WFCAM Commissioning Report

Andy Adamson & Paul Hirst

WFCAM first arrived in Hilo with a dummy, slot-shaped aluminium plate replacing the aspherical corrector plate which remained in Belgium for rework following its failure of acceptance tests. This rework in fact went very quickly and the original plan (to commission as far as possible with the dummy corrector in place) was rearranged to allow the corrector plate to be installed at the summit.

Commissioning observations of the Great Nebula in Orion, illustrating WFCAM's ability to image detail on large and small scales simultaneously. The left-hand image shows a full WFCAM tile; the image at right shows the many features seen only in the near-IR in the dark cloud to the north of OMC-1 and the Trapezium cluster. The small green square bottom-left shows the UIST field-of-view. Data courtesy of Chris Davis & Watson Varricatt. Final colour composite prtoduced by Douglas Pierce-Price.

WFCAM mounted on UKIRT during commissioning in autumn, 2004.

As reported at the time, a secondary support flexure bond broke when the piezos were initially switched on and tip-tilt correction was not restored until half way through the commissioning (12 November). This allowed time for the rebonded supports to cure, and some night time observing was put into unguided observing on UKIDSS fields, as insurance against the possibility that the supports broke a second time on powering back up (this would have brought commissioning to a swift end). As a result of this change, the engineering slot reserved for adjusting the external tilt shims was instead put into a necessary adjustment of the autoguider focus lenses, guided by results from the initial observing. A complete test of the external adjustment will be carried out during recommissioning in March.

Post-Tip-Tilt commissioning

Once tip-tilt was working and initial performance adjustments were made, commissioning observations were obtained to address the main on-sky characterisation areas. The CASU group have since produced initial estimates of sensitivity from selected datasets.

Commissioning was of course carried out with a tilted focal plane. A significant amount of time was put into obtaining, and understanding, comprehensive measurements of this tilt; the calculated tilts were checked three times independently, and adjustments were made in engineering time after the camera came off the telescope. (Stop press: on-sky measurements made in March 2005 indicate that the tilt has not been properly removed; progress with this will be tracked in the UKIRT web pages).

Overall performance

WFCAM Imaging appears fundamentally good; the corrector plate appears to be working well and 0.6 arcsecond images have been produced with tip-tilt. Even unguided, the median image size was about 1 arcsecond. More detailed information on residual aberrations will come from wavefront sensing information, which will become much more tractable with the focal plane tilt removed. Throughput and sensitivity appear as good as, and in some bands slightly better than, UFTI.

Issues currently being addressed

Autoguider performance clearly required improvement. Great progress has since been made on this by Derek Ives, who at the time of writing is again at the JAC working on both the guider and infrared readouts.

Interactions between the secondary mirror position and the internal instrument focus setting are complex and affect both the amount of residual spherical aberration and the ability to simultaneously focus the guider and infrared arrays. One of the key aspects of recommissioning will be to optimise the relative positions of all these optical elements.

Recommissioning

We planned to reinstall WFCAM on the telescope at the end of February; however, instruments like to keep life interesting and WFCAM threw a couple of small spanners into the works - a broken focus mechanism which required open-cryostat surgery by the JAC ETS team, and subsequently to closing up again, a dead array channel which caused a week's delay. At the time of writing we are just commencing the process of reinstalling the camera on the telescope, in preparation for on-sky observing on the 10th of March. Derek Ives (UK ATC array expert) is at JAC working on the array characteristics, as is Jim Lewis (CASU, working on the data reduction pipeline). We are again into a busy time and looking forward to recommissioning this fantastic instrument.

WFCAM Science Archive update from WFAU

Nigel Hambly

Work on the science archive for WFCAM data has proceeded steadily over the past 6 months, under the aegis of the VISTA Data Flow System (VDFS) project. Funding for increased archive developer effort was secured from UK PPARC e-Science round 2 monies in 2004; an application for funding to cover archive operations for the next five years has been made also to PPARC, as part of the WFAU grant renewal process.

A screen-shot showing the WFCAM science archive homepage

In lieu of WFCAM data, the prototype WFCAM/VISTA Science Archive was built to serve 2.5 Terabyte of scanned Southern Sky Schmidt survey data. This is the SuperCosmos Science Archive (SSA), and has been available to the community for a year now. We have worked hard to ensure that the archive platform is stable, scalable and fault tolerant with good performance under heavy loads. The SSA offers database querying facilities similar to those designed for the WFCAM Science Archive, including joining queries with the 2MASS and USNO-B catalogues, and releases of the SDSS, as they become available. This prototype is available at http://surveys.roe.ac.uk/ssa and UKIRT users are invited to try it out to get the look and feel of the WFCAM archive, which will be available at http://surveys.roe.ac.uk/wsa once we have WFCAM data to serve.

