institution Nova University F. Bretherton K. Hasselmann M. Hendershott, R. Davis, and W. Munk P. Rhines A. R. Robinson P. Welander The Johns Hopkins University University of Hamburg Scripps Institution of Oceanography Woods Hole Oceanographic Institution Harvard University University of Gothenburg H. Stommel and D. Moore (on leave from Nova University) Massachusetts Institute of Technology Additional associated projects H. Mofjeld Atlantic Oceanographic and Meteorological Laboratory J. Larsen R. Harvey R. Von Herzen University of Hawaii, Hawaii Institute of Geophysics University of Hawaii, Hawaii Institute of Geophysics Woods Hole Oceanographic Institution Source: "MODE-1: Ttie Program and the Plan, "Appendix 1. MODE-1 Executive Office, 54-141 7 MIT, Cambridge, Massachusetts. 51 Many of the results are discussed in other articles in this issue. The Outcome As already mentioned, MODE-1 was in many ways failure-proof. Barring some catastrophe, there was no way that the experiment could have failed to increase, by orders of magnitude, our knowledge of open-ocean variability. In its overall goals MODE-1 was thus an overwhelming success: it was confirmed that an open-ocean eddy field ("weather") existed; individual eddies, at least in that area, had lifetimes exceeding many months; and the simplest notions about eddy behavior were confirmed. The fact that the eddies were found to be exceedingly energetic strengthened the hypothesis that they probably have some profound effects on the mean ocean circulation. As the MODE-1 data are studied in the years ahead, more surprises about the ocean are certain to emerge. To a considerable extent, there has been a revolution of ideas. The notion of a slow, sluggish general ocean circulation driven directly by the climatological average winds and heating is gone forever. Most older models of global circulation have been reduced to mathematical curiosities- interesting and useful as they were in their day, no one any longer believes that the ocean works like that. The idea that eddies in some way control the movement of the water is rapidly being assimilated into studies of oceanic chemistry, biology, and meteorology in a way not even thought of a few years ago. Even at this comparatively early stage, it does seem fair to say that MODE-1 was worth all the travail of getting it done. Where do we go from here? The Next Step As noted earlier, MODE-1 was originally called Pre-MODE, the predecessor to a proper mid-ocean dynamics experiment, and it was not envisioned as in any sense completing the study of the role of eddies in ocean circulation. Even before the first results of MODE-1 had been completely assimilated, planning began for the next step. MODE-1 made very clear the magnitude of the problem now at hand. To observe anything like the full-time evolution of an eddy in the MODE-1 area means many months of intense observations in that region. But is the MODE-1 area in any way typical of the rest of the oceans? If one understood the MODE-1 area completely-and we are a long way from that— would that mean one understood the physics of eddies in all other parts of the ocean? It seems unlikely. Understanding the ultimate question of MODE-1 -the role that eddies play in the general circulation of the ocean— will require many years of measurement in many different parts of the oceans. The magnitude of the task is great, so great in fact that it seems too much for the resources of any one or two countries. The next phase seems to demand the resources of much of the entire international community of physical, and to some extent chemical, oceanographers. Recognizing this fact, the U.S., U.S.S.R., U.K., Canada, France, and West Germany are taking the next experimental and theoretical steps under the rubric POLYMODE (from the Soviet POLYGON and the U.S. MODE). Paradoxically, the planning for POLYMODE is much more difficult than it was for MODE-1. In the face of nearly total ignorance prior to MODE-1, the outline of that experiment was almost inevitable. But now we know a great deal, and that knowledge implies many second steps that could be taken, any one of which would provide significant data on some aspect of the eddy problem, but each of which would tend to use up so many of the limited resources that it would preclude immediate progress in other, equally promising directions. Thus, considerable discussion and heated argument have gone on since the end of MODE-1 concerning the choice among many appealing options. To the outside oceanographic community it has sometimes seemed as though the POLYMODE oceanographers cannot agree on anything and do not, in fact, have a program. But the problem arises through legitimate and productive disagreement about an abundance of scientific possibilities. Indeed, out of the wrangling there has emerged considerable agreement about POLYMODE, and the experiments planned for the next three years are quite definite. It is now generally agreed (there are still some dissenters) that the most pressing problem is finding out whether eddies are basically of one type, or whether there is an eddy "zoo." If one looks in a different part of the ocean, does one see motions like those of MODE-1, or something completely different? Until that question is answered, it will be difficult for theorists to come up with models of how the ocean works. 52 There are already hints. Gulf Stream rings (see pages 65 and 69 ) are a form of eddy. They look quite unlike the MODE-1 eddy, and there is some evidence that their physics may be at least somewhat different. We already know that as one moves out of the MODE-1 area, the energy of the eddy field changes markedly. But what does this change mean? To pursue our meteorological analogy a little further, MODE-1 may be thought of as a weather-observing program that managed to observe a three-day New England northeaster for a little over two days, and that this was the only storm ever really seen. One would then have a long string of questions: Do northeasters occur in New England all the time? If not, how often? What happens in between? Are these storms typical of weather in all parts of the vorld? During MODE-1 there was an oceanic "storm." Are there analogues to hurricanes— with all their climatic implications— in the MODE-1 area and elsewhere? How long would we have to measure to see one? Or are Gulf Stream rings an analogue to hurricanes? We do not know the answers to any of these questions. Thus the POLYMODE program is taking the first steps toward global answers. Because of the long time scales in the ocean, the difficulties of measurement, and the logistics, these answers will be many years in coming. If the true measure of the success of an experiment is whether it generates many more questions than it answers, MODE-1 was an overwhelming scientific success. The sociological, organizational aspect of MODE-1 may also be judged a success— it got the job done. On the other hand, working within a highly organized program did not appeal to all of the participants, many of whom flinch at the prospect of POLYMODE, which, because of its wider international participation, tends to be an even greater tax on time and patience for bureaucratic purposes. Many of the MODE-1 scientists have gone back to their old ways, back to "small science." Fortunately, oceanography still has many areas where "small science" continues to be the best way to do things. And there is no denying that it is much more fun. Carl Wunsch is a professor of physical oceanography in the Department of Earth and Planetary Sciences, Massachusetts Institute of Technology, Cambridge. r Tfie MODE-1 central mooring is prepared for launch to serve as a navigation reference and to gather weather data. One night during the experiment, the large buoy was "recovered" by an unknown ship and was never seen again. (Susan Tarbell) 53 Instrumentation for MODE-1 Recovery of a MODE-1 mooring by R/V Chain. Each of the units shown here is a glass sphere in a protective plastic case ("hard hat "). Each unit gives about 25 kilograms of buoyancy. (Susan Tarbell) Almost all of our knowledge of mid-ocean eddies has been acquired during the present decade. This may seem surprising, but the reason for the recent upsurge can be directly related to our ability to make certain kinds of measurements in the deep ocean. Eddies are dynamic features, and although they may be detected and identified by detailed temperature maps, a full understanding requires that dynamic measurements be made. We must be able to determine the velocity of water in the deep ocean for periods long enough, and over an area big enough to be able to "see" an eddy, to measure its energy and velocity structure. The ability to make such measurements has evolved over the past fifteen to twenty years, during what has seemed a long, slow, and often frustrating process. Our present skills can be put into perspective by considering the development of two important instruments in deep- sea measurement— the neutrally buoyant float and the current meter. In the mid-1950s our knowledge of deep- ocean currents, from direct measurements, was sparse. K. F. Bowden in 1954 could summarize by W John Gould all the previous measurements in a single paper in the journal Deep-Sea Research, and the total duration of all those records was counted in hours rather than days. Bowden's paper awakened the oceanographic community to the lack of direct information about ocean currents. In an attempt to alleviate the problem, H. Stommel, in a letter to the same journal, suggested the development of a new technique to measure ocean currents. During World War II M. Ewing found that there was an optimal depth for the transmission of sound horizontally over long distances underwater. Stommel therefore suggested that this SOFAR (SOund Fixing And Ranging) channel could be used to fix the positions of freely drifting neutrally buoyant floats by receiving their signals at listening stations on islands in the Atlantic. The first attempts to make neutrally buoyant floats for tracking deep currents were not so ambitious as Stommel's SOFAR float idea. J. Swallow at the British National Institute of Oceanography (now Institute of Oceanographic Sciences) set about developing a neutrally buoyant float that could be followed from an attendant ship. 54 The body of the float, which had to withstand the pressure at which it was to work, be sufficiently buoyant to carry all the necessary batteries and electronics, and be less compressible than seawater, was made of a pair of 3-meter-long aluminum scaffolding tubes— one to provide buoyancy, the other to hold the batteries and electronics. The floats were designed to be used to a depth of 4500 meters and for a lifetime of about 3 days, with acoustic signals transmitted every few seconds. In 1955 Swallow reported two float tracks of 2 and 3 days at 900 and 400 meters from an area west of Portugal. The measured speeds were low (5.7 and 2.4 centimeters per second) and thus were not inconsistent with the previously held ideas about the slow, steady movement of deep-ocean currents (see page 28). The first major experiment with neutrally buoyant floats was the attempt by Swallow and J. Crease, working from the Aries, to make long- term measurements of deep currents southwest of Bermuda. The measurements were made between June 1959 and August 1960 and revealed that deep currents in that part of the open ocean had energies much higher than originally thought and that much of the energy was contained in features with apparent wavelengths of typically 100-200 kilometers. This work, in addition to providing the first look at deep- ocean variability, showed that if the deep ocean in general was as variable as the measurements suggested, then the difficulties involved in the prolonged use of the existing tracking technique for neutrally buoyant floats (chasing each one by ship) would be great and that a larger-scale experiment would be prohibitively expensive in ship time (see page 20). The current speeds recorded during Swallow's early measurements stimulated a parallel line of attack. It was considered that such currents might be fast enough to be measured by recording instruments (current meters) attached to moored buoys in the deep ocean. Toward this end, W. Richardson, then at the Woods Hole Oceanographic Institution, began to develop the appropriate instruments and mooring techniques. The current meters used a vane to read the direction of current flow relative to an internal magnetic compass and a rotor made from "back-to-back" half cylinders to sense current speed. A test mooring was set near Bermuda in December 1960 that consisted of a toroidal surface float connected by polypropylene line to an anchor on the sea bed, with the recording current meters inserted between the lengths of rope. It was successfully recovered in late February 1961, having survived winter conditions for 79 days, and the design was quickly adopted. In early May 1961 four such moorings were set on the continental slope and shelf south of Cape Cod, but by the end of the month only one was completely recovered. At this point and in spite of the losses, the most ambitious part of the project was begun: the maintenance of a line of 13 moorings with 62 current meters between Cape Cod and Bermuda. Attempts to maintain this buoy line throughout the year that followed revealed that there were a multitude of hitherto unknown difficulties with both the mooring design and the instrumentation. Many moorings were lost, some totally; others were recovered adrift from their proper positions after mooring lines had parted. The current meters, which recorded their data on 16mm film, posed problems that were attributable to the dynamic response of the instruments in the mooring line; oscillations and rotations that could not be adequately resolved by the instrument made many records either difficult or impossible to read. Although solutions were found for most of the problems as they arose, a decision was made in 1962 to abandon the "Bermuda line," and in 1963 there was a major change in emphasis. Efforts were directed toward engineering studies of the mooring and current meter problems. Among many other developments, these studies led to the use of subsurface moorings that were unaffected by weather at the sea surface and thus stood a better chance of surviving winter storms. Current meters were redesigned to record their data on magnetic tape instead of film, greatly simplifying the subsequent data analysis. As problems were One of the instrumented SOFAR floats. At the bottom of the float are the two low-frequency sound projectors. Angled fins around the float body convert vertical motions relative to the water to rotations. (See SOFAR float tracks, pages 14 and 27 .) 