Flexible Scheduling at UKIRT in Semester 03A

Jane Buckle & Andy Adamson

As announced in the last Newsletter, UKIRT was flexibly scheduled for the first time in Semester 03A. At the time of writing, the semester is just complete, and it is time to review the success or otherwise of the undertaking and consider what has been learnt and what changes should be made. We first give a few statistical indications of the degree to which flexible observing has benefitted the science output of the telescope.

Flexing at UKIRT is based on a short list of principles, endorsed by the UKIRT Board:

• Completion of programmes, especially high-priority programmes, is the key goal.
• Programmes requiring the best observing conditions should have the best possible chance of being done in those conditions.
• Observers who come to the telescope should have a good chance to carry out their own observing programme.
• UKIRT programmes vary considerably in complexity, and there are a number of different atmospheric constraints. Summit observers must therefore retain a significant degree of discretion in selecting programmes for execution.

The first two principles are addressed through the TAG, who award summit occupancy to highly ranked projects and/or projects with tight weather constraints. Low-ranked projects, even if they have severe weather constraints, are unlikely to be awarded summit occupancy. The third principle is addressed directly by the UKIRT flexing model, and the fourth principle is addressed by procedures at UKIRT that are overseen by support astronomers. Summit observers can observe their own project unless the weather is not good enough for it, and can choose (according to their technical expertise) between the top few projects in the queue when not observing their own project.

FIGURE 1: The completion of all flexed projects for 03A, colour coded by conditions constraints.

Early indications are quite positive. Across all programmes in the flexed queue, the average completion rate is a creditable 77% (Fig. 1). Across the top 10 flexed programmes, however, the average completion rate is much higher, at 94%. A considerable amount of effort was put into monitoring progress through the semester; as an example, Fig. 2 plots the progress (in percentage completion) for each of the top ten programmes, weekly through the semester. Of these ten programmes, the ones not completed are those which, for a variety of reasons, did not always have MSBs (minimum schedulable blocks) in the observing queue although their targets were available. The completion rate of programmes not available until late in the semester was artificially lowered by the unexpected extension to the aluminizing shutdown.

FIGURE 2: The completion per week for the top ten ranked programmes.

These completion rates are a measure only of the telescope time accessed by each programme through the semester. PIs now have to judge how much of the science has been completed by the data taken. On the UKIRT PI report, which all PIs are now required to return, PIs are asked to estimate the scientific completion of their programme. Early signs are again good: almost all of the reports so far returned state that their science programme is as complete as their time allocation. There are a few exceptions to this, of course; these involve complex observing modes known to be testing the capabilities of the instruments.

Statistics such as these indicate that the essential goal of flexible scheduling - preferential completion of the best programmes - was met, even in the first semester of operation. However, it was also clear in the course of the semester that the speed of the feedback loop between the observatory and remote PIs was insufficient in some cases to guarantee the science quality of the data. This was clearly identified as a critical area for both telescopes. Semester 03B therefore brings a number of enhancements in the OMP software - the JCMT/UKIRT shared suite allowing us to operate in this mode at all. It is now possible for a remote PI to access the pipeline-reduced data products in essentially real time via their project home page, and so the "loop" should be much better closed in Semester 03B than in 03A. But PIs beware - while raw data are persistent for the whole semester, reduced data products remain on disk only for a short while; make sure you react quickly when you get the automatic announcement that data are available for your programme!

Primary Mirror Aluminising - Summer 2003

Marc Seigar

Between June 23rd and July 29th UKIRT was closed for aluminising the primary mirror and modification of the mirror plug and covers to support WFCAM. In addition a new facility UPS was installed.

Due to the crowded instrument delivery schedule over the past few years, it had been a long time since the mirror was last aluminised (almost five years) and, as a result, there were very few science staff at JAC who had seen the procedure done before.

FIGURE 1: Lowering the mirror cell				FIGURE 2: The mirror and (below it) the mirror-cell			FIGURE 3: The mirror-cell in full view

The first step in the process was to lower the mirror cell, complete with mirror, onto a trolley on which it could be wheeled out from under the telescope (Figure 1). This entire structure weighs about 20 tonnes. The mirror was then (carefully!) lifted out of the cell (Figure 2). This was only the second time this had been done since the installation of active optics actuators around the edge of the mirror - the anchorage points can be seen in Figure 2. Figure 3 shows the mirror cell after extraction of the mirror. The aluminium cylinders are the pneumatic supports which are connected together in three sectors and form the "pneumatic whiffle tree". There are 80 of these pneumatic supports, each of which supplies a pressure of 80 lbs per square inch to the mirror. During the shutdown each of these supports was tested for leaks. One of them was already known to have a leak and was repaired.

