STScI Logo

Hubble Space Telescope
WFC3 STAN Issue 4, June 2010

WFC3 Space Telescope Analysis Newsletter - Issue 4, June 2010

For new information about WFC3 visit the "New in the Last 45 Days" and "Late Breaking News" sections of the WFC3 website at

This and previous issues of the STAN can be found at /hst/wfc3/documents/newsletters.

I. Phase II Update for Cycle 18
1. Changes to APT for WFC3 General Observers
2. Changes in Apertures for Cycle 18
3. Dither Patterns for WFC3
4. Revision of the WFC3 Instrument Handbook
5. Using IR subarray apertures with grisms
6. Update on IR Subarray Image Artifact

II. Calibration and Data Processing
7. WFC3 Data Processing Updates
8. Reference Files
9. IR low frequency flat correction
10. Update to EXPTIME keyword for short UVIS Exposures
11. New Documentation

I. Phase II Update for Cycle 18

1. Changes to APT for WFC3 General Observers - Larry Petro

Since the Cycle 17 Phase 2 Proposal submission deadline, many changes have been made to APT that better support General Observers using WFC3. The changes are the following.

1. Apertures
Several apertures have been added and others removed, as described in the companion STAN item, "Changes in Apertures for Cycle 18."

The apertures removed are:
UVIS1-2K4-SUB, UVIS1-M512-SUB, UVIS1-C512A-SUB, and UVIS1-C512B-SUB.

Apertures added are:

Fiducial points of some apertures have been adjusted, which will be correctly displayed by Aladin. The adjustments are the result of the SMOV and Cycle 17 alignment programs and decisions to better position some fiducial points to optimally avoid image artifacts.

2. Aladin display
The large circular region of dead pixels in Q2 of the IR detector (also known as "Death Star") is displayed.

3. Convenience Patterns
Two three-point, line convenience patterns were added for UVIS and IR exposures (WFC3-UVIS-DITHER-LINE-3PT and WFC3-IR-DITHER-LINE-3PT). These are useful for orbit-packing for certain kinds of visits.

The default value of CR-SPLIT for UVIS exposures is changed from 2 to NO.

5. IR subarrays and grisms
Exposures using IR subarray apertures (IRSUB*) and grisms (G102 and G141) are now supported. Please see the companion STAN item, "Using IR subarray apertures with grisms" for further information.

6. IR subarrays and sample sequences
Some combinations of IR subarray apertures and samples are no longer supported. Generally, for a given sample sequence, support for the small subarrays has been removed. In all such cases, larger subarrays (or full-detector apertures) may be used with that sample sequence and will allow efficient, parallel dumping from the Science Data Buffer to the Solid State Recorder. The table of supported combinations, Table 7.9, is presented in the WFC3 Instrument Handbook, section 7.7.4, "MULTIACCUM Timing Sequences: Subrray Apertures."

2. Changes in Apertures for Cycle 18 - Larry Petro

The apertures for use with the UVIS and IR channels of WFC3 have been significantly revised for Cycle 18. While most of the apertures supported in Cycle 17 remain, new apertures have been added, fiducial points of apertures have been adjusted, and subarray apertures have been added. The changes are reviewed here, but observers are referred to the WFC3 Instrument Handbook, the Phase 2 Proposal Instructions, and the Observatory Support website (/hst/observatory) and especially to the links thereon to Apertures (/hst/observatory/apertures) and Science Instrument Aperture File (SIAF; /hst/observatory/apertures/siaf.html) for further information and specifics.

For the UVIS channel, subarray and IR-matched apertures have been added. The new subarray apertures provide 512x512, 1Kx1K, and 2Kx2K subarrays. Two each of the 512x512 and 1KxK subarrays are provided, all in the Amplifier C quadrant of Chip 2, which provides good photometric cosmetics and low sensitivity to orbital telescope breathing. The "C" aperture of each pair (UVIS2-C512C-SUB, UVIS2-C1K1C-SUB) is placed in the corner of the chip, which places it near the readout amplifier, thereby reducing CTE effects, and adjacent to the physical overscan pixels, which are included in the readout. The "M" aperture of each pair (UVIS2-M512C-SUB, UVIS2-M1K1C-SUB) is placed near the middle of the UVIS detector (contained on the Amplifier C quadrant), which provides the lowest sensitivity to orbital breathing, but precludes inclusion of physical overscan pixels in the readout. The 1Kx1K subarray apertures provide replacements for the ACS/HRC field of view. Four quadrant subarray apertures (UVIS1-2K2A-SUB, UVIS1-2K2B-SUB, UVIS2-2K2C-SUB, UVIS2-2K2D-SUB) for use with the full-FOV filters (the non-quadrant filters) are now available. For all the preceding 512x512, 1Kx1K, and 2Kx2K filters the fiducial pixel is near the center of the light sensitive subarray. A full-detector aperture whose fiducial point matches the IR-FIX aperture (UVIS-IR-FIX) was added for Cycle 18. Exposures in the IR and UVIS channels in visits using the two matched apertures will require no telescope motion between exposures. This will improve observing efficiency and will maintain the FOV of parallel Scientific Instruments.

