6.11 UVIS Observing Strategies

6.11.1 Dithering Strategies

For imaging programs, STScI generally recommends that observers employ dithering patterns. Dithering refers to the procedure of moving the telescope by pre-determined amounts between individual exposures on a target. The resulting images are subsequently combined via post-observation processing techniques using software packages such as DrizzlePac.

Use of dithering can provide improved sampling of the point spread function (PSF) and better correction of undesirable artifacts in the images (e.g., hot pixels, cosmic rays, the UVIS channel’s inter-chip gap, and the UVIS “droplets”). Cosmic ray removal is more effective if more than 2 images are obtained, especially for exposure times greater than 1000s. A sequence of offsets of a few pixels plus a fractional pixel in each coordinate is generally used to simultaneously remove hot pixels and cosmic rays and to sample the PSF. A larger offset along the image Y axis is needed to fill in the interchip gap in full-frame images (the WFC3-UVIS-MOS-DITH-LINE pattern uses a conservative step size of 2.4 arcsec). To ensure the best photometric accuracy,  consider dithering to compensate for droplets (Section 6.10.5).

Larger offsets, up to sizes approaching the detector’s field of view, can also be used to create mosaics. However, as a result of geometric distortion (Appendix B), some objects shift by an integer number of rows (or columns), while others shift by an integer plus some fraction of a pixel. The PSF is not resampled in that dimension in the former case, but is resampled in the latter case. Where the exposures overlap, the PSF is thus better sampled for some objects than for others. If PSF sampling is important, a combination of mosaic steps and small dither steps should therefore be used. Note that, in practice, mosaic steps must be contained within a diameter ~130 arcsec or less (depending on the availability of guide stars in the region) to use the same guide stars for all exposures. The r.m.s. pointing repeatability is significantly less accurate if different guide stars are used for some exposures (see Appendix B of the DrizzlePac Handbook).

The set of Pattern Parameters in the observing proposal provides a convenient means for specifying the desired dither pattern of offsets. The pre-defined mosaic and dither patterns that have been implemented in APT to meet many of the needs outlined above are described in detail in the Phase II Proposal Instructions. The WFC3 patterns in effect in APT at the time of publication of this Handbook are summarized in Appendix C. Observers can also define their own custom patterns to tailor them to the amount of allocated observing time and the desired science goals of the program. Alternatively, they can use POS TARGs to implement dither steps (Section 6.4.3). Observers should note that thermally-driven drift of the image on the detector, typically 0.1 to 0.2 pixels per coordinate within one orbit (WFC3 ISR 2012-14), will limit the accuracy of execution of dither patterns. Dither strategies for WFC3 are further discussed in WFC3 ISR 2010-09, which provides a decision tree for selecting patterns and combining them with subpatterns. WFC3 ISR 2020-07 provides compact patterns with up to 9 steps (in the form of POS TARGs) designed to preserve sub-pixel sampling as much as possible over the face of the UVIS CCDs, given the scale changes introduced by geometric distortion. In addition, WFC3 ISR 2023-05 presents new dithering patterns to optimize observations for proposals using WFC3 and ACS simultaneously (i.e. one instrument as prime and the other in parallel).

6.11.2 Parallel Observations

While the design of WFC3 prevents the simultaneous use of both the UVIS and IR channel, it is possible to use one or more of the other HST instruments in parallel with WFC3. Since each instrument covers a different location in the HST focal plane (see Figure 2.2), parallel observations typically sample an area of sky several arc minutes away from the WFC3 target. For extended targets such as nearby galaxies, parallel observations may be able to sample adjacent regions of the primary target. In other cases, the parallel observations may look at essentially random areas of sky.

For processing and scheduling purposes, HST parallel observations are divided into two groups: coordinated and pure.

coordinated parallel is an observation directly related to (i.e., coordinated with) a specific primary observation, such as in the extended galaxy example above. A pure parallel is an observation typically unrelated to the primary observation, for example, parallel imaging scheduled during long spectroscopic observations. The primary restriction on parallel observations, both coordinated and pure, is that they must not interfere with the primary observations: they may not cause the primary observations to be shortened; and they must not cause the stored-command capacity and data-volume limits to be exceeded. The proposal software (APT) enforces these rules and notifies the observer when a specified parallel is not permitted.

In order to prolong the life of the HST transmitters, the number of parallels acquired during each proposal cycle is limited. Proposers must provide clear and strong justification in their Phase I proposal in order to be granted parallel observing time. Please refer to the HST Call for Proposals for current policies and procedures concerning parallels.

6.11.3 Spatial Scans

Spatial scanning of stellar images on the UVIS detector creates the potential for astrometry of unprecedented precision as well as improved relative photometry. Two representative scientific examples of the former are parallax measurement of Cepheid variable stars (program 12679, Riess et al. 2021 and references therein) and a high-precision trigonometric parallax to NGC 6397 (programs 13817, 14336, 14773; Brown et al., ApJ L 856, L6, 2018).

