A Strategy for Ground-Based Optical and Infrared Astronomy

NOAO's mission is to provide national access to the sky by means of excellent optical and infrared observational facilities on outstanding sites in both hemispheres. Use of the facilities is determined by open peer-reviewed competition among the best scientific ideas from the entire astronomical community. The multiwavelength community has been very well served by the competitive access to NOAO facilities, and that must continue.

(Here, and throughout this report, the panel refers to NOAO's priorities and strategy for ground-based nighttime OIR astronomy. It does not discuss priorities for solar astronomy.)

Approximately half of the nation's active research astronomers have access to major telescopes at independent observatories, and approximately half are affiliated with institutions that lack such facilities. Many of the former group count on NOAO to provide observing options not available at their own observatories. For these astronomers, CTIO is especially valuable. The latter group includes many astronomers actively engaged in research at other wavelength bands (e.g., radio astronomy, space astronomy).

Why has NOAO been a success? NOAO has (1) built major telescopes in good sites, (2) equipped them with good instruments, and (3) provided excellent service to astronomers. NOAO has built up teams of skilled engineers and scientists in the areas of optical and infrared detectors, controllers, cryogenics, optical fibers, and data analysis software. The smaller telescopes have been equipped, not with a fleet of instruments, but with dedicated-purpose ones, particularly CCD imagers. These well-instrumented small telescopes have played a major role in the successful science conducted by NOAO users.

Despite these successes, NOAO has lagged in the construction of new telescopes. The 3.5- meter WIYN telescope is the first new facility in two decades. In the past, NOAO has tried to satisfy the entire user community with an extremely broad mix of user services and instrumentation on most of its telescopes. This effort has sapped resources that would have been better focused on the construction of more technically advanced facilities.

NOAO provides a level of service not found at the independent observatories. A substantial share of NSF dollars going toward NOAO is appropriate since it is the only observatory open to all astronomers. The question of balance between NOAO and the independent observatories is one that has been raised repeatedly and that is addressed below.

In addition to providing national access to telescopes, NOAO aims to provide leadership in the development and operations of major new telescopes, in developing instrumentation and software, and in scientific research. NOAO has had substantial success in each of these areas. However, the fact remains that NOAO does not now and will not be able to maintain pre-eminence in all aspects of OIR astronomy. To achieve leadership in a constrained budget environment, NOAO will have to make hard choices in distributing its resources. The panel suggests the following guiding principle: NOAO should concentrate its resources in those areas where it has the best chances to assert scientific leadership. It follows that NOAO will not be able to maintain preeminence in all aspects of OIR astronomy, nor should it try to do so.

The twin 8-meter Gemini telescopes are being built to permit the national scientific community competitive access to two of the world's largest telescopes in two of the world's best sites. NOAO should play the key role in determining how the United States interacts with the Gemini telescopes, instrumentation choices, and scientific support of the two telescopes. Gemini's science, instrumentation, and operations must be NOAO's highest priorities.

It will be a challenge for NOAO management to realign its observatories to accommodate the needs of the astronomy community in the Gemini era. These needs include community access to state-of-the-art instruments and high-performance 2- to 4-meter OIR telescopes, as well as the Gemini telescopes. This renewal of the NOAO infrastructure is crucial to the future of U.S. astronomy and was the highest-priority recommendation of the AASC report. However, to ensure effective operations and access by a wide community to these new facilities during a period of flat or declining budgets, major components of the present NOAO would by necessity vanish. Renewal of facilities must lead to a decrease of long-term operating costs so that more science can be supported within a fixed budget.

This realignment will be painful, for a different mix of talents and projects will be needed. Planning such a realignment must include (1) rethinking the role of the national observatory and (2) restructuring to optimize the effectiveness of that new role. The panel believes that reorganization will be required and that elements of that reorganization might include the following:

• Enabling scientific programs that require both Gemini and smaller telescopes, and that need to be conducted in either or both hemispheres. All stages of a project would be included, from possible pre-Gemini surveys to follow-up post-Gemini observations. The instrumentation on the smaller CTIO and KPNO telescopes and the national time allocation committee procedures may require modifications.

• As NOAO focuses its attention on fewer tasks of high priority, it should evolve to a leaner and more focused organization, with fewer employees and a different mix than at present, a smaller staff in Tucson, and a smaller core of tenured scientists.