The mirror was then placed into its transportation box and delivered from UKIRT to its destination at the Canada France Hawaii Telescope (CFHT - Figure 4).

At CFHT the mirror was cleaned and stripped of the old coating (Figures 5 and 6). Chemical safety regulations dictate that everyone had to be trained and specially clothed.

FIGURE 4: Transporting the mirror cell to the coating chamber at the CFHT			FIGURE 5: Cleaning the primamry begins...			FIGURE 6: Cleaning in full-swing

The next step was to coat the primary, which is done in a large vacuum tank. The evaporative coating process consists of simply boiling small samples of aluminium in carefully-selected places in the vacuum chamber; in the absence of air the atoms drift in a straight line and stick to the first thing they hit. The art comes in knowing just where to put the samples, and how much to use to obtain the right thickness of coating. On this occasion, the first attempt proved to be perfect, with a coating thickness of 850 nanometres. The amount of aluminium used was only about 2 mg - about one eighth the amount of aluminium in a typical soft drink can!

Back at UKIRT, the newly coated mirror was lowered back onto the mirror cell, as shown in Figure 7.

Two other projects were ongoing whilst the primary mirror was going through the various stages of cleaning and re-coating at the CFHT. Both tasks were performed in order to support the new Wide Field Camera (WFCAM), which will be delivered to UKIRT early next year. The first was modification of the mirror plug (at Bear Machinery in Honolulu); the new mirror plug is shown in Figure 8. The second task involved modification of the primary mirror covers on the telescope so that they will close while WFCAM is mounted on the telescope (Figure 9). Both tasks were completed successfully and in good time.

FIGURE 7: Reinstalling the primary back onto the mirror cell			FIGURE 8: Working on the mirror plug			FIGURE 9: Modifying the mirror covers

So, well done and a very big thank you to all those involved, especially the ETS group who spent around eight weeks at the summit whilst this heavy engineering was going on and being prepared. We are also grateful to the CFHT for the use of their aluminising equipment and the efficiency with which they accommodated this work.

A web page showing more photos of the shutdown can be found at the following URL: http://www.jach.hawaii.edu/JACpublic/JAC/ets/ukirt-aluminising-shutdown/webpage

The latest on WFCAM from the UKATC

Mark Casali

WFCAM sub-assembly integration and testing is proceeding on many fronts simultaneously. Among the important recent milestones are

One of the four science-grade detectors from Rockwell

• delivery of all four Hawaii-2 Science Grade detectors from Rockwell
• delivery of the window, field flattener lens and field lens from SESO
• delivery of all eight acquisition and data processing PCs
• manufacturing of all parts for all cryo-mechanisms
• achievement of operating temperature for the focal plane assembly in the final cryostat, after modifications to the thermal paths.
• successful cool-down with the large 480mm window
• the first successful SDSU controller operation from the high-level WFCAM software
• completion of the fast-autoguider CCD hardware and software

• difficulties in achieving the wideband coatings on the large optics: this has forced us to descope to simpler coatings on all surfaces
• bugs in the 32-channel SDSU controllers: these are being investigated and removed in consultation with Bob Leach
• delays with our test cryostat: essential for measuring and achieving detector coplanarity and alignment

At time of printing (September 2003), delivery of WFCAM to the JAC is expected for March 2004

See www.jach.hawaii.edu/UKIRT/instruments/wfcam/ for a description of the instrument, and http://www.roe.ac.uk/atc/projects/wfcam/ for an overview of the project and the latest news from the ATC.

Update on the UKIRT Deep Sky Survey (UKIDSS)

Steve Warren

The UKIRT Infrared Deep Sky Survey (UKIDSS) will begin after commissioning of the Wide-Field Camera (WFCAM). The anticipated start of the surveys is mid 2004. UKIDSS will use approximately 1000 nights of UKIRT time over 7 years (subject to continuing Board support!), taking about half of all the UK time on the telescope. A description of the scope and aims of UKIDSS was provided in Newsletter Issue 10 (Spring 2002). This brief note provides an update on UKIDSS activities since that time. The web page www.ukidss.org gives fuller details. As always, please contact Steve Warren, the Survey Scientist, at sjw4@ic.ac.uk, if you have any queries, comments, requests, or you have help to offer. See also the previous article by Mark Casali for an update on the status of WFCAM.