Results from non-proprietary data of program 12679 indicate that differential astrometry a few times less precise than that set by diffraction and Poisson statistics are attainable (Riess, priv. comm.). For HST, a 2.4-m telescope, operating at 600 nm, the diffraction limit is Θ ~ λ/D = 51 mas. In the theoretical limit, astrometry in one dimension is approximately equal to the FHWM Θ divided by the signal to noise ratio, \small{\sqrt{N}}, where N is the number of photo-electrons recorded. If we adopt N equal to the full well of the UVIS CCD, ~64,000 e, times a trail of length 4000 pixels, i.e., N = 128 million e, then the theoretical astrometric limit of scanned data is ~3 microarcsec per exposure. A more conservative estimate of ~13 microarcsec can be derived as follows: the nominal, state-of-the-art astrometric precision of a staring-mode exposure is ~0.01 pixel, so the astrometric precision of a 1000-pixel-long scan could be ~\small{\sqrt{1000}} or ~30 times smaller, which, for the 40 mas WFC3 UVIS pixels, is 13 microarcsec. In 2012 the TAC recommended programs which anticipated a per-exposure precision of 30 to 40 microarcsec (13101 and 12909).

Some data analysis tools for spatial scans are currently being developed by the WFC3 team to aid users in reducing spatial scan data, but for the most part users should expect to develop their own analysis software to reduce images and obtain useful astrometric and photometric results (e.g. Riess et al. 2018; Riess et al. 2014).

Scans can be made at any angle, although users typically orient the scans approximately, but not exactly, parallel either to rows or to columns of the detector. For example, in order to sample pixel phase, program 12679 prescribed an angle of 90.05 degrees; the extra 0.05 degrees corresponds to a shift of ~1 pixel every 1000 pixels along the trail. In the interest of observing efficiency, this program performed forward and reverse scans alternately. Observers are cautioned that there will be a spatial offset between forward and reverse scans in the scan direction. Forward scans are centered in the frame as predicted in the APT display; reverse scans are offset by an amount that is greater for faster scan rates.

Boustrophedonic (from the Greek, literally, “as an ox turns in plowing”) scans, are possible too. In boustrophedonic scans, a.k.a. serpentine scans, the user specifies a set of constant-speed scan lines separated by a specified angular distance, like rows in a farmer’s field. An example is illustrated in Figure 6.27. The advantage is that more scan lines are possible per exposure, which can improve observing efficiency. The trajectory of such scans has been modeled (WFC3 ISR 2017-06).

Figure 6.27: A vertical boustrophedonic scan of a star field on the UVIS focal plane (from program 12679).

The three thickest black lines in the center are the target star scanned upward, downward, and upward again in the same exposure. The U-shaped trails are from the turning points of nearby fainter stars.
Spatial scanning generally permits more precise photometry than staring mode, by collecting more photons and averaging over more detector pixels. However, actual photometric precision may not achieve the theoretical limits due to at least two factors: 1) flat field errors (typically ~0.5%) and 2) shutter-timing non-repeatability (~0.004 s r.m.s. for the UVIS shutter). WFC3 ISR 2022-04 determined that photometric repeatability of spatial scan observations was ~2.5 times better than staring mode observations across multiple cycles (five years) of data; WFC3 ISR 2017-21 suggests that photometric repeatability may be ~3-5 times better than staring mode for observations within the same cycle.

Two examples of utilizing spatial scans for precision time-series photometry are program 14621 (P.I. Wang) and program 15129 (P.I. Burke). These programs take advantage of the orbit-to-orbit and visit-to-visit stability of the UVIS detector. Attempts to obtain precise photometric time-series within a single exposure, by using the trailed image of a star to record its flux versus time, have not been successful, because the positional feedback loop of the FGS control introduces lateral and longitudinal displacements from an idealized, constant-velocity scan, which results in photometric “flicker” of a few per cent (Figure 6.28). Although differential photometry of two or more stars would mitigate the FGS-induced “flicker”, the two flat-field and shutter factors would remain.

For those preparing a phase II program description, we recommend WFC3 ISR 2012-08 and WFC3 ISR 2017-21. Also, IR imaging with spatial scanning is discussed in Section 7.10.4, and slitless spectroscopy with spatial scanning is discussed in Section 8.6. See Figure 8.12 for a diagram provided in APT to assist observers planning spatial scan observations. Note: starting in 2016 (Cycle 24), the Exposure Time Calculator (ETC) supports spatial scanning for UVIS and IR imaging and IR spectroscopy (WFC3 STAN issue 22).

Figure 6.28Photometric “flicker” appears in scans under FGS control.

