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.
4 Conclusions and Recommendations IN~ODUCl-lON The Physica! Oceanography Pane! of the Committee to Review the Outer Continental Shelf Environmental Studies Program was formed to evaluate physical oceanographic aspects of the OCS oil and gas leasing program. In completing its review, the panel considered four basic subject areas: 1. The acquisition and use of physical oceanographic information by the ESP and use of the information by BEM and in EISs; 2. A review of the state of knowledge of general physical oceanographic processes that are most important for understanding and modeling the motion and fate of oil spills in the ocean; 3. A review of the state of knowledge of the physical oceanography of each of the ESP regions, based on all available sources; and 4. An evaluation of the adequacy and applicability of each of the ESP regional physical oceanography programs and the Washington office generic studies program, as measured by (a) the success of the field programs and modeling efforts in meeting ESP needs, (b) contributions to the general state of physical oceanographic knowledge, and (c) interactions with other agencies and the scientific community in the region. This chapter presents a brief summary of the current role of physical oceanography in the ESP, followed by the conclusions of the panel s review and its recommendations for future ESP physical oceanography studies. THE ROLE OF PHYSICAL OCEANOGRAPHY IN THE ESP Physical oceanography and meteorology provide the basis for calculating estimates of the transport and fate of oil spills in the ocean. These calculations in turn provide the basis for estimating potential impacts of of} spills on resources. Projection of potential oil-spill impacts is based on results generated by the OSRA model that was developed by the USGS in 1975 to model oil-spill trajectories. Because potential risks are principally evaluated through OSRA modeling, the primary use of physical oceanographic information within the ESP has been in support of the OSRA model and in preparation of associated EIS documents. Physical oceanographic information has also been used to support biological and ecological studies, and to predict the transport of drilling muds and cuttings and other byproducts of oil exploration and production. The Physical Oceanography Pane! has concentrated its review on the needs and uses of physical oceanographic (and associated meteorological) information as input to the OSRA model. Most oil-spill models, including the OSRA model, require physical oceanographic and meteorological data for environmental input. Definitions of the wind, current, temperature, and ice fields (if present) in space and time constitute the typical environmental input parameters. 95
96 PHYSICAL OCEANOGRAPHY OF THE U.S. OUTER CONTINENTAL SHELF These data can be derived from other meteorological or oceanographic circulation models or from field observations, but in either case, accurate spill predictions require high-quality environmental data that are carefully integrated into the spill model. The environmental data are subsequently used in the algorithms that calculate probable spill trajectories and fates; the version of the OSRA mode} currently used to support OCS leasing decisions does not include any calculation of oil-spill fate, but an experimental version (under development) does incorporate fate calculations. In present usage, physical oceanographic input to the OSRA model is provided through the predictions of regional circulation models for all of the regions. These predictions have been provided in the form of mean climatological flows over various time scales at spatial resolutions of typically 15 to 30 km, with the principal forcings being the mean density field and the seasonal mean wind stress. Recently, MMS has begun to specify time-dependent velocity fields. In the Atlantic, Pacific, and Gulf of Mexico regions, until 1989, meteorological input has been provided in the form of transition-probability matrices of wind speed and direction, calculated from observed winds over long periods of time at selected coastal and offshore buoy stations. In 1989, MMS began to use the output from meteorological models for wind fields instead of transition-probability matrices (pers. comm., MMS, 1990~. The OSRA model then uses a Monte Carlo technique to calculate spill trajectories for selected launch points in a proposed lease area by season or month. In the Alaska region, meteorological input has recently been provided through the use of the- Fleet Numerical Oceanographic Center weather model, which estimates spatially and temporally varying winds in Alaskan waters, assimilating observed wind and pressure fields when these are available. In Alaskan waters, the method of predicting spill trajectories depends on the contractor. In all cases, the OSRA model then calculates the hits, or number of times a spill encounters an environmental resource target or shoreline segments, using historical data and the conditional impact probabilities on the resource within a preselected time. CONCLUSIONS Acquisition and Use of Physical Oceanographic Information MMS-funded studies have contributed to the dramatic increase in knowledge of the oceans that has occurred during the past decade. These contributions have included the development of circulation models and the observational study of circulation patterns. MMS- funded physical oceanography studies have fit in well with those of other agencies. This improved information has resulted in many excellent summaries of the relevant physical oceanography occurring in the "Description of the Affected Environment" section of the EISs. The panel has noted several discrepancies between the physical oceanographic information that is potentially available as input to the OSRA model, either through MMS-funded studies or through cooperation with other agencies, and the information that is actually used. The most important of these are summarized below: 1. In reviewing the oilspill trajectory and environmental resource impact modeling performed by or for MMS, it was evident that the general strategy is to rely extensively on the use of model-derived results to estimate the circulation for