STScI Logo

Hubble Space Telescope
WFC3 STAN - Issue 1, September 2009

**********************************************************
WFC3 Space Telescope Analysis Newsletter September 2009 **********************************************************
Contents

1. Welcome

2. Calibration of Geometric Distortion

3. CALWF3, Multidrizzle and the STScI Archive Pipeline

4. Reference Files

5. Photometric Zeropoints

6. UVIS User Defined Subarrays

**********************************************************
1. Welcome

Welcome to the first issue of the Wide Field Camera 3 Space Telescope Analysis Newsletter. These periodic announcements will highlight calibration and data analysis issues of interest to WFC3 observers. At this time, WFC3 is operating at and beyond our prelaunch expectations, the first Early Release Observations and some non-proprietary GO science data have been released, the SMOV observations have all been obtained,and the GO program is fully underway. Please visit http://www.stsci.edu/hst/wfc3 and look especially at the "Late Breaking News" section. We have posted (and will update as new results become available) detailed information on the results of the SMOV observations, a summary of instrument performance, and preliminary versions of useful reference files and other calibrations.

John MacKenty

**********************************************************
2. Calibration of Geometric Distortion

The near-term goal for the geometric distortion correction is to enable the observer to obtain images with rms errors at least as small as 0.05 pixel for exposures obtained within one visit. We are close to meeting this goal for the UVIS filter (F606W) and the IR filter (F160W) used in the initial solutions. We expect distortion-corrected products to be delivered by the pipeline starting some time in mid to late October. Up until that time, all drz files produced by the pipeline will continue to be empty products that are used only as placeholders in the HST archive.

Calibrated images can be corrected for geometric distortion using the Multidrizzle code. See Section 3 of this STAN and the Multidrizzle Handbook at http://www.stsci.edu/hst/HST_overview/documents/multidrizzle. Polynomial fits to the distortion are stored in the IDCTAB calibration file. 4th order fits have been obtained for the F606W filter (UVIS) and the F160W filter (IR) by comparing the positions of stars in SMOV calibration data to their positions in an astrometric catalog (ISR ACS 2007-08). The SMOV data consists of exposures of fields in 47 Tuc and the LMC taken in programs 11444 (UVIS) and 11445 (IR). On September 9, preliminary IDCTAB files with these 4th order fits were posted in the "Late Breaking News" section of the WFC3 webpage (http://www.stsci.edu/hst/wfc3). These solutions produce residuals of 0.08 pixels (0.003 arcsec) in F606W drizzled images relative to the astrometric catalog, and residuals of 0.08 pixels (0.010 arcsec) in F160W drizzled images relative to the astrometric catalog. The solutions for F606W have been entered for all filters in the UVIS IDCTAB, and those for F160W have been entered for all filters in the IR IDCTAB. Improvements will be made to these preliminary files by making filter-specific measurements, and by delivering DGEOFILEs along with the IDCTABs. The DGEOFILEs provide increments to the polynomial fits in the form of images of residuals in x and y. Errors in the World Coordinate System, possibly several arcseconds, will be also be quantified and reduced.

Filter-specific distortion solutions will be derived and delivered as exposures in more filters are acquired. This work will begin this fall with analysis of data acquired in other SMOV programs. Exposures of 47 Tuc and Omega Cen were made in SMOV program 11452 with filters F225W, F275W, F336W, F438W, and F814W in addition to F606W. Exposures of 47 Tuc were made in SMOV program 11453 with filters F110W, F125W, and F140W in addition to F160W.

More filters will be analyzed in the cycle 17 calibration programs 11911 and 11928. These programs have multiple visits to check the stability of the geometric distortion solutions. Two visits will be executed in 11911 (UVIS), with current plan windows of December 2009 and March-April 2010. Omega Cen will be observed with the 6 filters already included in 11452 (F225W, F275W, F336W, F438W, F606W, F814W) and in 4 more filters (F390W, F555W, F775W, F850LP). Three visits will be executed in 11911 (IR), with current plan windows of December 2009, March-April 2010, and August-September 2010. Omega Cen will be observed with 3 of the filters already included in 11453 (F110W, F125W, F160W) and in 2 more filters (F098M and F139M).

Linda Dressel, Vera Platais, Brian McLean, Colin Cox and Larry Petro

**********************************************************
3. CALWF3, Multidrizzle and the STScI Archive Pipeline

The STScI OTFR Pipeline currently employs the latest updates to CALWF3 (v1.6), but uses pre-flight calibration reference files. The schedule for producing and releasing in-flight calibration products is discussed in the next item.

