You are here: Home » Research & Monitoring » Seismic Hazard Mapping

2002 United States National Seismic Hazard Maps

Following the release of the 2002 United States Hazard Maps, two subsequent updates were released. The first in April of 2003 and the second in October of 2003. Descriptions of each release can be found below. What is currently made available on the web reflects all the updated data and image files.

April 2003 Revisions to 2002 California Maps

In April 2003, we made changes in the input parameters of seven California faults used in the national seismic hazard maps. This results in minor changes in the hazard maps for parts of California. These input changes are described below and were done in cooperation with the California Geological Survey.

In addition to the fault changes, we have now added a distance taper to the hazard calculation for California A-type faults. In the revised ground motions maps, we linearly taper the ground motion hazard (frequency of exceedance) for California A-faults at distances between 200 km and 250 km, for all periods. This tapering eliminates the sharp fall-off of 1 Hz S.A. probabilistic ground motions in western Arizona apparent in the original 2002 maps, which occurred at 200 km distance from the San Andreas and Imperial faults. This 200 km distance was the maximum distance used in the hazard calculation for WUS faults. In the revised maps, the 200 km maximum distance is retained for all other faults in the Western U.S.

Modifications to the 2002 California ground motion maps.

We have made modifications to seven California faults to account for differences between the way the parameters were calculated in 1996 and in 2002. These modifications are listed below:

  1. Cucamonga fault: In 1996 we used M 7.0 to account for 2 m average displacements along the fault. In 2002 we used the Ellsworth and Hanks and Bakun relations to calculate magnitudes of M 6.9 and 6.7. In the revision we fixed the magnitude at M 6.9 (consistent with the Ellsworth relation) to account for the fault scarps and the recurrence time of 650 years, as in the 1996 model. The new GR incremental a-value is 2.52973.
  2. Landers fault: In 1996 we used the magnitude of the 1992 Landers earthquake (M7.3) for the maximum magnitude of the fault. In 2002 we used the Ellsworth and Hanks and Bakun relations to calculate magnitudes of 7.2 and 7.1. In the revision we have reapplied the historical maximum magnitude of 7.3 with a recurrence time of 5000 years, as in 1996. The new GR incremental a-value is 1.47714.
  3. Owens Valley fault: In 1996 we used the historical M 7.6 from the 1872 earthquake to account for the maximum magnitude. In 2002 we used the Ellsworth and Hanks and Bakun relations to calculate magnitude 7.4. In the revision we have reapplied the historical maximum magnitude of 7.6 with a recurrence time of 4000 years, as in 1996. The new GR incremental a-value is 1.76868.
  4. Garberville fault: In 1996 we applied a M 6.9 with a 220 year recurrence. In 2002 we applied a M 6.7 (based on Ellsworth and Hanks and Bakun relations) with about a 100 year rate. In the revision we have all owed a M 7.5 to float anywhere along the Maacama-Garberville faults with a recurrence time of 220 years. The new GR incremental a-value is 2.85961.
  5. Honey Lake fault: In 1996 we characterized the fault using a M 6.9 earthquake with recurrence of 1493 years based on paleoseismic data. In 2002 we used M 7.0 and 6.8 with recurrence times of 748 and 375 years based on the magnitudes from Ellsworth and Hanks and Bakun. In the revision we have reverted back to the 1996 model, having a M 6.9 earthquake with recurrence of 1493 years. The new GR incremental a-value is 2.32690.
  6. Pleito fault: In 1996 we applied a M 7.2 earthquake, in 2002 we applied a M 7.4 earthquake based on an incorrect fault width. In the revision we applied a M 7.0 and 6.9 earthquake based on the relations of Ellsworth and Hanks and Bakun. The new GR incremental a-value is 2.13675 and 2.27285.
  7. Monte-Vista-Shannon fault: In 1996 we used a M 6.8 earthquake to characterize the fault. In 2002 we used a M 6.9 and 7.0 based the relations of Hanks and Bakun and Ellsworth calculated using an incorrect fault width. In the revision we have fixed the width and recalculated the magnitudes to obtain M 6.7 events. The new GR incremental a-value is 1.76127.

Explanation for the October 2003 Revision of the National Seismic Hazard Maps

In October 2003 we slightly revised the national maps in limited portions of California, Oregon and Washington. The changes only affect a limited area along the hanging wall of thrust or reverse faults. The changes involve at most an increase of about 2-3% in the probabilistic ground motions in the maps. These changes are not large enough to be visible in the maps posted in the website, so we did not post new maps.

It came to our attention that we used an incomplete definition of the hanging wall for the Abrahamson and Silva (1997) attenuation relation for thrust or reverse faults, one of the four attenuation relations used in the hazard maps for non-extensional areas. In that paper they state, “the hanging wall term should be used for sites on the hanging wall and within the edge of rupture.” (p.106). We interpreted this to mean that the hanging wall term was to be used only for sites within the surface projection of the rupture zone, i.e., the fault plane. We also had a miscommunication with the lead author on this point. In fact, the authors intended that the hanging wall term should also be invoked for sites behind the surface projection of the fault plane, although a distance taper makes this a limited area. This definition was presented in Abrahamson and Somerville (1996), although the angle used for the expanding zone behind the surface projection was not specified in that paper. This paper was not referenced in Abrahamson and Silva (1997), contributing to this misinterpretation.

For the October, 2003 revision we corrected the definition of the hanging wall in our hazard calculations to make it completely consistent with the definition of the authors of these attenuation relations. This leads to a small increase in probabilistic ground motions for thrust and reverse faults in limited areas behind the surface projection of the fault planes. We used an angle of 22.5 degrees in map view, as specified by N. Abrahamson (pers. comm., 2003) to describe the widening of the hanging wall zone in map view behind the projection of the fault plane. We did not apply the hanging wall term to sites behind the surface projection of thrust faults with dips less than 30 degrees, because the geometrical justification for the term being used behind the surface projection of the fault plane is less valid for very shallowly dipping faults. We still used the hanging wall term for sites over the surface projection of the fault plane for the shallowly dipping thrust faults and all the other thrust and reverse faults, as was done in the original 2002 maps.

References

  • Abrahamson, N.A. and P.G. Somerville (1996). Effects of the hanging wall and footwall on ground motions recorded during the Northridge, California, earthquake, Bull. Seism. Soc. Am., v. 86, no. 1B, pp. S93-99.
  • Abrahamson, N.A. and W.J. Silva (1997). Empirical response spectral attenuation relations for shallow crustal earthquakes, Seis Res. Letts., v. 68, no. 1, pp. 94-127.