Sky model bug fixDo not propagate the sky line amplitude scaling to the final sky mode… #1452
Conversation
|
Thanks for the quick fix. The sky model for night 20211014 r6-00104284 looks much better now (old left, this PR right): The resulting sframe also looks much better, though there are still some strong residuals for some sky lines (old sframe-r6-00104284 left, this PR right):
It does, however fix the z~4.3 excess the originally motivated this PR: Old (current master): New (this PR): This is definitely better than current master, but those strong sky residuals are concerning so let's investigate those a bit more before merging. |
|
Julien and I looked into those sky residuals and it appears to be due to the sky genuinely varying across the focal plane in a way that the model can't absorb, and it also is present when processing with older code so it isn't a leftover bug from the PCA adjustments. Side note: this plot is at H-alpha, but a similar structure appears at other sky lines (but not all of them!): For kicks: The same structure is also clearly visible in H-beta, and maybe in H-gamma and -delta. I don't immediately recognize what the wavelengths redward of H-alpha are (roughly 6585, 6719, 6733, ...) |
|
For this tile 7951, the Balmer series emission is from the Milky Way as one can see in the H-alpha sky maps: https://www.legacysurvey.org/viewer?ra=48.3831&dec=-6.7536&layer=halpha&zoom=7&desifoot=48.3570,-6.5600&tile=7951 This raises the separable question of if+how we may want to subtract Galactic nebular emission. The current pipeline will essentially be subtracting the mean of the nebular emission across each DESI petal. |
These lines are [NII] 6584 (the stronger of the two doublet lines with [NII] 6548) and the [SII] 6716,31 doublet. |
|
Nice! I had been looking for the H-alpha map on the viewer under "overlays" rather than "more surveys"; that's a good trick to know. I'm now imagining that some will ask us to subtract the spatially varying nebular emission, while others will ask us to purposefully not subtract it because that's their science signal... :) Thanks @moustakas for identifying the other lines. Screenshot from the viewer link that David provided, for comparison with the plot I posted above from DESI observations: |
|
What is the source of the halpha map? Seems like we would want to include it in our set of maps to test against for selection function variations?
Cheers,
Ashley
From: David Schlegel ***@***.***>
Date: Tuesday, October 19, 2021 at 1:04 PM
To: desihub/desispec ***@***.***>
Cc: Subscribed ***@***.***>
Subject: Re: [desihub/desispec] Sky model bug fixDo not propagate the sky line amplitude scaling to the final sky mode… (PR #1452)
For this tile 7951, the Balmer series emission is from the Milky Way as one can see in the H-alpha sky maps: https://www.legacysurvey.org/viewer?ra=48.3831&dec=-6.7536&layer=halpha&zoom=7&desifoot=48.3570,-6.5600&tile=7951
This raises the separable question of if+how we may want to subtract Galactic nebular emission. The current pipeline will essentially be subtracting the mean of the nebular emission across each DESI petal.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub<#1452 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADLSM5XA2MGFIJU6MWCRULLUHWQJTANCNFSM5GHDHBWA>.
Triage notifications on the go with GitHub Mobile for iOS<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
|
@ashleyjross the attribution is in the lower-right corner of the viewer in tiny tiny font. The map is on a composite made by Doug Finkbeiner from a few different surveys, including WHAM-- |
|
Thanks!
From: Moustakas ***@***.***>
Date: Tuesday, October 19, 2021 at 1:32 PM
To: desihub/desispec ***@***.***>
Cc: ashleyjross ***@***.***>, Mention ***@***.***>
Subject: Re: [desihub/desispec] Sky model bug fixDo not propagate the sky line amplitude scaling to the final sky mode… (PR #1452)
@ashleyjross<https://github.com/ashleyjross> the attribution is in the lower-right corner of the viewer in tiny tiny font. The map is on a composite made by Doug Finkbeiner from a few different surveys, including WHAM--
https://faun.rc.fas.harvard.edu/dfink/skymaps/halpha/
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub<#1452 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADLSM5UFEMCTYIMNWG2Q4QLUHWTRJANCNFSM5GHDHBWA>.
Triage notifications on the go with GitHub Mobile for iOS<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
When adjusting sky lines wavelength and width, an amplitude is fitted at the same time (and effectively marginalized). In the most recent modification to the sky model, those scale factors, interpolated across wavelength and fibers (with a median filter) were erroneously included in the final sky model. This is a bug that causes some discontinuities across fibers (because of the median filter). This term is removed in this PR.