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.
- Appendix B Workshop on Modeling in Physical Oceanography The Physical Oceanography Panel convened a workshop in Seattle on August 5-6, 1987 to assess the current state of modeling for MMS, to compare these efforts to other state-of-the-art physical oceanographic models, and to recommend future modeling developments for MMS. The attendees, listed below, were MMS-funded modelers and outside experts (in addition to the panel members). The panel is very grateful to these experts for their time and thoughtful discussions. The panel found the discussions and the workshop recommendations useful in the preparation of its report and in the development of its own recommendations. Physical Oceanography Panel Workshop on Oil-Spill Modeling August 5-6, 1987 Participants Kenneth Brink, Woods Hole Oceanographic Institution Jerry Gait, Hazardous Materials Response Branch, Seattle H. James Herring, Dynalysis Inc., Princeton Zygmunt Kowalik, University of Alaska-Fairbanks Jan Leendertse, RAND, Santa Monica David Liu, RAND, Santa Monica Mark Luther, Florida State University, Tallahassee Akira Okubo, State Univerity of New York (SUNY), Stony Brook Robert Pritchard,-Icecasting, Inc., Seattle Mark Reed, ASA Inc., Narragansett Allan Robinson, Harvard University, Cambridge Robert Smith, Oregon State University, Corvallis Allen Wallcraft, Naval Ocean Research and Development Activity, NSTL Station (Bay St. Louis) Dong-Ping Wang, SUNY, Stony Brook Workshop Recommendations Oil-spill motion: The MMS procedure for analysis of oil-spill risks used in the lower 48 states has four parts: 133
134 . . 2) APPENDIX B Estimation of spill probability, based on historical records. Monte-Carlo simulation of trajectories (~500 spills per season, 30-day trajectories). This step uses water motions predicted form a numerical model and a separate simulation of winds, based on 3-hour transition probabilities and velocities usually from shore stations. The contribution from the wind is calculated at 3.5% of the wind with a variable drift angle to the right of the wind. This is added to the model's surface currents. The probability of hitting various "environmentally sensitive" areas is calculated from these trajectories. The conditional impact probability for a particular resource is then the spill probability times the hit probability. In Alaska, the procedure has been different: more detailed consideration of the surface layer was made and trajectory calculations were carried out in conjunction with the circulation modeling. The workshop focused primarily upon the numerical mode! predictions of water motion; however, the participants did also make more general recommendations about oil motion calculations. 1) The calculations of oil-spill movement should be made consistent with the modeling of water motion. The current MMS practice of using wind patterns that are not related to the winds driving the water motions may lead to substantial errors in estimating the oil motion. We recommend that the calculations of oil movement be made as a part of the model run. Models may have difficulty calculating Lagrangian trajectories; we suggest that it may be possible to calculate the probability distribution for oil as an alternative to trajectory modeling. It also seems reasonable to include weathering/fate calculations at the same time. The meteorological effort needs to be more thorough. Representing wind changes by transition-probability matrices does not yield accurate temporal correlations. In addition, the spatial structure is not at all represented. Simulated winds, perhaps from an LFM (limited fine mesh model) with appropriate interpolation to even smaller scales, should be used whenever possible. The 3.5% rule needs more analysis or should be replaced by more accurate surface-layer models. 3) Extreme events and statistics must be considered. In any environmental problem, even extremely small impact probabilities may be important if an important resource is affected. The MMS procedure is probably least accurate at low levels of probability, right where the concern may be greatest. More consideration needs to be given to the extreme event e.g., hurricanes-that may lead both to higher spill probability and more rapid water and oil motion. Models: Realistic, applied modeling of ocean dynamics is a rapidly evolving field. In the view of the workshop participants, there are a number of state-of-the-art models for calculating water flows that have been made available to MMS by their contractors. However, participants recommended: 4) The ice modeling has not been state-of-the-art; more modern models include calculations of thickness distribution and use of an elastic-plastic or viscous-plastic constitutive relationship with the strength calculated from an energy balance statement. Some participants felt that such models may be more sophisticated than required for estimating oil-spill motion.
WORKSHOP ON MODELING IN PHYSICAL OCEANOGRAPHY 5) 6) 7) 135 Other physics must be assessed. A number of potentially important physical processes are not included in these models surface fronts and convergences, Stokes' drift, Langmuir cells, and other near-surface processes. It is not known how these factors might affect oil movement. In each region, MMS needs to take a careful scientific look at both modeling and data-gathering efforts to ensure that the work is matched to the kinematics and scales of the transport processes and to assess the necessity for including other processes. Sensitivity analyses should be carried out for all modeling work. There is essentially no information on the sensitivity of the model results or, even more seriously, of the final trajectories to the initial conditions, fluid and ice boundary conditions, forcing functions, etc. 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. Model intercomparison and verification cannot be neglected. In addition to better validation against data (discussed below), the group suggested more model-to-model comparisons, independent critiques, and more efforts on quality control. Relationship to field work: The workshop participants agreed that the relationship between MMS-funded field work and modeling activity is insufficient and does not reflect state-of-the-art practice. S) Verification against data must be much more thorough. A new program should be initiated to perform systematic model evaluation and verification against data not only for currents but also the trajectories/probability distributions themselves. 9) 10) Better cooperation between observers and modelers is essential in planning work in new regions. Physical oceanographers engaged in field work and modeling should work together closely 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. Data-assimilation techniques should be explored. In future MMS studies, efficient regional simulations may best be carried out by employing the methodology of data assimilation, which melds fields calculated from models and from data.