Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
on specific instrument needs, improved capability for numerical modeling and analyses of large data sets, a new look at oceanographic data archiving, and increased emphasis on training young scientists in the mathematical and computer skills needed in physical oceanography. RECOMMENDED RESEARCH INITIATIVES Climate-Related Studies Many countries are participating in the planning of a global research program, the World Climate Research Programme (WCRP), to address aspects of climate variation. This program seeks the understanding that will allow us to answer questions with important economic and social implications. For example, what are the mechanisms, if any, by which the ocean influences year-to-year variations in the earth's climate? Does the ocean play a role in producing climate anomalies, such as droughts, floods, heat waves, or abnormal frosts? If it does, can we understand the processes whereby this occurs? Can we develop a capability for predicting climate change? We know that the ocean plays a major role in determining the mean climate of the world. It is critical in controlling global patterns of precipitation and evaporation. The ocean absorbs energy from the sun and releases energy to the atmosphere at times and places distant from the point where the energy was received. The seasonal temperature range is reduced over land areas adjacent to the ocean because of the large thermal inertia of the ocean.
The poleward flux of heat of the oceans is of the same order of magnitude as that of the atmosphere, but the regional and temporal variations of the oceanic transport are not well known. To under- stand the mean climate state of the world, we must understand the ocean's role in establishing and maintaining the global heat balance. Unless oceanic variability can be defined in terms of its departure from some mean state, we may be unable to explain the influence of the ocean on global climate. Models of the atmosphere with and without a moving ocean show that the ocean influences the mean atmospheric temperature distri- bution. The circulation of the ocean appears to affect climate variability on all scales. Ocean heat transport and storage processes have lifetimes that are long in comparison with those of the atmosphere. The ability to predict atmospheric variability from solely atmospheric models is now limited to 1 or 2 weeks. We know that significant variability occurs over much longer time scales, however. Ocean heat storage, transport, and transfer to the atmosphere are variable and may be the principal oceanic factors controlling these longer time scales of climate variability. Thus, long-range climate forecasting must take ocean processes into account. Interannual climate variation is the largest nonseasonal climate variation, and it may be larger in some regions than annual or seasonal variations. Year-to-year variations in the earth's climate are of great economic importance. Unusual rainfall, drought, or heat waves can have significant impacts on agriculture. Oceanic thermal
variations related to climate variability can affect marine fisher- ies. Thus, there are economic incentives for the development of a predictive ability. Climate variations with scales of approximately a decade and that produce important economic effects are known to exist but are less well documented than those with an annual time scale. There is evidence that the ocean plays a role in decadal climate variability, and proposals for large-scale ocean experiments to understand decadal variability have been made. Long-period climate variability, with time scales between decades and centuries, is not well documented. We wish to study climate variation through models of the coevolution of the ocean and atmosphere over time scales from months to geologic time. Oceanographers and meteorologists have proposed research programs to enhance our understanding of climate variability (Board on Ocean Science and Policy, 1983; Committee on Climatic Changes and the Oceans, 1983; Webster, 1984a). The two large-scale research programs are TOGA and WOCE. Parts of these programs are already under way, though plans for neither are complete. Much of the discussion of these programs in this report is taken from Webster (1984b). Interannual Variability of the Tropical Oceans and Global Atmosphere Experiment The Southern Oscillation is a large-scale exchange of atmospheric mass between the Eastern and Western Hemispheres in the tropics. It can be detected in sea level atmospheric pressure records as a
see-saw cycle of high pressure in the South Pacific Ocean and low pressure in the Indian Ocean, alternating with the opposite conditions in the other phase of the cycle. It has a characteristic cycle length of several years and may occur at two- to ten-year intervals. It is the most obvious instance of interannual climate variability. Associated with the Southern Oscillation are sea surface tempera- ture anomalies in the Pacific, Indian, and Atlantic oceans. Changes in the equatorial current system and the heat content of the Pacific Ocean are particularly marked. The largest such oceanic anomaly yet documented is El Nino, a sea surface warming off South America which brings destruction to the fisheries off Peru and Ecuador. Plankton, fish, and birds, which are dependent on a chain of nutrients normally provided by the upwelling of cold seawater off the coast, die when El Nino's warming occurs. This has economic effects on the global markets for fish, poultry, and fertilizer. El Nino also brings heavy coastal rains that cause flooding and damage crops along the coast of South America. Because of its strong signal and time scale, the Southern Oscil- lation has a significant effect on the climate. Though the Southern Oscillation occurs regularly, its occurrence has associated manifest- ations that normally persist for nearly two years from first to last appearance. This duration offers the potential to develop a predic- tive capability of perhaps a few months. The component phenomena of the Southern Oscillation normally occur at specific seasons of the year.
