Apache Point Observatory 3.5-meter
telescope: operational design and issues
Bruce Gillespie
Apache Point Observatory, PO Box
59, Sunspot, NM 88349
ABSTRACT
The "new-technology"3.5-meter telescope at
Apache Point Observatory has been in routine operations since 1994.
Designed to enable nearly full remote operation via the Internet,
remote use of the telescope comprises two-thirds of all observing.
Rapid instrument change capabilities and flexible scheduling allow
for some optimized science utilization. Several science programs can
share the telescope on a given night, using more than one scientific
instrument. Remote users can also collaboratively use the telescope
simultaneously from different geographical locations. Synoptic
observing programs and rapid-response observations are routinely
accommodated. More than two hundred observers have used the telescope
remotely, and by the end of 1997 more than 60 scientific publications
based on telescope data have appeared in the journals.
Several scenarios for operating the
telescope have been explored. The current scheme is to schedule the
telescope by quarters based on prioritized proposals submitted by the
consortium member institutions. Except for short synoptic
observations and targets of opportunity, each night is divided into
halves. These half- night blocks provide adequate time on target plus
calibration time, and provide simplification of the scheduling
process which is done manually.
Enhancements to telescope performance and
efficiency are underway, which will provide broader scientific
potential and support more exotic scheduling and operational
paradigms. "Remote" queue and adaptive scheduling are conceivable,
and the use of existing AI- assisted scheduling aides is
also under consideration.
Keywords: remote observing,
Internet, 3.5-meter telescope, Astrophysical Research Consortium,
ARC, Apache Point Observatory
1. BACKGROUND
The Astrophysical Research Consortium (ARC)
was chartered in 1984 to build and operate observatory facilities at
Apache Point for the shared use of consortium astronomers and
students. The major projects at the site are the 3.5-meter telescope
which has been in operation since 1994, and the Sloan Digital Sky
Survey telescope whose installation is near completion. The
construction and ongoing operations of the 3.5-meter telescope are
funded by the following member institutions: the University of
Chicago, New Mexico State University, Princeton University, the
University of Washington, and Washington State University, with Johns
Hopkins University becoming a member shortly after the telescope was
commissioned. Telescope time is allocated to these institutions in
proportion to their level of project contributions. Partial
construction costs of the telescope were provided by a NSF
grant.
Apache Point Observatory (APO) is located
in the Sacramento Mountains at 2800 meters elevation near Sunspot,
New Mexico. Observing facilities at APO include a 3.5-meter telescope
used for general visible and IR imaging and spectroscopy, the SDSS
2.5- meter survey telescope, a smaller telescope which is used to
calibrate survey photometry, and a 1-meter telescope owned by NMSU.
The site also contains operations and support buildings, plus
dormitories. An aerial view of the observatory is shown in Figure
1.
Figure 1. Aerial view of Apache
Point Observatory, photograph by Dan Long,
The 3.5-meter telescope and its early
operations have been described elsewhere.1 This paper
gives an update on the operational design as it has evolved since the
early commissioning, and includes discussion of issues and future
enhancements. Although the telescope is still undergoing significant
engineering enhancement, the conduct of observing programs on the
telescope has become robust and scientifically productive. The
inherent flexibility in scheduling and operations&emdash;which are
dependent on the hardware design and remote observing
capabilities&emdash;has proven its worth. For example, a remote
synoptic program which makes quick periodic observations of
brightness changes in lensed quasars has derived an independent and
robust estimate for H0, the Hubble Constant. Other rapid- response
observations have led to the optical identification of a recent
gamma-ray burst object.
The telescope exterior and interior views
are shown in Figures 2 and 3.
Figure 2. Exterior view of the
3.5-meter telescope, photograph by Dan Long
Figure 3. View of the 3.5-meter
telescope from inside enclosure, photograph by Dan
Long
The telescope regularly uses an infrared
imager and spectrometer, a medium-resolution visible-light
spectrograph and imager, and a large-format visible-light CCD imager.
A high-speed infrared CCD imager, and a prototype adaptive optics
system have also been integrated with the telescope. Also, a number
of visitor-supplied instruments have been used with the telescope,
some for limited observing runs and others on a continuing basis. An
echelle spectrograph is scheduled for delivery at APO this year.
Planning for next-generation facility instrumentation has been
initiated.
2. OPERATIONAL
STRATEGY
The styles in which observatories support
users range from "help yourself" to "full-service. " The APO user
support model is somewhere in the middle of this range. The users are
"owners" of the observatory in a very direct sense; this helps
promote a shared and cooperative approach to operations. This
operational paradigm is often described as a partnership between
astronomers and the observatory, which is facilitated by the remote
observing capability and by on-line user information systems.