The fiducial points of the quadrant filter apertures (UVIS-QUAD, UVIS-QUAD-SUB) have been moved to the center of the useful image area, in order that extended targets will better avoid the cross-shaped overlapping passband region of the four quadrant filters in each filter slot. Note that, in contrast, the fiducial pixel of the full-FOV filter 2Kx2K subarray apertures are at the center of the quadrant.

Apertures that have been removed are the one-chip subarray aperture UVIS1-2K4-SUB and the 512x512 subarrays UVIS1-M512-SUB, UVIS1-C512A-SUB, and UVIS1-C512B-SUB. The 1-chip subarray was little-used, required dedicated calibrations, and greater observing overhead for single-amplifier readout. The three 512x512 subarray apertures have been replaced with equivalent apertures for Cycle 18.

For the IR channel, three full-detector apertures with fiducial points that match UVIS aperture fiducial points have been added. They are IR-UVIS, IR-UVIS-CENTER, and IR-UVIS-FIX. As for the similar UVIS aperture, these apertures support mixed-channel visits by easing planning of target positions, reduce observing overheads, and prevent telescope motion when switching channels.

3. Dither Patterns for WFC3 - Linda Dressel

Two new convenience patterns for WFC3 have been added to APT for cycle 18, one for each detector. They execute dithers which provide optimal 3-point sampling of the point spread function, as described in Section 2.5 of the MultiDrizzle Handbook. /hst/HST_overview/documents/multidrizzle These new 3-point patterns supplement the 2-point dither patterns and 4-point box patterns introduced for WFC3 in cycle 17. (See Appendix C of the WFC3 Instrument Handbook, version 2.1, /hst/wfc3/documents/handbooks/currentIHB/wfc3_cover.html for a full description of the cycle 17 patterns.) WFC3-UVIS-DITHER-LINE-3PT performs steps equivalent to [POSTARG X, POSTARG Y] values of [0,0], [0.092,0.099], [0.185,0.197] arcsec to step across the detector with steps of [0,0], [2.33,2.33], [4.67,4.67] pixels. WFC3-IR-DITHER-LINE-3PT performs steps equivalent to [POSTARG X, POSTARG Y] values of [0,0], [0.452,0.404], [0.903,0.807] arcsec to step across the detector with steps of [0,0], [3.33,3.33], [6.67,6.67] pixels.

A detailed discussion of dithering strategies and suggested patterns that the user can implement in APT are given in ISR WFC3 2010-09. This document includes descriptions of image artifacts that should be dithered over in certain cases, and decision trees to guide the user in the selection of a pattern.

4. Revision of the WFC3 Instrument Handbook - Linda Dressel

The previous edition of the WFC3 Instrument Handbook (Version 2.0) was produced in January 2010. The slightly revised Version 2.1 was released on June 24. /hst/wfc3/documents/handbooks/currentIHB/wfc3_cover.html

The principal revision concerns the full well depth of the pixels over the area of the CCD detector. On-orbit observations have shown that the onset of saturation varies from 67000 to 72000 electrons over the UVIS2 chip, and from 63000 to 71000 electrons over the UVIS1 chip. These levels are substantially lower than the levels reported previously, based on ground-based measurements: 75000 to 80000 electrons over the detector. Updates have been made accordingly to Table 5.1 and Sections 5.4.5 and 6.9.1, with advice on using ETC calculations to avoid saturation in an exposure.

Revisions have also been made to Sections 7.7.3, and 7.7.4, now called MULTIACCUM Timing Sequences: Full Array Apertures and MULTIACCUM Timing Sequences: Subrray Apertures. The intention of the revisions is to make it more apparent that the same IR sample sequence names correspond to different sets of times for different size apertures, and to make it easier to find the times in the sample sequences defined for the IR subarrays.

