"Drift-and-Shift" (DASH): A New Method for Wide-Field Near-IR WFC3 Surveys

Ivelina Momcheva, imomcheva@stsci.edu

Many areas of research require observations of wide fields in order to cover extended nearby objects (M31, Magellanic Clouds), identify rare bright sources over large areas, beat cosmic variance, etc. From the ground, large-format cameras are now doing wide optical and infrared surveys ranging from the 12 deg2 VIDEO and 155 deg2 CFHTLenS surveys, to the 4000 deg2 UKIDSS Large Area Survey in the infrared (IR) and the 5000 deg2 Dark Energy Survey in the optical, up to SDSS which covers ∼1/3 of the entire sky. Many of the science questions addressed by these surveys can be greatly advanced with imaging at the resolution of the Hubble Space Telescope, however, so far the areas imaged with Hubble have been quite small. The largest optical area covered is the 1.7 deg2 COSMOS field, observed with the Advanced Camera for Surveys (ACS) at a cost of 640 orbits (Scoville et al. 2007); the largest infrared areas are the 0.5 deg2 PHAT survey of M31 (828 orbits; Dalcanton et al. 2012), and the 0.24 deg2 CANDELS survey (795 orbits; Grogin et al. 2011; Koekemoer et al. 2011).

The reason for the lack of very wide Hubble surveys is its relatively small field of view for a single observation (the ACS and Wide Field Camera 3 [WFC3]/IR fields of view are 11.3 arcmin2 and 4.6 arcmin2, respectively) coupled with a time-consuming guide star acquisition—10 minutes are required for the Fine Guidance Sensors (FGSs) to lock on a guide star. In the best-case scenario, only four independent guide star acquisitions would fit within a 50-minute orbital-visibility window (current policy allows a maximum of two acquisitions per orbit), leaving just 160 seconds for science exposures at each position. As a result, tiling large areas with Hubble under fine guiding is very inefficient and many thousands of orbits are needed to cover large areas of the sky.

There is a way to circumvent the limitations imposed by the guide star acquisitions. If no new guide star is acquired between pointings, the overhead decreases dramatically and it is possible to fit a larger number of distinct pointings in a single orbit. Ever since the last servicing mission, Hubble has had three working gyros at all times, irrespective of whether a guide star is acquired or not. In a standard guided exposure, the pointing-control system receives continuous corrections from the FGSs. Turning off guiding merely stops the stream of corrections from the FGS and redirects the pointing control system to use information from the gyros alone. The effect of this switch is that the telescope begins to drift by typically 0.001 to 0.002 arcseconds per second and exposures longer than a minute in gyro-only mode suffer reduced image quality along the drift axis.

For CCD detectors, such images would be scientifically unusable. However, an exposure with the WFC3/IR detector is composed of multiple non-destructive, zero-overhead reads. The exposure time between reads can be set between 2.9 and 200 seconds, and for times up to 25 seconds, the drift between them is less than 0.05 arcsec, or less than half of a 0.13 arcsec WFC3/IR pixel. The image obtained in the interval between two reads is simply the difference between two consecutive reads. Therefore, an unguided, gyro-controlled, 300 s exposure with 25 s reads effectively consists of 12 independent exposures that can be shifted and combined into a full-resolution image with hot pixels and cosmic rays removed. We have dubbed this new method "drift-and-shift" (DASH). A number of different DASH implementations are possible, however we find that observing eight pointings per orbit provides the most efficient use of an orbit. In this way, WFC3/IR can cover an area as wide as 1 deg2 in just 100 orbits, much faster and less costly than previously possible.

The DASH technique opens an exciting new avenue for Hubble studies of galaxy evolution. Since the original Hubble Deep Field campaign in 1995 (Williams et al. 1996), the Hubble Space Telescope has imaged many "blank" fields at many wavelengths to obtain a census of galaxies over most of the history of the universe. The survey strategy of the extragalactic community has been to image a few individual Hubble pointings to great depth. Examples are the Ultra Deep Field and the Frontier Fields (Beckwith et al. 2006; Ellis et al. 2013; Illingworth et al. 2013; Lotz et al. 2016; also see the STScI Frontier Fields website; and for larger areas to shallower depth, the GEMS survey, the GOODS North and South fields, CANDELS, etc. (Rix et al. 2004; Giavalisco et al. 2004; Grogin et al. 2011; Koekemoer et al. 2011). The form of the luminosity function of most astronomical objects drives this "wedding cake" strategy of tiered surveys. The number density of faint objects is almost always larger than that of bright objects, which means that representative samples of faint objects can be obtained in deep, pencil-beam surveys and representative samples of bright objects require shallow, wide-area surveys.

