[Shake-dev] ShakeMap Update

[Bruce Worden]

Hello everyone,

There is a new update to ShakeMap available in the repository. The repository is at:

https://vault.gps.caltech.edu/repos/products/shakemap/tags/release-3.5/

To get the update, do 'svn update'. You will then need to do a 'make' to get the new code compiled and installed. There were a couple of changes to config files, so you'll have to review those, as well (the output from make will tell you which ones to review).

This update includes the following changes:

o Pete Lombard has updated the db2xml program and added a new supporting module, SNCLInfo.pm. This program is only used by networks using a CISN (California Integrated Seismic Network) database, but may be of interest to others who retrieve amps from a database. It uses full Station/Network/Channel /Location station identifiers: stations are given IDs like "Network.Station" and channels have IDs like "Location.Channel". There are numerous other improvements over the previous version of db2xml, as well.

o Pete also provided an updated version of zone_config. The file zone_config.conf now has meaningful documentation, and the zone_config program has better doc and cleaner code.

o Licia Faenza of INGV provided updates to the Italian GMPEs: ALPI_ORIEN_table, ALPI_OCCI_table, and APPEN_table, as well as the tables that support them.

o Licia also provided a revised way of converting from PS to JPEG. We had tried this approach before, but it resulted in maps that were all black on some systems. The new way uses a two step process to do the conversion, and it appears to work in all systems we have tested, and produces better looking text in the various maps. Please test on your system and let us know if you have any problems.

o We have revised the way we produce the MMI (or, more generally, 'intensity') site amplification tables. The previous approach over-amplified the very small intensities (MMI 1.0 to ~1.5). The new approach rolls off the amplification, reaching zero at MMI 1.0. This is a bit of a compromise solution, but allows the intensity maps to fade gracefully to white. You will need to rerun makeMMIsiteTable to generate the new tables. Please see the Software Guide for instructions on running makeMMIsiteTable -- the program now takes two arguments: a GMICE, and your site amplification table.

o There is now an optional flag to grind called '-directivity'. Using this flag will apply the directivity functions described in Rowshandel (2006, 2010) to the predicted amps from the GMPE. The effect is fairly small and limited to areas very close to the fault (especially for everything other than 3.0-second PSA). In cases where the fault does not reach the surface, the effect may not be noticeable at all.  Some notes:

- To use directivity, you must define a polygonal fault per the ShakeMap specification (see the Software Guide for information on defining a fault).

- You can use -directivity with any GMPE, but the functions were produced for the four NGA GMPEs. While it is likely pretty reasonable for use with other active tectonic GMPEs, I can't endorse it for other environments.

- While it will function for any magnitude, Rowshandel does not endorse these functions for earthquakes less than M6.0.

- Except for the DefaultIPE, directivity is not applied to intensity. If you are using the DefaultIPE, the directivity functions are applied to PGA and PGV before they are converted to intensity. I'm happy to revisit this as soon as there is research describing the application of directivity to intensity.

- There is a new program called 'plotdirect'. If you run it ('plotdirect -event <evid>'), it will produce contour plots of the directivity functions in .../data/<evid>/richter. (You must run grind with the -directivity flag, first.) One figure is called dparam.jpg -- it is the directivity parameter itself (i.e. the thing you get from Equation 3 in Rowshandel (2010)). The other figures are called 'direct_<param>.jpg' -- they are the directivity factors applied to the estimates themselves (i.e., the thing you get (after exponentiation) with Equation 9 of Rowshandel (2010)). All the figures also have the fault and epicenter plotted.

- Directivity in ShakeMap should be considered experimental until it has been more thoroughly tested. Please give it a try, but don't include it in your production environment until you are convinced it is working properly.


o As usual, this update contains numerous small enhancements and bug fixes.

Please let me know if you run into any problems with this update.

Thanks,
Bruce



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://geohazards.usgs.gov/pipermail/shake-dev/attachments/20110227/bfc3fa54/attachment.html>