WFC3 STAN - Issue 22, February 2016
1. Spatial Scan Modes in the ETC
L. Dressel, P. McCullough
A new version of the Exposure Time Calculator, ETC 24.1, is available at this page. Scan modes are being introduced for UVIS spatial scan imaging, IR spatial scan imaging, and IR spatial scan spectroscopy; see sections 6.12.3, 7.10.4, and 8.6 in the WFC3 Instrument Handbook. They are intended to help observers plan exposure times and scan rates, and avoid saturation in the peak pixels. The target is assumed to be a point source. The user specifies a scan length and either a scan rate or an exposure time. For purposes of computation, it is assumed that the scan proceeds along a column. The user can choose an extraction height of 1 pixel or the entire scan length. For the imaging modes, there is a choice of extraction width of 1, 3, 5, 7, or 9 pixels. The IR spectroscopic scan mode uses the same convention as the IR spectroscopic stare mode, reporting count rates for a width of 1 pixel and counts for a width of 1 resolution element (2 pixels).
2. Cycle 23 WFC3 Calibration Plan
The Cycle 23 calibration plan for the UVIS and IR channels of WFC3 has been defined, and a summary of the calibration programs can be found here. This is largely a continuation of the routine calibration and monitoring programs from Cycle 22 and earlier cycles (see WFC3 ISR 2015-07). There are two main new initiatives for Cycle 23: (1) Completion of the astrometric calibration for all UVIS filters, including all narrow band and quad filters (Program 14393) and (2) Programs to fully characterize the spatial behavior of IR persistence across the detector (Programs 14380 and 14381). Results from these programs will be advertised in future STANs and presented in ISRs from the WFC3 team.
3. Improved UVIS Dark Calibration
The forthcoming release of the CALWF3 version 3.3 software will feature several improvements to the WFC3/UVIS dark calibration. These changes are modifications to the previous algorithm, which is described in WFC3 ISR 2014-04. As a result of these improvements, new dark calibration reference files (i.e. superdarks) have been generated for all on-orbit data for use in the CALWF3 pipeline. Users can obtain the improved science data products by re-retrieving from the MAST archive or by manually reprocessing with the new dark reference files provided in the HST Calibration Reference Data System (CRDS).
The major improvements are:
1. CTE-corrected reference files. CTE-corrected superdarks (i.e. *_dkc.fits files) will be delivered to CRDS and used in the CALWF3 pipeline for generation of CTE-corrected data products (i.e. *_flc.fits, *_drc.fits files). The CTE-corrected superdarks help mitigate background signal introduced by CTE losses from the cosmic rays and hot pixels that are commonly found in WFC3/UVIS observations. Non-CTE-corrected superdarks (i.e. *_drk.fits) will continue to be provided for use in generating non-CTE corrected calibrated data (i.e. *_flt.fits, *_drz.fits).
2. Improved temporal accuracy of hot pixel populations. In the previous algorithm, hot pixels were flagged based upon an average-combined set of dark frames taken from non-overlapping 4-day windows, resulting in a new reference file roughly every four days. In the new algorithm, reference files are generated using a 'sliding' 4-day window instead, resulting in a new reference file each day. This yields a more accurate measure of the hot pixels affecting a given observation.
3. Improved accuracy of dark current measurement. The previous algorithm set non-hot pixels (i.e. those with dark current < 54 e-/hr) in the reference file to the frame's median value. However, in the new algorithm, each non-hot pixel is set to the pixel's 'masterdark' value, which is computed from an anneal-cycle average (i.e. all dark observations occurring between two UVIS anneal procedures, ~100 dark frames). This results in a more accurate measurement of the true dark current across the detector.
Further details on these improvements will be presented in forthcoming WFC3 Instrument Science Reports.
4. Sampling Sequence SPARS5 Full-Frame and Subarray Dark Calibration Files Released
C. Martlin, S. Baggett
The recently developed sample sequence SPARS5, which went operational on WFC3 on Sep 8, 2015, now has superdark reference files for all aperture sizes available for use in the calibration pipeline.
The superdarks for each array size were generated from a set of on-orbit dark frames taken during September 2015; the new reference files will replace the placeholder dummy superdark files previously in the pipeline.