User interface developments include a deployment of Astrogrid infrastructure software to enable science archive database products to be accessed through the evolving Virtual Observatory. We are currently implementing Astrogrid registry and community servers as a solution to user authentication for proprietorial WFCAM data.

Aside from the above work, WFCAM science archive software developments have concentrated on ingest, source merging, database driven image processing and general user interface applications. We are currently in the midst of final shake-down and testing of the archive system using the WFCAM phase-I commissioning data, pipeline processed by CASU. Over the next few months, releases of processed commissioning data will be made available increasingly widely to the UKIDSS user community for feedback on pipeline and archive functionality. Watch this space!

WFCAM Data Processing update from CASU

Mike Irwin and Jim Lewis

Final design, tuning and implementation of the summit processing and standard processing pipelines for WFCAM was awaiting the availability of WFCAM science array test data, and in particular on-sky test data to be taken during the first commissioning phase. During the commissioning observing, CASU supplied data processing software modules and recipes which were run under ORAC-DR to aid in real-time assessment of the data quality.

Over the last few months all of the WFCAM commissioning data (~2 Tbyte) taken in Autumn 2004 have been shipped to Cambridge on LTO-I tape and converted to Multi-Extension FITS (MEF) format. A brief summary of what we have learned so far from these data is given here (see also the CASU diary entry page on http://www.ast.cam.ac.uk/~wfcam for the very latest news and where to find more details).

A whole series of tests have been done to assess the stability of the darks/reset anomaly and the flatfield properties. There are some unresolved problems that will be revisited during the next phase of the commissioning, but the results are encouraging in terms of the stability of these detector properties which fits in with the general design of the pipeline processing strategy.

A variety of flatfield strategies have been examined, and the stability of the flats indicates that master twilight, or even dark sky flats, can be used to correct data. This removes the pressure to reobserve twilight flats every night in every filter band. The really good news is that using stacked high count level twilight flats to flatfield stacked dark sky flats fails to reveal any measurable fringing (ie. it is <<1% of the background)) in all of the passbands so far analysed (Y,J,H,K).

All of the detectors show measurable non-linearity (eg. ~1% 5K counts, ~2% 10k to ~4% at 20k counts) over their entire dynamic range. In CDS mode this can be corrected post-acquisition using an algorithm developed by CASU (see http://www.ast.cam.ac.uk/~wfcam/docs/). This non-linearity has been characterised using sequences of dome flats and will be corrected as one of the first steps in the pipeline.

The 2MASS point source catalog (PSC) has been used in two Galactic Plane regions to categorise the astrometric distortion and to calibrate several MSB sequences of science data, with very promising results. The proposed WCS projection gives an accurate description of the radial distortion and gives global residuals of less than 100 milli-arcsec over the entire WFCAM array. The PSC calibration on a frame-by-frame basis demonstrates that extinction monitoring at the few % level is achievable. It also shows that WFCAM performance is at least as good as UFTI in K and somewhat superior in the bluer passbands. Other quality control measures such as seeing, image ellipticity (trailing), sky brightness are also giving excellent results.

Comparison of the UIST/IFU and CGS4 Efficiencies using Uranus in the K-L' Bands

Laurence Trafton¹, Tom Geballe², Steve Miller³ and Tom Stallard³

The advent of the integral field unit (IFU) into infrared spectroscopy is beginning to enable a new type of astronomy, one where entire extended objects such as galactic nuclei, nebulae, and solar system objects can be dissected in one telescope pointing, rather than by a succession of pointings when a long slit (LS) spectrograph is in use. Yet one must use IFUs with some caution, as due to their complex optics their throughputs generally suffer compared to their longslit cousins. One can show that for an IFU of modest size, say 10x10 elements, and throughput half that of a LS instrument, the improvement in signal-to-noise ratio is no more than 1.6, which can be negligible.