Technical web page

There have been some minor technical changes to the instrument which are detailed on the technical page, www.ukidss.org/technical.html. The array spacing has increased from 90% to 94%. Two filters have been added, a Br-gamma filter and a Z filter (0.83-0.925 micron). The Z filter has a similar central wavelength to the Sloan z' filter, and takes advantage of the unexpectedly good short wavelength quantum efficiency (Q.E.) of the WFCAM arrays. Because of this good Q.E. the Galactic Clusters Survey have decided to expand the number of filters used to include the full broad-band set ZYJHK, at the expense of reduced areal coverage. Details of the revised fields will appear on the web pages once they have been finalised.

At a meeting of the consortium in November we finalised the list of image parameters that will be measured, which is provided on the technical page.

A set of UFTI observations to investigate issues of fringing, flat fielding, and sky subtraction has been designed, and are expected to be executed in September and October of this year. The results of the analysis of these observations will be reported on the UKIDSS web page shortly after completion of the observations.

Two-year plans

The consortium has been focusing on the 2-year plans, detailed at www.ukidss.org/surveys.html. Meetings were held in May and June to consider detailed implementation issues. Each working group drew up a list of detailed requirements, for example, filter sequence, field priorities, jitter sequence, microstepping, etc. The next step in this process will be filling and prioritising all the MSBs for the entire 2-year plan, once the Survey Definition Tool is complete. At the same time a more detailed scheduling model will be developed, to identify areas where the requirements of the different surveys may come into conflict.

Complementary optical data with the ESO VST

There has been a positive response from both ESO, officially, and from the ESO community to the suggestion that a substantial fraction of the ESO time for surveys with the VST (which is nominally 40% of the total VST time) be devoted to obtaining complementary optical data in UKIDSS fields over the first 3 years of operation of the VST. The VST is expected to become operational towards the end of 2004. Following input from the Large Area Survey (LAS), the Galactic Cluster Survey (GCS), and the Deep Extragalactic Survey (DXS) working groups an outline request to support these surveys was presented this Spring to the ESO Working Group for Surveys (WGS), and to an ESO workshop devoted to Large Programmes and future public surveys. The total amount of time required for UKIDSS VST support is estimated at between 100 and 155 nights. There was strong support for this proposal from both fora. The next step is for the ESO WGS to formulate a policy for issues such as the procedures for proposing surveys, the actual amount of time to devote to surveys, and the provision of pipeline and archive facilities. Later this year UKIDSS will need to write a scientific and technical case for the time.

World data release

The current consensus of the survey working groups is that the data from the first two-month WFCAM run should be released to the world about 15 months later, and that the data from the completed two-year plans should be released 12 months after the completion of observations. This is the position stated on the web page.

View from the Top

Thor Wold

It has certainly been yet another busy semester here at UKIRT, not least due to our July-long shutdown for primary mirror aluminizing and heavy engineering. However, this was also of course the first semester of fully queue-scheduled observing. This went rather well, overall, judging from the reports from the PIs - both those who spent time at the telescope and those that had their observations done fully via the queue.

There were some snags. Most noticeable was the lack of full information regarding location of targets, especially in crowded fields. Finding charts with different scales and even at different wavelengths can sometimes be unhelpful (optical images in particular are often not helpful). From the observer's viewpoint, every bit you can give us to help us to determine the location of sources is essential. Otherwise, we can sit here for quite a while, with your clock running, trying to figure out what is what...

During the July engineering, Tim Carroll and I spent the first week up-top, mostly working over at the Canada-France-Hawai'i (CFH) telescope in the aluminizing room dealing with the preparations, but also at UKIRT doing a lot of odds and ends and watching the dismantling of the telescope and transport of the mirror to CFH. This was quite an educational experience. It was also quite weird being up there and working during the daytime. I never had to shut my room window shade, for example; something we always have to do with our usual sleep-cycle.

Tim and I also had the job of creating the list of tasks and the proper chronology for coming back on-line after everything got put back together. I must say that our guys did a fabulous job! Everything came back to life sensibly and we were able to get everything re-aligned, etc., within a few nights' work, and without any major problems. Congratulations are in order to the ETS staff for a job very much well done!