a given area. The use of circulation data sets based solely on field observations or derived from a melding or assimilation of field observations and model results appears minimal at present. This same discrepancy was noted in reviewing the regional programs; in general, it was found that the physical oceanographic field studies carried out for MMS have been extensive, with the exception of those conducted in the Gulf of Mexico, but that the data collected have been underused. Circulation- and trajectory-model results are ultimately integrated into EISs for lease sales through OSRA predictions. How the results of the large-scale physical oceanographic field programs influence the EIS preparation process is not clear. The information derived from the observational programs appears to be used along with the model results to describe the circulation that is ultimately summarized in the EIS. The information also appears to be used by other investigators (biologists, chemists, and geologists, for example) to assist them in interpreting and
CONCL USIONS A ND RECOMMENDA TIONS analyzing their data. However, although observational data sets and associated interpretation could be used to improve, calibrate, or validate circulation model predictions or to provide an independent source of data to describe the circulation for input to the OSRA model, they have seldom been used for these purposes. 97 2. A related problem is that the best available circulation-mode! results in the OSRA calculations have not always been used, notably in several calculations done in the Atlantic region (including the OSRA calculation for the Georges Bank EIS). 3. Studies funded in physical oceanography have tended to be of two types: (a) large- scale, multiyear observational field programs with associated data analysis and interpretation, and (b) numerical modeling studies of the circulation of major shelf and adjacent deep-water areas. Rarely are the two study types combined. This separation has made the integration of modeling and field program results difficult and has hampered the most efficient use of either type of study. 4. The wind fields used in calculating spill trajectories for the Atlantic, Pacific, and Gulf of Mexico regions have been inconsistent with the wind fields used to drive the circulation models. The use, until recently, of transition-probability matrices based on observations at a limited number of stations could have led to inaccuracies in the resultant trajectory calculations, especially when spatial variability is addressed by selecting discrete zones over which a given station and its associated transition-probability matrix are assumed to apply. Recent detailed analyses of winds observed onshore and offshore found large spatial and temporal variability in the structure of the wind field and the coherence between onshore and offshore stations. This variability is greatest near the coast, where land-sea interaction is an important factor. Physical Oceanographic Models and Processes of Importance to the ESP MMS's present and probable future reliance on numerical circulation models for physical oceanographic input to the OSRA model makes it imperative that the strengths, weaknesses, and limitations of the modeling approach be fully understood. This is true whether the circulation models are used in a predictive mode or for spatial and temporal extrapolation of observed data. In addition to considerations of the accuracy of the models used, attention must be paid to the time and space scales of motion that are required for accurate trajectory simulation. Also, an accurate representation of the physical processes that contribute to weathering is needed for the development of oil-spill-fate models. In considering these problems, the panel reached several conclusions about the present state of numerical circulation modeling and the physical processes that must be represented for accurate modeling of the transport and fate of oil spills in the ocean: 1. MMS puts too much faith in the available circulation models. Although the models used by MMS contractors often represent the state of the art, the state of the art does not justify MMS's implicit trust. Verification, intermodel comparison, and sensitivity studies are needed. Areas for possible improvement that are common to most numerical models have been identified in this review, including the parameterization of subgrid-scale processes, in vertical and horizontal dimensions; construction and use of appropriate lateral open-boundary conditions; better incorporation of driving forces at the surface (e.g., wind, heating, and cooling) and at the coast (e.g., riverine inputs); and incorporation of data directly into the model, to derive full benefit from a set of observations. In all cases, it is important to test fully a model against observations and to understand its behavior. Techniques to accomplish these often difficult tasks quantitatively are becoming available but are too seldom used. Numerical general circulation models alone will not simulate the circulation with enough realism for trajectory prediction or estimation of trajectory statistics in the next few years. Model studies need to be supplemented with more field observations. Trajectory predictions or estimations of trajectory statistics realistic for use in risk analysis or in accident management cannot be obtained without new fieldwork, including drifter studies. Although these observations ultimately might help the development of circulation models and it is hoped that they will be in quantitative agreement with existing data MMS's immediate priority should be to take more field observations and to incorporate them into trajectory predictions. -~7-