From the viewpoint of the United States, the correlations of the Southern Oscillation with North American climate anomalies present an intriguing challenge. Can the Southern Oscillation be used to predict wintertime climate anomalies over the United States a season in advance? Correlations between the Southern Oscillation and North American climate anomalies were first described in the 1930s by Sir Gilbert Walker. Since then there has been growing evidence of their reality. Wintertime temperature anomalies are correlated with earlier atmospheric pressure anomalies over the South Pacific and with sea surface temperature anomalies in the equatorial Pacific Ocean. During the winter of 1982-1983, the strongest El Nino event ever observed took place. It was not forecast, nor was it generally recognized as an El Nino occurrence until it was well developed, and its subsequent evolution and duration were not anticipated. Considerable research has been stimulated by this event, which under- lines the incomplete state of our understanding. The link between the tropical Pacific Ocean and the atmosphere has attracted considerable scientific attention. Likewise, the Atlantic and Indian oceans provide interesting but different examples of large-scale interactions between the tropical ocean and the global atmosphere. Atlantic Ocean sea surface temperature anomalies corre- late with droughts in Brazil. In the Indian Ocean, there is large seasonal change in response to the monsoon; for example, the Somali Current reverses seasonally. The Indian Ocean thus provides a unique location for studying some kinds of large-scale interaction between 8
the ocean and the atmosphere. Indeed, the early evolution of the Southern Oscillation appears to occur in the atmospheric circulation over the southern Indian Ocean. TOGA is principally an oceanographic program focused on the upper tropical ocean and overlying atmosphere. Its goals are to determine the nature of long-period fluctuations and their relationship to the global atmospheric circulation and to understand oceanographic and atmospheric processes that determine interannual variability. TOGA includes an oceanographic observation program aimed at describing the month-by-month variability of the upper layer of the tropical oceans over a 10-year period that began in 1985. One aim is to provide a 10-year data set of the atmosphere-ocean exchanges of momentum, heat, and moisture utilizing [n situ and satellite-derived observations. The integration of these observational programs will be a modeling component aimed at developing tropical ocean models and assessing the sensitivity of the global atmosphere to forcing by the tropical ocean. TOGA is an exciting research program. The Southern Oscillation is a strong climate signal that promises a predictive capability for climate variations in temperate latitudes. The economic benefits that could be derived from this capability would be great. A number of excellent scientists are enthusiastically working on the problem. Progress is being made in data analysis, field experiments, and theory. The first fragments of a comprehensive theoretical framework exist, and some linking physical mechanisms have been hypothesized. However, development of theory and improvement of the descriptive
base must proceed in tandem to design a full TOGA experiment with assurance. The United States is supporting and should continue to support a major TOGA experiment in the Pacific Ocean. Complementary research activities in the Atlantic and Indian oceans also should be supported, although other nations may play the principal role there. World Ocean Circulation Experiment We must understand the global oceanic circulation to understand the ocean's role in maintaining the climate state and in influencing climate variability. Without this knowledge we are unlikely to be able to predict future climate variations. WOCE is being planned to examine global ocean circulation and its relation to ocean climate processes. WOCE will be directed at des- cribing the circulation of the ocean, defining the linking physical processes in the ocean-atmosphere climate system, and understanding the sensitivity of that system to forcing by changes in the atmosphere. Recent oceanographic observations have revealed a number of processes that could be important to the ocean's role in climate variability: mesoscale eddies, tropical waves, isopycnal mixing, seasonal variation of the mixed layer, and mixing in the interior of the oceans. Computer models of the large-scale ocean circulation underline the importance of some of these. Thus, to observe and understand the circulation of the ocean, we need to describe the processes that are responsible in enough detail to model them. A major obstacle to describing the ocean is the difficulty of 10
obtaining measurements over long time scales and over great distances. Recent technical developments and new means of making measurements have made it feasible to consider carrying out a global experiment to understand the role of ocean circulation in climate. Orbiting satellites give promise of regular global measurements of sea surface temperature, surface currents, and wind stress on the sea surface. If these observations are combined with subsurface measurements, it should be possible to develop a description of the ocean that, for the first time, would begin to be as complete as our description of the atmosphere. The WOCE is the oceanographic component of the study of long-term climate trends and sensitivity of the WCRP. It is being planned for the 1990s. At the international level, the formulation is being guided by a Scientific Steering Group under the auspices of the Committee on Climatic Changes and the Ocean and the Joint Scientific Committee of the WCRP. The goals for the international WOCE are (1) to collect the data necessary to develop and test ocean models useful for predicting climate change; (2) to determine the representativeness of the specific WOCE data sets for the long-term behavior of the ocean; and (3) to find methods for determining long-term changes in the ocean circulation. Since the focus is on the construction of ocean models and the collection of data sets necessary for demonstrating that these are useful models of the ocean circulation, the determination of what these data sets should be is the crux of the WOCE design. The large-scale circulation of the ocean includes the mean and 11