We have recently been in a position to compare the performance of the CGS4 long slit spectrograph to that of the UIST IFU in observations of Uranus in the K and L bands. CGS4 first detected the H₃⁺ ion in Uranus (Trafton et al. 1993, ApJ, 405, 761) and Saturn (Geballe at al. 1993 ApJ, 408, L109). Uranus would seem to be ideally matched to the UIST IFU because of its size (3.6"). The comparison is based on observations of Uranus at low dispersion by both instruments in the K-band continuum and K-, L-, and L'-band emission lines of H₂ and H₃⁺. Despite the relative ease of using the IFU, this is one case where the relatively low throughput of the IFU favors multiple exposures with a longslit spectrograph. We summarize the results here; see the UKIRT technical report, ``On the Mapping Efficiency of Integral Field vs Longslit Spectrographs: Comparison of the UIST/IFU and CGS4 at UKIRT using Uranus in the K-L Bands" by Trafton et al. (2005) for full details.

To accommodate making the comparison at similar spectral resolutions in each wavelength band, we used CGS4 spectra obtained along Uranus' diameter in conjunction with UIST/IFU spectra obtained with the HK, Short L, and Long L grisms (R=800-2300). During the CGS4 observations (June 2001), Uranus was nodded along the 0."61 wide (one pixel) slit. The spectrographic configuration employed was the frequently-used 40 l/mm grating (in 1st order) and the 300 mm ``long" camera. Separate observations were obtained for the K-band and for the combined L- and L'-bands. The merged spectrum was sampled every 1/3 pixel. Uranus' diameter at the time was 3."63. A single 1-D flux spectrum was extracted from Uranus' spectral image by summing the 7 brightest spectral rows (4."27) across the disk diameter.

The UIST/IFU observing program on Uranus was carried out during service ``queue" observations during late 2003 with clear skies and moderately good seeing. Separate UIST grisms having different dispersions were used to cover the different atmospheric windows. In the K-band and L-band, the effective spectral resolution was comparable to that of CGS4, but in the L' window it was significantly higher. The telescope was nodded between the object and sky. Uranus' apparent diameter varied from 3."70 to 3."43. These compare with the IFU field of 3."33 x 6."43, which has 14 usable image slitlets, each 0."24 wide (two pixels), spanning the width of the effective aperture. Each pixel row is 0."12 high. Six CGS4 observations are needed to cover the single-exposure IFU field. For IFU, an average 1-D intensity spectrum was extracted from the 54 x 1024 median spectral image by averaging, for each wavelength sampled, the intensity over a rectangular area defined by the rows spanning Uranus' image. The resulting spectra are shown in Figure 1 for the case of the K-band. The H₂ quadrupole S(1) and Q(1) lines are detected in both spectra; but the S(0) line (at 2.223 micron) and the Q(3) line are detected only with CGS4.

Figure 1: Comparison of the extracted UIST/IFU spectrum of Uranus in the K-band (heavy) with an extracted CGS4 spectrum (light, offset +5x10-6 at comparable resolution; smoothed by a Gaussian having FWHM=1.5 pixels). The region marked at a relatively flat section of the continuum indicates the wavelength interval over which the standard deviations were sampled.

Our basic approach in comparing the efficiencies of these two instruments was to determine which instrument would take less time to map Uranus' disk at useful S/N per common spectral and spatial resolution element. Defining the NEI to be the 1-&sigma standard deviation of the calibrated continuum, we measured it for each instrument using a section of the continuum in each of the three amospheric windows where planetary emission lines appeared to be absent and telluric absorption was known to be weak.

Since the observations were obtained at disparate exposure times, resolutions, and samplings, transforming them to a common basis was necessary for each wavelength band in order to compare instrumental performance. The disparate spatial resolution along these slits was neutralized by summing illuminated pixels along them, to yield the 1-D spectra above. Because CGS4 has the larger pixels and slit width, we used these modified CGS4 exposures, sampling, and original spectral and 1-D spatial resolutions as the common basis for comparing these instruments. The UIST/IFU spectra were, in effect, rebinned to match the CGS4 resolution elements, with two such IFU bins sampling each element.

The NEI values resulting from taking the standard deviation of the continuum regions &sigma_c and &sigma_u for CGS4 and UIST/IFU, respectively. The modified NEI noise &sigma_us, corresponding to a rebinning of the 0."24 wide IFU image slices to fill the 0."61 wide CGS4 slit, thus matching its spatial resolution with 2:1 sampling, was obtained by augmenting &sigma_u for the medianed 14 IFU slices by sqrt(14)/sqrt(0.61/0.24) = 2.35.