98 PHYSICAL OCEANOGRAPHY OF THE U.S. OUTER CONTINENTAL SHELF 2. Although physical oceanography does not usually include considerations of the behavior of oil and other contaminants per se, oil-spill-fate modeling is quite important for accurate prediction of oil-spill behavior. The panels review has concentrated on spreading and dispersion of oil, since they are the most closely tied to near-surface physical oceanographic processes. Spreading is one of the most important processes in oil-spill dynamics, because it determines the areal extent of spilled oil and affects the various weathering processes influenced by surface area. Dispersion is generally assumed to result from wind-generated breaking waves dispersing oil in the water column. Both processes are as yet poorly understood, but both depend critically on the interactions of the wind, the surface wave field, the response of near-surface waters, and the use of chemical dispersants. Thus, the incorporation of oil-spill-fate models into the OSRA framework will require accurate representation of physical processes occurring at and near the surface. 3. As a consequence of surface concentration of oil in oil spills and relatively rapid weathering, the principal physical oceanographic problems are understanding and predicting the motion and fate of of! in surface waters over periods of up to 30 days. During this period, several physical oceanographic factors are important to the net motion and variability of spill trajectories. Primary among these is the response of surface waters to wind forcing, including spatial and temporal variability of both the wind field and the surface response. Variability of underlying currents in this time frame also is important, but to an unknown degree relative to direct wind forcing. Mesoscale oceanic features are principal examples of this variability, affecting the motion of water over the OCS and slope and the exchange of water and material across the shelf break, although they do not appear to influence currents over the inner and midshelf. Important examples of mesoscale motions can be found on all U.S. margins: Gulf Stream meanders and filaments in the South Atlantic Bight; warm core rings in the North Atlantic Bight; the jets, squirts, and filaments of the California Current system; eddies of the Loop Current in the Gulf of Mexico; and (probable) eddies of the Bering Slope Current in the Bering Sea. The current practice of specifying only seasonal mean currents for input to the OSRA model does not properly account for the variability that may be associated with these mesoscale motions. 4. Mean flows do not contribute much to the variability of oil-spill trajectories but can be responsible for substantial advection of oil. Mean flows are often observed to run counter to the direction of the mean wind stress, often in the direction of coastal-trapped-wave propagation. There are several possible driving mechanisms, including undercurrents of the wind-driven circulation, density-driven flows due to input from rivers of fresh water at the coast, response to alongshore pressure gradients in the adjacent deep ocean (probably not very important), and nonlinear rectification of current fluctuations, among others. It is likely that different mechanisms are dominant in different locations, but it is not yet possible to determine causality for any given location. Prediction of seasonal mean currents based on the seasonal mean wind stress and the density field alone may not yield correct results. 5. At higher frequencies, surface and internal tides and higher-frequency internal waves generated by the interaction of tides and topography are a ubiquitous feature of continental shelf flows. Alongshore tidal currents may not be a dominant component of the alongshore variance, but can be highly amplified by local topography. Cross-shelf tidal currents are often the dominant component of the cross-shelf variability. Rectification of tidal currents is the dominant component of the mean flow in some locations. Tides and the internal motions generated by tides may be an important mechanism for mixing, especially over the inner shelf and at the shelf break. In addition, surface convergences associated with internal waves concentrate buoyant material quite effectively. The importance of these smaller-scale processes for the motion of oil spills needs to be investigated, especially near shore. 6. Strong storms produce major perturbations in transport and mixing in continental shelf waters, but storm-induced flow and mixing are poorly understood. According to Allen et al. (1987), The effect of the large currents, mixing, and the transport associated with the storms on the shelf budgets and on the transport of material are important unsolved coastal problems. Coupled meteorology and physical oceanography programs [are] needed to understand the detailed cyclogenesis and subsequent meteorological forcing. The effects of storms are particularly important for near surface-mixing and transport, including the motion and fate of of! spills.
CONCLUSIONS AND RECOMMENDS TIONS 7. The details of the structure and dynamics of the surface mixed-layer are very important to understanding and predicting the behavior of oil in near-surface waters, as well as the transfers of mass, momentum, and heat across the ocean surface and down into the ocean interior, and the structure of the surface and near-surface velocity fields. Research on the dynamics of the oceanic surface layer has resulted in significant advances over the past decade, but there are still important questions that have not been answered or even adequately addressed. Current parameterizations for the moisture, momentum, and heat fluxes across the air-sea interface are workable, but there is little consensus on the actual physical processes that control the fluxes. Candidate processes are surface wave breaking, Langmuir cells, and shear-generated turbulence, among others. It is also well known that upper-ocean processes are intrinsically three-dimensional, and that the three-dimensionality can have important consequences for all of the important transfer processes in the surface layer, especially in the vicinity of fronts associated with upwelling and other larger scale processes. Thus, two-dimensional models are of limited use. The models used for oil-spill transport and fate calculations need improved parameterizations for subgrid-scale variability in surface layer processes in both the vertical and horizontal dimensions. S. Cross-shelf flows, although particularly difficult to measure, are of direct importance to cross-shelf exchange processes and often provide a clearer diagnostic signal for model and data comparison than do alongshelf flows. 