LSC-Virgo Burst Analysis Working Group
Review Committee Meeting Monday November 23 2009 08:00 Pacific / 11:00 Eastern
Minutes: Monday November 23 2009 08:00 Pacific / 11:00 Eastern
Agenda and Contact Info
- All-sky paper
- On-line analysis: position reconstruction
- On-line analysis: all else
On EVO use "bersten" to get in, or phone bridge ID 39354 and "6519" to get in.
- Present: Owen, Hughey, Kanner, Gouaty, Searle, Katsavounidis, Klimenko, Yakushin, Salemi, Prodi, Robertson, Vedovato, Flaminio, Drago, Rollins, Daw, Hello, Mours, Whitcomb, Shawhan
- All-sky paper
- On Nov 19 the LSC Executive Committee approved the paper subject to some minor revisions, which have been made except:
- Author list must be finalized (Virgo needs to update). Shawhan will incorporate.
- Figure 1: Bigger fonts. Change LIGO calibration from v3 to v4. Check GEO and Virgo calibrations. People are working on this offline, Shawhan coordinating.
- Sec.VIII language vague on which results are from new numbers, which from combined w/S5y1. Owen will fix.
- On Nov 10 Virgo assigned a reviewer; report is pending.
- What next? Make all changes and circulate to LSC and Virgo for final cooling off period; then arXiv and another wait; then PRD.
- Omega position reconstruction walkthrough
- These are from 1000s of SW injections in JW1 (S5). However code version is what will run on-line now.
- How are position errors used? Too large errors (1 sq deg for Swift, few for most optical, 50 for QUEST) mean trigger is not passed, unless box is sparsely inhabited by nearby galaxies (each "pointing" is really a few).
- Accuracy depends, but hit 1 sq deg threshold for sine-gaussians with signal/noise about 10 (high frequency) or 30 (low frequency). So mostly high ones will make it, just life with wave mechanics.
- WNB areas are 30x bigger. Not surprising since pipeline was built on s-G's!
- Error bars are too optimistic by construction. For really loud WNBs this can be huge (5% in 90% box) but normally errors are accurate to few %.
- Calibration errors (at least for JW1) were fraction of degree, liveable.
- cWB position reconstruction
- Again this is JW1, but new code version.
- Look at Table 2. "B" is what will run on-line. Hard to read, errors look worse than Omega? That is line 12, WNB should be worse than SG injections.
- "Page" link ... sometimes error bars are optimistic, sometimes pessimistic. Not usually drastic.
- Searle: Line 20, SG injections, only 26% in 50% error box? Klimenko: Oh, it's a bug. Owen: Well fix it!
- Calibration errors: (pdf) We are talking about doing only the unmodeled search, so extra error is (red line - blue line) or a fraction of a degree. Liveable. This is fake S6 noise, also did JW1 and will do new noise.
- HOMEWORK (both analyses):
- Generate clear, concise plots like Searled did for Omega, for both analyses, for data now coming in. Principals say something partial can be shown next Monday, good job can be done in 2 weeks deadline.
- Run the whole thing NOW as if Stan Whitcomb is SWIFT (can send e-mail instead of 3am phone call, will check time stamp). He is worried Omega is not sending HW injections to LUMIN.
- EVO kicks us out as discussion turns to HW injections and possibly having more of them.
- What else is needed for the review team? - See e-mail.
$Id: minutes-2009-11-23.html,v 1.3 2009/11/23 23:05:31 owen Exp $