Minutes from 23 April 2008

2008/04/23 CSEP meeting

Attendees: Yan, Phil, Danijel, Jeremy, Jochen, John, Fabian, Masha

1. Introducing Western Pacific testing region into the CSEP:
   • Amit is working on Python class to automate Yan's forecasts within CSEP
   • Amit is out until Monday, April 28
   • Yan out of town April 28 - May 8

   • Masha is working on existing catalog PostProcess classes to make it data source dependable (ANSS, CMT):

     1. Since input catalog is shared by all models within forecast group, need to have generic filtering to be applied within that class.
        Currently, for all installed models within CSEP input catalog preparation consists of:

        • Cut to magnitude
        • Cut to collection area
        • Cut to time period

2. Fabian and Danijel presented Completeness web-site

3. Conclusions drawn from Monday's CSEP Global Collaboration and Testing Meeting:

   • Danijel said that non-uniform grid will be used for the testing
   • Introducing of Western Pacific testing region and getting CMT catalog data source into CSEP are first steps towards global testing
   • Interest from Russian team to install their models in CSEP:
     
     1. Jeremy worked on RTP model evaluation in the past:
        • Multi-region prediction experiment
        • Used catalogs: ANSS, CMT, JMA, catalog from Israel
   • For now will not have fault based forecast models in CSEP
   • Phil and Masha will work on a schedule
   • Danijel will invite Warner from Italy to install his model into SCEC Testing Center to compete with David/Yan's model
   • Jeremy plans on getting alarm based forecast to be installed for CSEP V8.10 release
   • Jeremy suggested to have grid down-sampling tool - would be useful for modelers
4. Software development:
   • There are some problems identified with V8.4 that would need an intermediate release of software:
     

     1. Trac ticket #81: list->set->list conversion for the list of forecasts files causes two sets of cumulative files for R-test

     2. Trac ticket #82: Cumulative R-test plot is missing on the results web-page
     3. Trac ticket #79: Forecasts maps don't display events with magnitude values less than 5.0
   • Everybody agreed to have release with fixes - scheduled for Monday, April 28.

5. Phil gave an overview of EarlyWarning reports and associated concerns