The same star was trailed across the WFC3 IR detector repeatedly, with rates of 1, 1, and 0.5 arcsec s-1 respectively. Photometry integrated transverse to each of the three trails, show intermittent oscillations with ~3% amplitude at ~1.3 Hz, due to image motion or “jitter” during the scans. The 2nd and 3rd trails have been offset vertically by 0.1 and 0.2 for clarity.

6.11.4 PSF Subtraction

UVIS imaging has been shown to be highly effective in detecting faint point sources near bright point sources (WFC3 ISR 2011-03). For a variety of narrow, medium, and wide filters, when a high signal-to-noise drizzled image of a star was scaled down by 10 magnitudes and shifted and added to the original image, the simulated faint companion could usually be seen for separations greater than 1.0 arcsec. Based on the annular signal-to-noise of the deep stellar image, 5 sigma detections of companions fainter by two magnitudes could be made at a separation of 0.1 arcsec. Theoretically, companions several magnitudes fainter could be detected at that separation in deeper images, but, in practice, variations in the PSF (point spread function) due to telescope breathing limit the detectability within about 0.3 arcsec of a bright star.

If observers want to use stellar images to subtract the PSF from a target comprised of a point source and an extended source to detect or measure the extended source, they should keep several points in mind:

  • UVIS pixels undersample the PSF (Section 6.6), so the stellar and target exposures must be dithered to produce good sampling of the PSF.
  • Position drift and reacquisition errors can broaden the PSF (WFC3 ISR 2009-32WFC3 ISR 2012-14).
  • If a single guide star is used for a visit, roll angle drift causes a rotation of the target around that star, which in turn introduces a small translational drift of the target on the detector. In recent years, as gyroscopes have failed and been replaced, the typical roll angle drift rate has increased from 1.5 mas/sec to ~17 mas/sec, producing a translation of up to 60 mas (1.5 UVIS pixel) in 1000 sec.
  • The characteristics of the PSF depend on the focus, which generally changes measurably during an orbit; its range in a particular orbit will not be known in advance (WFC3 ISR 2012-14, WFC3 ISR 2013-11).
  • The characteristics of the PSF vary with location on the detector (e.g., see WFC3 ISR 2013-11, ACS ISR 2003-06). PSFs near the A amplifier on UVIS1 are noticeably elongated by astigmatism (WFC3 ISR 2013-11WFC3 ISR 2013-13).
  • More than one exposure time may be needed to produce an image that is both unsaturated in the core and has good signal-to-noise to the desired radius.
  • For exposures shorter than about 10 seconds, the UVIS PSF will be affected by vibration of the shutter (WFC3 ISR 2009-20). In some cases, use of the APT exposure-level option BLADE=A may be justified (Section 6.10.4).

We do not recommend usage of the PSF modeling software Tiny Tim. While STScI continues to host the software as a courtesy to the community, it is no longer maintained or supported. In addition, its applicability for WFC3/UVIS images is poor, as it was not optimized to reproduce observed PSFs and relies on outdated optical models (WFC3 ISR 2008-14). Sub-sampled empirical PSFs are available on the WFC3 PSF webpage, and Section 6.6.4 discusses in greater detail ongoing work to provide PSF models to observers.

6.11.5 The Drift and Shift (DASH) Observing Strategy

The term DASH (“drift-and-shift”, Momcheva et al., 2016) was originally adopted to describe the observing strategy of taking a series of WFC3/IR exposures of many targets within one orbit while the telescope is being guided under gyroscope control, thus avoiding the overhead cost of acquiring a new pair of guide stars for every slew between targets of greater than about 2 arcmin. A WFC3/IR sample sequence comprised of short exposure times is selected to avoid image smearing within each time step, and the differential samples in one exposure are later aligned and combined to compensate for the greater drift due to gyroscope control. The technique was designed to allow users to carry out shallow large-scale mosaic observations with the WFC3/IR camera, but has since been adapted to efficiently observe a collection of bright targets within a field ~1 deg across with WFC3/IR subarray apertures (which have shorter time steps for a given sample sequence) and WFC3/UVIS subarray apertures. More detailed discussions of HST guiding and considerations for program design are given in Section 7.10.6. A brief discussion of WFC3/UVIS observations is given here.

Given that WFC3/UVIS is read-out as a traditional CCD, it is necessary to specify short exposure times in place of the non-destructive reads made in an IR sample sequence in a DASH program. Subarray apertures are used to increase the number of exposures that can be fit into the on-board buffer before a time-consuming serial buffer dump must be made. Note, however, that specifying small subarray apertures is risky, especially towards the end of an orbit, when the high drift rates could cause a target to fall outside the aperture. A mitigating strategy is to specify increasing subarray sizes as the orbit progresses, to allow for the uncertainty in predicting the drift. As discussed in Section 7.10.6, DASH visits should be limited to one orbit, and the exposures in an orbit should be grouped into a non-interruptible sequence container in APT. The first exposure should be taken using guide stars to start the observations with accurate positioning.