9. Direct measurements of Lagrangian motion with drifters, although few, have indicated that the Lagrangian mean transport may not match transports derived from Eulerian measurements. This is certainly the case in the presence of strong horizontal gradients in velocity, but the problem also may depend on season and location in some unknown way. Since oil spills are inherently Lagrangian, whereas the circulation fields used as input to the OSRA model are Eulerian, the problem is of direct relevance for spill trajectory calculations. 10. The presence and dynamics of ice are clearly an important feature for modeling oil trajectories in most Alaskan waters and some nearshore New England waters during especially cold winters; ice conditions are highly influential in determining the movement and final disposition of spilled oil. The panel concluded that the problems of modeling sea ice and ice-oil interaction were truly at the process level and not simply at a descriptive level for the Alaska region. Several well-developed ice models are currently available for application to the Alaskan OCS waters, but there is little sense of.how well these models actually perform in representing the range of ice conditions for the OCS areas; verification against data and sensitivity studies are needed. Interactions between oil spills and sea ice are extremely complex, depending on the percentage of ice cover, ice motion, temperature, wind, duration of ice cover, and the history and location of ice-of! contact. Several researchers have attempted to quantify and predict some aspect of oil-ice interactions, but most research has been carried out under laboratory conditions only and thus is limited in either dimension or scale; more research is needed. Extension of spill models to handle oil-ice interactions has been extremely limited. A fully coupled ice/hydrodynamics model should be applied to selected areas and times for which data are available, and a detailed model-data comparison should be done. In addition, an extensive series of sensitivity studies should be performed to determine those parameters critical in controlling the solution (stress-strain relationship, elastic, plastic, viro-elastic, type of yield surface, air-ice drag coefficient, ice-water drag coefficient, boundary conditions, initial conditions). Such studies will document the present ability to predict ice motion and probably, and most importantly, indicate where present understanding is weak and needs improvement. Some of the conditions that should be reproduced in ice models include the location of the pack-ice edge and the various percentages of ice coverage of areas versus time (including interannual, seasonal, and short-time-scale, wind-forced variability); occurrence, dynamics, and causes of breakout events through the Bering Strait; the movement of the pack ice; the ice thickness distribution; the freezing and thawing processes at the ice edge and in leads; and the response of ice in free drift motion to wind and current forcing. 11. Studies of bottom boundary-layer and sediment-transport processes on the continental shelf are germane to the overall goals of the ESP in two ways. First, the drag exerted by the bottom on the flow, and the structure of the flow within the bottom boundary layer, are important for understanding the overall structure and dynamics of flow on the continental shelf; 99
loo PHYSICAL OCEANOGRAPHY OF THE U.S. OUTER CONTINENTAL SHELF this is increasingly true as the depth decreases. Recent advances in boundary-layer modeling have indicated that the nonlinear nature of the turbulent bottom boundary layer necessitates inclusion of motions at higher frequencies and consideration of stratification effects for accurate calculation of bottom stresses. Second, sediment-transport studies support concerns about both the long-term effects of oil incorporated into bottom sediments and the motion of drilling muds and cuttings. In the view of the panel, however, bottom boundary-layer and sediment-transport studies are outside the immediate domain of this review. The focus of this review and of ESP physical oceanography in general is the motion and fate of oil in surface and near-surface waters. Accurate modeling of bottom stress and boundary-layer velocity remains an important consideration, particularly because of their effects under extreme conditions, but at a secondary level relative to understanding the processes that directly influence surface flow and mixing. The transport of drilling muds and cuttings is most likely limited to the immediate vicinity of drilling activity for the coarse fraction, which makes up about 90% of the effluent (see, e.g., NRC, 1983~. Long-term ecological effects of oil incorporated into the sediments may be important and may depend on sediment-transport processes, but the panel felt that these considerations are properly the domain of the Ecology Panel. Regional Oceanography The continental margins covered by the four regional offices of MMS's ESP are distinguished by far more than MMS's internal division of responsibility. The four regions have fundamental differences in their geology, topography, and bathymetry and in the processes that control the circulation of shelf waters and affect the motion of oil in surface waters of the regions. These differences are apparent even between subregions within each region. In general, the physical oceanography of all of the major continental margins is reasonably well understood, especially from a basinwide, descriptive point of view. There are exceptions for specific areas (e.g., the shelf of the Gulf of Mexico), and the details of specific processes that are active in the various regions are still under investigation. In addition to the mesoscale oceanic motions identified above, there are some general circulation processes in each of the regions that require particular attention. Briefly, these are as follows: 1. In the Alaska region, the presence of sea ice, which varies from season to season and from subregion to subregion, is an important factor. Large interannual variability associated with ENSO events has also been observed. Sea ice and interannual variability combine in the Bering Sea to create a large interannual and seasonal variability in ice cover, temperature, and winds. Two topics worthy of investigation are interannual variability in all the Alaskan waters (Gulf of Alaska, Bering Sea, and Beaufort Sea-Chukchi Sea-Arctic Ocean) and the circulation along the Bering Sea continental slope. The subsynoptic-scale weather associated with orographic effects is important, particularly in the Gulf of Alaska and along the Beaufort Sea coast, but no general algorithms and approaches exist to define the winds in these areas. The FNOC model does not adequately address this problem in that it does not include realistic topography. In addition, these winds tend to be at subsynoptic or subgrid scales. Further modeling of the effects of orographic winds would be useful. Freshwater runoff is a major forcing mechanism for circulation in the Gulf of Alaska. 