Minutes from 3 December 2008

2008/12/03 CSEP meeting Attendees: Phil M., Jeremy Z., Fabian E., Danijel S., Masha L.

1. csep-op:
   • Failures for test dates of 2008-10-13 and 2008-10-15 due to Matlab license problems were re-processed successfully
2. Upgrade of New Zealand testing center to CSEP V8.10:
   • Danijel provided new ForecastML template for the forecast with grid cells at x.x5 centers
   • Danijel will provide GMT script for forecasts maps
3. Wish list from Jeremy:
   • Masha generated declustered ANSS catalog with start date of 1932-01-01:
     
     
     1. Jeremy identified an extra event for 2008/04/26 - Masha traced it to the differences in raw ANSS catalog data
4. S("space") and M("magnitude") evaluation tests to be added in V9.1, but not invoked in operations until V9.4:
   • Jeremy sent Masha Matlab code and acceptance tests reference data
   • Masha will introduce Python classes for the tests
5. Masha implemented fixes and added acceptance tests for:
   • Trac ticket https://intensity.usc.edu/trac/csep/ticket/113
   • Trac ticket https://intensity.usc.edu/trac/csep/ticket/114
6. Global testing:
   • Data source: CMT
   • No collection or test area
   • Danijel provided master ForecastML template:
     
     
     1. 1.0 x 1.0 degree cells
     2. M5.95 (or M6.0) and above
   • Jeremy submitted Triple-S one-year rate-based and alarm-based model for testing
     
     
     1. Masha implemented a Python wrapper for the model, updated CSEP build files to include the model
     2. Masha updated configuration files for nightly crontab on csep-cert:
        • Added Global testing region with one-year-models and one-year-alarm-models forecasts groups
        • Generated forecasts with start date of 2008/11/02
        • Evaluated forecasts:
          
          
          1. Alarm-based: ROC test
          2. Rate-based: N, L, ROC tests
     3. Jeremy sent Masha reference data for an acceptance test
     4. Danijel will work with Jeremy to provide b-values for Triple-S forecast
   • Danijel will work on GMT script for forecasts maps
7. Masha has sent email to Yan summarizing rules for one-day and five-year models within California:
   • Waiting for Yan's response
8. Keck foundation meeting on Friday:
   • 4 speakers for 45 minutes presentation
   • Will send existing posters to Tran for printing