55 solved, the accumulated knowledge of deep-ocean currents increased, and by the late 1960s the deployment and recovery of moored current meters had become a relatively routine operation. The first attempt to study mesoscale eddy features over a large area was the Soviet POLYGON experiment of 1970 (see page 40). It required a great deal of effort by the ships and people involved to maintain the large array of 17 moorings for the 8-month period, each buoy having to be replaced at 1 -month intervals. However, the techniques used were well tried and thus did not involve a high risk of instrument failure. MODE-1 Instrumentation and Design From the outset the design philosophy of MODE-1 (Mid-Ocean Dynamics Experiment) differed from that of POLYGON in the types and complexity of instruments used. While some of the measuring techniques were well tried, others were entirely new. In many cases MODE-1 was a testing ground for new instruments. The main objective of the program was to map both the velocity and the density field over the MODE-1 region, although additional experiments were incorporated to provide other indirect recordings of the influence of mesoscale eddies. Primary measurements included the mapping of currents from moored buoys, neutrally buoyant floats, and vertical profiles, and the measuring of densities from CTD (conductivity/temperature/ depth) and STD (salinity/temperature/depth) profiles, XBT (expendable bathythermograph) profiles, and a towed CTD. Moored Arrays The mooring techniques used to map the horizontal velocity field were all directly related to those developed over the previous decade. The sixteen instrumented moorings all had subsurface buoyancy. In the planning stages it was thought that a combination of subsurface and surface moorings could be used, with the surface buoys serving as platforms from which meteorological measurements could be made. It was found at a relatively late stage that the measured values of currents were dependent on the type of mooring used (the vertical motions of the surface float are transferred down the mooring line and can have a large effect on the current meters, particularly in the deep water where velocities are low). The current meters used by Woods Hole Oceanographic Institution were of the then relatively new vector averaging type (VACMs). These had been developed at Woods Hole by R. Koehler and J. McCullough with a view to improving the performance of existing current meters in situations where there might be a large amount of high- frequency energy (such as on surface moorings). Even though all the MODE-1 moorings were subsurface, and not subject to such conditions, it was thought that the much simpler data processing made possible by VACM records would provide speedier dissemination of information among the MODE-1 scientists. With some 83 records to be processed, this was an important factor. In fact the most sophisticated and new parts of the VACM gave little trouble. These were the internal computer, which calculated the east and north components of velocity from the speeds and directions measured by the rotor and vane, and all the associated logic circuitry. The main failing of the VACM in that experiment came from a totally unexpected quarter. As the records were decoded it was found that the response of the rotors and vanes became progressively more sluggish as the experiment progressed. The problem was later traced to a buildup of carbonate compounds in the bearings. The stagnant water in the bearings had acted as an electrolytic cell driven by leakage currents through the sacrificial anode on the pressure case. Not all The recovery of a vector averaging current meter (VACM) during MODE-1. (IOS) 56 the VACMs suffered from this problem and since other, older types of current meter were also used in the experiment the result was not disastrous. Neutrally Buoyant Floats The other main method of current measurement used in MODE-1 employed neutrally buoyant floats that were tracked via the SOFAR axis by a technique not too far removed from Stommel's original concept. The floats were tracked from land- based listening stations via low frequencies (267-273 hertz at a nominal 1 -minute repetition rate and with a pulse length of 1 .667 seconds). The floats could also be located by surface ships via a separate high-frequency (10 kilohertz) system, which included an acoustic command recovery capability to release external ballast and return the float to the surface. The low-frequency shore-based tracking system obtained acoustic ranges to the floats from a maximum of four listening stations in Bermuda, Eleuthera, Puerto Rico, and Grand Turk Island. This was done by measuring the difference between the time of transmission at the float and that of the arrival at a pair of stations. The transmission signal was controlled by a temperature-stabilized quartz clock in each float having an accuracy of better than 1 part per million (approximately 1 second in 12 days). The float signals were distinguished from each other by having a possibility of 3 transmission frequencies and 7 repetition rates, thus giving 21 separate channels. Half of the twenty floats used in MODE-1 carried additional instrumentation to monitor vertical water movement, water pressure, and temperature. The vertical motions of the instrumented floats relative to the water were converted to a rotation by a set of angled fins around the circumference of the float, and the number of rotations, together with the temperature and pressure data, were recorded on a digital magnetic tape recorder every 256 seconds. All the floats were ballasted to stabilize at 1500 meters, and all of those instrumented showed initial depths of 1 525124 meters. However, the records indicated that the floats sank by 1 5 meters during their first day after launch and that the sinking rate stabilized at 0.85 meter per day after about 2 weeks. This effect has been traced to the slow compression of the aluminium float bodies under the sustained high pressure. This sinking had no serious effect on the value of the data. A rather more serious, though temporary, fault appeared early in the experiment. Several of the floats stopped transmitting their low-frequency signals after less than two weeks in the water. These floats were tracked and recovered via their high- frequency transmitters and were found to have a failure of the low-frequency sound projectors. Fortunately in almost all cases only one of the two projectors in each float had failed and then shorted the remaining good projector. Temporary repairs were made at sea and the floats relaunched; meanwhile a new projector seal was designed and was installed in all the floats as early as possible. The shore-based positioning of the floats was found to have a random route mean square error of about 500 meters (about half the design specification). At the end of the main MODE-1 field experiment (July 1973) all the floats were recovered, their batteries were recharged, and they were relaunched for an indefinite drift (the repetition rate was cut by one-third to prolong battery life). Many of these floats were still being tracked late in 1975. Swallow Floats The SOFAR floats were restricted to use near the sound axis at 1 500 meters, and thus no comparisons between Lagrangian (drifting) and Eulerian (moored) current measurements would have been possible at other depths. Lagrangian current measurements were made at scales smaller than the typical mooring separation of the main array. These were the sort of scales at which the ship- tracked neutrally buoyant floats developed by Swallow could be used. However, it was necessary to track about 4 or 5 floats simultaneously at each of the 4 standard current meter depths (500, 1500, 3000, and 4000 meters), and this would have been totally beyond the capacity of any system used by the National Institute of Oceanography (NIO) up to the time of MODE-1 planning. So a totally new system was designed. The floats, rather than transmitting sound continuously, would act as transponders (transmitting only when they received an interrogation pulse from the ship). They would be recoverable so that they could be reused and would be able to be tracked even when the ship was engaged in other work such as making STD or CTD observations. The technique for fixing a group of floats would be to transmit a signal from an interrogator attached beneath the CTD probe. All the floats within listening range would reply, and from the time difference between the outgoing and return pulses the horizontal range to each float could be calculated. Thus from the ship position each float would be known to lie on the arc of a circle. Floats that could not, due to the ray paths, be detected from one level might be contacted by 57 t 489 m Station 488 3/8" Dacron - 3/8" Dacron - One of the MODE-1 WHOI current meter moorings (not to scale). o 6 58 radio float with light 2 m 1/2" chain 2 m 3/8" chain 12 17" glass balls in hard hats on 1 2 m 3/8" chain VACM 488J 2 m 3/8" chain 96 m 3/16" wire 3 m 3/8" chain temp/depth recorder 4882 196 m 3/16" wire VACM 4883 2 m 3/8" chain 198 m 3/16" wire 1 m 3/8" chain 199 m 3/16" wire I m 3/8" chain 280 m 3/16" wire 8 1 7" glass balls in hard hats on 1 5 m 3/8" chain VACM 4884 500 m 3/16" wire 458m 458m 35m 5 1 7" glass balls in hard hats on 5 m 3/8" chain VACM 4885 455 m 456 m 34 m II m temp/depth recorder 4886 457 m 458 m 260m current meter 4557 42 m 13 m 13 17" glass balls in hard hats on 13 m 3/8" chain acoustic release, transponding 20 m 3/4" nylon 3 m 1/2" chain Stimson anchor, 2400 Ibs N T 200 km 200 km Moored current meter (left) and density survey (right) arrays for MODE-1. The inner 100-kilometer circle represents an "accurate mapping" region, the outer circle a "pattern recognition" area. transmitting from a shallower or deeper level. The ship would then go to another position, usually chosen to give a good "cut" between the two position circles. The ranges from the new position would be drawn and the intersection point would give the float position. The floats were each identified by a particular reply frequency. The method proved to be extremely successful and particularly efficient in ship time. An indication of this is that during the 40 days spent tracking floats in MODE-1, 714 float-days of track were accumulated; this is to be compared with the total amount of float tracking since the first float was launched by NIO and including all the Aries measurements of 867 float-days. Some floats were lost (1 1 out of a total of 52 launches) due to faulty manufacture of release units; none was lost through inability to track the floats or to locate them once they had surfaced. Vertical Profilers All of the foregoing velocity measurements were restricted to discrete, predetermined depth levels. The more detailed vertical structure of horizontal velocities in the MODE-1 region was revealed by measurements from two types of profiling instruments. The first, T. Sanford's electromagnetic profiler, sensed the very weak electric currents induced in the seawater by its motion through the earth's magnetic field. It does not measure the absolute current velocity but rather the departure from some arbitrary offset that is dependent on both the true mean value of the current profile and a zero offset associated with the instrument. The profiler records the data internally as it free falls to the sea bed (the total time for a profile to 6000 meters is about 1V£ hours), together with measurements of ambient pressure, seawater temperature, and electrical conductivity. The instrument resolution gives velocities to an accuracy of 1 centimeter per second at a vertical resolution of 10 meters. A somewhat simpler profiler that reveals absolute rather than relative profiles was used by T. Pochapsky. Here a slowly sinking float transmits to two transponders fixed on the sea floor. All the information from the floats