From: "Edgar V. Leyendecker" 
To: martini@virginia.edu
Subject: Central VA hazard maps
Date: Mon, 28 Jul 1997 10:46:52 -0400

Dear Kirk Martini:

I am responding to your e-mail to Nancy Dickman.  I am a structural engineer
on the hazard mapping project.  There are a number of reasons for the
differences between the 1994 USGS maps and the current maps.  One of these
is the source of the largest difference.

1.  The 1994 maps are minor modifications of our 1991 maps which were based
on data primarily from the mid to late 1980's.  Those maps were based on a
different modeling approach using source zones.  We are currently using a
concept referred to as smoothed seismicity but we add in faults and certain
geologic features.  This approach, including our entire methodology, is on
the web site as Open-file report 96-352.  We find little difference in the
answers from the two approaches when the same input data are used.  We feel
the current approach is less subjective than the prior one, although there
are still many advocates of the former approach. I've described some of the
difference below.

2.  In 1994 we an used a single attenuation function east of the Rockies
that was for soil. This was developed around 1990.  We currently use two
more recent functions that are for a firm rock.  This by itself would
produce somewhat lower values in the current map since the reference
material is different. The direct comparison is somewhat difficult due the
difference in attenuations.

3.  In the 1994 maps the historical earthquake magnitudes were based on
maximum epicentral intensities (there wasn't much else that could have been
done since there are no instrumental data).  Since then work has been done
on improving magnitude assignments. Many researchers have found that
assignments based on felt areas as determined by intensity contours is a
much better measure of magnitude as compared to a single epicentral value.
We used earthquake catalogues based on this newer approach.  This newer
approach, in general, resulted in smaller magnitudes (in the east) than were
used previously.  We did some comparisons using the same input data, except
for the magnitudes, and found that this change using newer research results
accounted for the largest differences in the maps in the central U. S. and
the east.


The differences in the maps are not uniform since there are other modeling
differences, but the magnitude reassignment is probably the single largest
factor.

Hope this helps.

E. V. Leyendecker