Analysis of the 10 February 2006: Gulf of Mexico Earthquake From Global and Regional Seismic Data

Abstract

Abstract The M?5.2 earthquake that occurred on February 10, 2006, in the Gulf of Mexico was unexpectedly large for this region of low seismic activity. It was also notable for the unusual characteristics of the teleseismic waveforms it generated. The teleseismic seismograms are depleted in high-frequency energy, and are not fit well by traditional double-couple source models. The seismograms are fit well by a model of sliding on a shallow, sub-horizontal surface within the thick layer of low-velocity sediments that blankets the Gulf of Mexico offshore region. These characteristics of the seismograms suggest a gravity-driven source rather than a tectonic-earthquake source. The earthquake was located in a region of active oil and gas production, making an understanding of its source mechanism important from both a scientific and economic perspective. Background The very unusual seismic event that occurred in the Gulf of Mexico on February 10, 2006, was widely recorded by global and regional seismic networks. The earthquake, of M?5.2, occurred off the coast of Louisiana, approximately 240 km south of New Orleans. The event was detected and located by the National Earthquake Information Center (NEIC) of the United States Geological Survey (USGS) using traditional P-wave-arrival methods, and by the Lamont-Doherty Earth Observatory (LDEO) using surface waves with periods around 50 seconds (see (1) for a description of the method). The earthquake was the largest to occur in the Gulf of Mexico since the M?5 event of July 24, 1978 (e.g., 2), which represents the best-recorded earthquake in the region prior to the February 10, 2006, event. A larger, M?5.8, earthquake occurred in the Gulf of Mexico on September 10, 2006; this event was also recorded well by global and regional seismic networks. Analysis The February 10 earthquake is large for this region of low seismic activity. It is also notable, however, for the unusual characteristics of the teleseismic waveforms it generated. Although many good recordings of surface waves in the period range 40 < T < 150 sec are available, attempts to model these data using standard centroid--moment-tensor (CMT) analysis (3, 4) failed. In contrast, CMT analysis provides good constraints on the faulting geometry and size of both the 1978 and September, 2006, earthquakes. The teleseismic data recorded from the February 10 earthquake are depleted in high-frequency energy, a characteristic often seen for seismic waves generated by a sliding mass rather than by tectonic faulting. We modeled the observed surface waves using a singleforce source (e.g., 5), often referred to in the academic literature as a "landslide" source. This type of source explains the observations well. We find a sliding direction of N244E and a mass X sliding-distance product of 5e13 kg-m. The sliding mass and sliding distance cannot be determined independently from this analysis, but correspond to, for example, 10 cubic km moving 2 m, or 1 cubic km moving 20 m, assuming sediment densities appropriate for the region.