The current public version of CALWF3 (1.4.1) was released in May 2009 as part of STSDAS v3.10. During the period of calibration and testing since SM4, a few relatively minor issues were addressed in updated versions 1.5 and 1.6. See the WFC3 Pipeline Release Notes website http://www.stsci.edu/hst/wfc3/pipeline/CALWF3ReleaseNotes.html) for information about each version. Future releases of CALWF3 will be advertised via WFC3 STANs and the WFC3 Pipeline Release Notes web pages. The latest version of CALWF3 will be included in the next release of STSDAS, currently scheduled for early November.

The most significant issue currently affecting calibrated images produced by the pipeline concerns the gain correction that is applied to IR images in the CALWF3 FLATCORR calibration step. The current calibration reference files result in an overcorrection of the quad-to-quad gain variations, which results in multiplicative offsets in the signal level of each detector quadrant - relative to one another - of 1-3%. While this is a relatively small error compared to the current accuracy of the on-orbit photometric calibrations, it can lead to noticeable systematic offsets in the background/sky signal level from one image quadrant to another in certain types of observations. Corrected images can be obtained by manually applying the following correction factors to each quadrant of calibrated flt images:

Quadrant 1 (upper left) = 1.004
Quadrant 2 (lower left) = 0.992
Quadrant 3 (lower right) = 1.017
Quadrant 4 (upper right) = 0.987

This issue only affects IR images taken through direct filters. It does not affect IR grism images or any UVIS images.

The WFC3 team has also identified two issues in the current public version of MultiDrizzle (v3.3.1) that affect the drizzling of WFC3 images. The World Coordinate System (WCS) assigned to drizzled UVIS images places the reference point in the wrong CCD chip, and it does not account for the fact that WFC3 IR flt images are in units of electrons/sec, as opposed to electrons. An alpha release of an updated version of MultiDrizzle will be made public by Sept. 18.

Howard Bushouse

**********************************************************
4. Reference Files

The WFC3 data processing pipeline is currently employing the best available reference files for calibrating science data. These reference files, installed in the Calibration Database System (CDBS) and available for retrieval via the HST archive should observers wish to inspect them, have been produced from ground-based thermal-vacuum test data. Over the next several months, we will be generating new reference files using on-orbit calibration data and installing these into CDBS for use in the data processing pipeline. The availability of the new CDBS files will be announced in future STAN(s). At that time, observers may wish to re-retrieve their science data from the archive in order to have the improved calibrations applied automatically by the pipeline or alternatively, manually recalibrate their data off-line using the new files. The WFC3 Data Handbook at http://www.stsci.edu/hst/HST_overview/documents/datahandbook/ provides details on how to manually recalibrate WFC3 images.

Tabulated lists of all WFC3 reference files installed in CDBS are available via the "Data Analysis - Reference Files" quick link on the left of the main WFC3 page (http://www.stsci.edu/hst/wfc3) or directly at http://www.stsci.edu/hst/observatory/cdbs/SIfileInfo/WFC3/reftablequeryindex These pages are generated via realtime queries of CDBS and thus present an up-to-date snapshot of the database contents at the time of the request.

Links to preliminary versions of reference files not yet in CDBS are available on the "Late Breaking News" section near the center of the main WFC3 page (http://www.stsci.edu/hst/wfc3) or directly at these links:

Pixel area maps for both UVIS and IR
http://www.stsci.edu/hst/wfc3/pam/pixel_area_maps

Photometric zeropoint tables
http://www.stsci.edu/hst/wfc3/phot_zp_lbn

UVIS on-orbit superbias file
http://www.stsci.edu/hst/wfc3/new_uvis_bias_lbn

Sylvia Baggett

**********************************************************
5. Photometric Zeropoints

The photometric sensitivity and stability of WFC3 has been measured through repeated observations of bright spectrophotometric standard stars during SMOV. The measured counts of these stars, as a function of aperture size, are compared to the Cycle 17 Exposure Time Calculator (ETC) estimates. In all cases, depending on the filter, the on-orbit sensitivities are higher by 5-20% compared to ground tests (for both the UVIS and IR channels). The ETC will be updated with the new throughput values for Cycle 18.