is transmitted acoustically and received by an overside transducer on the attendant ship from which the float position may be fixed at regular intervals. Both these techniques were expensive in terms of ship time required and thus were used for 59 only limited periods and in a few selected positions in the MODE-1 region. An attempt to obtain vertical profiles over a wider region and throughout the period of the experiment was made using a rather simple air-dropped profiler. The system could be used to reveal the vertically averaged velocity from sea surface to the bottom or to any intermediate depth. A dropping aircraft ran along a predetermined line, launching the probes at intervals of about 4 kilometers. As the instrument reached the surface it produced a dye patch that drifted with the surface current. The profilers for intermediate and bottom depths were then released and sank at a known rate until a ballast was released by an internal clock. On return to the surface the profile probe released a second dye patch. A second aircraft then flew along the track photographing the dye patches. From the relative locations of the dye patches, one could determine the surface currents together with the averaged currents down to the depths of the profiles. About 1000 usable photographs were obtained from 1370 drops and gave a total of 960 measurements of either surface current or integrated flow over same depth range. Malfunctions of the timing clocks and the inability (Top) San ford's electromagnetic current profiler. (Bottom) Some vertical profiles of horizontal currents as revealed by San ford's electromagnetic profiler. to -q •*> -» o a o o a; •» -n -o -» o' a c n to DROPS I94D 8 I960. NORTH ENO OF ROGE CROPS 1940 8 I9SO. NORTH OC V RCGC EAST, ( DROPS 1910 a I93D.9 NM SW OF RIOGE PEAK WOPS 1910 a I93O.9 NM SW OF RIDGE PEAK CAST,KM/S) NOfTTH, (CM/SI to-a-o-a o » oaao •?O -p -10 -5 0 5 0047 1 IT x an -IUO ••-MOW W42 if X IVT9 PROFILES OVER SMOOTH TOPOGRAPHY DROPS 1950 a I97D.BF. T1*EEN RCGES DROPS 195D a 197D. BETWEEN RIDGES EAST,(an/st farrnl (au/st •a> -a -o -5 o > 10 B x -20 -a -a -a o t a o to PROFILES OVER ROUGH TOPOGRAPHY 60 of the photographing plane to find the dye patches proved to be the major sources of data loss. STD and CTD Measurements The study of the density field in MODE-1 was accomplished for the most part by the lowering of STD or CTD probes on conducting wires from the ships. These instruments provide continuous profiles as a function of pressure. Instruments such as these have been in general use since the late 1960s; in MODE-1 the major problem in using them was to ensure that even with different instruments on a variety of ships, each with slightly different operating procedures, a uniform data set could be achieved while still retaining a useful data accuracy. The basic calibration technique is one not far removed from the classical water-bottle techniques. Samples of seawater were taken at a variety of points on each vertical CTD/STD profile, together with measurements of temperature and pressure determined by reversing thermometers. All the reversing thermometers had laboratory calibrations against some standard (typically quartz) thermometer. The conductivities of the water samples were measured on shipboard using laboratory salinometers. Two samples were drawn from each sample bottle, and about 25 percent of these were then interchanged with other ships so that the laboratory salinometers could be intercompared. The coordination and intercomparison of the density data, since it involved so many people, proved to be a long process, but a final data set emerged that was sufficiently accurate to enable the density field to be mapped. The CTD/STD stations were worked on a fixed grid of 77 stations at a typical spacing of 33 or 50 kilometers. The combined efforts of all the ships resulted in a complete survey of the density grid approximately once every 2 weeks, giving an adequate but not ideal amount of data from which to map the evolution of the density field. A large amount of ship time was required by this survey, each profile to 3000 meters and back to the surface taking between 2 and 3 hours. Some STD/CTD profiles were made to within a few meters of the sea bed; not all ships had long enough conducting cables to be able to do this, and on these vessels a separate water-bottle cast was worked to cover the lowest 2000 meters or so of the water column. The data from the density grid survey were capable of later detailed analysis to reveal the dynamic implications of that field. The two-week interval between the surveys, together with the fact that each survey took two weeks, meant that the A CTD (conductivity /temperature /depth) profiler. The sensors are within the protective cage around the lower circumference of the instrument. Above them is the pressure case containing associated electronics. Clustered sampling bottles with attached thermometer frames are for calibration purposes, while the cylinder with the downward- pointing mushroom-shaped transducer is an acoustic device to tell when the package is approaching the sea bed. (IOS) details of sudden changes and small horizontal scales (smaller than the density grid spacing) could not be observed with these data. There were, however, other measurements that, although not directly of the density field, enabled these more detailed structures to be studied. Temperature/Pressure Recorders Most of the current meters, and certainly all of the VACMs, recorded water temperature every 15 minutes throughout the experiment. Each mooring had additional temperature and pressure recorders at a variety of depths throughout the water column. This resulted in a detailed series of temperature records with, in the case of the central mooring of the array, a total of 1 1 temperature time-series throughout the water column. The importance of the pressure record 61 is that it allows temperature recordings to be corrected for mooring motion. Even moorings with subsurface buoyancy are not entirely stable. As the strength of the current profile increases, so does the drag on the mooring; it leans over (just as a tree does in a wind) with the result that a point on the mooring line does not stay at a fixed level but moves up and down through the water column. To a first approximation the density structure of the water column can be derived from temperature alone (since in any particular ocean area temperature and salinity have a quite well defined relationship to one another), and so these moored temperature records could be related to the CTD/STD measurements and could be used to supplement the density mapping. XBTs Although the moored T/P recorders could supplement the time evolution of the density field, the detailed horizontal structure on scales smaller than the density grid spacing had to be studied in other ways. Some ships were equipped with expendable bathythermograph (XBT) instruments that produce a profile of temperature as a function of depth in the uppermost 750 meters of the water column. The probes can be used from the ship as it steams at full speed, and although only temperature is measured-and not with the great accuracy of the CTD/STD systems-a detailed survey of a small area may be accomplished in a short space of time. Towed STD Just as the STD was used to make vertical profiles of the density structure, so it was also used on small horizontal scales to map the depth variations of a density surface. This was accomplished by towing a pair of underwater vehicles separated from one another by 10 or 20 meters, the upper one measuring pressure and temperature, the lower measuring pressure, temperature, and conductivity. By hauling in and paying out on the towing cable and by controlling the angle of a wing on the upper fish, the two towed sensors are controlled so that they straddle an isotherm. Four separate tows were made from the R/V Chain by E. Katz covering a total of 12 days. Each tow was worked on a regular closed pattern so that the depth of the density surface could be mapped over horizontal scales of up to 100 kilometers. Indirect Measurements In addition to the moored and shipboard instrumentation in MODE-1, there was a final category of measurements involving instrumentation that lay on the sea bed and recorded what might be regarded as "integrated" effects of mesoscale eddies. As the structure of the overlying water column changes, so does its mean density and hence the pressure at the sea bed. Because the variations are small and the total pressure signal large, the bottom-mounted pressure instrumentation had to be designed to detect and record variations in the pressure equivalent to 1 or 2 centimeters of water in the presence of an ambient pressure head of 5 kilometers (about 1 part in 105 or 106). Three types of pressure recorders were used in MODE-1 ; the experiment was designed in part as an intercomparison with all three types near to one another at the central mooring and partly as a scientific study of abyssal pressure fluctuations. W. Munk of the Institute of Geophysics and Planetary Physics (IGPP) at the University of California, San Diego, used three quartz-crystal A Harvard bottom pressure gauge. 62 pressure sensors; H. Mofjeld of the Atlantic Oceanographic and Meteorological Laboratory (AOML), three bourdon tube/optical lever sensors; and D. J. Baker, Jr., of Harvard University, five quartz bourdon tube/optical lever instruments. Of the three, the IGPP capsule is the only absolute pressure gauge. The pressure sensor is a quartz crystal whose resonant frequency is a function of pressure. The aim with the IGPP instruments was to see if pressure signals (excepting those due to tides) could be detected above the noise level of the instruments and if the difference between two such measurements separated by 200 kilometers was resolvable. All the types of pressure sensors used in MODE-1 are affected by even the small temperature fluctuations found in the water of the abyssal ocean; therefore, each instrument measured temperature as well as pressure so that the pressure values could later be corrected. In the case of the IGPP instrument, the temperature sensor was also a quartz crystal, but it was cut in such a way as to respond to temperature changes and to be insensitive to pressure. In order to obtain the required redundancy, two of the three IGPP capsules were set close to the central mooring; the third, which was to detect horizontal pressure gradients, had dual sensors. Both the AOML and Harvard recorders were similar in that each employed a mechanical pressure sensor that worked in a differential mode by measuring the departure from the pressure at the start of the record. Another similarity between the two types was that each used an optical lever arrangement to magnify the small pressure-generated mechanical distortions of the bourdon tube. Both the AOML and IGPP instruments had been used extensively before MODE-1, and each gave a high return of data. The Harvard instruments were new, and although there had been previous test deployments from which good data had been retrieved, all the pressure-measuring systems in the MODE-1 field experiment failed due to a hitherto undetected fault in the servo mechanism used to track the movements of the optical lever. The bottom-pressure experiment in MODE-1 did provide both useful insight into the performance of such instruments via the intercomparison at the central mooring and a new view of the long-term fluctuations of sea-bed pressure and their variability over distances of a few hundred kilometers. Inverted Echo-Sounders A novel new instrument, the inverted echo-sounder (IES), was used for the first time in MODE-1 in an attempt to study the temperature structure of the water column from an instrument on the sea bed. The principle on which this device works is as follows. The time for a pulse of sound to travel from the sea bed to the surface and back again is dependent on the total distance traveled (basically a constant apart from fluctuations in sea level due to the tides) and also on the mean value of the sound velocity in the water column. Thus a time series of the travel times may be corrected to represent the variations in the mean sound velocity. By regarding the water column, for the sake of simplicity, to be made up of an upper warm layer and a deep cold layer separated by a region of high temperature gradient, the thermocline, it can be seen that if the thermocline is displaced upwards, the relative amount of cold water in the column increases and the acoustic travel time increases. Nine inverted echo-sounders were set in MODE-1 , which, in spite of their newness, produced over 50 percent usable data. Variations in travel time were measured with an accuracy equivalent to a displacement of the 10°C isotherm of t4 meters, and comparisons with the data both from the moored temperature/pressure recorders and from CTD/STD stations later showed that an inverted echo-sounder was indeed capable of producing records that gave a good representation of the long period displacements of the main thermocline. The uncertainties in the IES data are of the same magnitude as the errors inherent in the CTD/STD data; for the study of integrated quantities over the entire water column the IES therefore represents a significant advance in being able to provide