Skip Links

USGS - science for a changing world

Open-File Report 96-532

National Seismic Hazard Maps: Documentation June 1996

By Arthur Frankel, Charles Mueller, Theodore Barnhard, David Perkins, E.V. Leyendecker, Nancy Dickman, Stanley Hanson, and Margaret Hopper

Special Zones

1. New Madrid. To calculate the hazard from large events in the New Madrid area we considered 3 parallel faults in an S-shaped pattern encompassing the area of highest historic seismicity (Figure 7). These are not meant to be actual faults; they are simply a way of expressing the uncertainty in the source locations of large earthquakes such as the 1811-12 sequence. The extent of these fictitious faults is similar to those used in Toro et al. (1992). We assumed a characteristic rupture model with a characteristic moment magnitude M of 8.0, similar to the estimated magnitudes of the largest events in 1811-12 (Johnston, 1996a,b). A recurrence time of 1000 years for such an event was used as an average value, considering the uncertainty in the magnitudes of pre-historic events. A recent compilation by Johnston and Schweig (1996) suggests a 500 year recurrence time for paleoliquefaction episodes (Johnston and Schweig, written comm., 1996). Tuttle and Schweig (1995) report two to three paleoliquefaction events over the past 5000-6000 years, but the earlier part of the record is likely incomplete. Of course some of these events could have had magnitudes less than 8.0.

Another possibility we tried was to use an exponential recurrence distribution starting at M7.5. We tried a run with such a distribution based on a cumulative recurrence time of 1500 years for events greater than M7.5. This run produced substantially lower probabilistic ground motions than the characteristic run.

When applying the Toro et al. (1993) and our attenuation tables to these finite faults, we measured distance to the closest point on the fault plane, with the shallowest part of the fault being at 10 km depth. We used the moment magnitude attenuation relations of Toro et al (1993) and directly applied our tables based on moment magnitude.

In the June maps, we used an areal source zone for New Madrid for models 1-3, rather than the spatially-smoothed historic seismicity. This area zone is shown in Figure 8 and is bounded by the borders of the Reelfoot Rift. This zone accounts for the hazard from New Madrid events with moment magnitudes less than 7.5. The a-value for this zone was calculated from the rate of mb 3.0 and larger events observed since 1976. This a-value yields a recurrence time of 540 years for events with mb>=6.5, when added to the rate of M 8.0 events in model 5.

2. Charleston, South Carolina. We used an areal source zone (Figure 9) to quantify the hazard from large earthquakes. This zone was drawn to encompass a narrow source zone defined by Pradeep Talwani (written comm. 1995; used in interim maps, see also Talwani, 1982; Marple et al., 1994) and a larger zone drawn by S. Obermeier and R. Weems (written comm., 1996). The extent of the areal source zone was constrained by the areal distribution of paleoliquefaction locations, although the source zone does not encompass all the paleoliquefaction sites. We assumed a characteristic rupture model of moment magnitude 7.3 earthquakes, based on the estimated magnitude of the 1886 event (Johnston, 1996b). Note that we lowered the Mmax to M7.2 for models 1-3 for this areal source zone, to avoid double counting events. For the M7.3 events we used a recurrence time of 650 years, based on dates of paleoliquefaction events (Amick and Gelinas, 1991; Obermeier et al., 1990, Johnston and Schweig, written comm., 1996). We used vertical faults with random strikes distributed throughout the areal source zone when calculating the hazard. Each of these fictitious faults is centered on a grid cell within the source zone.

3. Eastern Tennessee seismic zone. Several participants in the Memphis workshop wanted a special source zone for the eastern Tenn. seismic zone. This is a linear trend of seismicity that is most obvious for smaller events with magnitudes around 2 (see Powell et al., 1994 ). The magnitude 3 and larger earthquakes tend to cluster in one part of this linear trend, so that hazard maps based just on smoothed mb3 and larger events tend to be high in one portion of the zone. Therefore, we used a source zone suggested by Martin Chapman based on the microseismicity (Figure 10). When calculating a-values for model 1 we counted the number of events (mb>=3 since 1976) that occurred within this zone to determine an a-value for this zone. This was the a-value used for grid cells within the zone for models 1-3. Using this areal zone elongates the contours of high hazard, corresponding to the trend of the source zone.

4. Wabash Valley. Recent work has identified several paleoearthquakes in the areas of southern Indiana and Illinois based on widespread paleoliquefaction features (Obermeier et al., 1992). We considered an areal zone with a higher Mmax of 7.5 to account for such large events (Figure 4). We did not use this zone to determine a-values for the hazard calculation; the gridded seismicity was used instead (models 1-3). The sum of the gridded a-values in this zone calculated from model 1 produce a recurrence time of 2600 years for events with mb>= 6.5. The recurrence rate of M6.5 and greater events is estimated to be about 4,000 years from the paleoliquefaction dates (P. Munson and S. Obermeier, pers. comm., 1995), so it is not necessary to add additional large events to augment models 1-3. The Wabash Valley Mmax zone that we used in the maps is based on the Wabash Valley fault zone. Rus Wheeler drew the outline of the fault zone.

5. Meers Fault. We explicitly included the Meers fault in southwestern Oklahoma. We used the segment of the fault which has produced a Holocene scarp as described in Crone and Luza (1990). We considered a characteristic moment magnitude of 7.0 and a recurrence time of 4000 years based on the work of Crone and Luza (1990). Because of the long recurrence time, this fault has very little effect on the probabilistic grounds motions, event for the maps with 2% PE in 50 years.

6. Charlevoix, Quebec. As mentioned above, a 40 km by 70 km region surrounding this seismicity cluster was assigned a b-value of 0.76, based on the work of Adams, Halchuck and Weichert. This b-value was used in models 1-3.

7. Cheraw Fault. The June 1996 maps contain this eastern Colorado fault with Holocene faulting based on a study by Crone et al. (1996). We determined the recurrence rate of this fault from a slip rate of 0.5 mm/yr. A maximum magnitude of 7.1 was found from the fault length using the relations of Wells and Coppersmith (1994). As with the WUS faults, we used characteristic and Gutenberg-Richter recurrence models (see below) with equal weighting. CEUS attenuation relations were applied to calculate the hazard from this fault.

 

Part or all of this report is presented in Portable Document Format (PDF); the latest version of Adobe Reader or similar software is required to view it. Download the latest version of Adobe Reader, free of charge.

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
URL: http://pubsdata.usgs.gov/pubs/of/1996/532/Specialzones.html
Page Contact Information: GS Pubs Web Contact
Page Last Modified: Wednesday, 07-Dec-2016 16:06:19 EST