5. Using IR subarray apertures with grisms - Larry Petro

For Cycle 18, observers may use the IR grisms, G102 and G141, with the subarray apertures. It is expected that the measurement of time series of bright, isolated targets will be enabled by this new capability. Observers will need to position the first order spectra within a subarray using the POSition TARGet Special Requirement. The use of this new observing capability is described in this article.

The first order G102 spectrum is found in columns 59 - 209 to the right of the direct image. For the G141 grism, the corresponding columns are 34 - 167. The 151-pixel long G102 spectrum and the 134-pixel long G141 spectrum may be readily positioned within the IRSUB512 and the IRSUB256 apertures. By pointing HST to place the direct image of a target on the detector at pixel (505, 532) for the IRSUB512 aperture and (410, 532) for the IRSUB256 aperture, the first order spectra will be well positioned 20 pixels above the detector amplifier boundaries. Those positions achieve several other desirable goals, including avoiding image and detector artifacts (including CSM blobs and bad pixels) and inclusion of the direct image in filter exposures with the same subarray. The 512x512 image will also include the 0-order image.

To place the spectra at these locations, an observer should use POS TARG -2.303, +1.202 for the IRSUB512 aperture and POS TARG -15.172, 1.158 for the IRSUB256 aperture. In order to calibrate the wavelength offset, a matching direct, filter image should be taken with the same POS TARG specification. Wavelength offsets from direct images have been calibrated for F098M and F105W (G102) and for F140W and F160W (G141). For bright targets, narrowband filters may be preferable. Secondary calibrations of the F132N (for G141), and F128N and F130N (for G102) are available.

The same limitations on combinations of subarray apertures and sample sequences are in force as for direct, filter images. Those limits are documented in the WFC3 Instrument Handbook, Section 7.7.4. Efficient operation may be achieved with parallel dumping of data to the Solid State Recorder, which requires using adequately long sample times (i.e., an appropriate sample sequence).

6. Update on IR Subarray Image Artifact - Howard Bushouse

As reported in the Dec. 2009 WFC3 STAN, certain combinations of IR subarrays and sample sequences give rise to images containing an artifact. The artifact appears as a sudden jump in the overall background level of the image, with an amplitude of 3-5 DN. The jump occurs exactly at the vertical center of each quadrant of the image, such that the lower and upper quarters of the image have a higher overall level than the middle of the image.

We have now examined IR darks taken in all of the supported combinations of subarray size and readout sample sequence and find that the anomaly appears in the cases listed in the table below. "No" indicates that the anomaly does not appear, "Yes" indicates that it does, and "N/A" indicates that the mode is not supported.


The cause of this artifact is still under investigation. Our analysis to date has revealed the following. First, for a given mode in which we have seen the anomaly, it is always present in dark exposures for that mode, but not always obviously present in external images for that mode. Hence the anomaly can sometimes be imprinted on an otherwise good external image during the process of dark subtraction in calwf3. Therefore if users see the anomaly in their calibrated images, they may want to try reprocessing the data through calwf3 with DARKCORR set to OMIT and see if it disappears. Second, for modes in which we see the anomaly in both darks and external images, the amplitude of the effect in the dark reference file subtracted by calwf3 does not always exactly match that in the external science image, and therefore some residual of the effect may remain in the calibrated images.

II. Calibration and Data Processing

7. WFC3 Data Processing Updates - Howard Bushouse

Several updates to WFC3 calibration reference files have been made over the past few months that have a very direct effect on the results of IR image processing. First, a new IR bad pixel table (BPIXTAB) was released in April that not only updates the populations of pixels flagged as dead, hot, and unstable, but also includes a new population of IR pixels that are affected by the "blobs" (small-scale regions of reduced sensitivity). The IR blobs are flagged with DQ=512. The default processing performed in the STScI pipeline flags pixels affected by the blobs, but does not otherwise reject or handle them in any special way. Users who have sources that may be affected by the blobs and want to reject the affected pixels should utilize the DQ=512 values in the DQ arrays of calibrated image files. Second, an updated IR MutliDrizzle parameters table (MDRIZTAB) was also released in April, which changes the way CR hits detected by MultiDrizzle are handled for IR images. CR's detected during MultiDrizzle processing are rejected from the output drizzled image (_drz.fits files), but are no longer flagged in the DQ arrays of the individual _flt.fits files.