at least part of the data that would be collected from a CTD/STD survey but without the prolonged use of ships. Shipboard Computers In recent years shipboard computers have helped the oceanographer considerably in the collection, reduction, and analysis of his data while still at sea. Naturally, in MODE-1 computers played an important role. In all cases ships with CTD instruments recorded their data via a computer interface, and through this several derived quantities were available to the scientists while they were still at sea. Some of the ships employed computers in the routine navigation. Summary If we look beyond MODE-1 to future mid-ocean dynamics experiments, we shall perhaps see a change of emphasis. MODE-1 and POLYGON each studied one mesoscale feature for sufficient time to be able 63 to observe its propagation through a fixed moored array. We might expect future ocean dynamics experiments to involve the study of several features and over much longer time scales. It is not possible to contemplate such experiments relying on the continual availability of research ships. The demand for such ships for all kinds of research is so great, and the requirements for large numbers of scientists to remain at sea for long periods is unattractive. It is inevitable therefore that progress will be made towards the greater, and perhaps almost exclusive, use of remote recording techniques coupled with measurements that can be made from nonspecialized "ships of opportunity." It has already been demonstrated that the horizontal velocity field can be mapped well over a large horizontal area by the SOFAR floats. The development of moored tracking stations would mean that such a system could be used anywhere in the ocean. Recent studies have shown that the restriction of placing the floats in the sound axis can be relaxed to allow floats to run at depths between 700 and 2000 meters. The inverted echo-sounders give the prospect of being able to map the gross features of the density field from an extensive array of instruments that, being on the sea bed, are free from disturbance by bad weather. Moorings will continue to play their part by carrying temperature and velocity recorders, but we shall perhaps see the introduction of current meters without the rotors and vanes that are vulnerable to damage and that have a threshold velocity often as great as the currents that the scientist is interested in observing. Perhaps in MODE-1 we had not only an experiment that used the most advanced equipment available but also one in which the forerunners of a new generation of oceanographic instrumentation proved their worth. W. John Gould is a principal scientific officer at the U.K. Institute of Oceanographic Sciences. Deployment of an inverted echo-sounder (IES). Buoyancy is provided by two glass spheres in "hard hats. " 64 by Philip Richardson The Gulf Stream is the swiftest and most energetic current in the North Atlantic. One of the most interesting features of the Gulf Stream is the horizontal wave motions, or meanders, of its path; frequently these become sufficiently large to pinch off from the main current and form large eddies. The formation process is analogous to the cut-off highs and lows formed by the atmospheric jet stream. Gulf Stream eddies— their formation, movement, and decay— are vital in redistributing water, biological life (see page 69), and momentum and energy in the Gulf Stream system and western North Atlantic. F. C. Fuglister has suggested the name Gulf Stream rings for these eddies because, during their formation, segments of the Gulf Stream form closed rings. Those formed to the south of the Gulf Stream have cyclonic (counterclockwise) circulation and contain cold Slope Water in their centers. Rings formed to the north are anticyclonic and contain warm Sargasso Sea water. Although we have known about the existence of rings for almost 40 years, it is only during the last several years that we have learned about their distribution, long-term movement, and decay. Recent satellite infrared measurements, as well as ship and aircraft surveys, indicate that rings can be found in greater numbers in the western North Atlantic than we thought. Advances in satellite measurement techniques have made it possible to view the Gulf Stream system from space over a wide region during a short period of time (less than a day). One of the best satellite infrared photographs is Figure 1 , taken just off the U.S. East Coast, showing the Gulf Stream, two anticyclonic rings, and two cyclonic rings. The evidence suggests that the two anticyclonic rings that were first observed in January 1974 north of the Gulf Stream, moved west and south to coalesce with the Gulf Stream during the summer of 1974; the cyclonic rings south of the Gulf Stream both moved southwest, as other rings have been observed to do, and may have coalesced with the Gulf Stream off Florida during mid-1975; the Gulf Stream meander formed a cyclonic ring in June 1974, which was reabsorbed three months later near the place of its formation. Using successive satellite photographs and additional measurements, investigators at several oceanographic laboratories (including the National Oceanographic and Atmospheric Administration, the Naval Oceanographic Office, Research Triangle Institute, Texas A&M University, the University of Rhode Island, and the Woods Hole Oceanographic Institution) are trying to follow the evolution of these and other rings. Distribution Figure 2 is a synoptic representation of the Gulf Stream and rings at an instant in time. Since there has been no large-scale survey or good satellite coverage of the entire region, the figure is, for the most part, contrived, based on a wide assortment of data including recent satellite, aircraft, and ship surveys of portions of the area. North of the Gulf Stream we frequently see three rings at a time; south of the Gulf Stream and west of 50°W we can find approximately 8-14 rings. Some data suggest that rings can be found east of 50°W, but no thorough investigation has been made in this area. Only limited data exist for the region between 50°W and 60°W. Although rings have been observed to form only between 60°W and 70°W, their location east of 60°W and their westward movement indicate that they must also form east of60°W. The early stage of cyclonic ring formation is shown in Figure 2 near 38°N and 58° W where a large meander has trapped Slope Water in its "pocket"; the sides of the meander are closing, and the ring will separate from the Gulf Stream as shown 65 •'• f - by the dotted lines. The southern boundary of the ring region is not well known. There is a lack of long-term ring tracking and detailed hydrographic surveys, and the decay of rings makes them more difficult to find when they are old. In any case, no rings have been documented south of about 30°N except for the extreme western region. Just north of the Bahamas several rings were seen very close to the Gulf Stream, apparently coalescing with it. The initial sizes of rings differ; those north of the Gulf Stream are generally smaller than those to the south. The outer limits of rings are as difficult to determine as those of the Gulf Stream itself. In Figure 2 the Gulf Stream is shown as a band almost 100 kilometers wide, and the rings are from 150 to 300 kilometers in diameter. These limits represent, approximately, the locations where the main thermocline (the transition zone between warm surface water and cold, deep water) becomes horizontal and the current vanishes. The usual shape of rings is nearly circular, although significant variations are often found. There are limited but tempting data suggesting that rings may merge as well as break up into smaller pieces. Neither of these processes has been observed in detail, but they offer the simplest explanations of some observations. Movement Evidence for the movement of rings comes from the real-time tracking of a few rings by several techniques and from inferred trajectories based on an analysis of the National Oceanographic Data Center files of XBT (expendable bathythermographs) and hydrographic data. Rings north of the Gulf Stream move generally toward Cape Hatteras, with average speeds of 3-7 kilometers per day, where they have been observed to coalesce with the Gulf Stream. There is little variation from this mean movement because the rings are confined by the continental slope to the north and the Gulf Stream to the south. Cyclonic rings move south, away from the Gulf Stream, and then in a west and southwest direction. There appears to be a path offshore of the Gulf Stream between Florida and North Carolina along which rings typically travel. Approximately two rings per year follow this path Figure 1. NOAA-3 infrared photograph of the Gulf Stream region off the U.S. East Coast from Florida to Massachusetts, April 28, 1974. Warm temperatures appear as dark shades, cold as light shades. Tire Gulf Stream is depicted as a dark band sweeping across the photograph. Two rings can be seen north of the Gulf Stream and two south of it. One ring to the south is outlined by warm Gulf Stream water that has been entrained by the ring's strong cyclonic flow. The second ring, farther south, is faint, but its existence was verified by ship measurement. Clouds appearing white can be seen in the southeast region. The dramatic effect of rings on at least the near-surface circulation of the ocean is clearly shown by the satellite data. (Courtesy of NOAA/NESS) 67 Figure 2. A schematic representation of the path of the Gulf Stream and the distribution and movement of rings. It is an attempt to summarize a number of studies that were made at different times and that usually focused on a smaller region. at an average speed of 2 kilometers per day. The motion described here is long-term or mean motion; rings exhibit considerable variation from the mean over periods up to several months, and some coalesce with the Gulf Stream only a few months after formation. Rings sometimes move in complicated trajectories— for example, a looping clockwise motion with a speed of 10 kilometers per day, period of 60 days, and amplitude of 75 kilometers, as observed by Fuglister (personal communication). The processes causing rings to move southwest have not been determined. Most mathematical models of rings predict a westward movement due to the Coriolis effect (see page 28). Rings may also be carried passively along with the mean ocean flow; their movement is consistent with what is known about the speed and direction of the return flow from the Gulf Stream. As the Gulf Stream flows north and then east, its volume transport increases dramatically, from about 30 x 10 cubic meters per second off Miami to approximately 150 x 106 cubic meters per second north of Bermuda. The Gulf Stream's transport then decreases as it flows toward the Grand Banks. Although there is at present a debate on just how the water is recirculated in order to account for the transport variation of the Gulf Stream, rings are clearly an integral part of this circulation. Each ring consists of a sizable portion of the Gulf Stream, approximately 500 kilometers in length and roughly 20 percent of the mean path from Cape Hatteras to the Grand Banks. Thus the formation, movement, and subsequent entrainment of rings by the Gulf Stream represents an important part of the return transport of Gulf Stream water, especially when one considers the large number of rings forming each year, estimated to be about 5-8 per year on each side of the Gulf Stream. For example, the volume transport associated with the formation and movement of 13 rings per year, each one consisting of a 500-by-100-by-2-kilometer section of the Gulf Stream is 41 x 10 cubic meters per second. Decay When rings were first seen, they were thought to have short lives of only a few months. Several recent time-series measurements continuing for more than a year indicate that some rings may last as long as 2 years. This long life is possible since most (95 percent) of the energy is in the form of potential energy; only a small portion is in the form of kinetic energy, which can be dissipated by friction. A young cyclonic ring has the main thermocline raised 500-600 meters in its cold core. This potential energy, on the order of 1024 ergs, is "available" to be released as the thermocline slowly subsides in the decay process to the mean Sargasso Sea background level. During decay, the peak tangential velocities in the high-velocity core remain strong, at approximately 100 centimeters per second, but move radially inward as the core subsides. The initial distinctive Slope Water characteristics of the core gradually disappear, indicating that the ring mixes with surrounding Sargasso Sea water. Conclusions Rings are interesting and worthy of study in their own right, but their important, though poorly understood, role in general ocean circulation makes it imperative that we learn more about them. A number of investigators are planning a large-scale study of rings, including surveys of distribution, tracking of trajectories, and detailed measurements of decay. We hope to combine theoretical modeling and field experimentation from a variety of disciplines into a unified study of rings. Philip Richardson is an assistant scientist in the Department of Physical Oceanography, Woods Hole Oceanographic Institution. Suggested Readings Barrett, J. R. 1971. Available potential energy of Gulf Stream rings. Deep-Sea Res. 18:1221-31. Cheney, R. E., and P. L. Richardson. 