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

Central and Eastern U.S.

Note: For simplicity in nomenclature we have included southeastern Canada and northeastern Mexico in our definition of CEUS. Similarly, parts of Mexico and Canada are included in the section entitled "Western U.S."

The basic procedure for constructing the CEUS portion of the hazard maps is diagrammed in Figure 3. On the left side we considered four models of hazard. Model 1 is based on mb 3.0 and larger earthquakes since 1924. Model 2 is derived from mb 4.0 and larger earthquakes since 1860. Model 3 is produced from mb 5.0 and larger events since 1700. In constructing the hazard maps we assigned model 1 a weight twice that of models 2 and 3. The weighting of the alternative models is discussed below.

For models 1-3, we followed the procedure in Frankel (1995) to construct the hazard maps directly from the historic seismicity. The number of events greater than the minimum magnitude are counted on a grid with spacing of 0.1 degrees in latitude and longitude. The logarithm of this number represents the maximum likelihood a-value for each grid cell. Note that the maximum likelihood method counts a mb 5 event the same as a mb 3 event in the determination of a-value. Then the gridded a-values are smoothed using a Gaussian function. We used a Gaussian with a correlation distance of 50 km for model 1 and 75 km for models 2 and 3. The 50 km distance was chosen because it is similar in width to many of the trends in historic seismicity in the CEUS. In addition, it is comparable to the error in location of mb 3 events in the period of 1924-1975, before the advent of local seismic networks. We have also made trial hazard maps using correlation distances of 25 and 100 km for mb >= 3 events. The 25 km case produced very grainy-looking maps whose peak values were not much larger than the 50 km case. The 100 km case tended to over-smooth the results obscuring localized areas of higher seismicity. One problem with using such a large correlation distance occurs when two areas of high seismicity are separated by about one correlation distance. After smoothing, one can get higher values in the region between the active zones than in the zones themselves. This is not a desirable result. We used a larger correlation distance for models 2 and 3 since they include earthquakes further back in time with poorer estimates of locations.

Model 4 (Figure 3) consists of large background source zones. This alternative is meant to quantify hazard in areas with little historical seismicity but with the potential to generate damaging earthquakes. These background zones are detailed in a later section of this text. The sum of the weights of models 1-4 is one. For a weighting scheme that is uniform in space, this ensures that the total seismicity rate in the combined model equals the historic seismicity rate. We will describe later our spatially-varying weighting scheme which slightly exceeds the historic seismicity rate.

We used a regional b-value of 0.95 for models 1-4 in all of the CEUS except Charlevoix, Quebec (see below). We determined this b-value from our catalog for events east of 105 degrees W (see below). We have also produced hazard maps based on locally-derived b-values. We showed these at the Memphis workshop and most participants felt they should not be used, largely because of the relatively short time of the catalog compared with the local rate of mb 4 and greater events. It is the local rate of mb 4 and larger events relative to the mb 3 events that controls the locally-determined b-values. It is important to note, however, that by combining models 1-3 we are essentially including the effects of locally-variable b-values. For example, model 3 will have high values where there have been mb >= 5 events historically. These will tend to be areas of low b-values derived on a local basis. The same can be said of model 2 and areas with mb 4 events. Conversely, areas with high b-values will be lacking in mb 4 and mb 5 events and will have low values for models 2 and 3. By adding the PE's from models 1-3 together with weighting we are accounting, to some extent, for the local variations in b-value derived from historic seismicity.

Figure 4 shows a map of the CEUS Mmax values used for models 1-4 (bold M refers to moment magnitude). These values were discussed at two of the workshops and represent the consensus of the Memphis workshop. These Mmax zones correspond to the background zones used in model 4. Most of the CEUS is divided into a cratonic region and a region of extended crust. Rus Wheeler (see Wheeler, 1995) drew the dividing line based on the landward limit of rifting of Grenville age and older crust during the opening of the Iapetan (proto-Atlantic) ocean about 500 m.a. We assigned an Mmax of 6.5 for the cratonic area. For the Wabash Valley zone we used an Mmax of 7.5 in keeping with magnitudes derived from paleoliquefaction evidence (Obermeier et al., 1992). We used a Mmax of 7.5 in the zone of extended crust outboard of the craton. This large Mmax was motivated by the magnitude of the 1886 Charleston event (M7.3, Johnston, 1996b), since the workshop participants felt such a large event could not be ruled out in other areas of the extended crust. An Mmax of 6.5 was used for the Rocky Mountain zone and the Colorado Plateau, consistent with the magnitude of the largest historic events in these regions. An Mmax of 7.2 was used for the gridded seismicity within the Charleston areal source zone (see below), so that there would be no overlap between the magnitudes used in the gridded seismicity hazard calculation and that from the areal zone, where M7.3 events are included. Extending the Mmax used for the gridded seismicity to M7.5 in Charleston would make no significant difference to the hazard values. All of the above Mmax values are moment magnitudes. A minimum mb of 5.0 was used in all the hazard calculations for the CEUS.

Model 5 (Figure 3, right) consists of the contribution from large earthquakes (M>7.0) in four specific areas of the CEUS: 1) New Madrid, 2) Charleston, SC, 3) the Meers fault in southwest Oklahoma, and 4) the Cheraw Fault in eastern Colorado. This model has a weight of 1. We describe how we treated these special areas in a later section. There are three other areas in the CEUS that we call special zones: eastern Tennessee, Wabash Valley, and Charlevoix. They are detailed in a later section. Frankel (1995) shows hazard maps based on earlier versions of models 1,3,4, and 5, for reference.

 

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/CEUS.html
Page Contact Information: GS Pubs Web Contact
Page Last Modified: Wednesday, 07-Dec-2016 16:06:21 EST