• More open two-way links with the community--from contributed software and hardware to active participation in all phases of the observatory--would help spread the burden and the responsibility.

• New instruments might often be developed through competitive selection among NOAO itself and other optimal groups or suppliers, who would then collaborate closely with NOAO. A mechanism to stimulate this process could be an annual U.S. workshop for astronomical instrumentation.

• A reduced selection of instruments, with more instruments permanently installed on each telescope, should require fewer personnel for operations and maintenance and result in lower costs.

• Some observational programs would be done more efficiently by remote, queue, and service observing than by hands-on operation of the telescopes.

• NOAO site directors should have as much authority as possible to operate their sites in a scientifically cost-effective manner.

• Older telescopes of all sizes that are expensive and/or inefficient to operate should be retired.

To save costs, the international Gemini project intends to utilize as much of the existing CTIO infrastructure as possible for Gemini South and as much of the Joint Astronomy Center (JAC) infrastructure in Hilo as possible for Gemini North. The current plan is for IGP to buy services, in cash, from CTIO and JAC. NOAO will have a minimal presence on Hawaii, at least in the initial stages of Gemini operations.

While IGP has the responsibility to build, operate, maintain, and upgrade the Gemini telescopes and their instruments, it will not support Gemini science. Each participating nation is expected to provide the scientific, technical, and administrative infrastructure required for its astronomers to use the telescopes. For the United States, that role must be filled by the U.S. Gemini Project Office (USGPO), a division of NOAO located at Tucson. For U.S. users of Gemini, the USGPO will manage the national time allocation committee and telescope scheduling and will provide scientific support to astronomers, including advice concerning instruments, data, observing requirements, and access to archives.

The panel anticipates that a substantial fraction of Gemini science may be carried out most efficiently through queue observing and, possibly, by remote observing. Therefore, the USGPO must be prepared to support hardware and software interfaces for U.S. users to wideband telecommunication links with Gemini North and South. As is the case with software development for data analysis (see the subsection "Instrumentation" below), the USGPO effort to develop remote observing capability should be part of a national effort, taking maximum advantage of expertise outside of NOAO.

According to current estimates, the IGP will require approximately $11 M per year to operate and manage its telescopes, including $3 M per year for instruments and major telescope upgrades. The NSF will therefore be obliged to pay approximately $5.5 M per year to the IGP for its 50% share. The panel examined these costs and has found them reasonable. They do not, however, include the costs for NOAO to support the U.S. interface to the IGP, as described above. NOAO estimates that the latter activities will cost approximately $2.5 M per year and plans to absorb those costs by reallocating resources within its present budget. The panel endorses this plan. However, as discussed below, NOAO cannot absorb the additional $5.5 M per year charge without making severe cutbacks in its present operations, including a major downsizing of the Tucson operations, and probably the closing or privatizing of most of its older telescopes on KPNO.

After the direct support of the Gemini telescopes, the second priority of NOAO must be the support of moderate (2- to 4-meter-class) telescopes with the best possible capabilities. NOAO needs a variety of such telescopes to (1) support the Gemini scientific programs and instrument development, (2) provide other unique national capabilities, and (3) support the scientific programs of the best researchers and students throughout the nation. Items (2) and (3) are chief among NOAO's current activities, and the need to support them will not diminish with the coming of Gemini.

The new-technology 3.5-meter WIYN telescope at KPNO is an excellent example of a modest-class telescope. WIYN has already achieved an image quality better than 0.5''. It will complement Gemini North in the intermediate field of view, high-resolution imaging regime; it will provide wide-field multiobject spectroscopic capabilities; and it will provide access to the near ultraviolet (UV). The latter capability will be important because the majority of the faintest and most distant objects are in fact UV-bright.

The WIYN experience is a very promising model for NOAO. Not only does WIYN yield much better image quality than any other KPNO telescope, but it also requires roughly only half as many FTEs (7 versus 16) to maintain, compared to the KPNO and CTIO 4-meter telescopes. Replication of the WIYN telescope is estimated to cost approximately $12 M and would pay for itself, in terms of reduced maintenance cost relative to the current 4-meter telescopes on KPNO and CTIO, in less than 20 years.