1976. Observed decay of a cyclonic Gulf Stream ring. Deep-Sea Res. 23:143-55. Fuglister, F. C. 1972. Cyclonic rings formed by the Gulf Stream 1965-66. In Studies in physical oceanography, ed. A. Gordon, pp. 137-68. New York: Gordon and Breach. Parker, C. E. 1971. Gulf Stream rings in the Sargasso Sea. Deep-Sea Res. 18:981-93. Stommel, H. 1965. The Gulf Stream. Berkeley and Los Angeles: University of California Press. 68 The Biology of Cold-Core Rings by Peter Wiebe The Sargasso Sea, until quite recently, was considered a fairly homogeneous body of water. Properties such as temperature and salinity were thought to be uniform over a vast area, with only minor seasonal changes, and these only within 100 to 200 meters of the surface. In addition, it has been said that "volume for volume the Sargasso Sea is the clearest, purest, and biologically poorest ocean water ever studied" (Ryther, 1956). This view of the uniformity in physical, chemical, and biological properties and the paucity of biological life is now changing rapidly, in part because of the discovery that Gulf Stream rings are a ubiquitous feature of the western North Atlantic, especially the northern Sargasso Sea (Parker, 1971). For many years it was known that large cyclonic (counterclockwise) eddies frequently occur south of the Gulf Stream, but the formation of one such eddy was not observed until 1950, during "operation Cabot," the first multiship survey of the Gulf Stream. An eddy, or ring, is created when a large meander of the Gulf Stream breaks away forming a ring of swiftly moving Gulf Stream water (150-250 centimeters per second) around a core of seawater of different origin (see Figure 2, page 68 ). Rings forming to the south or east of the Gulf Stream entrap cold water from the continental slope and are known as cold-core rings. Those forming to the north or west of the Gulf Stream contain warm core water of Sargasso Sea origin. The cyclonic or cold-core rings are estimated to form 5 to 8 times a year (Fuglister, 1971). They are truly massive structures ranging horizontally from 150 to 300 kilometers and vertically from 2500 to 3500 meters when newly formed. Rings generally move south or southwest, gradually decaying as they travel. They have been known to persist as physically identifiable structures for two years or so. Recent unpublished data suggest that there may be as many as 1 5 cold-core rings wandering through the Sargasso Sea west of Bermuda at any given time. Anticyclonic or warm-core rings are believed to form with equal frequency, but they are shallower structures (about 1000 meters deep) and they generally have a shorter existence (approximately 6 months) before coalescing with the Gulf Stream near Cape Hatteras(Saunders, 1971 ). For the biologist, rings are of particular interest because during the process of ring formation, organisms originating in Slope Water or Subarctic Water are isolated within the cold-core structure. Since many of the Slope Water organisms are distinct from species living in the northern Sargasso Sea, the formation of a ring can be viewed as the beginning of a large-scale invasion of one oceanic community by another, with the concomitant intercommunity interaction. In fact, we now believe that the time-dependent events associated with the formation and decay of a cyclonic ring can be conceived of as a large-scale natural ecological experiment that offers the possibility of being able to separate the major effects of the physical- chemical environment on the structure and function of an oceanic community from the biological interactions among species. This is a problem of major importance today, stemming from the fact that in spite of our extensive knowledge about the large-scale dependence patterns of many oceanic organisms, we still cannot specify the factors that ultimately limit the distribution of any oceanic, planktonic organism. Although both cold- and warm-core rings are of interest, we have concentrated on the cold- core structures because of their deeper vertical extent and longer duration. Biological Effects of Changing Ring Structure The first biological study of a cold-core ring was made in September 1972 on R/V Atlantis II cruise 71 . During subsequent cruises, we surveyed five other rings and a large Gulf Stream meander (Figure 1), and measured the biomass (standing crop) of phytoplankton, zooplankton, and midwater fish, 69 85° 80° 75° 70° 65° 60° 55° 0*iS - • -A A A A • Slope Stations A N Sargasso Sea Stations x Rings Surveyed M Gulf Stream Meander 45C 40' 35C 30C 25' 85° 80° 75° 70° 65° 60° 55° Figure 1. Distribution of cold- core rings surveyed and the location of Sargasso Sea and Slope Water stations used in comparisons. Numbered rings are those discussed in detail in the article. 1: Knorr 35, November 1973; 2: Chain 111, February 1973; 3: Atlantis II 71, September 1972; 4: Knorr 38, March 1974. M represents the position of a Gulf Stream meander surveyed on Atlantis II 85, October 1974. Some data from this meander are presented in Figure 7. and the primary productivity (turnover rate) of phytoplankton. At the same sampling locations, observations of temperature, salinity, and plant nutrients were made to define the rings and to assist in the interpretation of the biological data. In the discussion that follows, only data from the first four rings surveyed will be discussed in detail. As rings age, they decay. The process of physical and chemical change can be seen in the warming and increasing salinity of the water, in the lessening nutrient content of the water column, and in the deepening of the oxygen minimum zone. (This zone occurs between 700 and 1000 meters in the northern Sargasso Sea but is much shallower, between 100 and 200 meters, in the Slope Water.) Similar changes are evident in the biota. Although we have yet to sample a ring immediately after its formation, our work in a large meander and in two rings approximately 3 months old has suggested that the plant and animal biomass and species composition in newly found rings is nearly identical to the Slope Water that gave rise to the ring's core (Figure 2). Indeed, even after 3 months, the amount of chlorophyll a, a measure of phytoplankton biomass, can still exceed that found in the adjacent Sargasso Sea water by 50-60 percent. This was the case for a ring formed in late summer and sampled in mid-November of the same year. A number of factors, however, appear to influence the degree of contrast observed, and, as a result, rings of similar age differ substantially in standing crop of plant material. Thus, for example, on the February cruise, we found virtually no difference in plant biomass between a 3 ^-month-old ring and the surrounding area. This, we believe, was largely due to strong vertical mixing of the surface waters (upper 200 meters) caused by winter storm activity. Such disparate results are not as evident in the zooplankton. The biomass of zooplankton in the two young rings was higher than in the Sargasso Sea by 40-90 percent (Figure 1), but the difference between the two rings was less extreme. The two older rings (8-12 months) sampled present a similar picture. Zooplankton biomass was generally lower than in the younger rings but higher than in the Sargasso Sea. In contrast, the distribution of plant chlorophyll was again one of extremes— essentially the same in the ring as in the surrounding waters on one cruise and approximately 65 percent higher on the other. The differences in plant biomass in these older rings also appeared to be a reflection of seasonally influenced processes. In this case the ring with the higher amount of chlorophyll was sampled in the spring, at a time when the "spring bloom" was underway; more nutrients were available within the ring, and primary productivity was higher. The other ring was sampled in late summer, when phytoplankton activity and standing crops generally reach a seasonal low in most temperate areas. Fish biomass data from the various rings are more limited, but show a pattern similar to that of the zooplankton— one of decrease in biomass toward Sargasso Sea levels with increasing age of the rings. Data on the effect of changing ring structure on species composition, while less extensive, exhibit a similar pattern of decrease in RING BIOMASS SARGASSO SEA BIOMASS 10 15 0 05 Knorr 35 November Chain 111 February Atlantis S 71 September Knorr 38 March PHYTOPLANKTON CHLOROPHYLL ZOOPLANKTON DRY WEIGHT Figure 2. Ratios comparing cold-core ring and Sargasso Sea phytoplankton chlorophyll and zooplankton dry weight for rings of different age and time of year sampled. 70 the number of Slope Water forms with ring age. Phytoplankton species in samples from the 3-month- old November ring showed not only greater biomass, but also significant differences between the species composition in the ring and that in the Sargasso Sea. However, the phytoplankton species composition in the ring also differed significantly from that observed in the Slope Water. In fact, the ring species had a closer affinity with the Sargasso Sea species than with those in the Slope Water. Contrasting sharply was the 10-month-old September ring in which the phytoplankton biomass differed little from that of the Sargasso Sea. In this case there were no significant differences in the number of species or their abundance, or in the species composition. For the zooplankton, analysis of species composition has focused on one particular group, the euphausiids, which are small shrimplike crustaceans (Figure 3). In sampling the slope, ring, and Sargasso Sea areas, we have encountered 32 species of euphausiids. Five species, Euphausia krohnii, Meganyctiphanes norvegica, Nematoscelis megalops, Thysanoessa gregaria, and T. longicaudata, are characteristic Slope Water forms and were numerically dominant in all Slope Water and in many ring samples thus far collected and counted (Figure 4). There were few of these species in the Sargasso Sea samples. Subtropical-tropical forms such as E. brevis, N. microps, N. tenella, Stylocheiron affine, and S. suhmii exhibit the opposite pattern, being most abundant in the Sargasso Sea and older rings, and much less evident in the Slope Water. In contrast, the more cosmopolitan species, T. parva, S. carinatum, E. tenera, and E. hemigibba, were found in considerable abundance in each type of water. Individual patterns of abundance are reflected in the overall changes in species composition that seem to occur as a ring ages. In the younger rings the euphausiids show strong overall compositional similarity to those in the Slope Water, and generally weak similarity to those in the Sargasso Sea. The older rings, however, show a reverse pattern. Moreover, the younger rings with strong similarity to the Slope Water contain, on the average, twice as many individuals in the upper 800 meters of the water column as do the older rings with strong affinity to the Sargasso Sea species. Sampling Instrumentation Although it is almost certain that the changes in species composition and abundance are linked to the gradual change in the environmental conditions in the ring, we still do not know specifically what the causal factors are. This stems in part from the fact that our descriptions of the changes are very gross and lack the detail required to determine cause and effect. While rings appear to be very large hydrographic features, we, for example, have reached the point where simple towed nets and trawls can no longer provide sufficient resolution to enable us to elucidate the spatial relationships of the organisms. Furthermore, these instruments do not permit simultaneous collection of environmental information such as pressure, temperature, salinity, light, and oxygen, the small-scale fluctuations of which are considered to be important elements in shaping the distributional patterns of the organisms. The need for more sophisticated sampling gear became especially acute as we began to study changes with time in the vertical distribution of key Slope Water indicator species in the ring. (An indicator species is one that is generally restricted to a particular area or water mass and when found elsewhere indicates the water in which it typically lives is also present.) Substantial effort was therefore devoted to improving the zooplankton sampling instrumentation. This work resulted in the construction of a Multiple Opening/Closing Net and Environmental Sensing System (MOCNESS), which has significantly enhanced our sampling capability (Figure 5). The system carries nine nets that can be opened and closed sequentially on command through conducting cable from the surface. Environmental sensors to measure conductivity (±.001 percent), temperature (+ .0005°C), and depth (±.01 meter) are attached to the net support frame. In addition, there are sensors to monitor flow past the net and the angle of the net assembly from the vertical, as well as indicators to record the electrical and mechanical function of the opening/ closing mechanism. All data are transmitted up the cable to the shipboard computer for real-time processing. Migration Patterns and Ring Decay We have now successfully used this system (Figure 6) on three cruises-one to a Gulf Stream meander in October 1974 and two (August 1975, November 1975) to the same ring— to examine the diel (24-hour) vertical migration patterns of zooplankton in the rings, Sargasso Sea, and Slope Water (Figure 7). The euphausiids in the Gulf Stream meander cruise samples clearly show migration patterns for various species of the genus Euphausia. These species migrate downward at sunrise to depths of 2 700 meters, and upward at sunset to within 100 meters of the surface, with many appearing at 71 Warm-Water Species Euphausia brevis Stylocheiron affine 5 mm Stylocheiron carinatum 5 mm Euphausia hemigibba 5 mm Stylocheiron suhmii Euphausia ten era 5 mm Nematoscelis tenella Nematoscelis microps Thysanoessa parva Figure 3. Euphausiid species more frequently encountered in the western North Atlantic where cold-core rings occur. 