The mapping efficiency E of CGS4 relative to UIST/IFU can be expressed in terms of the relative time it takes the two instruments to obtain an equivalent map of Uranus; i.e., t_us/t_map. Here, t_us is the UIST/IFU exposure time that would result if its pixels were binned to sample the CGS4 resolution elements twice, and the exposure time were reduced to yield the same count or NEI per bin as originally per pixel. t_map is the time for CGS4 to map the IFU field (equal to 6-times the CGS4 exposure time that gives the same S/N as an IFU binning that samples the CGS4 slit twice.) Thus, the formula for the efficiency of CGS4 relative to UIST/IFU for mapping Uranus is

   E = (1/6)(tus/tcs)(&sigmaus/&sigmac)2

where &sigma_c is the standard deviation of the CGS4 continuum, &sigma_us is the standard deviation of a sub-element of the IFU field equal to the width of the CGS4 spectrum over which IFU slices have been binned to match the CGS4 spectral resolution, and t_cs is the total exposure time for the CGS4 observation if it were sampled 2:1 instead of 3:1.

The results show conclusively that CGS4 is more efficient than UIST/IFU for observing Uranus at moderate spectral resolution for all three telluric windows tested. The values of E are greater than unity; i.e., equal to 1.6, 1.4, and 2.5 for the K-, L-, and L'-bands, respectively. A given S/N ratio is reached more quickly either by mapping Uranus' disk with multiple CGS4 observations or by setting the slit along the planet's CM and collecting observations while the planet rotates than by exposing once with UIST/IFU. Based on these results, the throughput of the IFU instrument relative to CGS4 is substantially lower than one would expect for an instrument that slices the aperture into 14 images. This lower throughput may be the underlying cause of the poorer performance.

When deciding in light of science objectives which of these instruments to use for observing a more general extended object, one must evaluate the specific target properties carefully against the opposing instrumental capabilities to determine which would be more efficient.

View from the Top

Thor Wold

Back last fall, UKIRT and IRTF held a joint 25th anniversary celebration at UH-Hilo. A fine time was had by all, although due to the short lead-time and financing, it was difficult to get some of the old-timers out here. Tom Geballe, Richard Ellis and Terry Lee gave some talks about the old days. Many, though, were greatly missed by those of us who have been here for most of these past 25yrs. Some, of course have passed away and others are suffering health problems that come with aging, and others just could not make it. I missed seeing many of the veterans - I won't mention names because I might forget someone. This was very unfortunate, but as I said, a fine time was had by all.

UKIRT at the Grubb Parson's factory in the UK.		Construction of the UKIRT dome on Mauna Kea.

The main event was drinks followed by a buffet dinner with speeches regaling where we have come from and where we are headed. There were many luminaries in attendance on the IRTF/IfA side, some of which had been around since the site survey days of the mid to late '60s, so the history did reach back beyond 1978 at times. The food actually was quite good. Alas, the audio for the talks was quite bad for those of us seated a distance from the podium, so some of the talks were quite difficult to follow.

The UKIRT control-room in 1979.

The next day, the IfA had a session with two talks followed by a panel discussion. The talks were great - Terry Lee, on our side, with a historical overview of the planning and construction of UKIRT.

Construction on the Mauna Kea Astronomy Education Center (MKAEC) continues on schedule. For pictures and updates, see http://maunakea.hawaii.edu/menu.html and click on construction. The scheduled completion is the end of this year and the facility will have a state-of-the-art planetarium, a restaurant and astronomical and cultural displays. The distinctive titanium cones are taking shape, and a contract has been secured for the restaurant.

Up top, your Beloved Vacation Resort's continuing exterior work on Building B continues...ever so slowly. The Chinese Water Torture Syndrome is still happening. After having finished the south wall at the end of October, the one-man crew turned the corner...then stopped. It appears that we can expect this to continue for quite some time, as very little has happened in the past three months, and the scaffolding has even been removed. Perhaps we ought to start a betting pool on a completion date?

Late summer and fall produced an incredible infestation of mice, which seemingly began at Hale Pohaku and eventually infected all the telescopes. We had mice audaciously running around the control room floor and up on the countertops - they would even run across the computer keyboards. It took quite some time to finally get some control over this, as the mice were quite plentiful - perhaps caused by the wet spring and early summer and an abundance of grass seed? Occasionally, mice could even be seen out in the open running over the cinders. It seemed they really were everywhere.

The situation in our control room was at times rather annoying. You never knew if they were going to run up your pants' leg. One tried to run up my sweater sleeve. They were VERY brave and seemed not to pay attention to human presence. One female observer was quite taken aback when a mouse tried to sit on her lap!

