Version of Seismic Hazard Software

USGS Bulletin 1772, "SEISRISK III: A Computer Program for Seismic Hazard Estimation" (Bender and Perkins), describes the program and usage. This publication does not have any accompanying program source or executable. The program was originally written for use on a DEC VAX 11/780 minicomputer with the VAX/VMS Operating System using VAX/VMS Fortran.

A version of the program adapted for the IBM PC and compatible micro- computers using either DOS or OS/2 operating systems is documented in Open-File Report 89-557. This documentation describes the process of porting the program to the microcomputer environment and diskettes containing the program source and executables are available. The USGS Bulletin 1772 is necessary for an understanding of the program usage.

A version of SEISRISK-III called SEISRISK-III.DP for the Apple Macintosh system is documented in Open-File Report 92-529. This version performs calculations for a simple dipping plane, as with a rectangular Benioff zone. We find that this version uses a default source-area calculation for a magnitude 8 earthquake and needs a revision to bring the magnitude- dependent rupture areas more in line with proper seismic moments.

To obtain copies of the publications and software refer to GENERAL:How to Order.

MODIFICATIONS:

We are aware that Dario Slejko at Trieste has modified Risk4A, an earlier version, to use elliptical rather than circular isoseismals - that is, the distance attenuation varies with azimuth. We presume that he has made a similar modification to SEISRISK III, which he now uses.

A "bug fix" exists that has not yet been incorporated in any published version of SEISRISK III. This fix is to ensure a proper interpolation for values which involve the "last box" in the ground-motion annual-rate accumulators. This bug does not usually cause a problem, because most users scale the ground motions such that the interpolation primarily is taking place in the lower range of accumulator boxes.

A "loose end" exists in subroutine inside. The subroutine has a trap which does not permit a calculation when a site point lies too close to a quad- rilateral corner point. When such a point is discovered in the course of a run, the run is terminated. If this happens, it is easy to adjust the "equator" by a very slight amount, thus moving the site point away from the corner point. Because rather a good deal of calculation can go past before such a site is discovered by the program, it would be desirable to permit the program to continue rather than terminate. We have not worked on a version to do this, because we encounter the problem rather infrequently in our usage. However, users with complex sources or small site-grid spacing may find the simple adjustment merely induces a failure at a different site.

At the moment, we have not decided on whether to make the above fixes available in an updated version.