72 Cold-Water Species Tliysanoessa gregaria Euphausia krohnii Thysanoessa longicaudata Nematoscelis megalops Meganyctiphanes norvegica 5 mm (Drawings by Nancy Barnes, after B. P. Boden, M. W. Johnson, and E. Brinton, 1955, The Euphausiacea (Crustacea) of the North Pacific, Berkeley and Los Angeles: University of California Press; E. Brinton, 1975, Euphausiids of Southeast Waters, NAG A Rept., vol. 4, pt. 5, University of California, Scripps Institution of Oceanography; and H. Einarsson, i Euphausiacea 1. Northern Atlantic Species, Dana Rept. No. 27.) 73 25% 50% SLOPE WATER SPECIES 25% 50% COSMOPOLITAN SPECIES 25% 50% SARGASSO SEA SPECIES ABUNDANCE OF DOMINANT EUPHAUSIIOS Figure 4. Relative abundance (%) of the dominant cold- water (Slope Water species) and warm-water (cosmopolitan and Sargasso Sea species) euphausiids in the Slope Water, ring, and Sargasso Sea. the surface. In contrast, members of the genus Stylocheiron do not exhibit migratory behavior. These species, as shown in Figure 7, live in separate parts of the water column and overlap very little vertically. The data on presence and abundance of species in the Sargasso Sea, in combination with similar data from the meander and slope stations on this cruise, show the very strong contrast that exists across the faunal boundary marked by the Gulf Stream. There is a decrease in the abundance of warm-water species and an increase in the number of cold-water species, which coincides with the abrupt change in water properties. The daytime depth distribution for several typically warm-water species that were found at the slope and meander stations tended to be shallower in the colder, less saline water. This kind of data forms a baseline, or background, and is necessary if we are to understand the changes in vertical distribution with ring decay that are distinct from basic patterns evident in home-range populations. It is the documentation ' TOGGLE RELEASE AND MOTOR DRIVE ASSEMBLY ) '' Figure 6. (Top) MOCNESS ready for launch over the stern of R/V Chain (August 1975). Information from the net system sensors is transmitted through the cable on the winch in the foreground to the shipboard computer for processing. (Bottom) After a MOCNESS haul, each of the nine nets is washed to assure that no organisms are left on the meshes. After wash down, the animals in the cod- end buckets are transferred to glass jars, preserved, and taken back to the laboratory for measurement of biomass and enumeration of species. 1m x14mNET Figure 5. Perspective drawing of the Multiple Opening/ Closing Net and Environmental Sensing System. Although only one net is illustrated, nine nets are carried on the support frame. A command from the surface causes the motor drive and toggle release to drop a net bar, thus closing one net and opening the next. (Drawing by Ttwmas Aldrich) 74 SARGASSO SEA GULF STREAM MEANDER Euphausia hemigibba NIGHT DAY . 1P SLOPE WATER DAY SARGASSO SEA GULF STREAM MEANDER SLOPE WATER . Euphausia tenera 800 Oi 200 400 600 Euphausia americana L_..i ^^^^^^^^^^^^^^J_J I 1 1 r r Stylocheiron carinatum DAY NIGHT DAY 100 10 ,1 1,0 , 100 10,00 1 1,0 DAY 1 10 Stylocheiron affini Sylocheiron elongatum r 0 200 400 600 800 0 200 400 600 800 0 200 400 600 GULF STREAM SLOPE WATER MEANDER Nematoscelis megalops DAY DAY GULF STREAM SLOPE WATER MEANDER Euphausia krohnii DAY DAY 0 200- 400-1 600- 800- 1 10 1 10 100 1000 1 10 100 1 10 100 1000 GULF STREAM SLOPE WATER GULF STREAM SLOPE WATER MEANDER MEANDER Thysanoessa parva Thysanoessa longicaudata DAY DAY DAY DAY 10 100 1 10 100 1 10 100 1 10 100 0 200 400 600 800 Figure 7. Vertical distribution of the more abundant euphausiids caught in four MOCNESS hauls (Sargasso Sea, day, night; Gulf Stream meander, day; Slope Water, day) taken on R/V Atlantis II cruise 85. Warm-water species above; cold-water species below. Triangles and stars indicate maximum depths sampled. of changes in basic behavioral patterns, together with information about changes in the environmental setting, that should provide clues about the reasons for the change. For example, although our field and laboratory sampling have not progressed to the point where we can say definitively that migration patterns change with ring decay, there are suggestions in our data that this is the case. In particular, it appears that as a ring ages, the Slope Water forms live deeper in the water column and that their migrations to the surface are inhibited. Such changes could have a pronounced effect on the survival of the species. It is widely believed that one manifestation of the migration behavior is that migrators feed in the relatively food-rich surface waters at night and that little or no feeding takes place where they reside during the day (several hundred meters below the surface). If this is the case, and if the ring surface- water properties become unsuitable and the migrators no longer swim to the surface at night, then the individuals could be effectively excluding themselves from sufficient food. Their ultimate fate may be starvation. Thus, in addition to examining the migrations for changes in pattern, we are currently measuring the total lipid content of individual euphausiids caught in the MOCNESS tows in an effort to assess changes in the nutritional status of the expatriated ring populations. Here our preliminary results tend to support the contention discussed above. Lipid levels in ring-expatriated euphausiids in older rings are indeed much lower than those in individuals of the same species in the Slope Water. Adding to the difficulty of interpreting the changes we have observed in species composition and abundance of the ring populations are the very large, natural variations in all oceanic populations. For example, a Slope Water indicator species has always been found to be the most abundant at some point in a ring, but no one species appears consistently as the numerical dominant. The species in the Sargasso Sea and in the Slope Water exhibit a similar pattern of shifting dominance. The statuses of the Sargasso Sea and Slope Water populations thus appear as critical determinants of 75 the initial biological structure of rings and their subsequent evolution. An obvious means to reduce the effect that the great variability in time and space has on our data is to obtain time-series measurements from a single ring starting from the time of formation. This we plan to do. Until recently our measurements have been obtained from different rings of different ages. Of potential evolutionary significance is the fact that some cold-core rings coalesce with the Gulf Stream after a period of weeks to months. This may be a mechanism by which expatriated individuals are reunited with similar forms in their home range. If, as appears to be the case, populations of Slope Water species living in the rings are under increasing environmental stress, this stress may provide a progressive selection mechanism. Cold-core rings may therefore be a means by which genetically altered populations are introduced into the parent population. Since paleocirculation studies (Luyendyk, Forsyth, and Phillips, 1972; Berggren and Hollister, 1974) suggest that the Gulf Stream has been in its present position for millions of years, the process of ring formation and decay has probably been of importance for about the same length of time. Thus we may be observing a phenomenon that helped determine the limits of present-day biogeographic distributions of oceanic populations. Conclusion The discovery of rings and the evidence for their widespread occurrence in the western North Atlantic have provided oceanographers with a new and exciting area for research. Rings clearly offer a unique opportunity to study processes that are important to a determination of distribution and abundance of oceanic plants and animals. Although rings are best known in the western North Atlantic, similar hydrographic features are likely to be found in other western boundary current areas. Their importance to the physics, chemistry, and biology of the oceans is only beginning to be understood. Peter Wiebe is an associate scientist in the Department of Biology, Woods Hole Oceanographic Institution. References Berggren, W. A., and C. D. Hollister. 1974. Paleogeography, paleobiogeography and the history of circulation in the Atlantic Ocean. In Studies in paleo-oceanography, ed. W. W. Hay, pp. 126-86. Soc. Econ. Paleontologists and Mineralogists, Spec. Pub. No. 20. Fuglister, F. C. 1971. Studies in physical oceanography. In A tribute to George Wust on his 80th birthday, ed. A. Gordon, pp. 137-68. New York: Gordon and Breach. Luyendyk, B. P., D. Forsyth, and J. D. Phillips. 1972. Experimental approach to the paleocirculation of the oceanic surface waters. Geol. Soc. Amer. Bull. 83:2649-64. Parker, C. E. 1971. Gulf Stream rings in the Sargasso Sea. Deep-Sea Res. 18:981-93. Ryther, J. H. 1956. The Sargasso Sea. Sci. A m. January, pp. 77-81. Saunders, P. M. 1971. Anticyclonic eddies formed from the Gulf Stream. Deep-Sea Res. 18:1207-19. 76 Mapping the Mfeather in the Sea by James C. McWilliams We live under the daily influence of the weather. It dictates many of our activities and some, perhaps many, of our moods. At the most primitive level, we can anticipate its influence simply by looking towards the horizon to see what might be coming. An extension of this, based only on the technology of being able to communicate rapidly over large distances, is drawing a large-scale weather map from a set of observations taken at nearly the same time. Such maps have been possible for over a century, and since 1870 they have been routinely drawn by the Army Signal Corps of the U.S. Weather Bureau (now called the National Weather Service). Meteorology has become quite a sophisticated science: the observational archives are, by oceanographic standards, enormous; theoretical arguments are abundant; and forecasting by computer integration of the equations of motion has been performed for over twenty years. High hopes have been held for automated forecasting, but its performance has thus far been somewhat disappointing. Even today, an experienced synoptician can forecast from a weather map almost as well as a computer can. In practice, of course, weathermen now make their guesses after examining both synoptic maps and computer weather forecasts. There can be no doubt, though, about the utility of mapping atmospheric synoptic-scale motions (the familiar "lows" and "highs" in pressure, cyclones, and anticyclones). Recently we have begun transferring these techniques to the field of physical oceanography. From MODE-1 (Mid-Ocean Dynamics Experiment), discussed throughout this issue, synoptic maps have been drawn for mesoscale ocean eddies (eddies with diameters of a few hundred kilometers). Since the data required for this exercise were both expensive and exhausting to obtain, these maps will be rare for some time to come. Furthermore, even if they could be used for forecasting, no one at present would see much value in this: the eddies directly influence few, if any, commercial activities, and their role in the broader issue of the earth's climate is years away from being understood. New skills generate their own demand, however, and I for one would be reluctant to predict what future uses may be made of these kinds of maps. I am purposefully making an analogy between atmospheric synoptic-scale winds (what flows around a pressure anomaly on a weather map) and oceanic mesoscale currents. Many similarities do occur between them. Both are essentially geostrophic (that is, the fluid motions are directed parallel to lines of constant pressure). Both have nearly the maximum vertical scale possible, with currents related to each other over the whole of either the tropospheric height or the ocean depth. Both are the most energetic currents of any that occur in their respective fluids (for example, mesoscale eddies are more energetic than the ocean tides). Finally, it is plausible that both may arise spontaneously by tapping the energy of the permanent currents that are present— such as the prevailing westerlies or the Gulf Stream. Obvious differences exist as well. The ocean eddies are relatively much slower (changing in months rather than days) and smaller (extending over hundreds, not thousands, of kilometers). However, we may argue that this is mostly due to the difference between water and air, particularly the way they are stratified. A useful manner of expressing the character of a fluid, based