2008/07/23 CSEP meeting

Attendees: Yan, Phil, John, Masha

1. Western-Pacific testing regions:
   • One-day forecasts are generated with 30-days delay
     1. Yan will compare forecasts maps to the ones at UCLA and will email his observations
   • One-year forecasts are to begin on July 1, 2008 - will be generated August 1, 2008 (+- 1day)
2. Yan will introduce forecast model for California some time in August
   • Will use ANSS catalog
   • Forecast will be generated for 0.1 degree cells within rectangle area of California and Nevada
   • Phil asked to provide reference data for the model acceptance test: input catalog and expected forecast
   • Phil asked if Yan would like to get a subscription to CSEP authorized data sets:
     1. Yan most likely will stop generating forecasts at UCLA once CSEP generated forecasts are proved to be consistent with UCLA-generated ones
   • Yan asked about Max's short-term model for California to be installed in CSEP
     1. Masha has not heard from Max
3. Software development:
   • Jeremy has alarm-based forecasts and tests in Java
     1. Masha will ask Jeremy if original Java code can be automated within CSEP for Version 8.10 release
   • CSEP failures during Masha's vacation:
     1. 2008-06-11 test date failed for all testing regions due to Matlab license issue
     2. 2008-06-18, 2008-06-19 test dates failed for Western Pacific regions due to CMT web-site problems
     3. All failues have been successfully reprocessed
     4. Masha should show Fabian how to reprocess failed dates in case Masha is not available
     5. Phil suggested to have other than email-based visibility for system failures
   • Skype meeting is scheduled for 4pm on Thursday, July 24, 2008, to discuss CSEP/Japan testing center
   • Outstanding Trac tickets
     1. #93: CMT web site changed directory layout for final catalog data of year 2008
        • Side effect is extra disk usage
     2. #96: Cumulative R-test result plot should switch curves labels
     3. Danijel suggested to have intermediate 8.7.1 release with fix for ticket #96
        • Will follow the same procedure as for scheduled releases
        • Will test release candidate for one week on csep-cert machine
   • Masha started batch processing of remaining test dates with Version 8.7.0
     1. Should be complete within one week
4. Hardware:
   • Phil suggested to look into new hardware issues
5. CSEP article update by Jeremy

   • Hey guys, while I won't make today's call, I wanted to mention some CSEP documents that either are in progress or may be soon. I'll be submitting a paper on the statistical properties of the area skill score by 1 August, & Danijel and I will submit a paper on the RELM testing results thus far soon thereafter. We might also consider a CSEP computational article for the upcoming ACES special issue. We also have a long-in-the-works Eos article sitting on Tom's plate. Cheers, Jeremy