Not surprisingly, the WIYN Queue Observing Experiment ran much more smoothly during the spring 1996 semester. By now, all active programs have been closed out and all datasets have been shipped. If you expected to receive data and did not, please contact us at wiynq@noao.edu.
Note that all spring 1996 WIYN queue programs have been officially terminated. To continue your program, to start a new program, or to try to get your old program started, you MUST re-apply for observing time using the normal KPNO observing proposal submission procedures.
NOAO was assigned 52 nights during the spring 1996 semester by the WIYN Consortium. These nights were split between classical and queue nights as follows:
Total Dark Bright Classical Nights 24 18 6 Queue Nights 28 16 12 Total 52 34 18
Programs scheduled classically were typically on-going survey projects where it either seemed advantageous for the PI to be on-site or an optimal match between observing conditions and observing program was not critical. This scheduling strategy eliminated the "program runaway" problem experienced during fall 1995 where a few programs greatly exceeded their TAC allocated number of hours. On the other hand, the relatively high percentage of classically scheduled nights appeared to reduce both queue execution sequence flexibility and our ability to match optimally queue programs to delivered observing conditions. During fall 1996, only 16% of the available NOAO dark nights and none of the bright nights have been scheduled classically. This change in the queue/classical ratio will allow us to test whether or not queue efficiency is increased when the queue has a higher percentage of nights in a given semester.
Of the hours available to the queue, 60% of the time was used by specific programs assigned to the queue (as delineated on our Web site, http://www.noao.edu/wiyn/obsprog), 30% was lost to weather, and 10% was lost to a variety of technical difficulties. This amount of downtime due to technical difficulties is still rather high (the WIYN goal is 2%) but not unreasonable for only the second semester in full operation for a very complex facility. Nevertheless, we continue to work with the operations staff to try to reduce such downtime.
A main goal of the WIYN queue observing experiment is to test empirically the hypothesis that, in the face of a high oversubscription rate (about 3.5:1 at WIYN), the science throughput of WIYN can be maximized by executing the most highly ranked science programs first, completing datasets in a timely manner, allowing a larger range of program lengths, and matching the observing program to the observing conditions. The table shows how we did in spring 1996.
Number of Programs
In Complete Partial None Deactivated
Long Programs: High Priority 7 2 3 0 2
Long Programs: HP, Synoptic 3 3 0 0 0
Long Programs: Best Effort 12 0 3 9 0
2Hr Queue: High Priority 9 4 1 2 2
2Hr Queue: Best Effort 4 0 0 4 0
We were clearly successful in executing the high scientific priority programs first, in executing a range of program lengths, and in facilitating synoptic observing in a way previously unavailable to KPNO users. We also successfully matched observing conditions to observing programs in an unexpected way. The queue program was only allocated 16 dark nights this semester, very few of which had sustained seeing better than 0.8 arcsec, although the median seeing for the entire semester (NOAO and University time combined) was 0.8". Thus, we never executed several highly ranked scientific programs that required sustained delivered image quality less than 0.8", but instead executed nearly as highly ranked proposals that did not require such stringent seeing conditions. Hopefully, we will have more nights of sustained sub-median seeing during fall 1996.
Two things prevented us from completing more high priority programs. First and foremost was the advent of Comet C/1996 B2 Hyakutake. This event ended up consuming about 9 hours of prime dark time in a critical period. When combined with the normal impact of bad weather, we ran into a significant scheduling conflict around 13 hours in right ascension. We regretfully elected to drop one high priority program to maximize the amount of data we could provide to other active high priority programs, reasoning that providing more data to few programs was better than providing a partial dataset to one more program. This is in keeping with our goal to provide scientifically useful partial datasets when 100% program completion is not possible. We continue to work closely with the PIs to assure we meet this goal. A second high priority program was dropped when the space mission it was supporting was not launched. For fall 1996, we have been careful to resolve right ascension conflicts before assigning programs to the high priority queue.
The 2hr queue program was not as productive as we had hoped: only 50% of the high priority programs were completed. Again, this arose from an overly optimistic scheduling plan with respect to celestial coordinates and likely weather. While we cannot control the weather, we have been more conservative about phasing 2hr queue programs with the long programs during fall 1996.
Our Web site (http://www.noao.edu/wiyn/obsprog) was a big success, and was clearly watched carefully by queue participants. How do we know? Whenever we made a mistake, we always got e-mail! We did not, however, update the Web site after every queue night as promised. During one period, we did not update the Web site for almost six weeks. Our goal during fall 1996 is to update the Web site after every queue night.
We also need to formalize our data distribution method. Right now, we are mostly shipping FITS files written on Exabyte tapes and associated observing logs. We plan on making this process somewhat more formal during fall 1996.
Please continue to send us feedback on any aspect of the WIYN queue experiment. Our e-mail address is wiynq@noao.edu. Your input is very important to guiding the evolution of this experiment. Our most important goal remains to produce the best quality data possible.
Dave Silva, for the WIYN Queue Team
(Di Harmer, Alex Macdonald, Paul Smith, Daryl Willmarth)