Minutes from 24 September 2008

Attendees: Tom J., Phil M., Jeremy Z., Danijel S., John Y., Masha L.

1. CSEP V8.10 release candidate:

   • Tagged in repository with tags/Version_8_9_21, nightly build succeeded

   • Includes MASS and ROC alarm-based tests from Jeremy, to be applied to:
     
     1. RELM 5-year rate-based forecasts for California
     2. Jeremy's 5-year Triple_S forecasts for California
     3. One-year forecasts for NW and SW Pacific testing regions
   • Includes 4 new models from Jeremy:
     
     1. 5-year rate-based and alarm-based forecasts for California
     2. 1-year alarm-based forecasts for NW and SW Pacific testing regions
   • Warner responded to Masha's e-mail:
     
     1. Would prefer to have a single version of the Double Branching Model code installed within CSEP
     2. Will fix the code for SW Pacific region
     3. Will be installed within CSEP for V9.1.0 release
   • Nightly Dispatcher processes exercise code base of the release candidate
   • John is working on the web-site to display new forecasts groups and alarm-based tests results:
     
     1. Jeremy will send interpretation for alarm-based tests - to be posted on the web-site
   • Fixed remaining tasks for the release:
     
     1. matplotlib: trajectory label doesn't display markers in the label - fixed by removing negative offset for the legend
     2. Jeremy submitted a fix for negative longitude for SW Pacific models
   • log10(forecast rate) for the map display to get more detailed forecasts maps is postponed until next release
   • Masha is updating Trac documentation pages
   • Danijel will send a V8.10.0 release announcement to the csep mailing list when release is installed on operational server
   • Phil suggested to have a release status meeting on Tuesday, September 30, at 1:30pm
2. Hardware:
   • New RAM is available for csep-op: will upgrade on October 1, 2008
3. Keck report:
   • Phil and Danijel distributed the draft to the group
   • Tom will contribute to the draft
   • Site review at USC
4. Tom asked is CSEP is capable of retrospective testing if new models become available:
   • Masha confirmed that functionality is in place
   • The only development that would be required for the file-based models is to convert them into the CSEP forecast format
5. Tom mentioned about possible collaboration with Chinese team:
   • North South Seismic Belt
6. ACES technical paper:
   • Jeremy is working on the draft
   • Jeremy asked for extension to submit the paper
7. Tom and Danijel are working on the CSEP paper with possible submission to:
   • Earth and Planetary Science Letters
   • EOS
   • Nature Geoscience