Operational costs are kept down by promoting the direct involvement
of the astronomers. Also, a side-benefit of this partnership is ARC
astronomers have a first-hand understanding of the data acquisition
process and hence are directly involved with the quality of their
data. The APO remote observing systems allow this "hands-on"
observing approach while enabling flexible and reactive telescope
use, and saves significant amounts of astronomers time (and
funds) by limiting the amount of travel necessary to observe. Since
astronomer travel costs are mitigated, new types of observations are
made easier such as synoptic observing programs.
The 3.5-meter telescope has a user
community estimated to be more than 200 astronomers and students.
Principal Investigators must be from ARC-affiliate institutions;
collaborators from outside ARC institutions are welcome and numerous.
Although not as large a constituency as a national facility, APO
serves enough astronomers and students that the user support systems
are a similar to some of the practices used at both small private
observatories and the larger federally funded
institutions.
3. SCIENCE PROGRAM
LIFECYCLE
3.1 Proposal support and
selection
Each ARC astronomy department selects a
faculty "scheduler" and a Users Committee representative. The
scheduler conducts the local proposal selection process, and delivers
a list of prioritized observing programs to the Director six to seven
weeks before the next scheduling quarter. The telescope was
originally scheduled on a monthly basis, but a quarterly system has
been adopted as more amenable to the users. The Users Committee
members facilitates two-way communication between the observatory and
the institutional users. User information is also furnished through a
web page at APO, or through direct phone or e-mail contact with site
staff or other ARC scientists. The APO web home page is shown below
in Figure 4, and contains information about the site and visitor
logistics as well as technical information for planning observing
programs.
Figure 4. APO World Wide Web home
page, at www.apo.nmsu.edu
The ARC proposal processing system is
relatively simple, inexpensive, and expeditious. The Director issues
a call for proposals by e-mail directly to the schedulers at the
respective ARC astronomy departments. The call for proposals contains
a short section on updated policies, time allocation quotas for each
institution, scheduling constraints, and a blank ASCII proposal form.
A typical proposal is less than two pages long, partly because only
the summary observing parameters need to be specified in the
proposal&emdash;the PI or collaborator will generally conduct the
program hands-on, so these details and not required in the proposal
[Proposal details are available by opening message 254 at
http://www.astro.princeton.edu/APO/apo35-
general/INDEX.html].
After selection and prioritization, the
institutional schedulers send the proposals to the Director
electronically, who hand-crafts a three-month schedule. Proposal
priorities, lunar and other scheduling constraints, balance between
institutional allocations, inclusion of engineering time, etc. are
reasonably well accommodated. The majority of the programs are given
half-nights unless they require longer or shorter observing duration.
Multi- night campaign observing programs as well as half-hour
synoptic or target-of-opportunity programs are in the minority but do
comprise a significant amount of observing time.
3.2 Observing
Typically, two to three programs are
scheduled for separate time intervals in the same night. These
science programs often involve different instruments, observers, and
institutional affiliations. Observers can either observe at the
telescope in person, or remotely through the Internet or modem
backup. Multiple remote users can also connect simultaneously from
different geographical locations, allowing "eavesdropping" or
collaboration at a distance. Routine remote operation of the
telescope has been conducted by ARC astronomers from Israel, Canada,
the United Kingdom, the Kuiper Airborne Observatory, and once via
satellite from the South Pole. Roughly two-thirds of all observing is
done remotely, with the remainder being on-site observations or
service observing. The observer interface software screen is shown
below, Figure 5.
Figure 5. Remote observing
control screen, "Remark" - click to enlarge
Instrument changes are done at night by a
single person, often in less than 5 minutes. The basic design of the
telescope will enable these instrument changes to be performed
without human assistance in the future, and allow several instruments
to be mounted on the telescope simultaneously and kept in "standby
mode. "
All telescope operation is supported by an
on-duty Observing Specialist, who is responsible for telescope and
human safety, as well as providing technological assistance to the
on-site or remote astronomer using the telescope. For a small number
of programs, the Observing Specialists are active collaborators and
conduct the observing.
Visits to the site by astronomers are
mainly for installation and testing of new instruments, or for
training purposes. Observing functions, which include complete
telescope control, instrument control, quick-look quality assurance,
and data retrieval, are all accomplished by the remote observer.
Remote observing may only be undertaken by, or with the direct help
and supervision of, observers with on-site experience and training.
