[Shake-dev] Another update
Bruce Worden
cbworden at caltech.edu
Tue Nov 2 21:30:08 UTC 2010
Hi all,
I've again updated the repository. There are only a few changes, but
one of them is relatively important for some users. As usual, 'svn
update' will get the new code, and 'make' will install it. There are
no special install instructions.
o Youngs97: This GMPE has had an error in one of the coefficients for
some time. We have fixed the error. The result will be a significant
decrease in PGV and PSA 1.0 sec amplitudes. At the same time, I have
updated the way the amplitudes are scaled to peak component (the
Youngs et al. relation gives a mean component). We previously used a
factor of 1.15, but Beyer & Bommer (2006) indicate that 1.10 is
appropriate for PGA and PSA, and 1.0 for PGV. Thus, there will be an
additional ~15% reduction in PGV amps, and about 5% for everything
else. In many cases these differences will have been erased by the
bias correction, but it is probably worth rerunning your important
events with the updated module.
o AK07: This GMICE uses distance and magnitude correction terms. The
distance measure is specified as fault or hypocenter distance, but the
module was using Rjb. This has been corrected. The effect is likely
minimal for most situations.
o grind: A minor fix to the code that generates output for plotregr.
The bug would only have affected plotregr plots for users of Youngs97,
and was fairly subtle, but would have resulted in the plotted
amplitudes being slightly low. Again, this only affected the plotregr
output -- the maps, grids, etc. were unaffected.
o Software Guide: Some old, obsolete instructions were deleted. A bit
of text about plotregr was added.
Let me know if you have any problems.
Bruce
More information about the Shake-dev
mailing list