on a variety of theoretical models, is in terms of a horizontal length scale R, called the internal deformation radius. We define it by the formula R = g sin where g is the gravitational acceleration on the earth, £1 the angular frequency of the earth's rotation, A the latitude, Ap/p the relative vertical change in 77 density (that is, the stratification strength), and h the vertical distance over which it occurs. For the ocean RQ = 50 kilometers, whereas for the atmosphere RA = 800 kilometers; these are also typical scales for oceanic mesoscale and atmospheric synoptic-scale eddies. The fact that aerodynamicists reproduce the real behavior of large airplane shapes in small wind tunnels demonstrates the success of mimicking a process through dynamically correct rescaling. We shall do the same thing here for atmospheric and ocean maps, where the rescaling is based on the deformation radius: ocean lengths should be smaller by the ratio RO/R< , and ocean time intervals should be longer by its inverse, RA/RO. (The rescaling in time is also based on a theory; namely, we assume that the actual currents are related to a phenomenon called Rossby waves [for Carl Rossby, a pioneer in theoretical meteorology] . In such waves, frequency is proportional to length, and length we have assumed proportional to R.) Parallel sequences of maps are shown in Figure 1 for both the atmospheric ground level pressure and the oceanic pressure at 150 meters depth (minus the mean hydrostatic pressure at 150 meters, which can drive no currents). Actual distances on the page are equivalent by the relation described above (100 kilometers in the ocean equals 1600 kilometers in the atmosphere). Similarly, the time intervals between maps are approximately comparable (30 days and 2 days, respectively). The geographical domains have also been chosen to be comparable in size. The atmospheric region is narrower in latitude than longitude to avoid the tropical and polar regions— the motions there are phenomenologically distinct from mid-latitude cyclones and anticyclones. These maps allow us to better see how good the analogy between atmospheric and oceanic motions really is. They expose structural details and their time evolution, and thus contain a different kind of information than the summary characteristics mentioned above for oceanic and atmospheric eddies. Similar numbers of eddies appear in the two sets of maps; there also seems to be a comparable partitioning between high- and low- pressure centers. These centers retain their identities from one mapping period to the next. For some of the eddies, there seems to be a systematic motion towards the west (for example, the northwestern atmospheric low and the central oceanic high), but for others this is not true. Surely it would be foolish to try to generalize eddy behavior on the basis of the study of so few eddies for so brief an interval. The meteorologists have plenty of other maps; unfortunately, the maps of Figure 1 nearly span MODE-1, which is presently the only completed, four-dimensional mapping experiment for mesoscale eddies. There are, of course, important structural differences between the two sets of maps (after all, it only rains in the atmosphere). Some are due to the different manners in which they were constructed-the number and distribution of observations as well as the techniques for drawing contour lines. (For example, the apparently greater amount of atmospheric small-scale structure, or contour wiggles, is strongly related to mapping technique.) Others may be related to the chance selection of which periods were mapped. No one eddy on the ocean maps can be expected to match identically any of the atmospheric eddies (disregarding the unlikely possibility that one set of eddies directly forces the other). In the face of these difficulties, I shall not attempt here to define the important inadequacies in the ocean/atmosphere analogy. For the remainder of this article, I shall neglect questions concerning techniques for drawing maps and the accuracy with which the end results are known— even though these are crucial issues for translating a pretty map into a scientific fact. Many other maps can be drawn from MODE-1 , for other depths and other times intermediate to the extremes shown in Figure 1 . I would now like to illustrate from some of them various characteristics of mesoscale eddies. Figure 2 is a schematic drawing of the vertical profiles of either the horizontal velocities or pressures (as mentioned above, these two quantities are proportional by the geostrophic relation) associated with ocean eddies. Two profiles have been drawn: we learned from MODE-1 that the vertical structure of currents was approximately a combination of these two simple structures. There are distinct patterns in time and horizontal position that are associated with each of these structures; during MODE-1 the patterns were reasonably, but not completely, independent of each other. The proportional magnitudes of the two profiles have been chosen in Figure 2 to represent the average conditions observed during MODE-1. We can see, therefore, that in and above the thermocline (above 750 meters), the observed currents were dominated by the structure that has vertical variation, while, at Figure 1. A comparison of maps of atmospheric synoptic- scale eddies (left), as drawn by the National Weather Service, and oceanic mesoscale eddies. The maps are of surface pressure and of pressure at 150 meters depth, respectively. Tfie atmospheric region spans the continental U.S.; the oceanic region is that of MODE-1. C. I. = contour interval on the maps. 78 C.I. = 8 mb 1600 km 7 p.m. EST 12/13/73 100 WEST LONGITUDE 7 p.m. EST 12/15/73 100 WEST LONGITUDE C.I. = 4 mb 100 km 4/15/73 30 29 28 CE O 27 26 I I 72 71 70 69 WEST LONGITUDE 68 5/15/73 7 70 69 WEST LONGITUDE 68 6/14/73 7p.m. EST 12/17/73 'So 100 WEST LONGITUDE 26 - L 72 71 70 69 WEST LONGITUDE 79 TYPICAL VELOCITY (cm/sec) IOOO-- 2000 -- 3000 4000 5000 -- DEPTH (m) Figure 2. Schematic vertical profiles of horizontal velocity (or pressure] for the two predominant mesoscale eddy structures. greater depths, the depth-independent structure dominated. The patterns shown in the oceanic maps of Figure 1 illustrate the characteristics of the depth- variable structure. The type of observation that made the greatest contribution to exposing these patterns was measurements of water density (in a hydrostatic fluid, the pressure at a level is caused by the weight of fluid above it, and weight is volume times density). In contrast, the maps of Figure 3 are from 1500 meters depth, a region where the depth-variable contribution is virtually nil. These maps were constructed on the basis of trajectories of neutrally buoyant SOFAR (SOund Fixing And Ranging) floats (see page 57). The interval between the 1500-meter maps was chosen to be shorter (ten days), because the typical time required for a synoptic feature to change was less for the depth- independent structure. The time changes were perhaps less systematic at this level; however, the C. I. = 0.4mb I 1 100 km 30 - 29 < 28 o: o 27 26 5/5/73 I I 72 71 70 69 WEST LONGITUDE 30 29 LoJ Q 28 CC O 27 26 68 5/15/73 J_ _L 72 71 70 69 WEST LONGITUDE 30 29 28 cr O 27 26 68 5/25/73 72 71 68 70 69 WEST LONGITUDE Figure 3. Maps of the ocean pressure at 1500 meters depth on three different days during MODE-1. 80 100 km 30 - 30- 1500m 5/15/73 C.I.=4mb LiJ O IT O 27 - 26 - 72 71 70 69 WEST LONGITUDE 30 - 29 LJ Q 28 X I- o 27 26 750-l500m i 72 71 70 69 WEST LONGITUDE 30 - 29 LJ Q 28 CC. O 27 26 I500-3000m 68 5/15/73 C.I.=0.4mb 1 vv_^ 1 1 1 1 72 68 71 70 69 WEST LONGITUDE Figure 4. Maps of the vertical difference of pressure for three depth intervals on a particular day during MODE-1. westward propagation of eddies seen in Figure 1 can also be detected here, though one gains confidence in this conclusion only after seeing a good many more maps than just those of Figure 3. The diameters of the 1 500-meter eddies were relatively smaller as well, and the current speeds were much slower than at 150 meters. The preceding claim about the ocean eddies involving only two vertical structures is, of course, an idealization, though it provides a useful summary of a number of complicated measurements. We can find some of the discrepancies from this claim in Figure 4, and at the same time see specific details of eddy patterns from top to (nearly) bottom. This figure presents patterns of the vertical differences in pressures between several depths on a particular day. If the ocean structure truly were as shown in Figure 2, then each of these patterns should be the same except for different overall magnitudes. This is because any vertical subtraction associated with the depth-independent structure would contribute nothing, leaving only the depth-variable profile. As one can see, the several maps in Figure 4 are similar, but not identical. They are, however, much more similar to each other than to the maps in Figure 3. We obtained from MODE-1 the best set of mesoscale eddy maps yet available. They indicated eddy structures that were approximately a combination of the two profiles shown in Figure 2, with the corresponding horizontal patterns shown in Figures 1 and 3. What instruction can be taken from these maps, whether for the purpose of forecasting or allowing us to better understand the physical nature of the eddies, is yet unknown. There is certainly the atmospheric precedent, however, to encourage us that they may have great value. James C. McWilliams currently works at the National Center for Atmospheric Research, Boulder, Colorado. Suggested Readings Gould, J., and H. Freeland. 1975. Objective analysis of mesoscale ocean currents. Deep-Sea Res. (submitted). McWilliams, J. 1975. Maps from the MODE Experiment. I. Geostrophic streamfunction. /. Phys. Ocean, (submitted). 81 Sea Surface temperature During MODE-1 by Arthur Voorhis and Elizabeth Schroeder One of the most important oceanographic problems is the study of long-term variations in mean sea surface temperature. This temperature is an important factor in the energy exchange between ocean and atmosphere, and we must know more about it in order to assess past and future changes in world climate. Because surface temperature depends on many variable physical processes such as local solar and atmospheric heating and cooling, and turbulent heat exchange with underlying ocean layers, its measurement requires averaging a great many observations over a large area and over a long period of time. Major programs such as the North Pacific Experiment (NORPAX) in the United States and the Joint Air-Sea Interaction Experiment (JASIN) in the United Kingdom are active in this area. The surface temperature at any fixed location is also affected by advection, that is, by surface currents that bring in warmer or cooler water from distant areas. This process can generate important changes in surface temperature, especially if large horizontal variations in temperature are present (for example, near the Gulf Stream) and if the currents are persistent and large-scale. Unfortunately, too little is known about such currents, particularly those on a scale of 100 to 500 kilometers. One of the few experiments designed to study currents in this range was MODE-1 (Mid-Ocean Dynamics Experiment), which was conducted in the area of the North Atlantic subtropical convergence during the spring of 1973 (see page 45). Recently we have correlated the currents with surface temperatures measured during MODE-1 and have concluded that the large-scale surface temperature pattern is largely the result of surface advection. The subtropical convergence is one of the classical transition zones separating two meteorological wind regimes. In the western North Atlantic, it lies roughly between 22°N and 32°N latitude and separates the prevailing westerlies to the north from the easterly trades to the south. Maps of monthly mean sea surface temperature averaged over years of data present a relatively uncomplicated picture in the area eastward of the Gulf Stream's influence. This can be seen in the map for March in Figure 1. In general, the temperature decreases northward at all times of the year by an amount that varies seasonally. The maximum decrease occurs in late winter (approximately 0.5°C per degree of latitude, as shown in the figure) and the minimum in late summer (about 0.1 °C per degree of latitude). The zonal (east-west) temperature variation is always very small. There are, however, major differences between this average picture and the synoptic temperature distribution (that is, the actual temperature distribution at any one time). For over 10 years it has been known that the subtropical convergence is a region of pronounced surface frontogenesis. These fronts are often quite visible from shipboard or from the air (Figure 2). Observations show that these features separate adjacent surface water masses of differing temperatures (and densities). Although the distance across a front is usually quite short (less than 100 meters), with a horizontal temperature gradient two orders of magnitude greater than the mean meridional (north-south) gradient, they have been tracked as they meander along the sea surface for distances exceeding several hundred kilometers. More recent and more dramatic evidence for large-scale variation in surface temperature comes from specially processed infrared satellite imagery of the sea surface, such as that in Figure 3, which clearly shows surface temperature dominated by meridional and zonal variations on a scale of several hundred kilometers. This is a marked contrast to the mean monthly picture (Figure 1). 