LSST 2017 PSF Session


(Jmeyers314) #1

Hey community,

I’m working on the agenda for the PSF Session at LSST 2017, and thought it would be a good idea to solicit input from the LSST community, especially since this session wasn’t previously featured on community.lsst.org.

The main items on the agenda so far are:

  • Updating LSST’s PSF modeling requirements. For this item I’m broadly aware of requirements for weak lensing science, but less so for other kinds of science. If you have PSF model requirements you’d like to discuss, please let me know!

  • Current thoughts on separately modeling the optics+sensor PSF contribution and atmospheric PSF contribution. I’ll present what I think this might look like and what I think are the most worrisome stumbling blocks and potential research questions.

  • Chromatic PSF corrections. Meyers&Burchat15 made suggestions for how to implement some such corrections, and I believe Alerts Production has some independent ideas on how to handle differential chromatic refraction for difference imaging. We should compare these.

If you’re working on PSF modeling, and you’d like to contribute to this session, please let me know.

Thanks!
-Josh Meyers (your LSST DM PSF modeler)


#2

Looking forward to this!


(Melissa Graham) #3

Hi Josh,

I’ve been working on a DM study that includes an investigation of how data from special programs – which might have a shorter/longer exposure time than the 15-30 second images from the main survey – can be integrated into the Level 1 and 2 pipelines and products. I became interested in the question of what is the minimum exposure time that an image can have and still be successfully run through the DIA pipeline, where “successfully run” means the image’s PSF is well-enough formed for PSF-matching and differencing, and therefore alert generation. @bxin gave me some Arroyo atmospheric LSST-like PSF models for a variety of exposure times, and I started with some very simple tests of just e.g., fitting a 2D Gaussian and plotting the centroid and its error vs. exposure time. This work is not progressed farther than that, but I do plan to attend this session to learn more about what has perhaps already been done in this area and/or how best to proceed. Over the coming year the community will be invited to submit white paper proposals for mini-surveys, and that’s why I’m interested in this kind of image processing constraint on exposure time.

See you in Tucson!
Melissa

P.S., This posting might be of interest also to: @connolly, @bxin, @CStubbs, @reiss, @isullivan, @ctslater, @ebellm