Survey Acronyms Decoded

CANDELS: Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey

CFHTLenS: Canada-France-Hawaii Telescope Lensing Survey

COSMOS: Cosmic Evolution Survey

GEMS: Galaxy Evolution from Morphology and SEDs (Spectral Energy Distribution) Survey

GOODS: Great Observatories Origins Deep Survey

PHAT: Panchromatic Hubble Andromeda Treasury

SDSS: Sloan Digital Sky Survey

UKIDSS: UKIRT (UK Infrared Telescope) Infrared Deep Sky Survey

UltraVISTA: Ultra [Deep] Visible and Infrared Survey Telescope for Astronomy (VISTA)

VIDEO: VISTA Deep Extragalactic Observations Survey

Figure 1: Illustration of the "drift and shift" (DASH) method of restoring unguided Hubble images. (a) Shows the standard data product (the FLT file) of an unguided, gyro-controlled exposure. The objects are smeared due to the lack of fine guidance sensor corrections. (b) Shows an example of an individual sample that ultimately comprises the final exposure. The smearing is small in each individual sample. (c) Shows the reconstructed image after shifting the samples to a common frame and adding them.

The first use of the DASH technique is the 57-orbit GO-14114 program (PI: van Dokkum) targeting the COSMOS field and adding a new wide/shallow tier to the extragalactic wedding cake of surveys. The program is being executed in Cycle 23 and will cover 0.6 deg2 of the (UltraVISTA) deep stripes (McCracken et al. 2012); these regions have deep complementary ground-based Y, J, H, and K imaging as well as deep Spitzer IRAC imaging from the SMUVS Exploration Science program (Spitzer GO-11016, PI: K. Caputi). Observations are carried out in the longest wavelength WFC3/IR filter, F160W, to maximize the color baseline at Hubble resolution. The first observations were obtained in October 2015 and initial results were presented in Momcheva et al. 2016.

Figure 2: Examples of objects observed in the COSMOS-DASH program (GO-14114, PI: van Dokkum) demonstrate that final images are well corrected and recover the full image quality of WFC3. (Top panel) Upper left shows the drifted WFC3 image of a known gravitational lens in the field (Faure et al. 2008); the lower left shows the corrected image. On the right are color composites of the ACS F814W (blue) and the COSMOS-DASH WFC3 F160W observations (red) of the lens. The Einstein ring has a radius of 0.74 arcsec. (Bottom panel) Examples of high-redshift mergers observed with UltraVISTA (left) and WFC3 and ACS (right).

The DASH method was made available to users starting in Cycle 24 and as long as the telescope has three operational gyros, we expect this method to be available in future cycles. This new technique is not just aimed at large surveys, but also can make small programs more efficient. Furthermore, DASH may also be used with IR slitless spectroscopy, providing a pathfinder for the future WFIRST mission that will eventually cover thousands of square degrees with both imaging and spectroscopy. Taking advantage of the gyro-only guiding, users have also come up with ways to carry out observations using WFC3/UVIS (ultraviolet and visible wavelengths).

The DASH technique is unique to Hubble and is unlikely to be used in future missions such as Webb and WFIRST. Guide star acquisitions for both missions will be much shorter, and furthermore, since they will be located at L2, visibility windows will be much longer than the 60 minutes in low-Earth orbit. As a result, both missions, especially WFIRST with its 0.28 deg2 field of view, will be much more efficient at tiling large areas. In the meantime, DASH is opening new scientific opportunities with Hubble, demonstrating that even after 26 years in orbit, the observatory still has a few tricks hidden up its sleeve.


Beckwith, S. V. W., et al. 2006, AJ, 132, 1729

Dalcanton, J. J., et al. 2012, ApJS, 200, 18

Ellis, R. S., et al. 2013, American Astronomical Society, AAS Meeting #221

Faure, C., et al. 2008, ApJS, 176, 19

Giavalisco, M., et al. 2004, ApJL, 600, 93

Grogin, N. A., et al. 2011, ApJS, 197, 35

Illingworth, G. D., et al. 2013, ApJS, 209, 6

Koekemoer, A. M., et al. 2011, ApJS, 197, 36

Lotz, J. M., et al. 2016, submitted to ApJ; arXiv:1605.06567v1

McCracken, H. J., et al. 2012, A&A, 544, A156

Momcheva, I. G., et al. 2016, submitted to PASP; arXiv:1603.00465v1

Rix, H.-W., et al. 2004, ApJS, 152, 163

Scoville, N., et al. 2007, ApJS, 172, 38

Williams, R. E., et al. 1996, AJ, 112, 1335