variability of the physical properties of the ocean: velocity, density, heat, and chemical constituents. In order to define a feasible experiment, priorities must be established among the many possible measurements. Exchange with the atmosphere is important as a primary source or sink for most physical properties in the ocean. Atmosphere-ocean exchanges of momentum and density provide the primary dynamic forcing of the ocean. On the other hand, the forcing of the atmospheric general circulation by the ocean is primarily through exchanges of heat and moisture and, on longer time scales, radiatively active chemical constituents. Understanding the interaction of the ocean and the atmosphere is thus a significant part of WOCE. Large scale means space scales ranging from eddies to basins and time scales from seasons to the length of the measurement period. Keep in mind, however, that some important smaller-scale processes may have to be included by either parameterization or direct measure- ment. Thus, any claim to understanding must include knowledge of the sources and sinks, or forcing and dissipation, and the redistribution by internal fluxes of these physical properties on these scales. Measurement techniques that encompass large scales are thus essential to WOCE. WOCE will have a satellite observation program. This will include missions to measure sea surface topography by radar altimetry and surface wind stress by scatterometry. A global hydrographic and chemical survey by research vessels will provide data to define and constrain the global models. Chemical constituents, in addition to 12
being components of the general circulation in their own right, have many close associations with the field of motion. (They can be tracers of ocean currents and mixing, constituents of ocean density, and quantities whose storage and exchange affect long-term changes in the atmosphere; thus, chemical constituents are considered to be physical properties of high priority for understanding the long-term circulation.) Also included in WOCE will be strategically located jn situ instruments, for example, current meters and a global network of sea level tide gauge stations. Arrays of subsurface floats and surface drifters will contribute to the understanding of both surface currents and deep interior motions. Finally, measurements from volunteer observing ships will add necessary data on the thermal and velocity fields of the upper ocean and meteorological parameters for atmosphere-ocean exchanges. The large-scale circulation is complex, involving many space and time scales. It is generally agreed that quantitative and accurate simulations and predictions can be achieved only with four-dimen- sional (three space dimensions and time) numerical models running on the fastest computers. Therefore, the development and evaluation of such models will be a fundamental integrating component of WOCE. Planning for the U.S. component of WOCE is well under way. With- in the U.S., the primary scientific objective is to understand the general circulation of the global ocean well enough to be able to model its present state and predict its evolution in relation to long-term changes in the atmosphere. The specific objectives are as follows: 13
1. to complete a basic description of the general circulation of the ocean; 2. to determine seasonal and interannual oceanic variability on a global scale and the effect of such variability on ocean measure- ment strategies and the coevolution with the atmosphere; 3. to improve the basic description of the surface boundary conditions and the exchanges of physical properties with the atmosphere, and to establish their uncertainties; 4. to determine the interbasin exchanges in the global ocean circulation; 5. to obtain quantitative estimates of the large-scale exchange of buoyancy and chemical constituents between the upper boundary layer and the ocean interior by adequately describing the properties of the surface layer, including its horizontal mass transport and divergence; 6. to determine oceanic heat transport and storage in relation to the heat budget of the earth; 7. to determine the large-scale transport capacity for ideal tracers in the ocean; 8. to determine the important processes and balances for the dynamics of the general circulation; and 9. to improve numerical models for the diagnosis, simulation, and prediction of the general circulation of the ocean. The primary practical objective is to provide the scientific background for designing an observing system for long-term 14