82 95 90 85° 45 40' 45* SURFACE TEMPERATURE DEGREES FAHRENHEIT MARCH 95° 90° 85° 80° 75° 70° 65' 60* Figure 1. Mean March sea surface temperature in the western Atlantic and the Caribbean Sea. (After Fuglister, 1947) 83 Figure 2. Aerial photograph of surface debris collected along a surface front observed near 26° N 66° W in February 1965. (After Voorhis, 1969) Eddy Surface Currents Except for results of the Aries expedition in 1959-60 (see page 20), little was learned about the subsurface currents in this region until MODE-1 . It is now known that these currents are dominated by large eddylike motions, which are for the most part confined to the upper 1 500 meters of the water column. We asked ourselves whether these motions could be responsible for the large anomalies in the sea surface temperature field. Unfortunately, no usable satellite images of the sea surface during MODE-1 were available. However, continuous surface temperatures were recorded from three ships (R/V Chain and R/S Researcher from the U.S., and RRV Discovery from the U.K.) as they wandered about the area during the four months of the experiment. With these data plus results from more than 800 temperature and salinity soundings, we were able to construct temperature maps having reasonable spatial coverage for successive 1 5-day periods from the beginning to the end of MODE-1. These were then compared with maps that describe the eddy surface currents during the same time periods. The sea surface temperature map for the period May 1 5-29 is shown in the top half of Figure 4, the corresponding (eddy) surface motion in the bottom half. The solid contours on the bottom map represent the surface dynamic topography relative to 1 500 decibars that was computed from the temperature and salinity lowerings at the positions indicated by small dots. The eddy surface current direction is shown by the arrows, and its magnitude can be estimated from the speed scale in the lower right-hand corner of the figure. To facilitate the comparison between the two maps, we have shaded on the bottom drawing all those areas of the sea surface that are cooler than the mean temperature on the top illustration. The temperature map clearly shows a large-scale structure superimposed on a mean meridional trend of about 0.5°C per degree of latitude, with cooler water to the north. The resemblance, especially as to scale, between it and the satellite surface image in Figure 2 is striking. The map of dynamic topography shows the famous MODE eddy that dominated the central region during most of the experiment. It appears on the figure as a hill on the surface having a height of about 10 centimeters and an average radius of about 150 kilometers. Around it flows an anticyclonic (clockwise) current with a speed of about 20 centimeters per second. It takes almost one month for surface water to make a circuit of the eddy. A visual comparison between the maps of temperature and dynamic topography strongly suggests that eddy surface currents are primarily responsible for the large-scale structure seen in the temperature map. Note the warm southern water advected northward along the western side of the eddy and the cooler northern water carried to the south along its eastern side. All of the temperature maps of the MODE area (and those of dynamic height) show similar large-scale patterns, but they change with time. Temperature maps are usually dominated by the long (200 kilometers) alternate intrusions of warm and cool water seen in Figure 4. These intrusions seem to last 20 to 30 days, which is approximately the circuit time for an eddy but much less than the lifetime of the eddy itself (about 90 days). It is for this reason that temprature maps show very little spiral structure; that is, the long tongues of warm or cool water appear to be carried by surface currents only once around an eddy. The duration of an eddy is, of course, controlled by the internal dynamics of the ocean. On the other hand, we believe that the persistence time for the long advected features of the surface temperature field is controlled mainly by heat exchange with the atmosphere. 84 Surface fronts similar to those in Figure 2, were often seen from shipboard during MODE-1 , and the continuous surface temperature records show numerous frontal crossings. Because the fronts were carried along rather rapidly by the eddy surface currents, it was simply not possible to map their position in any detail from the rather random ship crossings. In retrospect, it seems unfortunate that none was tracked by ship or from the air. Nevertheless, our data suggest that most of the fronts occurred in areas where there were substantial horizontal changes, or gradients, of surface temperature and where the eddy surface flow was such as to maintain these gradients. This situation usually occurred along the boundaries of the long lateral intrusions seen in the temperature maps. Such frontogenic areas in Figure 4 are in the vicinity of 27.5°N 7 1.0°W and 28.5°N 69 .0°W. This rather vague observation is supported by the theoretical work of Hoskins and Bretherton (1972), who have worked with a density-stratified fluid to study surface frontogenesis by a large-scale current field. One finds, using the Hoskins-Bretherton analysis, that surface fronts can form in some areas in three to five days. Figure 3. Enhanced infrared satellite image of the sea surface in the western North Atlantic showing large-scale advective patterns on April 1, 1974. Dark regions are warm water, grey are cool water; white areas are clouds and should be ignored. (Courtesy ofR. Legechis.NOAA/NESS) 85 tO KM '/DAY , 30° N 29° 28° 27° 26° 30° N 29° 28° 27° 26° 72°W 71 c 70" 69° 68° 67° Figure 4. Sea surface temperature map (top) and dynamic topography (bottom), relative to 1500 decibars, of the MODE area for the period May 15-29, 1 9 73. The shaded area on the bottom map shows surface water that is cooler than the mean temperature of the top map. The eddy surface current speed is inversely proportional to the spacing between the contours of dynamic height and is equal to 10 kilometers per day when the spacing is equal to the scale shown in the lower right-hand corner. (After Voorhis and Schroeder, 1976) that surface isotherms do not coincide with contours of dynamic height. This is mainly due to the fact that the eddy currents are continually changing with time. Therefore, in order to infer something about the currents from temperature, we need a rather complete knowledge of all the relative time scales involved. This includes the variation of mean surface temperature gradients, of the eddy field, and of the temperature structure created by this field. Summarizing our results, we conclude that surface currents from large-scale eddies are responsible for large-scale, slowly changing patterns in the sea surface temperature, both in the MODE area and, probably, in many other areas of the world ocean. For the oceanographer and meteorologist interested in annual or longer-term changes in the mean sea surface and their effect on climate, these varying temperature patterns are unwanted noise and introduce a further uncertainty into their measurements. Arthur Voorhis is an associate scientist in the Department of Physical Oceanography, Woods Hole Oceanographic Institution. Elizabeth Schroeder is a research associate in the same department. References Fuglister, F. C. 1947. Average monthly sea surface temperatures of the western north Atlantic Ocean. Pap. Phys. Oc. and Meteor. 10(2), 25 pp. Hoskins, B. J., and F. P. Bretherton. 1972. Atmospheric frontogenesis models: mathematical formulation and solution. J. Atmos. Sci. 29:11-37. Voorhis, A. D. 1969. The horizontal extent and persistence of thermal fronts in the Sargasso Sea. Deep-Sea Res. 16:331-37. Voorhis, A. D., and E. H. Schroeder. 1976. The influence of the deep mesoscale eddies on sea surface temperature in the North Atlantic convergence. J. Phys. Ocean, (in press). Conclusion A question arises from the material presented here that has important implications for oceanographers who are interested not only in the large-scale eddy field covered by MODE-1 but also in other parts of the world ocean: Can the knowledge of the time- varying spatial structure of surface temperature (or other properties), measured from satellite or by rapid ship surveys, contribute significantly to the understanding of these motions? We believe the answer is yes, although a great deal more work must be done. For example, it is clear from our maps 86 Oceanus welcomes Oceanus Latest addition to the research vessels of Woods Hole Oceanographic Institution is R/V Oceanus. Designed to be economical as well as efficient, she is 177 feet in overall length. Range at maximum cruising speed of 14.5 knots is 8000 nautical miles. Complement is 13 officers and crew, and 13 members of the scientific party. She and we take our name from the Titan who fathered the gods of the rivers and the nymphs of the sea. (Photo by Peterson Builders, Inc.) 87 Back Issues Limited quantities are available at $3.00 per copy; a 40 percent discount is offered on orders of five or more copies. We can accept only prepaid orders, and checks should be made payable to Woods Hole Oceanographic Institution. Orders should be sent to: Oceanus Back Issues, Woods Hole Oceanographic Institution, Woods Hole, MA 02543. SEA-FLOOR SPREADING, Winter 1974-Plate tectonics is turning out to be one of the most important theories in modern science, and nowhere is its testing and development more intensive than at sea. Eight articles by marine scientists explore continental drift and the energy that drives it, the changes it brings about in ocean basins and currents, and its role in the generation of earthquakes and of minerals useful to man. AIR-SEA INTERACTION, Spring 1974-Air and sea work with and against each other, mixing the upper ocean, setting currents in motion, building the world's weather, influencing our lives in the surge of a storm or a sudden change in patterns of circulation. Seven authors explain research in wave generation, hurricanes, sea ice, mixing of surface waters, upwelling, long-range weather prediction, and the effect of wind on circulation. ENERGY AND THE SEA, Summer 1974-One of our most popular issues. The energy crisis is merely a prelude to what will surely come in the absence of efforts to husband nonrenewable resources while developing new ones. The seas offer great promise in this context. There is extractable energy in their tides, currents, and temperature differences; in the winds that blow over them; in the very waters themselves. Eight articles explore these topics as well as the likelihood of finding oil under the deep ocean floor and of locating nuclear plants offshore. MARINE POLLUTION, Fall 1974-Popular controversies, such as the one over whether or not the seas are "dying," tend to obscure responsible scientific effort to determine what substances we flush into the ultimate sink, in what amounts, and with what effects. Some progress is being made in the investigation of radioactive wastes, DDT and PCB, heavy metals, plastics and petroleum. Eleven authors discuss this work as well as economic and regulatory aspects of marine pollution. FOOD FROM THE SEA, Winter 1975-Fisheries biologists and managers are dealing with the hard realities of dwindling stocks and increasing international competition for what is left. Seven articles explore these problems and point to ways in which harvests can be increased through mariculture, utilization of unconventional species, and other approaches. DEEP-SEA PHOTOGRAPHY, Spring 1975-A good deal has been written about the use of hand-held cameras along reefs and in shallow seas. Here eight professionals look at what the camera has done and can do in the abyssal depths. Topics include the early history of underwater photography, present equipment and techniques, biological applications, TV in deep-ocean surveys, the role of photography aboard the submersible Alvin along the Mid-Atlantic Ridge, and future developments in deep-sea imaging. THE SOUTHERN OCEAN, Summer 1975 The first of a regional series (in planning are issues on the Mediterranean and Caribbean) examining important marine areas from the standpoint of Oceanographic disciplines most interested in them. Physical, chemical, and biological oceanographers discuss research in antarctic waters, while a geologist looks at the ocean floor, meteorologists explain the effect of antarctic weather on global climate, and a policy expert sets forth the strengths and weaknesses of international scientific and political relations in the area. SEAWARD EXPANSION, Fall 1975-We have become more sophisticated in our offshore activities than is generally realized. Eight articles examine trends in offshore oil production and its onshore impacts; floating oil ports; artificial islands; floating platforms; marine sand and gravel mining. National and local policy needs oo are examined in the light of the move seaward. MBL WHOI LIBRARY UH IfllH X --,