Finally, and perhaps most importantly, an updated IR Cosmic-Ray Rejection parameters table (CRREJTAB) was released and put into use in the STScI pipeline on June 11. This table is used to control the behavior of the calwf3 IR "up-the-ramp" fitting and CR rejection process (CRCORR step). The updated table changes the value of the "badinpdq" parameter so that pixels that have all samples (readouts) flagged as dead (dq=4), zero-read deviant (dq=8), hot (dq=16), or unstable (dq=32) are no longer zeroed out in the SCI array of the calwf3 output _flt.fits file. They are, however, still flagged in the DQ array of the _flt.fits file. So users should pay particular attention to the _flt.fits file DQ array when performing subsequent processing or analysis in order to decide whether or not certain pixels should be used.

There have also been several WFC3 data processing software updates. Updates to the STScI WFC3 pipeline that were made on April 19, 2010 included calculation of the velocity aberration factor for all WFC3 images, which is stored in the image header keyword "VAFACTOR". This value is used within MultiDrizzle processing of WFC3 images to compensate for changes in plate scale due to HST's orbital motion. WFC3 images processed before April 19 had VAFACTOR set to a default value of 1.0. Users who require very accurate astrometry or drizzle combination of their images should consider retrieving older WFC3 images from the data archive, which will cause them to be reprocessed and compute the VAFACTOR.

Also of interest to MultiDrizzle users is a bug fix that has been made to the MultiDrizzle code that fixes a problem encountered when using MultiDrizzle off-line to process WFC3 IR images with the IVM (Inverse Variance Map) weighting scheme.

A few minor updates have been made to the calwf3 calibration software in the past few months, which are included in calwf3 v2.1. These include having the IR non-linearity correction and saturation checking (NLINCORR step) peformed before dark subtraction (DARKCORR) and a fix to the UVIS "wf3rej" processing so that DQ flags from detected CR's show up in the individual images _flt.fits files. For details on all the changes included in calwf3 v2.1 (and earlier versions), see the calwf3 release notes on the web at: /hst/wfc3/pipeline/CALWF3ReleaseNotes.html/

All of the above updates to MultiDrizzle and calwf3 have been included in the most recent release of the STSDAS package (v3.12), which was made available to the community on June 21.

8. Reference Files - Bryan Hilbert

New WFC3 Reference Files since the last STAN (March 2010):

IR and UVIS Bad Pixel Tables
On April 12, a new IR bad pixel table was delivered to the calibration database system. This bad pixel table was created using SMOV and Cycle 17 observations, and provides a better, more recent map of the various types of bad pixels on the IR detector compared to the previous, ground testing-derived file. This table includes flags for dead, bad in 0th read, and unstable pixels, and now also includes flags to mark pixels affected by the IR "blobs" (dq=512). An ISR detailing the method used to create this bad pixel table is forthcoming. See the accompanying article in this STAN "WFC3 Data Processing Updates" for more information on the effects to data processing from these changes, as well as changes described below for the IR CRREJTAB and MDRIZTAB.

On May 13, a new UVIS bad pixel table was also delivered. This bad pixel table contains a small update to the table previously in the calibration database system. In the old version of the bad pixel table, the flags for the three bad rows on each side of the gap between the two chips did not extend all the way across the detector. This update fixes that problem, extending the bad pixel flags across the entire length of the two chips.

IR and UVIS Dark Current Files
One new dark current reference file was delivered for each of the 15 sample sequences in the IR channel on April 9. Each reference file was created from between 11 and 63 individual SMOV and Cycle 17 dark current observations. These are the first IR dark reference files derived from on-orbit data and provide a much better dark current subtraction compared to the previous, ground-derived dark current files. Each file also provides an updated map of hot pixels, which is present in the data quality array.

On June 11, the WFC3 team also delivered the initial set of UVIS on orbit-derived dark current reference files. This set of 17 files provides measure of the UVIS channel's dark current between mid-August and mid-October 2009, with each file providing a snapshot of the dark current behavior over a different 4-5 day period.