After the last issue of this Newsletter, we finally began the WFCAM engineering and shakedown, which was very much hampered by foul weather, as well as the usual expected and unexpected kinks. We had no guiding at all for most of the time, and could not properly focus, yet we produced an awesome series of images of Orion (see accompanying article on WFCAM). Imagine what this might look like when we can function normally! We are about to start the second (hopefully final) engineering and shakedown and then begin the surveys in earnest.

During the December engineering downtime, Tim Carroll, Marge Dougherty and I took it upon ourselves to have a go at refurbishing the downstairs crew room; beginning by removing the old furniture and prepping and painting the walls. Marge is handling the purchasing of new furnishings and all the equipment for the project, which will be a rather long-term situation, as we can only work on it when we have the time. The hope is to have the new furniture in soon and decorate the walls with some historic images - inspired by the photos and images that were presented at the 25th anniversary celebration.

We also thoroughly cleaned the control room...you would not believe how much dust can settle in just a few years - and we replaced some ceiling panels that were damaged by leaks last winter. A contractor has done a great job of repairing our roof; the 'traditional' leaks are gone, and now a contractor is painting the stairwells. We hope our visitors find our improvements to their liking...

It has been 25 years and an incredible journey, especially when you consider the science and compare how it was done in 1978 to how it is done now. Our goal, as it has always been throughout the years, is to provide our users with the best observing experience and data possible!

Aloha!

Arrivals and departures

Its been another busy year at JAC in terms of staff changes and donut consumption. We've seen the departure of support astronomers Marc Seigar and Jane Buckle who have taken up post doctoral positions in California and Cambridge (U.K.), respectively. Jane's duties as scheduler will shortly be taken up by Mark Rawlings who (all being well) will arrive in Hilo this Spring from the U.K. Olga Kuhn, who spent over four years at JAC as a Research/Operations Support Astronomer, took up a support scientist position at the LBT late last year; her seat at the summit will be filled by Lucas Fuhrman, who moved to UKIRT and the JAC from the University of Northern Arizona, where he was an Astronomy/Physics Lab Instructor while earning his Masters Degree in Astronomy.

Nick Rees, head of software at the JAC, also headed east late last year to work at Diamond Light Source Ltd, in Oxford, U.K.. His skill, experience and alround knowledge of software, hardware and general astronomy (particularly in the wee small hours of the night - I bet he'll miss the phone calls!) will be greatly missed.

On the Engineering side, David Laird returned to Edinburgh after three years at UKIRT as an electronic engineer, while Chris Yamasaki and Kevin O'Connell headed to pastures new just across the street, moving to Gemini and the SMA, respectively. And Desi Okinaka left the administration department of the JAC to spend more time with her family. We wish them all good luck in their new endeavors, and hope to see those that left the Island back in Hilo soon.

At the same time, we're happy to welcome - in addition to Lucas - Lenwood Jack, James Kaulukukui and John Kuroda to our ETS team. Lenwood moved to Hilo from Lake Charles, Louisiana; before joining the JAC, James was a self-employed welding contractor; and John hails from the Big Island and previously worked at Gemini as an Electronic/Instrumentation Technician. Dylan Terry, a student at U.H. Hilo, has also been working at JAC as an Intern on outreach projects with our outreach specialist, Douglas Pierce-Price.

Lastly, we note that Alan Pickup of the U.K. ATC retired recently after a distinguished career. Alan's contribution spans virtually the whole history of UKIRT and his software has been integral to all of the beautiful science that has been done using the telescope over the years. We wish Alan a happy retirement and promise not to bug him too often!

And finally....

As part of UKIRT's and the IRTF's 25th anniversary celebrations, it was felt that some sort of sporting event was in order. The JAC footie squad immediately offered to take up the challenge and defend the honour of telescope, Queen and country. But, alas, the opposition couldn't muster a team! A canoe race was then proposed. Boats were borrowed, paddles were polished, and on an uncharacteristically overcast afternoon both teams took to the water in Hilo Bay. After a grueling afternoon the JAC team, shown below, were of course victorious. Better luck in 2030, IRTF....

With victory (and an alcoholic beverage) in sight, the JAC team paddle ferociously across Hilo Bay...	The JAC paddlers. Back row: Erik Starman, Ian Midson, Dylan Terry, Tomas Chylek, John Vierra, Matt Rippa, Craig Walther, and Iain Coulson. Front row: Matthew Doyle, Ming Zhu, Angie Midson, Connie Larson, Sandy Leggett, Nick Rees and Andy Adamson. (Ringers not in picture: Matthew Pincet and Steve Redgrave)