Normally, this is taken to be at least 3 nights of time at APO. At
the site, some help for experienced observers can be provided by the
Observatory staff but training of graduate students or other
inexperienced observers is not available; it is the responsibility of
ARC institution faculty and staff.
3.3 Post-observation support and user feedback
To help keep operating costs down, data
analysis and archive functions are assumed to be provided by the
users home institutions. IRAF and IDL are available on-site for
quick-look inspection, and sufficient disk storage space is provided
to allow user data files to be stored at the site for up to a week
before automatic erasure.
Nightly user feedback is collected by the
Observing Specialist and summarized in the night logs. If technical
problems are encountered, a web-based problem reporting systems is in
place, which is managed and attended to by the daytime staff. Monthly
telephone meetings of the 3.5-meter Users Committee are held during
the year and the minutes are electronically published through e-mail
distribution, to which users subscribe. There is no newsletter, but a
yearly users meeting is held at APO to discuss recent science and
engineering progress with the telescope, and to canvass opinion on
long -term planning and observatory priorities.
4. SUPPORTING
SYSTEMS
4.1 Observatory staff
The on-site APO operations staff are NMSU
employees whose time is split evenly between support of the 3.5-meter
telescope and the SDSS, with the exception of the Observing
Specialists and Electronics Technician who work exclusively on the
3.5-meter. On- and off-site support for the 3.5-meter telescope is
provided by slightly more than 10 full-time equivalents (FTEs),
consisting of the following job titles and
levels-of-effort:
- Site Director (0.1 FTE)
- Site Operations Manager (0.5
FTE)
- Deputy Site Manager, Observatory
Engineer (0.5 FTE)
- Telescope Systems Engineer (0.5
FTE)
- Electronics Technician (1.0
FTE)
- Observatory Computer Systems Manager
(0.5 FTE)
- Observing Specialists (3.5
FTE)
- Technical Writer (0.5
FTE)
- Data Aid/Housekeeping (0.5
FTE)
- Maintenance (0.5 FTE)
- Accounting Technician (0.45
FTE)
- Clerical Aide (0.22 FTE)
Substantial support is provided by off-site
personnel, including:
- Observatory Director (0.1
FTE)
- 3.5-meter Director (0.25
FTE)
- Telescope Scientist (0.3
FTE)
- Operations s/w maintenance (0.1
FTE)
- Instrument Engineering (0.2
FTE)
- ARC administration (0.25
FTE)
- Mechanical Engineering (0.25
FTE)
The site operations staff generally work a
5-day Monday through Friday work shift. For serious technical
problems, nighttime and weekend emergency coverage is arranged in an
ad hoc fashion. Authorized staffing levels do not support scheduled
on-call engineering support. The Observing Specialists who oversee
nighttime telescope operations work an innovative 21-day work cycle:
7 nights consecutive, 7 days off, 5 days mid-shift (noon to 8 p.m.),
two days off. An issue with this schedule is fatigue in the winter
months when the nights are longest. To minimize the effects of
fatigue on efficiency and safety, consideration is being given to
splitting responsibility for the longest nights between two Observing
Specialists.
4.2 Engineering time
About 7% of telescope time is nominally
scheduled for routine engineering tasks, which include periodic
corrections to pointing models, optical collimation, calibrations,
system throughput measurements, operating software upgrades,
training, and small repairs and tune-ups that can wait until a
scheduled night or half-night of engineering time. Site engineering
staff work in conjunction with off-site ARC astronomers and engineers
to plan and conduct these activities. The engineering observing is
often conducted remotely with the help of an on-site Observing
Specialist.
Major engineering shutdowns are planned and
executed on a "when ready" basis. These shutdowns have been allocated
as much as 25% of available telescope time during a year, the time
being used to effect major upgrades and repairs to the telescope.
When all the prerequisites are in place to install a major system
upgrade (e.g., a new guider) or perform substantial maintenance
(e.g., realuminize the primary mirror), the Director pre-empts the
planned observing schedule with at least two weeks advance warning.
This practice stems from the difficulties of planning large
engineering shutdowns months in advance because of manpower
limitations and the uncertainties related to various R&D aspects of
the work and its preparation.
4.3 Telecommunications
APO has a dedicated T1 circuit to NMSU
enabling remote operation of the 3.5-meter telescope through the
Internet, nominally at about 10% of T1 data rates due to Internet
bottlenecks elsewhere. A modem backup system is in place. Because new
instruments with larger CCDs have begun to tax the existing network
systems, plans for increasing the data bandwidth to the site are in
progress. Many of the ARC institutions already have or have proposed
to acquire vBNS or Internet2 network capability.