At CTIO, the current 4-meter telescope can provide wide-field imaging and some spectroscopic capability, but a new-technology 4-meter-class telescope is very much needed. Given the outstanding conditions available at Cerro Pachon, where the median seeing is roughly 0.4'', a telescope with superb imaging capabilities would be exceptionally productive scientifically. The current 4-meter telescopes, built over 20 years ago with old technology, likely cannot be upgraded to better than 0.5'' optics. At CTIO, a new-technology telescope would complement or replace the existing 4- meter telescope, just as WIYN complements the KPNO 4-meter.

In the Gemini era, CTIO will have the responsibility for supporting operations at Gemini South, as well as the support of existing CTIO telescopes. CTIO, with or without Gemini, is the only access to the Southern Hemisphere skies for the vast majority of U.S. astronomers, and its smaller telescopes should be kept open until they can be replaced in a cost-effective manner. As discussed above, an enormous amount of valuable science can, and should, be done on moderate-size telescopes. Closing the smaller telescopes on CTIO, particularly the 1.5-meter and the Schmidt telescopes, should be done only as a last resort.

In order for CTIO staff to fulfill its obligations, the observatory should maintain an adequate engineering staff. Because of its remoteness, CTIO needs to be more self-sufficient than observatories in the Northern Hemisphere. The panel thus questions the wisdom of across-the-board cuts of the various components of NOAO, since these cuts have forced CTIO to gradually reduce its engineering staff to the point that it can no longer build facility instruments and has had to struggle to develop a CCD controller (ARCON). With subcritical staffing, it will be impossible for CTIO to maintain its current instruments, let alone even assist in the development of further instruments.

As Gemini comes on line, NOAO will need to reduce the operations costs of KPNO. This cannot be accomplished simply by closing the smaller telescopes. The costs of the entire infrastructure of Kitt Peak must be reduced.

At present, the distribution of observing runs at the NOAO 4-meter telescopes is sharply peaked at three nights. Such a distribution function may be the one that maximizes the number of astronomers who use the telescopes in a hands-on fashion. However, it is not the distribution function that will maximize the scientific productivity of the telescopes. Short runs of instruments and exclusively short observing runs increase costs for supporting observers, instruments, and telescopes. Dedicated instruments on telescopes and key program collaborations will reduce operations, instrument maintenance, and travel costs. The panel recommends a broader distribution of observing run lengths, ranging from longer runs to observations that are much shorter in time but have better frequency coverage, such as an hour a night for several days or weeks.

KPNO should strive to provide scientific access to its telescopes through queue-scheduled and remote observing for observations requiring short allocations of telescope time, and should restrict hands-on use of the telescopes mainly to longer observing runs. The experience with the ARC telescope is worth watching; it may demonstrate that remote observing through wideband telecommunications links can be very efficient. Since NOAO must develop the technical infrastructure to support remote observing at both Gemini telescopes, it should be able to provide similar capabilities for the telescopes at KPNO.

The WIYN experience has shown that Kitt Peak can deliver excellent seeing, and KPNO should strive to support programs that take advantage of this capability to provide scientific capabilities complementary to observations by Gemini North. Toward this end, it may be wise for KPNO to close its smaller telescopes, such as the 2.1-meter, the 1.3-meter, and the 0.9- meter telescopes, especially if they can be replaced by a modern 2-meter-class telescope. If in fact the operations costs of WIYN are as low as they have been estimated to be, it would be sensible to consider replacing the existing 4- meter telescope by a twin of WIYN.

However, the fact remains that KPNO does not provide unique access to the northern sky for the majority of U.S. astronomers. In a severely constrained budget, keeping KPNO open must be given lower priority than maintaining CTIO. Of the existing telescopes on KPNO, WIYN is clearly the highest priority. If budgets force a cutback of operations on KPNO, some of KPNO's current users will lose their access to telescopes.

The panel identified the following guiding principles:

• New-technology telescopes and their instruments are increasingly interdependent. Special-purpose telescopes with dedicated instruments are highly efficient.

• Time trading with non-NOAO telescopes, leading to less duplication of instruments, will save costs and enable a more efficient distribution of run lengths.

• A core group of engineers and instrument scientists must be retained near each NOAO site. This staffing is necessary independent of how the facility instruments are acquired.

• Facility instrument development should be science-driven, rather than engineering-opportunity-driven.

