[SPRINT] DC2 mean shear / seagull challenge

[chihway]

DC2 mean shear / seagull challenge

Explore DC2 catalog to understand why we get a mean g1~-0.003 and mean g2~0. Also the seagull-like correlation between g1 and g2.

Contacts: Erin, Mike, Matt, Chris, Francois, Chihway Day/Time: Throughout the week, no specific time yet but maybe will call a meeting at the end of the week to regroup. Main communication channel: #desc-wl-shear GitHub repo:

Goals and deliverable

Figure out why mean g1 is not zero.

Resources and skills needed

Anyone who's interested in exploring data and plotting seagulls.

Detailed description

Some directions for exploration:

• Select subsets of the shear catalog based on different quantities (sub-pixel centroid location, isolated galaxies etc.)
• Stack expert check erin’s code for psf reconstruction, using deblender outputs, postage stamps etc.
• Look at HSM to see if it shows the same
• Reuse Emily's code to check for potential effects from optics misalignment

[esheldon]

I'll also add as goals

• figure out why there is a kink in the mean g1 vs true g1 plot
• figure out why when selecting +g1 true shear (in the original convention) we get a small m but when selecting -g1 true shear we get 35%

[cwwalter]

A few more:

• Figure out why there is a kink in mean true g1 vs binned true g2 [must be the same reason as above?]
• Check for dependence on true lensing vs observed position
• Study behavior as a function of size up to the full 5000 sq degrees input catalog

[esheldon]

I think the new thing I learned first day of the sprint was that the bias for objects with true shear g1 < 0 is huge, but fairly small for g1 > 0. I consider this work, done sunday night as part of the sprint.