Project

General

Profile

Agenda And Minutes for 1111 10am CST » History » Version 13

Eli Rykoff, 11/11/2013 12:13 PM

1 1 Eli Rykoff
h1. Agenda And Minutes for 11/11 10am CST
2 1 Eli Rykoff
3 1 Eli Rykoff
h2. Call-in Info
4 1 Eli Rykoff
5 1 Eli Rykoff
Call-in instructions:
6 1 Eli Rykoff
7 1 Eli Rykoff
1. Dial Toll-Free Number: 866-740-1260 (U.S. & Canada)
8 1 Eli Rykoff
2. International participants dial:
9 1 Eli Rykoff
Toll Number: 303-248-0285
10 1 Eli Rykoff
Or International Toll-Free Number: http://www.readytalk.com/intl
11 1 Eli Rykoff
3. Enter your 7-digit access code, followed by “#”
12 1 Eli Rykoff
13 1 Eli Rykoff
Your access code: 7691719
14 1 Eli Rykoff
15 1 Eli Rykoff
h2. Agenda
16 1 Eli Rykoff
17 1 Eli Rykoff
# Should there be an "SVA1 Catalog" Paper?
18 10 Eli Rykoff
# Bad Photometry Healpix Maps: Eli [[Bad Photometry Maps]]
19 11 Keith Bechtol
# SLR Updates: What to do about reddening? Keith ("redmine":https://cdcvs.fnal.gov/redmine/projects/descalibration/wiki/Stellar_Locus_Regression_for_SVA1_on_HEALPix_Grid, "slides":http://des-docdb.fnal.gov:8080/cgi-bin/ShowDocument?docid=7598)
20 1 Eli Rykoff
# MAG_MODEL tests?
21 12 Diego Capozzi
# Mangle -> depth: Diego [[Some tests on depth with Mangle Mask]] (Paragraphs 2 & 3)
22 13 Eli Rykoff
23 13 Eli Rykoff
h2. Minutes
24 13 Eli Rykoff
25 13 Eli Rykoff
Attending: Eli, William W., Tom D., Robert G., Felipe M., Ricardo, Gary B., Huan, Nacho, Anne, Steve K., Keith, Diego, Manda, Maayanne, Douglas, others?
26 13 Eli Rykoff
27 13 Eli Rykoff
1. We discussed if there will be an SVA1 catalog paper in, e.g, ApJ Supplements.  It was noted that (a) the DESDM paper is being worked on, and will definitely *not* include anything about data release, and (b) there are no plans or intents for an SVA1 catalog data release.  Nevertheless, Douglas (and Eli) thought it would make sense to have a paper compiling our tests and value-added catalogs.  If there were a targeted ApJ issue with DES SVA1 this would be a no-brainer, but I don't believe that we are planning something like this.
28 13 Eli Rykoff
29 13 Eli Rykoff
I note that the alternatives to having an SVA1 catalog paper are either piecemeal discussions of the issues in SVA1 science papers, or an additional section in the DESDM paper.  This may be significantly broader than the purview of the DESDM paper however.
30 13 Eli Rykoff
31 13 Eli Rykoff
2. A digression on star/galaxy separation.  Maayanne said that we should expect updates on her neural network S/G separation in 2 weeks.  Meanwhile, we also discussed how SPREAD_MODEL is limited at the faint end, where it misclassifies lots of unresolved galaxies as stars.  At the very least, we need a cut fainter than which we don't trust spread_model (though where to put this in and the testing of it are still TBD.)  Question for Huan: what is the status of the COSMOS full-depth coadds?
32 13 Eli Rykoff
33 13 Eli Rykoff
3. Eli discussed the [[Bad Photometry Maps]].  These will be released as HEALPIX maps of 0s and 1s.  The key here is that these are to be used in addition to "sanitizing" the input catalog of bad objects with crazy colors, by masking out entire regions that are bad we can more accurately create random points to indicate where we cannot detect any galaxies.  
34 13 Eli Rykoff
35 13 Eli Rykoff
Felipe mentioned that there are people also working on masking out NGC galaxies (which had to be done for BOSS-DR8) which are not "bad regions" per se, but are instead simply objects that mask background galaxies.  Similarly, very bright stars (from, eg, the Yale Bright Star Catalog, with R<6) have to be masked outside of the mangle depth maps.  For SVA1, we will join forces and have one mask file that will incorporate bad regions, bright very extended galaxies, and very bright stars.
36 13 Eli Rykoff
37 13 Eli Rykoff
In the future, Y1+, we should work to incorporate the list of bright galaxies and stars into Mangle directly.  This was the mode in DC6+, but found to be wanting when confronted with real data.  However, SVA1 data can and should be used to get a radius-catalog magnitude scaling relation for these brightest stars.
38 13 Eli Rykoff
39 13 Eli Rykoff
Question for Aurelien: can we flag these molygons so that it's clear where the mask came from (bright star, bright galaxy, etc).  
40 13 Eli Rykoff
41 13 Eli Rykoff
4. Keith presented his SLR "slides":http://des-docdb.fnal.gov:8080/cgi-bin/ShowDocument?docid=7598 .  Strongly recommended!  Gary suggested that the issues with the failures in the LMC may be helped by making additional color cuts to the input stellar catalog fed into the code.  The maps are now available, along with python code to do the linear interpolation.
42 13 Eli Rykoff
43 13 Eli Rykoff
5. We discussed how to distribute the SLR (and other related products) to avoid people duplicating effort and running the offsets manually.  Huan mentioned that there would be great value in having a global catalog that is downloadable, perhaps via the Portal.  Eli thinks the best option for everybody will be to (a) upload the SLR results to a database table (OBJID, OFFSET_G, _R, _I, _Z, FLAG); (b) create a standard query (including SLR joins, crazy color cuts, star/galaxy separation TBD, etc); (c) make the standard query and resulting flat file available.  Thus, those who wish to use the database to get specific columns will be able to do so, and those who wish to quickly grab a flat file and skip the database will have easy access to it.
44 13 Eli Rykoff
45 13 Eli Rykoff
Nacho has sent around instructions for creating and uploading to a personal database table.
46 13 Eli Rykoff
47 13 Eli Rykoff
6. Diego discussed the mangle depth: [[Some tests on depth with Mangle Mask]].  An issue was found in that some molygons were said by mangle to have very deep depth (i>25) even though they don't have much exposure time.  These molygons are all in the same region, so clearly there was a mangle failure here.  He also noted that the "galaxy number count" depth was significantly shallower than the mangle "error-based" depth.  Eli believes this is due to incompleteness at the faint end, which will tend to truncate the number counts.  In addition, the SPREAD_MODEL cut for star/galaxy separation may be contributing to the incompleteness at the faint end as faint unresolved galaxies are getting rejected as stars.