measurement of the large-scale circulation of the ocean as a basis for operational climate prediction. Specific objectives are the following: (1) to identify those parameters and indices that are essential for long-term measurement in a climate observing system, and (2) to develop strategies for a cost-effective operational system for describing the large-scale circulation, including approaches for merging satellite data with jjn situ information. Internal Wave and Microstructure Studies Models of water mass formation and modification in the world's oceans are very sensitive to the magnitude and nature of diapycnal mixing, despite its apparent slow mean rate in the ocean interior. A prime example of this is the abyssal circulation deduced from the classical Stommel-Arons theory. Variation of the dependence of the diapycnal mixing coefficient on the stratification can actually reverse the direction of the mean flow. Based mainly on theoretical conjectures, there appears to be significant regional and vertical variations in the types and rates of mixing. Some of the leading candidate mechanisms are (1) intermittent billow turbulence due to shear instability induced by internal waves; (2) intrusions from benthic boundary layers mixed by tides and breaking internal waves; (3) deep convective mixing in the wintertime mixed layers; (4) fre- quent billow turbulence in strong current shear; (5) double-diffusive convection, and (6) sloping convection across fronts. The difficulty is that there are few unambiguous observations of diapycnal mixing. The statistical approach through microstructure measurements is 15
somewhat controversial. In some cases it has proved possible to diagnose the presence of mixing events from fine and microstructure measurements and even from large-scale (T and S) properties measured by CTDs. The following is a summary of emerging trends in internal wave and microstructure research adapted in part from D'Asaro and Muller (1984) and Caldwell (1983). Internal waves are viewed as an important link in the overall oceanic energy cascade from the large scales of generation to the small scales of dissipation. Although the dominant sources and sinks for internal waves have not been identified, the following concept is generally accepted: Energy enters the internal wave field at large scales and cascades down to small scales by nonlinear wave-wave interactions. When the shear reaches a critical value, the waves break and generate small-scale turbulence and microstructure. At microscales, energy is dissipated by molecular processes. Research that led to this picture has been dominated by the concept of a universal internal wave spectrum, an idea introduced a decade and a half ago by Garrett and Munk (1972). This concept of a universal spectrum is being challenged, whereas the link between internal waves and microstructure has been substantiated. Observed internal wave spectra usually fit the universal spectrum to within a factor of three for frequencies significantly above the inertial frequency and less so in the near-inertial band. The variations are likely the dynamic signatures of the sources, sinks, and internal transfers of the internal wave field. Numerous sources 16
and sinks have been proposed. Observational ly, however, no dominant generation or dissipation mechanism has been identified. Theoretical and observational evidence is emerging that the wind generates near- inertial frequency waves at large vertical scales and that internal waves and the mesoscale flow strongly interact. Classically, it has been assumed that the internal wave field dissipates its energy predominantly in the interior of the ocean, through small-scale turbulence. Calculations indicate, however, that the loss of internal wave energy at a sloping boundary might be substantial and could be the major energy sink of internal waves. Significant sinks of energy also may occur in critical layers when near-inertial waves become trapped within fronts or eddies. These losses would be concentrated at particular locations in the ocean and not spread uniformly throughout its volume. Studies of the relationship between small-scale turbulence and the internal wave field require measurements of both the turbulence, using microstructure instruments, and the internal wave shear and density fields, using larger-scale measurements. Existing evidence suggests that the internal wave field is highly random and, therefore, many measurements are required. There is hope, however, that systematic relationships between fine-scale Richardson numbers and microscale dissipation rates can be established. Systems capable of repeated measurements of both microstructure and internal wave scales have only recently become available and some are limited to use in only the upper few hundred meters of the ocean. Arguments developed in the last decade allow vertical diffusivity 17
for mass to be estimated from microstructure parameters; however, a clear picture of the three-dimensional structure and evolution of these mixing events has not yet emerged. Generally, mixing is envisioned as being caused by Kelvin-Helmholtz billows, which have been extensively studied in the laboratory and observed at one location in the upper ocean. A variety of other stratified shear flow instabilities with structures distinct from Kelvin-Helmholtz billows, such as wave breaking and critical layer absorption, have been observed in the laboratory and may also occur in the ocean. Turbulence research in other fields has benefited greatly from flow visualization studies that aim to identify the dominant structures of the turbulent flow. Once the structures of a flow have been identified in this way, they usually can be seen in point measurements. It seems likely that similar studies using dye or high-frequency acoustics would likewise increase our understanding of oceanic turbulence. One of the more enigmatic products of the direct microstructure measurement programs is the suggestion of a low level of turbulent mixing in the ocean. Effective vertical diffusivities estimated from microstructure data are of the order of 0.1 cm2/s or less, a result consistent with predictions based on universal internal wave models. The puzzle arises when comparison is made with tracer data from the deep ocean which suggest that the effective diffusivity is 1 cm2/s or larger (e.g., Munk, 1966). In fairness, it should be stressed that the direct microstructure observations have been collected in the upper 1 km of the ocean only, while the tracer study results reflect the abyssal water column. 18