4.4 Environmental monitoring
Clouds, wind, dust, condensing humidity,
and precipitation are routinely monitored because of their effect on
data quality and possible risk to the telescope. Real time
weather-sensing transducers and a cloud monitor are in place, and
their outputs are collected and served through a web page, showing
current as well as past conditions. Criteria for telescope closure
are posted and enforced unilaterally by the Observing Specialists.
The APO weather web page is shown in Figure 6. Figure 7 shows a
thermal infra-red all-sky image which is taken by a prototype a
10-micron scanner in routine use at the site.
Figure 6. Sample APO web
page for weather monitoring
Figure 7. Sample 10-micron
all-sky images from cloud scanner
Another environmental issue is light
pollution. In order to preserve the relatively dark skies at APO,
observatory management conducts an active program which promotes
dark-sky legislation and community education. Outdoor lighting
ordinances are in place in the local communities of Alamogordo and
Cloudcroft, and legislation or guidelines are in preparation for
Otero County, the state of New Mexico, and local federal facilities
such as Holloman Air Force Base.
5. UPGRADES AND
ENHANCEMENTS
5.1 Performance
Although scientifically capable, our early
experience with the telescope has uncovered various deficiencies
which prevent users from benefiting from the full performance and
efficiency capabilities inherent in the telescope design. Defective
enclosure wheels were replaced. The primary mirror pneumatic support
system was redesigned and implemented. Improved primary ventilation
hardware was installed. A larger, more sensitive guide camera was
integrated into the off-axis Nasmyth instrument port. Advanced means
of aligning the optics were devised and used to improve image
quality. Computer and network systems were maintained and upgraded to
give higher capacity and better reliability for operations. Extra
funds were earmarked to accomplish a substantial list of improvements
to the observatory, including the replacement of the secondary
mirror, enabling automatic tertiary mirror rotation and mirror
covers, improving instrument capabilities, modernization of
operational software, etc.
The resources for major enhancements to the
telescope, existing instruments, and the development of new
instruments is not wholly included in the operational budgets.
However, the current level of site engineering support is much higher
than the early operational concepts, partly because ongoing
maintenance, improvements, and enhancement projects require a
substantial engineering presence at the observatory. Improvements and
enhancement work is generally treated as special projects and
undertaken by one or more ARC affiliates, working with the site staff
and in concert with the priorities of the Director and ARC Board of
Governors. Significant input to this process is derived from the
community of users, using mechanisms given above.
5.2 Efficiency
As the intrinsic performance of the
telescope continues to improve, the efficiency of obtaining
observations becomes increasingly important to the users. Planned
upgrades which will improve observing efficiency include:
- tertiary mirror automation-enables
faster instrument changes during the night;
- autofocus with guider-improves
throughput and eliminates need for periodic
refocus;
- "AI" and reactive scheduling-better
optimization of telescope use to available time and
conditions;
- telescope engineering monitoring
system-telescope problem diagnostics and efficiency
metrics;
- network upgrades-provide better
latency and bandwidth for remote observing and large format
CCDs;
- queue-remote observing-better
matches observing conditions to observing
requirements;
- service observing-enables science
observing without presence of principal
investigators;
- adaptive optics-more flux per pixel
on faint objects (needs laser beacon), plus higher resolution;
and
- new instruments-greater throughput
and science capabilities, lower noise detectors,
etc.
5.3 Future directions
As the APO 3.5-meter telescope approaches
its second half-decade of operation, the early commissioning problems
are largely understood and either fixed or in process of being fixed.
The telescope and instruments routinely provide the user community
with useful and often exciting data. Looking forward, the ARC
community has begun discussion on this question: is the 3.5-meter to
become an excellent general-purpose visible/IR telescope, with
current technology instruments, or do we concentrate limited
resources toward development of a unique scientific capability.
Perhaps a balance between these two paths is possible. In the coming
decades, telescopes of this size with large-field capabilities, high
resolution imaging, state-of-the art CCD imagers and spectrometers,
flexible scheduling, remote operations, and easy access for users,
will undoubtedly play a significant part toward future advancements
in astronomy.
6.
ACKNOWLEDGMENTS
My appreciation to Ed Turner and Kurt
Anderson for reviewing the manuscript, and to Dan Long for furnishing
the photographs and graphics.
7. REFERENCES
- B. Gillespie, R.F. Loewenstein, and D.
York, "Remote Observing at Apache Point", New Observing Modes
for the Next Century, ASP Conference Series, 87, pp.
97-108, 1996.
Further author information -
Email: gillespi/apo.nmsu.edu;
WWW: http://www.apo.nmsu.edu;
Telephone: 505-437-6822; Fax: 505-434-5555