stsci-hack
commented
2 years ago Moved to being a Jira ticket HLA-699.
Closed stsci-hack closed 2 years ago
stsci-hack
commented
2 years ago Moved to being a Jira ticket HLA-699.
stscijgbot-hstdp
commented
6 months ago This issue is tracked on JIRA as HLA-699.
stscijgbot-hstdp
commented
6 months ago Comment by Michele De La Pena on JIRA:
Here is another example of a dataset for this ticket. The filter-level IR image look OK, but the total is bad. The UVIS images may be OK.
File: ibig07qwq_flt.fits detector: IR subarray: True filter: F110W pctecorr: n/a File: ibig07qxq_flt.fits detector: IR subarray: True filter: F110W pctecorr: n/a File: ibig07qyq_flt.fits detector: IR subarray: True filter: F110W pctecorr: n/a File: ibig07qzq_flt.fits detector: IR subarray: True filter: F110W pctecorr: n/a File: ibig07r0q_flt.fits detector: IR subarray: True filter: F160W pctecorr: n/a File: ibig07r1q_flt.fits detector: IR subarray: True filter: F160W pctecorr: n/a File: ibig07r2q_flt.fits detector: IR subarray: True filter: F160W pctecorr: n/a File: ibig07r3q_flt.fits detector: IR subarray: True filter: F160W pctecorr: n/a File: ibig07r4q_flc.fits detector: UVIS subarray: False filter: F336W pctecorr: COMPLETE File: ibig07raq_flc.fits detector: UVIS subarray: False filter: F336W pctecorr: COMPLETE File: ibig07rcq_flc.fits detector: UVIS subarray: False filter: F336W pctecorr: COMPLETE File: ibig07rfq_flc.fits detector: UVIS subarray: False filter: F438W pctecorr: COMPLETE File: ibig07rhq_flc.fits detector: UVIS subarray: False filter: F438W pctecorr: COMPLETE File: ibig07rvq_flc.fits detector: UVIS subarray: False filter: F438W pctecorr: COMPLETE File: ibig07rjq_flc.fits detector: UVIS subarray: False filter: F814W pctecorr: COMPLETE File: ibig07roq_flc.fits detector: UVIS subarray: False filter: F814W pctecorr: COMPLETE File: ibig07rqq_flc.fits detector: UVIS subarray: False filter: F814W pctecorr: COMPLETE File: ibig07rsq_flc.fits detector: UVIS subarray: False filter: F814W pctecorr: COMPLETE File: ibig07riq_flt.fits detector: UVIS subarray: True filter: F814W pctecorr: OMIT
The "n/a" means the keyword PCTECORR was not found in the primary header.
The HLA has identified a use-case where multiple apertures (subarray formats) were used along with full frame exposures to observe a moving target. Not surprisingly, drizzlepac was unable to correctly use the WCS information to correctly combine these exposures. It is not clear what can be done, but this issue will provide a reminder to review this use-case to see what might possibly be done to generalize even further how these sets of exposures should be combined.
The following details came from a help call filed by R. White (HLA):
_Here are a couple of examples. Visit 17 of proposal 11559 has these 10 WFC3/UVIS exposures:
filename aperture ccdchip ib4v17plq_flc.fits UVIS-QUAD-SUB 2 ib4v17pmq_flc.fits UVIS-QUAD-SUB 2 ib4v17pnq_flc.fits UVIS-QUAD-SUB 1 ib4v17poq_flc.fits UVIS-QUAD-SUB 2 ib4v17ppq_flc.fits UVIS-QUAD-SUB 2 ib4v17pqq_flc.fits UVIS-QUAD-SUB 2 ib4v17psq_flc.fits UVIS-QUAD-SUB 1 ib4v17puq_flc.fits UVIS1-2K4-SUB 1 ib4v17pvq_flc.fits UVIS-QUAD-SUB 2 ib4v17pwq_flc.fits UVIS-QUAD-SUB 2
These are all subarray exposures of Jupiter. Most of the observations use the UVIS-QUAD-SUB aperture, but that apparently moves from chip to chip depending on the quad filter being used. And the ib4v17puq_flc.fits observation uses the UVIS1-2K4-SUB aperture, which is a much larger subarray. Applying the WCS from one image to another leads to the weird result shown in the attached preview.
Another example is visit 11 of proposal 13675, which has 11 WFC3/UVIS exposures:
filename aperture ccdchip(s) icoc11fmq_flt.fits UVIS 2 1 icoc11fnq_flt.fits UVIS 2 1 icoc11foq_flc.fits UVIS2-2K2C-SUB 2 icoc11fpq_flc.fits UVIS-QUAD-SUB 1 icoc11frq_flc.fits UVIS2-2K2C-SUB 2 icoc11fsq_flc.fits UVIS2-2K2C-SUB 2 icoc11ftq_flc.fits UVIS-QUAD-SUB 1 icoc11fuq_flc.fits UVIS-QUAD-SUB 1 icoc11fvq_flc.fits UVIS2-2K2C-SUB 2 icoc11fwq_flt.fits UVIS 2 1 icoc11fxq_flt.fits UVIS 2 1
This visit has several different apertures, including some UVIS apertures that are not subarrays (and so have both CCD chips) and some subarrays (with only 1 chip).
For this visit the results are really weird. I'm pretty sure that what happens is that one of the 2-chip observations is used as the reference image for the WCS. It supplies a WCS for both chips. When it gets to one of the images that is a subarray, there is no second chip. The 4th extension in those files is the D2IMARR extension, which is a 64x32 pixel distortion array. Apparently the astrodrizzle code blithely proceeds to apply the WCS to that tiny coefficient image and drizzles it into the output! The results for exposure icoc11foqflc.fits are shown in the second image.