In the absence of direct microstructure observations, several models have been put forward to account for the apparent higher level of mixing in the abyssal ocean than in the main thermocline. In one such model, it is argued that if the fine-scale shear field is limited by Richardson number, the effective diffusivity produced by shear mixing increases with weakening stratification. An effective diffusivity of the order of 1 cm^/s may then be possible in the weakly stratified abyssal ocean. A second school of thought suggests that boundary mixing is the dominant mechanism for cross-isopycnal fluxes, either by breaking internal waves or boundary layer turbu- lence. Through a combination of strong vertical mixing at sloping boundaries and lateral mixing in the interior, it has been suggested that apparent vertical diffusivities of 1 cm^/s are possible in the abyss. A challenge for the future is to collect direct microstruc- ture and internal wave measurements in the deep ocean to test these ideas. Internal wave research is presently undergoing a transition from a dominantly kinematic study of spectral slopes to a dominantly dynamic study of sources, sinks, and internal fluxes. The link between internal waves and oceanic turbulence is becoming more apparent, and the glimmers of a dynamic understanding are emerging. The parameterization of the internal wave and turbulent fluxes is a major goal of these studies. Further progress will come from simultaneous measurements of internal waves and microstructure and from a detailed comparison of experimental data with the results of numerical models. This work will require collaboration between workers in different areas of research. 19
Internal waves are only one candidate for generating small-scale mixing, and future attention can be expected to shift to several different convective mechanisms for mixing. One such topic will be the very intense mixing found in the deep, strongly cooled winter mixed layers. Warm core rings of the Gulf Stream have proven to be a good laboratory for studying this process. Recent work with rings has shown that the turbulence levels in actively cooled layers obey scaling laws similar to those of convective systems in the atmosphere, and that the heat and salt budgets of a cooling warm ring appear to require lateral mixing with surrounding waters. Hints of anticyclogenesis also are seen in association with convective cooling in rings. What are the dynamics of this process, and does it have implications for gyre-scale circulation? Since this process of deep convection is one that helps to establish the temperature and salinity of "mode" waters subducted into the thermocline, we must determine the relative roles of vertical entrainment and lateral mixing in our attempts to piece together a theory of the thermocline. In addition, much work will be required to discern the structure of convective plumes within these mixed layers. What are their morphologies and scales? Do they generate or interact with internal waves in the underlying thermocline? How is the horizontal momentum redistributed by these plumes? Other convective systems under active investigation are found in thermohaline staircases in the main thermocline. These series of thin, high-gradient sheets separated by well-mixed layers some tens 20
of meters thick are caused by double-diffusive mixing at the inter- faces. This is a field which has been well served by laboratory experiments which, in many cases, demonstrated the phenomenon that was later to be discovered in the ocean. The salt finger process has attracted more interest because a great volume of the ocean is unstable in the fingering sense (both temperature and salinity decrease with depth). It appears that there are regions where the vertical exchange in the staircases is sufficiently strong to be the dominant term in the heat and salt budgets. In such areas (e.g., the Caribbean Sea and the Mediterranean outflow), the strong vertical mixing appears to cause rapid horizontal changes in water mass struc- ture. However, these calculations are based on the extrapolation of laboratory flux laws down to the small salinity differences observed in the ocean. This could be a reasonable assumption, except for the fact that the ocean salt fingers must surely be affected by the ocean internal wave field. There is some evidence that the inertial wave field is much reduced in thermohaline staircases, but detailed measurements confirming this have yet to be made. In addition to field work in areas of strong fingering staircases, efforts can be expected to describe the weaker, more intermittent fingers found over great expanses of the gyres. Instrumentation is being developed that will be capable of detecting salt fingers on a more routine basis. As we increase our understanding of how fingers behave under distinct hydrographic conditions, we will be able to predict with some confi- dence the vertical exchange rates from simple T and S profiles alone. The other form of double-diffusive instability arises when both T 21
