The 10 February 2006 "Green Canyon" Earthquake: a Case History of an Unusual Seismic Event

Dellinger, Joe (BP America) | Blum, John (Scripps Institution of Oceanography) | Nettles, Meredith (Lamont Doherty Earth Observatory)



The 10 February 2006 MS = 5.3 “Green Canyon” earthquake has been of interest to the oil industry because of its location in an area of active hydrocarbon exploration and production. At the special workshop at the 2006 SEG dedicated to the event, three possible mechanisms were proposed: 1) an earthquake in the crystalline basement, 2) a shallow crustal earthquake in the sedimentary cover, or 3) a large undersea landslide. No consensus could be reached among the workshop participants as to which of these scenarios was the most plausible. As part of a BP hazard-implications review, we attempted to better constrain its location and mechanism.

Dewey and Dellinger (2008) noted that the event’s unusual seismic signature, with much weaker P-wave body waves than would be expected for an event that size, ruled out an earthquake in the crystalline basement. To refine the event’s location, we located and provided to the USGS seismic recordings of the event from two oil-industry arrays, the Atlantis ocean-bottom-node survey and the CGG Green-Canyon Phase VIII multi-client streamer survey. These were used to update the USGS location for the event (Dewey and Dellinger, 2008).

Nettles (2007) was able to fit the event’s long-period surface waves using either a shallow earthquake or a gravity-driven landslide source. To test the landslide hypothesis, we performed slope-stability modeling, which showed that a failure on the prominent SouthWest-facing-slopesWest of the USGS epicenter could be consistent with that source mechanism. In October 2008, the Scripps research vessel Roger Revelle performed a multi-beam bathymetry survey over the Western epicentral area. We differenced the 2008 Scripps bathymetry with 1999 bathymetry derived from the CGG Green-Canyon Phase II multi-client streamer survey. No large differences were seen that would be indicative of a large landslide source. We conclude that the event was most likely a shallow earthquake in the sedimentary cover.


The two earthquakes of magnitude > 5 in the Northern Gulf of Mexico in 2006 generated sufficient interest that there was a special workshop at the 2006 SEG convention and a special session at the 2007 OTC convention where these events were discussed by scientists from academia, government, and the oil industry. After examining the published USGS locations for the events on detailed proprietary geological maps, oil-industry scientists generally favored a mechanism involving faulting of crystalline rock beneath the sedimentary cover for both events. This interpretation is now the consensus one for the 10 Sept. earthquake ( usslav/, accessed March 2008), which was widely felt in Florida. Its mechanism is similar to that of another earthquake that occurred in the region in 1978 (Frohlich, 1982).

Closer investigation of the 10 Feb. “Green Canyon” event, however, revealed that its seismic signature was unusual (see Figure 1). Especially notable was the weak short-period (1s) energy of the event compared with the its long-period (20s) energy. Disproportionately weak generation of short-period energy is characteristic of faulting of weak rock or of landslide processes. Nettles (2007) showed that the long-period surface waves generated by the 10 Feb. event could be explained by a shallow double-couple source (M0 = 6x1016N m) in a medium with a low shear-wave velocity (based on Gregory, 1977), either with a slip at an azimuth of 244East of North on a sub-horizontal plane, or a vertical slip on a vertical plane with a strike of 152.