2. In the Pacific region, the process of wind-driven coastal upwelling is particularly important, because the winds are predominantly southward throughout the summer over most of the coast. Interannual variability associated with ENSO events in the tropical Pacific has been identified, but interconnections between coastal, oceanic, and atmospheric responses are not well understood as yet. 3. In the Gulf of Mexico region, 21 major estuaries are found on the U.S. coast. A number of rivers dominated by the Mississippi River provide sufficient freshwater input so that the basin may be classified as a positive estuary. An additional, very important factor for
CONCLUSIONS AND RECOMMENDATIONS 101 oil-spill-risk analysis is the regular presence in the gulf of major tropical storms and hurricanes that can cause great damage and may result in rapid, extreme, wind-forced surface transport. 4. In the Atlantic region, storms are common in the winter, with cyclogenesis occurring south of Cape Hatteras and intense nor'easters occurring over the Mid-Atlantic Bight. Tropical cyclones and hurricanes may strike many of the offshore sites along the Atiantic coast. A distinct feature of the Atlantic shelf regions is a seasonal shelf break front, which separates well-mixed water over the shelf in the winter from the slope waters in the Mid-Atlantic Bight and from Gulf Stream waters in the South Atlantic Bight. The front is weakened in summer by uniform thermal stratification in the surface layers. The contributions of shelf-break frontal dynamics to cross-shelf exchange and to the concentration and transport of surface buoyant material are poorly understood. Evaluation of Regional Programs In evaluating the physical oceanographic components of the ESP of the four regional offices and the WO, the panel noted several general tendencies. Those not mentioned above are summarized below: 1. In general, MMS-funded physical oceanography studies have fit in well with studies funded by other agencies (e.g., USGS, NOAA, and NSF) and have contributed to the increase in our knowledge of the coastal oceans that has occurred over the past decade. 2. Although there are variations from region to region, too little of the work carried out for MMS has been published in the open, refereed literature. This is particularly true of the modeling and model-data intercomparison studies. Publication in the open literature would both improve quality control of MMS-funded efforts through the peer-review process and substantially increase the body of knowledge available to the oceanographic community, at little extra cost. MMS's recent efforts in this direction are commendable. Specific conclusions for the regional evaluations are as follows: 3. In the Alaska region, more of the observational study results appear to have been published in the open literature compared to other regions; some of the modeling work has been published, but not enough. The general state of knowledge appears to be good, given the enormous size of the region. More site-specific observational studies will probably be required once of! production starts. The recent Alaska region modeling efforts have incorporated improved techniques, especially with regard to use of temporally and spatially varying wind fields that are consistent between the hydrodynamic model and the spill trajectory model; this is due partially to the practice of the contractor performing spill trajectory calculations. 4. There have been a number of large studies in the Pacific region, both observational and modeling, but to date the contributions by the West Coast programs to state-of-the-art knowledge, as measured by the number of refereed publications, have not been impressive. Only one refereed journal article on an observational study has appeared, and the lead author on that paper was funded by NSF, not MMS (Brink and Muench, 1986~. The potential return from these experiments has not yet been realized, possibly because contractors have not employed sufficient outside (university) personnel and have not spent sufficient resources on detailed analyses. The modeling work, if published, could make a contribution by illustrating the inadequacies of a state-of-the-art numerical model; that is, deficiencies can provide important clues as to the physics that might be missing and other model shortcomings. The circulation model used in all West Coast modeling studies is a state-of-the-art model, but the modeling studies have had serious inadequacies. 5. In contrast to the other regions, there have not been many extensive basic research programs in the Gulf of Mexico, in spite of the fact that the gulf has had by far the greatest amount of acreage leased and the greatest amount of oil produced. In particular, no major shelf circulation studies have been conducted, although there have been several small-scale investigations. There have been few useful results of MMS-funded modeling efforts, except for