Any SPARS5 science data taken after Nov 03, 2015 will automatically have these new superdarks applied by the pipeline. Observers with data taken before that date can re-retrieve their files from the MAST archive to have the calibration performed or download the appropriate superdark reference file, listed and retrievable below, and reprocess their data manually with CALWF3.
Observing Mode Dark Reference File
SPARS5 Full-Frame zb21929si_drk.fits
SPARS5 SQ512SUB zb219301i_drk.fits
SPARS5 SQ256SUB zb219300i_drk.fits
SPARS5 SQ128SUB zb21929ti_drk.fits
SPARS5 SQ64SUB zb219302i_drk.fits
For more detail on the SPARS5 mode, see the article by McCullough in the WFC3 STAN Issue 20.
5. Scan Data Processing
S. Baggett, M. Sosey
A recent update to the WFC3 data processing pipeline and calibration software includes two improvements for WFC3 scan data, where the target position in the detector field of view is continuously changing during the multi-accum exposure.
(1) All the UVIS and IR scan-related keywords, formerly accessible only via the engineering file headers (*spt.fits), will now be present in the calibrated science data headers (e.g. *flt.fits). These keywords include:
SCAN_TYP= 'C ' / C:bostrophidon; D:C with dwell; N:N/A SCAN_WID= 0.000000000000E+00 / scan width (arcsec) ANG_SIDE= 8.971199798584E+01 / angle between sides of parallelogram (deg) DWELL_LN= 0 / dwell pts/line for scan pointing (1-99,0 if NA) DWELL_TM= 0.000000000000E+00 / wait time (duration) at each dwell point (sec) SCAN_ANG= 4.467850112915E+01 / position angle of scan line (deg) SCAN_RAT= 3.459574654698E-02 / commanded rate of the line scan (arcsec/sec) NO_LINES= 1 / number of lines per scan (1-99,0 if NA) SCAN_LEN= 1.139756870270E+01 / scan length (arcsec) SCAN_COR= 'V ' / scan coordinate frame of ref: celestial,vehicle
(2) In IR scan data, the default for the cosmic-ray correction calibration step (keyword CRCORR) is now OMIT, disabling the up-the-ramp fitting in CALWF3. The result is that the calibrated science data (*flt.fits) will consist of the first-minus-last read of the the multiaccum readout, a more reasonable representation of the image than the up-the-ramp fit. Observers with pre-existing scan data may re-retrieve their files from the MAST archive to obtain the improved products.
6. CALWF3 Version 3.3
The new release of the WFC3 calibration pipeline, version 3.3, comprises two fundamental changes to the processing of raw WFC3/UVIS data by CALWF3. First, photometric calibrations are determined independently for each CCD. Second, the processing pipeline, CALWF3, applies pixel-based CTE (Charge Transfer Efficiency) corrections.
The two UVIS CCDs were manufactured on different wafers and foundry production runs. Consequently, their physical properties, e.g. quantum efficiency and thickness, are different, and they may age differently. For example, UVIS2 is ~30% more sensitive in the UV than UVIS1; between 3500 and 7000 �, the two CCDs have similar sensitivity; and at wavelengths longer than 7000 �, UVIS1 is slightly more sensitive; see Figure 5.2 in the WFC3 Instrument Handbook. Thus motivated, photometric calibrations are now determined independently for each CCD in the UVIS channel.
UVIS Filters Affected by Version 3.3
Long Pass and Extremely Wide: F200LP, F300X, F350LP, F475X, F600LP, F850LP.
Wide: F218W, F225W, F275W, F336W, F390W, F438W, F475W, F555W, F606W, F625W, F775W, F814W.
Medium: F390M, F410M, F467M, F547M, F621M, F689M, F763M, F845M.
Narrow: F280N, F343N, F373N, F395N, F469N, F487N, F502N, F631N, F645N, F656N, F657N, F658N, F665N, F673N, F680N, F953N.
None of the quad filters are affected by this change.
New flat fields for each filter+CCD combination have been recomputed and normalized to the median value for each detector. Thus for each filter there are now two flat fields. That is, the flat field for chip 1 has been normalized to chip1 and the flat field for chip 2 normalized to chip 2. In the previous implementation of CALWF3, one flat field per filter was generated, with each chip normalized to the same selected region on chip 1, preserving the sensitivity offset between chips. In version 3.3, the chip-dependent sensitivity correction is instead applied as part of the photometry calibration step in CALWF3.
