Determining tectonic motion across the New Madrid Seismic Zone
On a project with Seth Stein, John Weber (Grand Valley State), Joe Engeln (Missouri), Tim Dixon and Ailin Mao (both at Univ. of Miami), I am working to determine the velocities of plate motion across the New Madrid Seismic Zone. We are achieving this by doing intermitant GPS (Global Position System) surveys across the region. So far, sites have been measured three times (1991,93,97). We hope that with these data, we can determine whether there is significant differential motion, especially between the near and far fields. This work is instrumental in determining the recurrence rate of large earthquakes (s.a. the 3 assumed magnitude 8's in 1811-12) in the region.
The below figures are the results from the three recent GPS surveys performed by Northwestern. The results are quite surprising.

*click on illustration for a larger view
This is a map view of the New Madrid Seismic Zone and surrounding areas. The fault zone (yellow) is located in the upper middle of the diagram. The red arrows show the annual motion of each site with respect to the ITRF96 North American Plate motion.

*click on illustration for a larger view
Left panels: Profiles of site velocities, with 1-sigma error bars, parallel to the approximate strike direction of the major right-lateral strike-slip faults in the NMSZ. Panels show the entire network and its far- and near- field subsets. For each, the profile mean is removed. Also shown are best-fitting profiles computed for a locked vertical strike-slip fault driven by far-field motions. Misfit as a function of far-field velocity is shown on the right, with 2 sigma range colored yellow. None of the best fitting velocities differ significantly from zero. It is very interesting that motion across the fault is indistinguishable from zero, within errors. This signifies that if the current motions are representative of long-term dynamics, then the recurrence time for a magnitude 8 earthquake (assumed magnitude of 1811-2 series, with 5-10 meters of slip), is greater than 3,000 years!

*click on illustration for a larger view
Locations of continuously recording GPS sites used to estimate an Euler vector for the presumably-stable portion of North America. For each, the misfit between the observed velocity and that predicted for a single plate is shown. These misfits have a mean of 1 mm/yr. Inset: Euler poles for the eastern and western subsets of the sites (divided by dashed line) compared to that for the entire set. Because the poles for the east and west data overlap at 95% confidence, the platewide GPS data show no resolvable motion across the NMSZ.

*click on illustration for a larger view
Relation between inter-seismic motion and the recurrence time of large New Madrid earthquakes. For an assumed horizontal slip in 1811-1812 of 5-10 m, the geodetically observed inter-seismic motion of less than 2 mm/yr discussed here implies recurrence times greater than 2500 yr. Also shown are recurrence estimates from paleoseismic studies. The paleoseismic and geodetic data are jointly consistent with slip in 1811-1812 being about 1m, corresponding to a magnitude 7 earthquake.

*click on illustration for a larger view
Top: Earthquake frequency-magnitude data for the New Madrid zone. Both the recent and historic (1816-1984) data have slopes close to one, and predict a recurrence interval exceeding 1,000 yr for magnitude 7 earthquakes and 10,000 yr for magnitude 8 earthquakes. Bottom: Earthquake history for the NMSZ. Since 1816, there have been 16 earthquakes with magnitude greater than 5 (about a 10-yr recurrence), and two with magnitude greater than 6 (about an 100-yr recurrence). We thus expect magnitude 7 and 8 earthquakes to have about 1,000 and 10,000 yr recurrence.

*click on illustration for a larger view
National Seismic Hazard map showing predicted peak ground acceleration expected in 50 yrs at 2% probability [Frankel et al., U.S.G.S. OFR 96-532]. Because these maps incorporate assumptions of rapid deformation and frequent great earthquakes, the predicted acceleration for the New Madrid zone exceeds that in San Francisco, and the predicted very high acceleration (exceeding 1.2g) area for the NMSZ is larger than for Los Angeles or San Francisco. Our results imply that the seismic hazard for the New Madrid area is lower and that such maps should be revised.

    References:
  • Newman, A. V., S. Stein, J. C. Weber, J. F. Engeln, A. Mao & T. H. Dixon, Slow Deformation and Implied Long Earthquake Recurrence Intervals From GPS Surveys Across the New Madrid Seismic Zone, Science, 284, 23 April 1999 ( PDF version )
  • Newman, A. V., J. Schneider J., S. Stein & A. Mendez, Uncertainties in New Madrid Seismic Zone Seismic Hazard, Seism. Res. Lett., 72 6, 653-667, 2001

Updated Mon Apr 26 09:29:49 CDT 1999 | my current research page