102 PHYSICAL =~N~ OF THE U.S. OUTER CO=INE=^ SHELF recent efforts using the Hurlburt and Thompson two-layer model. These modeling efforts have led to an improved understanding of the major circulation features in the gulf, and of the boundary conditions required for modeling the shelf circulation, a task that remains to be done. In addition, the long history of offshore drilling and large numbers of active leases in the Gulf of Mexico make long-term impact studies crucial. It is clear that the MMS Gulf of Mexico regional office cannot afford to fund such extensive studies without significant additional support. 6. MMS has funded many large physical oceanographic studies in the Atlantic region, both observational studies and modeling. In most cases, the observational studies have provided fairly high-quality data on the mean and variable circulations in the various areas. The analyses of the individual data sets by the subcontractors are competent; however, there is no quantitative synthesis of the various types of data to produce a dynamically consistent description of the flow, and it does not appear that much of this information (or any synthesis work) has been published in the reviewed literature. There appears to be little relationship between the modeling work and the data collection. The reports do not reference each other and the modeling does not appear as a subcontract in the same project as the observational work (with the exception of the Mid- Atlantic Slope and Rise Study). The data used in the models comes from the historical record: verification of the models has been minimal. 7. The efforts of the WO, in contrast to the regional offices' goals of data collection, analysis, and synthesis, are focused on supporting regional studies, addressing generic process and modeling issues, and summarizing or documenting previous studies programs. The number of physical oceanographic studies funded under the WO is extremely limited, but according to the material available, those studies completed under the WO have addressed areas of real concern and have been completed with quality products in a timely manner. An important question is why the WO budget is so small compared to its regional counterparts. There are several important generic research efforts that would clearly seem appropriate for the WO under the generic studies category that have in the past been carried out by the regional offices. It appears that these studies could have been better directed and more efficiently executed, with results that would have been more widely useful, if they had been conducted under the aegis of the WO. The mandate to complete these overview or generic efforts belongs with the WO, and the management structure and funding allocations should clearly reflect that. RECOMMENDATIONS There are three general recommendations of the panel for future ESP physical oceanography and oil-spill studies, each of which has several associated specific recommendations: 1. The Minerals Management Service should support continuing research on relevant physical oceanographic and meteorological processes anti features that are poorly understood, poorly parameterized in existing models, or poorly represented by e~cistiI'g modeling methodology. Improvements should continue to be incorporated into the OSRA model. a. MMS should support continuing investigations of surface-layer physics, aimed at improving basic understanding and modeling. The detailed physics of the surface layer is poorly understood, and many potentially important physical processes are not included in oil-spill-trajectory models as a result. These include surface and subsurface fronts and convergence zones, Stokes drift, Langmuir cells, shingles, interleaving, and breaking wave dynamics. It is not known how these factors might affect oil movement, but it is time that the "3.5% rule" or one of its variants be replaced by a more accurate description of near-surface drift. (MMS accounts for wind-induced drift by assuming it to be 3.5% of surface wind velocity with a variable drift angle to the right). Advances in this area will also help to better approximate near- surface physical processes that control the dispersion of oil into the water column. The dispersion models currently available give order-of-magnitude estimates for oft incorporation into the water column but are not sufficient to provide accurate, quantitative
CONCL USIONS AND RECOMMENDS TIONS 103 predictions. New fundamental insights into the near-surface processes are required to provide improved estimates of these fluxes. b. Understanding and modeling oil-spill-fate processes are very important to the leasing and environmental assessment studies program. MMS should continue to support research in these areas. Although considerable MMS funds have been invested in such studies to date, the return in terms of improved algorithms to estimate oil-spill fate has been minimal. Fate processes that deserve particular attention are drifting, spreading, dispersion (naturally or chemically enhanced), and dissolution. Studies on actual large spills, such as the Exxon Valdez spill in Prince William Sound, would be useful in this regard. The interaction of oil and ice, particularly in freezing, thawing, and partial ice-cover conditions, also needs further research. c. Additional studies of sea ice, both modeling and observation, are needed. Better representations of ice thickness ant! constitutive relationships need to be explored to determine whether these give better estimates of ice motion. Detailec! model-data comparisons need to be done, and sensitivity studies are required to determine the parameters that are critical in controlling mode! solutions. d. Calculations of oil-spill trajectories at the sea surface and those for the underlying currents should be consistent. The MMS practice, until recently, of using wind patterns that are not related to the winds driving the water motions may have led to substantial errors in estimating the oil motion. The same considerations apply to calculations of ice motion in Alaskan waters. In addition, MMS should be sure that sufficient trajectory simulations are performed to develop stable, robust statistics for impacts on environmental resources. This requires that sufficient trajectories be calculated such that the distribution of impacts is independent of the number of trajectories simulated. e. The meteorological input to oil-spill-trajectory simulations needs to be improved to account correctly for the spatial and temporal structure of the wind field. The use of transition-probability matrices, either based on wind time series or weather patterns, as used until recently for many OCS areas, does not yield accurate temporal correlation and has little or no spatial structure. A meteorological model (e.g., limited-area fine-mesh), with