SV Cluster Mass & Light Telecon 08.01.2013¶
participants: Armstrong, Jarvis, Clampitt, Vikram, Plazas, Patton, Suchyta, Huff, Melchior
new UPenn shape measurements¶
Bob summarized that increasing the shapelet order to 10, and the PSF polynomial order to 8 on the entire coadd yielded good results in terms of D1/D2. Further increases in either parameter were not helping.
He also indicated that the star-galaxy cross-correlation seemed not not vanish, and we agreed that this might not necessarily be the case in a cluster field. The next steps here are to compare the S/G correlation function to see whether the two pipelines agree in a non-vanishing result; if so we need a theoretical calculation to tell us if this is to be expected.
new data products¶
Eric Suchyta explained that the new coadds comprise old and newer exposures (the latter in g and r band under dark conditions plus deep Y-band coverage for El Gordo). Following Huan's coadd script, Eric will produce a consistent set of coadds on a larger area (probably around 2 deg^2) using the new r-band data as detection band as it is deeper and with a smaller PSF than the previous i-band. Also, the PSF ellipticity is much smaller then before, indicating that the engineering run in between was actually helpful.
Mike expressed concerns that we are not using the DM coadds, which are not yet available for these fields (Bob). Peter argued that this data has a different observations strategy (both depth and dithers) than the main survey such that we don't have to use the DM coadds. We agreed to run on the DM coadds if they become available in time (Bob asks DM to produce what we want) and report back to Dm if we find problems.
comparison to X-rays¶
Vinu said that he started to collect references for X-ray analysis. Peter promised to help on the SZ side.
Peter argued that X-ray comparison will be rather coarse: centroid and M_500, and that these cluster are not relaxed, some badly violating the hydrostatic equilibrium assumption.
see new draft for changes to content.
Other issues that came up when discussing the draft:
- Are we required to publish in a particular journal?
- For the shear testing section, we agreed on a galsim-based script to produce a realistic cluster field, assuming a spherical NFW halo of mass and redshift matching the observed ones (one for the z=0.3 and one for El Gordo sufficient), using the PSF shape (from PSFEx) and the same distributions of galaxy S/N and sizes as observed in the data. This test data provide two things: (weak) check for shear calibrations, and a realistic shape catalog error budget.
- For this purpose, Mike wants to incorporate a PSFEx module in galsim to generate spatially varying PSFs that actually look like the ones in the observations. Peter offered a code her got from Daniel for this purpose and has asked Daniel if he agrees with public release (answer pending).
- Vinu mentioned that cross-matching OSU and UPenn shape catalogs was not working well. Peter promised to provide a catalog with actual RA/Dec, instead of pseudo coordinates.
- We agreed on Mike suggestion adding tangential shear profiles from the two pipelines to the paper because they offer simple error estimates on both statistical (error bars) and systematic errors (cross shear).
- Since data comes from SV, Gary should be included in discussions.
In terms of job assignments, everyone was happy. The time scales for some critical jobs were then discussed:
- Eric Suchyta can provide the new coadds in the next days.
- The bulk of the shear accuracy testing is to be done during the UPenn meeting.
- The galsim script should be ready before that, but most building blocks are already in place
- All color-selection and photo-z work will commence with the new coadd catalogs.
- The redmapper cluster-member catalog and the mass maps from Portsmouth, both based on the old coadds, are coming in during the next days.