Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-21881

Reproduce PSF discrepancy on coadds

    XMLWordPrintable

Details

    • Story
    • Status: Done
    • Resolution: Done
    • None
    • None
    • None
    • 36
    • DRP F19-5, DRP F19-6 (Nov), DRP S20-2 (Jan), DRP S20-3 (Feb), DRP S20-5 (Apr), DRP S20-6 (May)
    • Data Release Production

    Description

      The size of PSF measured on the coadds was significantly different from those measured on single visits and warp (see the plot from rearmstr attached ). This ticket is to process the HSC data and re-make similar plots from the latest repo.

      Attachments

        Issue Links

          Activity

            The discrepancy in size residuals between stars and PSF models appears to have been a temporary issue. This is seen in S16A and in PDR2 data, but has diminished amplitude in the recent HSC re-processing runs done on lsst-dev machines. Potential contributing factors for size residuals such as interpolation errors, background subtraction schemes are ruled out as possible explanations. The notebooks below summarize the analysis done. While the problem has largely been taken care of, there will be an on-going inspection to identify the source of the size residuals to avoid such mismatches in the future.

            https://github.com/arunkannawadi/lsst-dm/blob/master/check_interpolation.ipynb

            https://github.com/arunkannawadi/lsst-dm/blob/master/WarpPSF.ipynb

            https://github.com/arunkannawadi/lsst-dm/blob/master/simulating_stars.ipynb

            https://github.com/arunkannawadi/lsst-dm/blob/master/analysis.ipynb

            https://github.com/arunkannawadi/lsst-dm/blob/master/RhoStatistics.ipynb

            kannawad Arun Kannawadi added a comment - The discrepancy in size residuals between stars and PSF models appears to have been a temporary issue. This is seen in S16A and in PDR2 data, but has diminished amplitude in the recent HSC re-processing runs done on lsst-dev machines. Potential contributing factors for size residuals such as interpolation errors, background subtraction schemes are ruled out as possible explanations. The notebooks below summarize the analysis done. While the problem has largely been taken care of, there will be an on-going inspection to identify the source of the size residuals to avoid such mismatches in the future. https://github.com/arunkannawadi/lsst-dm/blob/master/check_interpolation.ipynb https://github.com/arunkannawadi/lsst-dm/blob/master/WarpPSF.ipynb https://github.com/arunkannawadi/lsst-dm/blob/master/simulating_stars.ipynb https://github.com/arunkannawadi/lsst-dm/blob/master/analysis.ipynb https://github.com/arunkannawadi/lsst-dm/blob/master/RhoStatistics.ipynb

            Can you pull out one plot that says "The PSFs in the LSST-reprocessing of HSC PDR2 shows the problem Bob demo'ed went away"? If it's clear enough for people outside DRP to understand, swinbank will include it in his monthly report.

            yusra Yusra AlSayyad added a comment - Can you pull out one plot that says "The PSFs in the LSST-reprocessing of HSC PDR2 shows the problem Bob demo'ed went away"? If it's clear enough for people outside DRP to understand, swinbank will include it in his monthly report.

            Is this a simple enough plot for swinbank to demonstrate that our PSF modelling is much better? I can figures from S16A and/or visit-level plots if preferred.

            kannawad Arun Kannawadi added a comment - Is this a simple enough plot for swinbank to demonstrate that our PSF modelling is much better? I can figures from S16A and/or visit-level plots if preferred.

            This looks great — thank you kannawad!

            swinbank John Swinbank added a comment - This looks great — thank you kannawad !

            People

              kannawad Arun Kannawadi
              kannawad Arun Kannawadi
              Yusra AlSayyad
              Arun Kannawadi, John Swinbank, Yusra AlSayyad
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Jenkins

                  No builds found.