and S increase with depth. This is termed diffusive convection. Significant regions of the oceans near both poles are stratified in this sense, and the role of diffusive staircases in the heat budget of these areas may be significant. Most observations of diffusive steps have been made under the ice. The layers tend to be smaller than salt finger layers (a few meters versus a few tens of meters), and more modest vertical diffusivities are calculated (on the order of 1 cm^/s versus 10 cm^/s for the finger steps). Recent analysis suggests that it will be possible to model the mixing rate solely as a function of the ratio of the temperature and salinity gradients. Direct measurements of turbulence levels in such a system would be very valuable. As our parameter!zations of the double-diffusive processes become more accurate, we may be able to incorporate these processes into our numerical models. Simple one-dimensional models of the effects of salt fingers on the T-S relation already exist. Similar modeling efforts for the diffusive case can be expected. It should be possible to extend the models to two and three dimensions, and even up to the coarse vertical resolution of basin or global circulation models. Feedback from theoretical and numerical results will be an important check on the accuracy of suggested mixing rules. We also can expect further progress to be made in laboratory experiments on the structure and behavior of double-diffusive systems. Many experiments have worked with more strongly driven convection than is found in the ocean, but there is no fundamental difficulty in achieving conditions that more closely match those in 22
the ocean. The very sensitive microstructure instruments that have been developed for ocean work could be applied profitably to laboratory experiments. Another double-diffusive phenomenon that recently has attracted much attention is the thermohaline intrusion. Laboratory studies have demonstrated the ability of double-diffusive mixing to drive horizontal intrusions from water mass fronts. A variety of fine and microstructure studies of intrusions have demonstrated some of the predicted effects. Double-diffusion interleaving appears to be an important cross-frontal mixing mechanism, providing the necessary link between the mesoscale and the microscale. While there are already reasonable models of the large-scale effects of thermohaline staircases, proper parameterization of the intrusion process will require a great deal of further work. It is clear that, in addition to the acquisition of regional statistics on the intensity of interleaving, a substantial effort must be made in exploring the dynamics of the intrusions themselves. This will include further laboratory, theoretical, and modeling work, as well as large, well-coordinated field experiments using the most advanced mapping and microstructure instruments. Such field experiments could be served well by the use of Lagrangian techniques such as neutrally buoyant floats and tracers, both natural and introduced. In summary, internal gravity waves, direct convection, double diffusion, and small-scale turbulence are the motions by which the ocean mixes momentum and mass. The specific way in which this mixing is done has pronounced effects on geostrophic eddies, tracer 23
distributions, and the general circulation. To understand these grander scales of motions, we must understand the smaller-scale mixing processes. Continental Shelf and Slope Circulation A regional approach is required to understand shelf and slope circulation regimes because the relative importance of the various physical processes that influence shelf circulation generally vary from region to region. Observations made with modern instrumentation during the last decade have resulted in much-improved descriptions of the dominant physical processes that influence circulation over continental shelves (e.g., Allen et al., 1983), but much research remains to be done to understand and model accurately the physical processes that govern these regional flow regimes. Continued analysis of existing data, development of improved theoretical and numerical models, and further field experiments will be required. The circulation over the continental shelf and slope is forced primarily by surface wind stress, continental runoff, tides, large-scale currents, eddies, and wave motions from deep sea regimes which impinge on the continental margins. Continental shelves provide the sink for much of the ocean's tidal energy. In some regions tidal currents are sufficiently strong to cause local vertical mixing throughout the water column, to dominate sediment transport and the benthic community, and to drive vigorous residual currents. Continental runoff influences regional water properties and drives alongshore flows. Recent experimental programs have
focused on the role of local wind stress, particularly the alongshore component, in driving energetic low-frequency currents over the shelf. Initial results indicate that the degree of coupling between local wind and local current may vary considerably on a regional and seasonal basis. The ability of the continental margin to act as a waveguide for coastally trapped waves has been verified observation- ally. These waves can be excited by several mechanisms and may propagate large distances away from the generation region. The general subject of cross-shelf transports of, for example, mass, momentum, energy, salts, and nutrients remains poorly under- stood, though this is one of the most important aspects of shelf circulation. Cross-shelf surface Ekman transport driven by the alongshore wind stress is a fundamental element in dynamic models of the generation of wind-driven currents over the continental shelf, yet observations rarely show such a simple picture. Recent Eulerian and Lagrangian current observations obtained off California show a rich mesoscale structure in the flow field over the shelf, even during periods of strong alongshore winds and coastal upwelling (The CODE Group, 1983). The Lagrangian drifter observations suggest that the mesoscale circulation is much more energetic than previously imagined, and the estimated cross-shelf eddy diffusivity on the order of 10^ cm^/s indicates that cross-shelf eddy processes play a very important role in the cross-shelf flux of mass, heat, and nutrients (Davis, 1985). The coupling between flow on the outer shelf and the offshore current or eddy regime over the slope certainly needs additional study. Other important unknowns related 25
