The UKSERVice Program at JCMT

An analysis is made of the UKSERVice observing program at JCMT. Most users are satisfied with the quality of their data, although improvements are suggested and need to be addressed. A rough estimate shows service observing could save up to 10% of JCMT telescope time.

Introduction

Service observing programs have been operating at JCMT for 3-4 years. Applications for UKSERV time are similar to those to PATT. They are evaluated for scientific value and technical feasibility, and for conformity to the approved use of the service program (see the JCMT Newsletter of July 1993, p12). The observations are performed by JCMT support staff, in a flexible way depending upon weather conditions, and the data are made available to the PIs is for electronic collection and reduction at their home institutes.

With the conference 'New Modes of Observing for the Next Century' scheduled for June 1995 in Hilo, it seemed timely to determine whether the community was getting the most from this service. The results were presented as a poster paper at that conference, the proceedings of which, including this contribution, are published in the Conference Proceedings series of the Astronomical Society of the Pacific.

The investigation was in three parts :

POLL 1 - of PIs who had obtained UKSERVice data,

POLL 2 - of UK astronomers who had obtained regular PATT time at JCMT, but who had not used the UKSERV program as PI,

an analysis of observing reports for both service programs and regular PATT programs to derive efficiency of time usage.

Poll 1

PIs of JCMT UKSERV programs were surveyed by e-mail during March and April of 1995. The questions, responses and other comments are presented here as Appendices 1 and 2.

The distribution of programs by semester and instrument ( U= the common user bolometer UKT14, A/B/C= the A-/B-/C-band spectroscopic receivers), and the distribution of responses by the same criteria are shown in Table 1.

	<------------ programs ------------->	<------------- responses ----------->

Sem	U	A	B	C	Total	U	A	B	C	Total	%

V	4	6	4	-	14	4	1	3	-	8	57

W	10	3	4	-	17	3	0	1	-	4	24

X	10	1	1	1	13	3	1	0	0	4	31

Y	5	2	4	2	13	0	1	2	0	3	23

94a	12	9	3	1	25	7	3	2	1	13	56

94b	6	9	5	4	24	5	5	2	2	14	58
  
Total	47	30	21	8	106	22	11	10	3	46      

						47	37	48	38	43	%

Were the observations performed as requested ?

		All programs	U		A+B+C		Point-sources	Maps

Yes		31	84%	12	67%	19	100%	22	81%	9	90%

Sort of		4	11%	4	22%	0	0%	4	15%	0	0%

No		2	5%	2	11%	0	0%	1	4%	1	10%

		All programs	U		A+B+C		Point-sources	Maps

satisfactory	26	70%	9	50%	16	84%	19	70%	7	70%

useful		9	25%	7	39%	3	16%	7	26%	2	20%

unsatisfactory	2	5%	2	11%	0	0%	1	4%	1	10%

		All programs	U		A+B+C		Point-sources	Maps

excellent	15	40%	5	28%	10.5	55%	9.5	36%	5	50%

satisfactory	18	49%	9	50%	8.5	45%	13.5	52%	4	40%

unsatisfactory	4	11%	4	22%	0	0%	3	12%	1	10%

Client type		UK		International 

Access Easy		20	91%	3	60%

Not easy		2	9%	2	0%

Reduction Easy		19	90%	3	60%

Not easy		1	5%	2	40%

not attempted		1	5%	0	0%

Publications			Impression			Use again

Already		10	27%	favourable	28	90%	Yes	28	90%
 
likely		24	65%	non-committal	2	7%	Maybe	0	0% 

unlikely	3	8%	otherwise	1	3%	No	3	10% 

Of 52 PIs polled on their 106 programs (average = 2.0 programs per PI), we received responses from 32 PIs regarding 46 programs. Tables 2-4 show the breakdown of responses to some of the questions regarding satisfaction with the observing and data, ease of access to the data, resulting publications, and general impressions.

Of 25 discharged UKSERV programs, 21 (84%) were also described by respondents as being 100% completed, and 4 were described as 0%, 33%, 50% and 95% completed. The latter was considered successfully concluded by the PI, and the third was deemed to be unviable as a service proposal and was terminated. Only in two cases did respondents feel that they had been let down :

1. only 1 hour of integration was done cf. 5 hours requested. The target, an asteroid, moved away from Earth and so the loss was irrecoverable (we note that the conditions were not favourable for photometry during the only two UKSERV shifts available).

2. because 'observing instructions were not followed fully' integration times were shorter than requested : 0.5 hours cf 2 hours for 2 of 3 targets (we note that 450 mm photometry was good on only one of the 6 nights that it was attempted. The program was not resubmitted for completion).

Poll 2

Seven replies were received from 22 questionnaires. Respondents typically perceived that restrictions on time and complexity would prevent acceptance of their likely proposals. A need was expressed for more regular scheduling and a faster response time. One respondent was adamant that he wouldn't want, wouldn't trust, anyone else to do his observing for him.

Analysis of Observing Reports

Records for 1993 & 1994 show that 20% of time allocated to primary observing programs was lost to equal proportions of weather that was 'poor' (photometrically unstable or high opacity) or 'bad' (leading to closure). The poor weather was then used for a backup program. In the same period such weather claimed only 11% of service time.

Results

The responses to the questions in the polls, the other comments of the respondents, and the analysis of the observing records may be summarized as follows:

1. At about the 90% level, UK astronomers are satisfied with UKSERV,

2. Service observing makes good use of the telescope. It has an implicit flexibility that allows good use to be made of almost any weather.

3. For about 5% of the proposals the observations and calibrations were not performed satisfactorily. This crack is large enough for those PIs who fell into it, and UKSERV must attempt to close it. This may be done by

- better communication between client and observer, and/or more detailed specifications of the observations and desired noise levels;

- having observers who understand the needs of astronomers. (We think our current observers meet this criterion already.)

4. There is no obvious difference in perceived performance between proposals requiring maps and those requiring single point observations, although our sub-mm photometry was less well received than our spectroscopy. In this context sub-mm photometry should perhaps be considered as more difficult than spectroscopy. How these perceived performance levels compare with those of 'classical' observers needs to be assessed. We mention here that multi-shift PATT proposals have also been performed at JCMT in serviced mode with reasonable success.

5. Our data transport mechanism uses VMS backup, a STARLINK compression and FTP. Our international clients had some trouble accessing STARLINK software, and increasing numbers of UK- based scientists in Unix environments are requesting their data in appropriate format.

6. PIs need more feedback concerning the status of their programs. These problems may best be solved by a noticeboard system such as the World Wide Web.

7. A facility is needed for service clients to access JCMT utilities in order to better analyse their data.

Conclusion

1. UKSERV provides UK astronomers with high-quality data that might be difficult to obtain through normal channels.

2. Service observing is an acceptable means of obtaining data. While the characterization of program by complexity is unclear, there seems no limit yet to the possible scope of service observing.

3. Service observing is an efficient way to use telescope time, saving perhaps 10% of the available time currently lost to 'poor' weather.

UKSERV may be improved by :

better communications with the clientele, perhaps using WWW, particularly in order to reassure them that their proposals are still in the queue

easier access for clients to their data and to JCMT utilities

convincing that part of the community still unsure of the feasibility of having their data obtained for them.

Iain M. Coulson, JAC