During SMOV we obtained UVIS observations of the hot white dwarf standard star GD153 in 37 filters. The observations were taken with a 512x512 pixel subarray on Amp A of UVIS1 with four dither positions. The exposure times were set to ensure a S/N > 100 measurement in each of the individual exposures. Repeat observations in the same format were taken at time intervals of 1 day, 1 week, and 1 month in each of the broadband filters F225W, F275W, F336W, F390W, F438W, F555W, F606W, and F814W. The standard deviation in these repeat observations is measured to be <1%. Relative to current ETC predictions, our on-orbit observations suggest a higher throughput of the instrument by 15-20% at central wavelengths and 5-10% at the blue and red ends of the UVIS spectral range. For the ETC update, we first fit these ratios of measured counts to predicted counts in the wide and medium band filters with a smooth 2nd order polynomial, and account for a 3% offset in the pixel area map for the location of our subarray relative to the center of the UVIS field. We note that F336W observations were not included as the ground flats appear to be affected by (this investigation is ongoing). We then calculated synthetic photometry by taking the integral of the new throughput with the bandpass of each of the wide and medium band filters, and defined a new correction that is multiplied by the old curve. This process is iterated several times until the standard deviation between the predicted and observed counts is stable (<1% in the final comparison). The final correction as a function of wavelength, is f = 0.838 + 1.279*(w) - 1.191*(w)**2, where w is the wavelength in micrometers. We note that the F845M filter represents the reddest data point in our comparison and therefore the curve is extrapolated to longer wavelengths. New photometric zeropoints in each of the 37 observed filters have been calculated in the ABMAG, STMAG, and VEGAMAG systems and are available in the UVIS table on the "Late Breaking News" section of the WFC3 webpage.

For the IR channel we obtained observations of both GD153 and the solar analogue star P330E in all 15 filters. The 128x128 pixel subarray was used and the star was placed at two dither positions. The exposure times and sampling sequences were selected to ensure a high S/N detection in all filters. Repeat observations were obtained in all filters at time intervals of 1 day, 1 week, and 1 month,and demonstrate that the throughputs of the wide and medium band filters are stable to <1% and the narrowband filters to 1-2%. Relative to the current ETC, the on-orbit measurements indicate that the instrument throughput is 10-15% higher in all filters. We used a similar procedure on the IR data as was applied to the UVIS data with the measured counts over current synphot predictions defining a correction factor that is folded into the ETC. The observations of the two separate stars were averaged for wide and medium band filters, across all dither positions and visits, to define the fitting function. The polynomial used is a simple 3rd order function, iterated several times with the output synthetic photometry from synphot as described above. For wavelengths shorter than 9000 Angstroms, the correction applied was defined from the IR grism sensitivity. The new photometric zeropoints in each of the filters have been calculated in the ABMAG, STMAG, and VEGAMAG systems and are available in the IR table on the "Late Breaking News" section of the WFC3 webpage (http://www.stsci.edu/hst/wfc3).

Jason Kalirai, Abhijith Rajan, and Adam Riess

**********************************************************
6. UVIS User Defined Subarrays

Subarrays are a rectangular portion of a full WFC3 detector that are read out to the science data buffer. They are useful for reducing data volume and (non-exposing) overhead time. Predefined subarrays are provided to users as SUB apertures that specify the portion of the detector to be read out and the reference position of the target within that subarray. These SUB-aperture subarrays are fully supported in Cycle 17 and their use is preferred if data volume or overhead time present limitations in achieving the science goals of a program. However, these predefined subarrays may not meet the needs of all users. User-defined, custom subarrays may be specified in those cases. There are some noteworthy characteristics of user-defined subarrays. First, they may be defined only for UVIS exposures - not for IR exposures. Second, the user must assure that the target is placed in the subarray - the placement is not automatic in all cases. Third, pixels from the non-light sensitive overscan portion of the CCD may be included in the subarray if appropriately specified.

Finally, the subarray must meet several restrictions on the location and size of the subarray. Specific direction can be found in Chapter 6 of the Instrument Handbook, http://www.stsci.edu/hst/wfc3/documents/handbooks/currentIHB/wfc3_cover.html in the Cycle 17 Phase 2 Proposal Instructions (Engineering Version) that is available through help@stsci.edu, and in the Instrument Science Report "Operational Definitions and Implementation of WFC3 UVIS Subarrays," O. Lupie et al. (ISR-WFC3-2002-14).

The calibration of UVIS subarray data requires superbiases that have been constructed from bias frames read out through the same amp as the subarray. A subarray contained within a single quadrant is read out through the closest amp. In this case, the best superbias is the standard 4-amp full-frame superbias, from which the pipeline extracts the matching subarray section and applies it to the science data. Subarrays which can use the standard 4-amp full-frame superbias can be located on either chip; these include the predefined QUAD filter subarray apertures as well as any user-defined subarrays positioned entirely within a single quadrant.

However, for subarrays of any size that span two quadrants (e.g., either a custom subarray or the predefined M512 or 2Kx4K subarrays), the appropriate superbias must be constructed from a stack of bias frames read out through the same amp as the subarray. To minimize the observing time required to accumulate sufficient bias frame exposures to support calibration of quadrant-spanning subarrays, thereby maximizing HST time available for science observations, superbiases will be provided only for custom subarrays that 1) are located on chip 1 and 2) are read out through amp B. For any questions or concerns about subarray calibrations, please email your contact scientist or help@stsci.edu.

Larry Petro and Sylvia Baggett