IR Cosmic Ray Rejection Table
A new IR channel cosmic ray rejection table (CRREJTAB) was delivered on June 11. This file changes the default way in which calwf3 handles bad pixels in science data. Previously, pixels that were flagged as bad in any way in all reads of an observation were given a value of zero in the calwf3-output flt image. With this new table, calwf3 will instead perform line-fitting and produce non-zero flux values in the flt output file for pixels flagged as bad. In this way, observers can examine the data quality arrays to identify flagged pixels, decide which flags are relevant for their analyses, and choose to use or ignore each type of bad pixel. The only exception to this behavior is for pixels which are flagged as saturated in all reads of an observation. These pixels will continue to be given a value of zero in the final flt file, as there are no reliable flux measurements with which to perform line-fitting.

IR Multidrizzle Table
A new IR multidrizzle parameter table (MDRIZTAB) was delivered on April 22. The new parameter settings in this table are based on experience from the application of MultiDrizzle to IR images from the first 10 months of on-orbit science.

Changes in this new table include:
Please see the Multidrizzle Handbook (/hst/HST_overview/documents/multidrizzle) for details on the keywords listed above.

IR Grism Gain Conversion Flats
Two new pixel-to-pixel flat field reference files were delivered on April 22. These reference files are for use with IR grism data only. These flats contain constant values in each quadrant. Applying these flats to grism images result in a quadrant-based gain correction only. Flat fielding corrections on the grism data are applied during processing with the aXe software.

9. IR low frequency flat correction - Elena Sabbi

An alpha release of updated flat-field calibration files (LP-flat) for IR images was made available to the community on May 26, 2010. The files combine ground-based pixel-to-pixel flat field (P-flat) and the low-frequency correction (L-flat) as determined by comparing the flux of high signal to noise stars in Omega Centauri over different position of the detector. These files, with extension _lpflt.fits can be downloaded from the WFC3 webpage at the address /hst/wfc3/analysis/ir_flats.

The L-flat solutions were derived from images obtained in the filters F110W, F125W, and F160W. For the remaining 12 IR imaging filters the L-flats were obtained from a linear interpolation of these filters, based on the filter pivot wavelength.

Tests carried out during the Servicing Mission Observatory Verification (SMOV) indicate that the ground based P-flats do not fully remove the low-frequency structures, leaving spatial variations up to 1.5% (Hilbert et al., ISR 2009-39). By applying the L-flat correction, the residuals in the photometry are reduced to a maximum of 0.6%.

At the moment data downloaded from the HST archive are not calibrated with these LP-flats. Users who are interested in applying the new calibration files to their data should edit the header of the raw images to update the keyword PFLTFILE to the new LP-flat name, while the DFLTFILE and LFLTFILE keyword should remain unchanged (=N/A), and then run CALWF3.

10. Update to EXPTIME keyword for short UVIS Exposures - Bryan Hilbert

A recent analysis of short (< 1 sec) UVIS exposures shows that the difference between the commanded exposure time and the amount of time that the shutter is truly open is significant for commanded exposure times of 0.5 and 0.7 sec (see ISR 2009-25 for details). As a result, the WFC3 team is in the process of making a change to STScI OPUS data processing software so that the actual shutter open time is reflected in the value of the EXPTIME image header keyword for these very short UVIS exposures. This is similar to the shutter timing fix already in place for short ACS exposures. This change will be implemented in the near future. In the meantime, the values shown in the table below can be used to make your own corrections to EXPTIME values. For all other exposure times, the actual exposure time is indistinguishable from the commanded exposure time and therefore the EXPTIME header keyword will continue to contain the commanded exposure time.

Commanded Exposure Time (sec) Actual Exposure Time (sec) % Shorter than Commanded
0.5 0.480 4.0%
0.7 0.695 0.7%

In order to accurately convert UVIS images from units of electrons to electrons per second, observers should divide the measured signal by the actual exposure times listed above.

11. New Documentation

These new ISRs have been published since the last STAN (March 2010):

ISR 2010-09 Dithering Strategies for WFC3, Dahlen, Dressel, and Kalirai 24 Jun 2010
ISR 2010-08 WFC3 Pixel Area Maps, J. Kalirai et al. 27 Apr 2010
ISR 2008-33 WFC3 TV3 Testing: IR Persistence, P. McCullough, S. Deustua 22 Jun 2010

The complete WFC3 ISR archive is at:

An update of the Instrument Handbook can be found at: (see revision of the WFC3 Instrument Handbook in this STAN).

Need help?

To subscribe or unsubscribe to the STAN, send a message to with a blank subject line and the following in the body: [un]subscribe wfc3_news
The Space Telescope Science Institute is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555.