Furthermore, the UV flats, which were obtained under ambient conditions during ground testing, have now been corrected for sensitivity variations in a cross-hatch pattern on spatial scales of 50-100 pixels. White dwarf standards stepped across the two UVIS chips with the old flats showed photometric variations of +/-3% peak-to-peak (1.5% rms). With the new chip dependent flat fields, these are now corrected to +/-1.6% peak-to-peak (0.7% rms) with the new solutions.
Inverse Sensitivity (PHOTFLAM)
The inverse sensitivity is calculated for each filter/CCD combination using all observations of the white dwarf standard stars G191B2B, GD153, and GD71 obtained between 2009 and 2015 for an aperture of radius r = 0.4 arcsec. The resulting photometry has been improved compared to prior photometric calibrations. These new values agree with ACS photometry in the ST magnitude system. Users may notice that the version 3.3 inverse sensitivity values are systematically ~3% different, with some filters up to 6% higher (F200LP, F953N) compared to the published 2012 zeropoints. This effect is due to how AstroDrizzle handles cosmic-ray rejection for exposure-weighted images when exposure times are very different. Noting this, special care was taken in flagging cosmic rays for the 2016 PSFs, resulting in more accurate zeropoints. Users may wish to reprocess their data, particularly those who are interested in multi-epoch comparisons.
Photometry-Related Changes to CALWF3
One new calibration switch has been added: FLUXCORR; when this keyword and PHOTCORR are set to PERFORM, CALWF3 will compute the chip-independent sensitivities and scale chip 2 to chip 1. To support the new photometric approach, three new keywords are now added to the image headers: PHTFLAM1 = inverse sensitivity for UVIS1, PHTFLAM2 = inverse sensitivity for UVIS2 and PHTRATIO = PHTFLAM2/PHTFLAM1. The original inverse sensitivity keyword, PHOTFLAM is retained and has the same value as PHTFLAM1.
In addition to new flat fields applied to each CCD, there is a change in the structure of the photometry table (IMPHTTAB). The new IMPHTTAB has two new extensions, for a total of five. Extensions 1, 2, and 3 are the same as before, i.e. they contain the sensitivity values, pivot wavelengths, and bandwidths (PHOTFLAM, PHOTPLAM, PHOTBW), respectively, for all the configurations of the detector. Extensions 4 and 5 contain the observing mode and values for the independent inverse sensitivities: PHTFLAM1 for chip 1 (UVIS1) and PHTFLAM2 for chip 2 (UVIS2). Note that with UVIS2.0, PHOTFLAM is the same as PHOTFLAM1.
FLT and DRZ, FLC and DRC, and Drizzle
The application of the new zeropoints (PHOTFLAM, PHTFLAM1, PHTFLAM2) by CALWF3 will be transparent to the user. Most users will be able to use the *_flt.fits and *_drz.fits data products as before. Photometry in drizzled data products should now be vastly improved for UV filters, with source photometry of blue objects now continuous across the two CCD chips. For visible/red filters, the drizzled data products will be nearly identical for broadband filters but improved for narrow, medium, and long-pass filters where inflight flat-field corrections were based on interpolated solutions.
Charge Transfer Efficiency Corrections
The pixel-based CTE correction, previously available as a stand-alone FORTRAN program (see this page) has now been incorporated into the CALWF3 pipeline for all UVIS full-frame images. Controlled via a new calibration switch (PCTECORR = PERFORM) and associated calibration table (PCTETAB), CALWF3 will produce two sets of products: the standard non-CTE-corrected (e.g. *_raw.fits, *_flt.fits, *_drz.fits) files as well as the new CTE-corrected results (*_flc.fits, *_drc.fits). Users will be able to use the *_flc.fits and *_drc.fits data products in the same way as *_flt.fits and *_drz.fits files.
The UVIS CTE model algorithm is different from that operating in the ACS pipeline. Whereas the ACS model allowed traps to affect fractional pixel levels and worked on real-number pixel arrays, the WFC3 model explicitly deals only with integer numbers of electrons and specifies the cumulative number of traps as a function of packet size in electrons. The WFC3 model includes an improved readnoise-mitigation algorithm to identify the smoothest possible image that is consistent with being the observed image plus readnoise. The reconstruction algorithm operates on this smooth image to make a conservative estimate of how charge may have been transferred from one pixel to another in the real image during readout. This transfer is then subtracted from the original image.