transports are the effects of density fronts and topographic irregularities and the differences between such transports in the frictional surface layer and the underlying water column. The alongshore variations on large (on the order of 100 km) scales of low-frequency, seasonal mean, and annual mean currents are likewise not well understood, but are of major importance. For some shelf regions, recent observations have yielded clear descriptions of the annual mean and seasonal variations in the alongshore currents. The driving mechanisms for even the mean circulation, however, have not been identified unambiguously. (For example, it is unclear to what extent the mean flow in the Middle Atlantic Bight is driven by an imposed offshore pressure gradient or an upstream source of shelf water.) Related to the dynamics of the mean and seasonal variation in alongshore flow are specific questions regarding the role in the alongshore momentum balance of the alongshore pressure gradient and bottom stress, two fundamental forces that are important in even the simplest theoretical and numerical models. Recent advances, both theoretical and experimental, in our understanding of the bottom boundary layer have changed our concept of the role of bottom friction. The generation of shelf currents by density differences caused by river runoff and the interaction of these density-driven currents with wind-driven currents are two related subjects that need additional study. In some cases, coastally trapped waves transport low-frequency kinetic energy along the margin. The effects of stratification, irregular topography, coastline orientation, mixing, and energy dissipation need further examination. 26
In summary, a few key questions were successfully pursued during the past decade: What is the physical mechanism of coastal upwelling? What role do continental shelf waves play in shelf circulation? By what mechanism is a steady-state pressure field maintained over an open continental shelf? What is the relative importance of wind, freshwater inflow, and coupling to deep water phenomena in driving shelf circulation? What is the role of a shelf-sea front in shelf-slope exchange processes? While none of these questions have been answered fully, the broad outlines of the answers are known as a result of the explosive growth of coastal oceanography. Follow-up work is essential in order to exploit fully the results of recent labors. Moreover, key questions remain to be answered; these have already prompted new research initiatives and are likely to guide the work of the near future. It seems clear that the main emphasis in the next phase of coastal physical oceanographic research will be the interaction of shelf circulation with water movements on the upper continental slope. The most vigorous kind of interaction occurs when a western boundary current occupies the upper slope â as is the case off the South Atlantic Bight and (intermittently) off the west Florida shelf. In the South Atlantic Bight much has already been learned. Generally speaking, the Gulf Stream imposes the pressure gradient which drives the outer shelf mean circulation. Important perturbations are caused by cyclonic eddies traveling along the Gulf Stream front. The properties of these eddies are known descriptively, but they are not dynamically understood. Similar eddies have been discovered in the 27
Gulf of Mexico. The circulation of the outer west Florida shelf is under the influence of these eddies, and it cannot be conceptualized without first understanding the cyclonic eddies. Large anticyclonic eddies (warm core rings) impinge on the continental slope in the Middle Atlantic Bight and in the western Gulf of Mexico. Their behavior over sharply sloping bottom and their interaction with shelf circulation are not understood. We do know that they generate strong currents ( ~1 m/s), carry small cyclonic eddies on their perimeter (which on satellite pictures look like the arms of a spiral galaxy), and give rise to strong bursts of inertial oscillations in their neighborhood. More generally, seas over continental slopes are likely to prove to be a unique oceanic environment because of the strong vorticity constraints that a steeply sloping seafloor imposes. One hypothe- sized consequence is the relative quiescence of the current over mid-slope (depth range, 1-2 km), except of course in the presence of a western boundary current. Quiescent mid-slope regions might well constitute an important worldwide sink for anthropogenic pollutants, as well as for the fine particulate matter that is naturally present in the ocean. Improved descriptions of the foregoing phenomena and processes will likely require additional field measurements. To rationalize the descriptions, there are critical needs for the development of (1) improved theoretical and numerical models incorporating the different dynamic processes important in shelf circulation, and (2) objective methods to test model predictions against observations 28