appropriate interpolation to smaller scales, is one way to address the need for an accurate and consistent definition of the wind field. Such a model would help to accurately reproduce mean, storm, and extreme events that characterize the meteorology of an area. In Alaskan waters, orographic effects require particular attention. MMS's recent moves to adopt such procedures are commended. f. More consideration needs to be given to extreme events (e.g., hurricanes) that may leaf! to both higher spill probability and more rapid water and oil motion. The present MMS procedure is least accurate at low probability levels associated with extreme events. However even extremely small impact probabilities may be critical if important resources are affected. MMS might consider studying several extreme-event scenarios as special case studies for each lease area. g. Future trajectory simulations should incorporate a methodology to address the inherent variability in both the wind field and the current field at subgrid space and time scales. For each realization, the resulting random contribution to the trajectory and effective dispersal of the oil needs to be included or evaluated. h. MMS's study of the use of drifters to represent oil spills should be extended to actual field trials in varied areas of the OCS, and similar regions worldwide as the opportunities occur. The drifters must reasonably represent oil movements. Experiments should be fairly extensive, in order to acquire sufficiently large data sets to test models. Multiple deployments of ARGOS-tracked drifters from selected high-risk launch points should be compared with model-predicted trajectories. This comparison will provide an excellent test of the ability of the model to predict trajectories and, of equal or greater importance, help to identify model weaknesses. 2. MMS should reduce its present overreliance on model results until the models can be more fully tested and verified; such testing will require sensitivity analyses and model intercomparisons. MMS should use its extensive observational data base more fully. Verification will require close cooperation between field scientists and modelers in all future MMS programs.
104 PHYSICAL OCEANOGRAPHY OF THE U.S. OUTER CONTINENTAL SHELF a. It is recommended that MMS use a more balanced integration of model and field-data products for future trajectory calculations. MMS has adopted a de facto procedure of using hydrodynamic model simulation results to describe the current fields for input to oil-spill-trajectory calculations. Unfortunately, this strategy has ignored the extensive data collected in physical oceanographic field programs and the subsequent analysis and interpretation of that information. The possibility that numerical models may never provide a simulation of the circulation adequate for determining trajectory statistics needs to be considered. b. For scientific credibility, it is imperative that detailed sensitivity studies be carried out for all modeling work. Hydrodynamic model parameters that should be studied include boundary conditions, initial conditions, eddy viscosity and turbulence representations, grid size, bottom friction coefficients, and forcing functions. For oil-spill-trajectory models, the relative roles of wind and current forcing should be assessed, at a minimum. Without some understanding of the factors and processes that limit the oil-spill-motion calculation, it is difficult to either assess or improve the modeling. These sensitivity studies will also allow insight into model strengths and weaknesses. c. Systematic model intercomparison and verification must also be carried out for scientific credibility. MMS should have an intercomparison study performed between the various hydrodynamic models used for predicting current fields, to allow model strengths and weaknesses to be documented and to enable rational scientific evaluation of the various approaches. Verification against data-including data from actual spills must be much more thorough. Systematic model evaluation and verification against all available data, for both currents and oil-spill trajectories, should be routine. More attention needs to be paid to cross-shelf currents in this process, because cross-shelf currents are more sensitive to model assumptions than are alongshelf flows. d. Future MMS-sponsored physical oceanographic field programs should require close cooperation between field scientists and modelers. More input from field scientists is also needed in model design, application, and verification. Physical oceanographers engaged in field work and modeling should work closely together to design observational and numerical experiments so that necessary and sufficient initial, boundary, and updating data, as well as critical data sets for calibration and verification, are all obtained. Consideration should be given to the latest techniques for obtaining the required amount of observational data. For example, use of aircraft and spacecraft data from NASA and NOAA could prove of value. e. Given the present level of understanding for most shelf regions of interest to the OCS leasing program, a carefully integrated program using field observations ant! numerical hydrodynamic modeling is suggested to provide a description of the circulation necessary as input to the OSRA model. The exact details of a modeling strategy to be used are critically dependent on the principal forcing and circulation features of the shelf area of interest. A practical general strategy to achieve representative circulation data would include the following: (1) Predict the circulation associated with known forcing (tides, winds, and density) using separate simulations for each forcing. The dimensionality of the model should be selected to provide the most accurate and efficient computation for each forcing parameter. As an example, two-dimensional, vertically averaged simulations might be adequate to model tides. Three-dimensional simulations are probably required for the density and wind-driven flows. The duration and temporal resolution of the predictions will depend on the particular forcing parameters of interest. (2) Assess the potential interactions between the various forced flows by model-sensitivity analyses. This