Attendees : ajc, Kem, Kirk, MWV, Peter, Richard, Scott, Yusra, Cathy, Simon, Jim, John, Debbie
1. DESC meeting
- 100 galaxy/cosmology simulations generated to date (most small). Issue is with getting simulated observed catalogs not the n-body work. 2 areas need to be addressed: fit SEDs so the data can be pushed through phosim and then shape and size properties for the galaxies
- Can use SED fitting from the catalog framework and cosmology group will generate shapes
- Debbie has the DES catalogs with mags from DES that could be used
- We need to improve the SEDs for photoz (PCA implementation and emission lines) to help the photoz groups
- MAF - lots of interest and Alex Kim already generated a set of metrics.
- Github has DESC organization and they are hosting metrics. Question about whether other working groups (non DESC) will have access (yes).
- Implementing RA bounded rolling cadences and swiss cheese model in opsim
- Not going to change rolling cadence name (MWV made that request)
- John - Stubbs requested variable exposure times to go to fixed S/N depths, Kem - opsim can vary exposure per filter but not across proposals
- 2 sessions with a number of technical talks. Lot of different talks on detailed physics or different systems being simulated, grids of PSFs are being generated to understand the PSF varies across the focal plane, and testing blending and deblending algorithms.
- Have to continue to make the documentation and interfaces clean. Lots of phosim experts could send those to external groups
- DESC mainly interested in the analysis algorithms. Mix of use cases (e.g. whether people used DM to process data)
- Getting enthusiasm for the data challenges (short and long term)
- Long term for 2018 (take a couple of years to build these).
- Short term with analysis challenges. LSS will be the prototype. Could do this with catalogs and apply errors initially and then move to phosim
- Will need to work on coming up with the goals for these data challenges
- arc - GAIA ran 4 data challenges in 1.5 years (incrementally added to the simulation data), people who didn't use the data challenges are now in a more difficult position (with data coming in) than those that did work with the data challenges.