While our crew was so efficient and well-organized, the same surely cannot be said about the ongoing construction at your beloved Vacation Resort Hale Pohaku. The project, to bring HP into compliance with the Americans With Disabilities Act, was to be a 'few months' long (starting in mid-January). But unfortunately, at this writing, it is nowhere near being finished. Activity is truly on-again, off-again, on-again... with periods of construction crews working and periods of no activity at all.

The worst of this situation is the reconstruction of the lavatories in the Commons Building, a project that was supposed to be finished by around March, but which was delayed and delayed and is now due for completion in late August. In the meantime, if you do not have a room, you have to put up with the portable facilities, which were put in place 'way back in March outside the west door. These naturally went unused until utterly necessary, and then it had to be pointed out that they did not have facilities for hand-washing -- so a garden hose was set up with a sink! I think it was a while before anyone realized that it would be nice if at least one of the units had an electric light in it!

In the meantime, it was decided that the outer walls of the dorms needed to be replaced, as I mentioned in the previous Newsletter. They began by working on building C, all the while still putting people in rooms, but on the opposite side of the building from the work. Well, this has not worked out too well at all, causing people to lose sleep from the racket. We have repeatedly had to send reminders about the contractors not working in the mornings, but it always seems that after a few weeks they are right back at it, from 7:30AM.

Then, in June, work ceased. The scaffolding is still in place on the Mauna Kea side of building C and the suspicion is that work will at some point re-commence. I have always been of the opinion that the folks that run Hale Pohaku often forget what it is like to have to contend with working all night and trying to sleep in the day. The worry now with HP regulars is that they will next target building B, which houses (by far) most of the daily tenants. If this does happen, it will definitely not be a thing to look forward to at all.

Back in Hilo, construction is nearly finished on the new Puainako Street Extension. I've mentioned this project in the past, because it may well change how people commute between JAC and Hale Pohaku. The project ought to be finished by around October if the weather cooperates. It may then be most efficient to turn left out of the University Park for about a block, then drive up the new road, which will re-connect near the 6 mile mark of Kaumana Drive, thereby avoiding all of the twisty, narrow part of Kaumana. The only real reason to follow the current route (up Mohouli) would be if one had to fill up on gas on the way. We will be sure to let our visitors know of the new situation when access becomes available. The Puainako extension will eventually also connect into the re-alignment of the Saddle Road, so it will make your commute not only quicker but much safer.

The Silversword plant on Mauna Kea

While this past semester was loaded down with an unusual amount of engineering time, the current one will not be too much better, since we still have some urgent projects we have to fix in addition to preparations for the arrival of WFCAM, currently slated for 04a. WFCAM will also require a large block of engineering when it is first installed on the telescope. I do hope folks will be patient with us through this next year. Please remember we do strive to give you the best service possible, and we do have a reputation for just that, a reputation that we need to keep up.

The rare Mauna Kea Silversword plant that I wrote about in the August 2001 article has indeed bloomed at last (see the attached figure). I took this photo back at the end of May when the flower spike was just starting to blossom (from the bottom up). There was relatively high humidity at the time - cloud tops were lurking nearby and I had to get this photo while I had decent light before the next bit of cloud came in. What surprised me, though, was the aroma! I had no idea they carried a scent that humans could detect - of course, they have to have some sort of aroma to attract their pollinators. It may have been the time of day (evening, on the way up to work) and/or the humidity that caused the plant to release the scent, but I must say it was certainly awesome and very pleasant. As for a description, this is difficult. If I could make it a visual experience, I would say "amber honey".

The tour vans were oblivious to the Silversword's presence until very recently. This was good, I suspect, for at least the poor plant was mostly left alone. Now, however, every time we come up, there are literally hordes of tourists cluttered around the thing..

The bloom is about finished now. If the plant survives the waves of "specimens of touri", it ought to start to produce side buds and develop into more of a clump (at the moment it looks like a single plant). I expect someone is monitoring the situation and will probably be collecting seeds for propagation, as well as re-positioning the protective fence to keep the ungulates from eating it (this was the fate of the other silversword plants that were planted in the same area). Then, perhaps, we can hope to enjoy more blooms in a few years time. I, for one, would surely love to smell that wonderful aroma again.

Aloha!