minutes of computer processing unit time was required to run the oil advec- 

 tion model with 10 years of historical wind data. 



The results of simulated spills at the site are presented as probability 

 diagrams. Figure 2-10 shows the results for no currents and winter winds, 

 while Figure VII-15 (appendix VII) is the same except for spring winds. 

 Figures VII-16 and VII-17 show winter and spring winds, respectively, com- 

 bined with a sea current of 0.25 knots in a westerly direction. About 450 

 spills were simulated for the spill site. These diagrams indicate the im- 

 pact, in percent, that a 10-mile by 10-mile offshore area would receive by 

 the calculated oil trajectories on a seasonal basis. As indicated by these 

 diagrams, oil movements from the spill site shows a strong offshore tendency. 



2.3.4 U. S. Geological Survey, Systems Analysis Group 



A fourth numerical modeling study was done by T. Wyant, D. A. Smith, and 

 J. Slack of the USGS Systems Analysis Group in Reston, Virginia. The results 

 of this study were not part of the on-scene effort, but they do provide an 

 opportunity to verify, apply, and extend an oilspill trajectory model pre- 

 viously developed as part of an oil spill risk analysis for the proposed 

 North Atlantic Outer Continental Shelf lease area. The latter analysis was 

 done to determine environmental hazards of developing offshore oil in the 

 region and has been described in detail by Smith et al. (1976). In the risk 

 analysis, model runs were used to estimate probabilities that spills occur- 

 ring at anticipated production sites would have an impact on certain biologi- 

 cal and recreational resources in the North Atlantic coastal region. When 

 the Argo Mevohant broke up in this area, it became possible to compare model 

 output with observed movements of the spill. Also, since model input for the 

 area was fully prepared before the incident, it was possible to make short- 

 and long-term forecasts of slick behavior from the moment the tanker went 

 aground. 



The USGS oil spill trajectory model was constructed and used to simulate 

 oil slick movement on a digital map of the North Atlantic between 38°N and 

 45°N latitude and 65 °W longitude and the North American coast. The funda- 

 mental transport equation in the model expresses oil slick movement as the 

 vector sum of residual surface current velocity, tidally averaged, and 3.5% 

 of wind velocity. Slick movement was simulated as a series of straight-line 

 displacements, each representing the joint influence of wind and current over 

 a 3-hour period. Monthly surface current velocity fields were provided by 

 the Bureau of Land Management and were based in part on drift bottle studies 

 conducted by Bumpus (1973) . 



Wind velocities were provided for the simulations in one of two ways, 

 depending on the mode of operation of the model. First, for purposes of 

 verification, wind observations were input directly and updated every 3 

 hours. Wind reports were received for two locations: Nantucket Light Ship, 

 and the USCGC Vig-iZant, which was on the scene of the grounding. Second, for 

 purposes of forecasting, prevailing wind velocity was used for the first 3- 

 hour step, and all subsequent 3-hourly winds were randomly generated from a 



51 



