Location of the Green Canyon (Offshore Southern Louisiana) Seismic Event of February 10, 2006

We calculated an epicenter for the Offshore Southern Louisiana seismic event of February 10, 2006 (the 'Green Canyon event') that was adopted as the preferred epicenter for the event by the USGS/NEIC. The event is held at a focal depth of 5 km; the focal depth could not be reliably calculated but was most likely between 1 km and 15 km beneath sea level. The epicenter was calculated with a radially symmetric global Earth model similar to that routinely used at the USGS/NEIC for all earthquakes worldwide. The location was calculated using P-waves recorded by seismographic stations from which the USGS/NEIC routinely obtains seismological data, plus data from two seismic exploration arrays, the Atlantis ocean-bottom node array, operated by BP in partnership with BHP Billiton Limited, and the CGG Green Canyon phase VIII multi-client towed-streamer survey. The preferred epicenter is approximately 26 km north of an epicenter earlier published by the USGS/NEIC, which was obtained without benefit of the seismic exploration arrays. We estimate that the preferred epicenter is accurate to within 15 km.

We selected the preferred epicenter from a suite of trial calculations that attempted to fit arrival times of seismic energy associated with the Green Canyon event and that explored the effect of errors in the velocity model used to calculate the preferred epicenter. The various trials were helpful in confirming the approximate correctness of the preferred epicenter and in assessing the accuracy of the preferred epicenter, but none of the trial calculations, including that of the preferred epicenter, was able to reconcile arrival-time observations and assumed velocity model as well as is typical for the vast majority of earthquakes in and near the continental United States. We believe that remaining misfits between the preferred solution and the observations reflect errors in interpreted arrival times of emergent seismic phases that are due partly to a temporally extended source-time function and partly to failure of our travel-time model to account for the extremely complicated velocity structure of the sedimentary section in which the event occurred.