Virtual Vessels Evading Historical Hurricanes in the Gulf of Mexico

When dynamically-positioned (DP) drilling vessels operating in the Gulf of Mexico are forecast to be near the path of a hurricane, they need to secure their well, pull their riser and evade to a safe position where wind and wave conditions are within acceptable limits. A vessel can experience undesirable met-ocean conditions if it begins its evasion too late or does not allow enough extra time to overcome significant forecast errors in the track, intensity or extent of high wind and wave conditions. Simulations of virtual DP vessels evading the tracks of historical hurricanes in the Gulf of Mexico were conducted in order to better understand the outcome of different evasion techniques and serve as a statistical basis for future hurricane avoidance practices.

A database of wind and wave conditions across the Gulf of Mexico at different timesteps was generated in order to determine the maximum met-ocean conditions experienced by virtual vessels during each evasion attempt. Data sources included the Extended Best Track Database (1988-2010) and NOAA WAVEWATCH III reanalysis data (1999-2010). Only those tropical cyclones containing hurricane-force (33 ms^-1) winds were studied. A total of 684 National Hurricane Center (NHC) forecast tracks of tropical cyclones from 37 tropical cyclones were used as the input that the virtual vessels evaded.

The starting grid consisted of 419 virtual vessels distributed on a 0.25° x 0.25° grid in the U.S. deepwater (>305 m) areas of the Gulf of Mexico. A simulated vessel was positioned at each of the sites and programmed to evade the historical hurricanes as they traversed the Gulf of Mexico. Two possible evasion directions practiced by mariners were evaluated: a direction between south and west (180° to 270°) that was approximately perpendicular to the storm path at its closest approach and an eastward (90°) heading. Representative vessel velocities ranging from 1.5 to 4.1 ms^-1 were simulated and ships and semi-submersibles were evaluated separately due to their different sensitivities to met-ocean conditions. Changes in vessel direction during the evasion and those limited by shallow water in the desired direction were not evaluated.

Graphics of each evasion attempt were created and certain cases were studied. Evasion outcomes of each combination of vessel type and speed with similar starting characteristics relative to the forecast path's closest point of approach (e.g. 36 to 42 hrs from the arrival of the forecast center of the storm and 25 to 50 NM to the right of the forecast path) were grouped together and evaluated for all evasion attempts. Statistics were generated and evaluated as measures of successful evasion outcomes where the highest wind and wave conditions were below the set limits for each vessel type.

This study considers a number of factors affecting the success of vessel evasions including forecast error (intensity, along-track, across-track), vessel characteristics (speed of advance, sensitivity to wind and wave conditions), and evasion decision characteristics (starting position relative to the forecast track and direction of evasion). Simulations using perfect forecasts instead of the NHC operational forecasts were also conducted to determine the role of forecast error. The results for only the largest storms as measured by the Hurricane Severity Index (HSI) were evaluated separately because of their large swath of harmful met-ocean conditions.

The results provide a statistical basis for the Hurricane Evacuation Plan that stakeholders can agree on to achieve acceptable risk during operations. The simulations can also be used for training captains to assist in tactical decisions about when to leave the site and in which direction to go based on the position relative to the forecast track, size of the approaching hurricane and vessel characteristics.