One new type of reference file has been added to CALWF3: a sink-pixel file, which allows for the flagging of sink pixels in the science data quality (DQ) extension; no change is made to science data pixels. A type of bad pixel, sink pixels register systematically low, presumably due to a large number of traps within the pixel and can generate trails very similar to CTE trails (see WFC3 ISRs 2014-19 and 2014-22). The sinks and their trails are now flagged as part of the DQICORR correction step with DQ bit values of 1024 in both the CTE- and non-CTE-corrected data products.
The new calibration pipeline for WFC3/UVIS (CALWF3 version 3.3) achieves the following goals:
- Zeropoint accuracy is improved by ~3%.
- CTE correction will improve photometry of faint sources.
- MAST will archive CTE corrected data products, *_flc.fits and *_drc.fits, in addition to the standard *_flt.fits and *_drz.fits files.
The new CALWF3 version 3.3 was ingested into the MAST archive on Feb 23, 2016. The new software is described in detail in the CALWF3 version 3.3 Reference Guide, and 2-chip photometry is described at this webpage. Currently, users who wish to process their own data with version 3.3 will need to download the development version of UREKA (SSBX) and the new reference files from MAST. More details on reprocessing may be found in the CALWF3 version 3.3 Cookbook. The new CALWF3 software will move to the official UREKA (SSB) release in 1-2 months.
7. LINEAR: A Grism Simulation Code for WFC3/IR
R. Ryan, S. Casertano
We present LINEAR, a suite of routines in IDL and C to simulate noiseless grism images for a given scene and instrumental configuration. The software can predict which spectra will be affected by other sources and the average wavelength for an object in the dispersed imaging. These simulation modules may be useful for observers who are preparing cycle 24 proposals using WFC3/IR grism modes, and the current package is available through the WFC3 grism webpage. In the near future, we plan to enhance the capabilities of LINEAR by adding modules for extracting 1-d spectra and working with other instruments (such as WFC3/UVIS, ACS, JWST, and WFIRST). The key distinctions between LINEAR and existing tools (such as aXe) are:
1. Direct image inputs. The primary inputs to LINEAR are a direct image and segmentation map of the field, so there is no need to specify any input catalog. These data products can be constructed from the pre-imaging usually obtained as part of the grism observations; however, existing data beyond the field limits are highly desirable. These direct image products can be at any orientation or position, such as using archival mosaics as the direct imaging (e.g. CANDELS or Frontier Fields).
2. Contamination. Since LINEAR uses all the data simultaneously, there is no notion of contamination.
3. Full dataset reconstruction. LINEAR attempts to solve for the optimal non-parametric spectrum that can reproduce the entire canon of dispersed imaging (regardless of the orientation or position of the frames). The extraction modules are still in the testing phase, but we expect to release these tools very soon.
For more information, see the User's Manual for LINEAR from the WFC3 grism webpage and forth-coming reports and publications by R. Ryan and S. Casertano.
8. New Documentation
ISR 2015-13: WFC3 UVIS Read Noise -- H. Khandrika, S. Baggett
ISR 2015-14: WFC3 IR Gain from 2010 to 2015 – C. M. Gosmeyer, S. Baggett
ISR 2015-15: Persistence in the WFC3 IR Detector: An Improved Model Incorporating the Effects of Exposure Time – K. S. Long, S. M. Baggett, J. W. MacKenty
ISR 2015-16: Persistence in the WFC3 IR Detector: Spatial Variations – K. S. Long, S. M. Baggett, J. W. MacKenty
ISR 2015-17: Source-Dependent Master Sky Images for the WFC3/IR Grisms – G. Brammer, R. Ryan, N. Pirzka
ISR 2015-18: Spatial Accuracy of the UVIS Flat Fields – J. Mack, A. Rajan, A. Bowers
ISR 2016-01: The Updated Calibration Pipeline for WFC3/UVIS: A Reference Guide to Calwf3 (version 3.3) – R. E. Ryan, S. Deustua, J. Anderson, et al.
ISR 2016-02: The Updated Calibration Pipeline for WFC3/UVIS: A Cookbook to Calwf3 3.3 – V. Bajaj
Version 4 of the Data Handbook is now available from the WFC3 webpage.
The complete WFC3 ISR archive is here.
Need help? Start here.