might lead to modifications of the simulation procedures for the individually forced flows noted above. (3) Compare the model predictions with observations of the current (Eulerian and Lagrangian), sea-surface elevation, and hydrographic structure. (4) Use data- ass~milation procedures and modifications of the hydrodynamic model parameters and open-boundary conditions to optimize agreement with observations. (5) Assemble circulation data sets that accurately define the flow field and its associated spatial and temporal variability for input to the OSRA model. The circulation data set should cover a long enough time (about S° 10 years) to represent the important time scales of variability in the circulation so as to assure accurate oil-spill-trajectory predictions. The recommended modeling strategy allows a practical, cost-effective methodical approach to describing the circulation as input to the OSRA model. The model can also be used as a research tool to investigate the sensitivity of predicted currents
CONCLUSIONS AND RECOMMENDS TIONS and spill trajectories to interactions between the variously forced flows, poorly known model parameters (e.g., vertical eddy viscosity), and open-boundary conditions. This approach should also serve to advance knowledge of the circulation dynamics and assist in planning observational programs. f. The Minerals Management Service should strengthen its ability to respond scientifically to accidental oil spills, such as the Exxon Valdez spill. MMS should maintain the capability to initiate studies of oil spills rapidly and should have workable plans to coordinate such studies with those of other federal agencies, in particular the Fish and Wildlife Service, - NOAA, and EPA; state agencies; and industry. Oil spills are rare scientific opportunities for MMS to assess its current understanding of oil-spill transport and fate. These opportunities should allow MMS to validate models of oil-spill fate, and oil-shoreline interaction. In physical sciences, such studies should include (but not be limited to) the deployment of drifters as MMS did in the Prince William Sound spill, chemical studies of oil and seawater, and verification of models. 105 g. The recent oil spill in Prince William Sound illustrates the importance of analyzing worst-case scenarios. For individual lease sales, MMS has historically used a probabilistic representation to assess the potential impact associated with development and production scenarios. Although this appears to be a reasonable approach, MMS should supplement this approach by studying (in a non-probabilistic manner) several worst-case scenarios as well, as was done for the siting of the Alyeska terminal in Valdez (U.S. DOI, 1972~. These scenarios should include spill location, volume, rate of release, oil type, and environmental conditions and should address a variety of likely cleanup responses. h. Although technically MMS's jurisdiction covers the OCS, MMS should consider oil spills occurring shoreward of the OCS. The panel notes that oil spilled inshore of the OCS could well be OCS oil. 3. Program priorities and operating procedures in the ESP should be modified as necessary to ensure that improved scientific input Is obtained at all stages of ESP operation in all regions, that available data from cooperating agencies are used, that development of a better-integrated national program continues, and that study results are published in the open literature. a. Appropriate balance between national and regional priorities is needed. Physical oceanographic and hydrodynamic modeling studies have been performed for all the major offshore lease areas, with particular focus on basin-wide studies (South Atlantic, Gulf of Mexico, Southern California shelf, Bering and Chukchi seas, and the Beaufort Sea). Some shelf areas, however, have not been adequately studied, the most striking example being the Gulf of Mexico. Long-term studies are also crucial. b. Improved scientific input needs to be obtained and used more fully in the formulation of 5-year plans, in the preparation of RFPs, and in program assessment, including the synthesis of field data and models and production of final probabilities for oil-spill distribution. Possibilities for improved scientific input include: . · Having university research scientists and faculty and government agency researchers (e.g., EPA, NOAA, ONR9 NASA' U.S. Army Corps of Engineers, Coast Guard) work with the ESP staff in a visiting scientist capacity. Similar programs have been successfully used by ONR and NSF. · Individually or jointly sponsoring a specialty conference or special sessions at an existing conference (OCEANS, Offshore Technology Conference, Oil Spill Conference, American Society for Testing and Materials) on modeling of oil-spill fate and impact assessment. · Continuing to expand the use of peer review for REP preparation, proposal evaluation, and critique of contract work products. Peer review, properly used, is an excellent way to maintain the quality and scientific integrity of a contracted scientific research program.
106 PHYSICAL OCEANOGRAPHY OF THE U.S. OUTER CONTINENTAL SHELF c. MMS's cooperation with other agencies is commended. These collaborations can provide much valuable information at low cost. Efforts to synthesize data from experiments of other agencies into a form usable by MMS should be encouraged. d. The establishment of the ESP was accomplished historically by assembling, and to a limited extent integrating, the study programs of the regional offices. Recently the WO has taken a more active role in shaping the program by developing a 5-year plan and fully integrating the regional study efforts. This is a positive step in managing the overall program. Efforts should continue to develop a better-integrated national program while maintaining the regional office structure. e. MMS should continue to strengthen its program and have the results of its studies presented at scientific meetings and published in the open, refereed literature. This includes funding for manuscript preparation, publication charges, and travel to attend meetings. Interagency agreements that MMS has with NOAA, USGS, and others should also include this requirement when specifying contract deliverables.