• It is healthy for the field to support a wide range of instrumentation (and observing) styles, from experimental to user-friendly. NOAO should incorporate the best ideas and technologies in its facility instruments, whether built in-house or externally.

• Facility instrument development might proceed via cooperative agreements that guarantee some telescope time. Initial science operations could involve key programs open to the community.

Applications of large-array technology to OIR instrumentation continue to generate new instruments with corresponding science opportunities. Fundamental advances can be made with innovative instruments on telescopes of all apertures. For the first time in the history of astronomy, nearly all the photons in the focal plane will be effectively used, and even modest-size telescopes, properly instrumented, can make important contributions.

A common problem for both national and private observatories is large detector array development and supply. The special requirements of OIR astronomy dictate detector specifications vastly different from the specifications for non-astronomical uses. Opportunities for collaboration among observatories, instrument builders, and space astronomy missions should be exploited. Detector and readout technology sharing and transfer are already in place in OIR astronomy, and NOAO has played a major role in this process.

Some of the greatest scientific gains in OIR astronomy will come from achieving near-diffraction-limited image quality, using single telescopes and adaptive optics and distributed arrays. Therefore, NOAO should look for opportunities to purchase such technologies for their telescopes as they become available.

What is the best way to develop innovative and effective OIR instrumentation? The process works best when directed by a scientist with a strong motivation to use the instrument to do his or her own science. Access to telescopes for testing is necessary. Innovative instruments do not usually come into being because of a diffusely perceived need. Rather, innovative instruments are most often developed to address a particular problem in science. The motivation comes from individuals with the freedom to design and build leading-edge and experimental instruments.

The best, most innovative, and most productive instruments should be supported regardless of origin. NOAO should seek opportunities to leverage NSF support with nonfederal funding to provide facility-class instruments for its telescopes. Groups would receive guaranteed time in addition to partial funding in exchange for delivery of a facility-class instrument that would become available to the community. Such arrangements should be regarded as collaborations between NOAO instrument scientists and engineers and those outside NOAO, rather than as subcontracts. Without an in-house champion, no instrument will succeed. Therefore, it is important that NOAO staff, together with the user community, maintain a strong say in what instruments are "right" for NOAO telescopes.

Such collaborations are currently under way, but the community may not recognize them as such because they have come from individual contacts rather than a community-wide announcement. NOAO should actively encourage any sort of proposal to provide instruments and should inform the community of its intent to do so.

Ranking high among the many benefits of such arrangements would be new opportunities for involving graduate students in instrument development. One example of a model for future instrument development is the Fabry-Perot instrument built at Rutgers University and used extensively at CTIO. It is important that the universities maintain instrument development capabilities, since the universities are where graduate students are trained.

NOAO should take advantage of the opportunity to tap a much larger pool of experienced instrument builders across the nation. There are a number of physics, astronomy, and space science research laboratories well equipped and experienced in sophisticated astronomical instrumentation. By inviting these institutions to collaborate in major instrument developments, NOAO can ensure that each project has a focused, dedicated team of scientists and engineers, and will be able to provide leadership in instrument development in a constrained budget environment.

The panel is concerned that NOAO Tucson operations may be too large and ineffectively utilized, and may have the wrong mix of personnel. The panel examined the NOAO Engineering and Technical Support Division and found that the number of engineering projects currently exceeds the number of instrument scientists, creating a pileup of projects for certain staff. For the FY 1993 to FY 1994 period there appeared to be an imbalance between the number of optical and infrared projects, and there was no clear user pressure for some projects. The panel found no consistent records of true project costs and personnel utilization within the Tucson office of NOAO; this was particularly true of KPNO and the Central Services at NOAO headquarters.

More rigorous project management tools should be used to track costs and schedules of NOAO departments. The panel recommends that a reorganized NOAO make use of focused teams of scientists and engineers to work on a given project from conception to completion. (The panel found examples of this team approach in two new autonomous teams: the GONG group of NOAO and the engineering group of the international Gemini project.) Focused teams will be particularly useful in collaborative instrumentation projects and should further improve the accounting of project costs. It would be helpful to identify a "customer" for each new instrument before development.

Finally, the panel found evidence for a wide range of motivation among the service, engineering, and scientific staff. The newer staff appeared overworked (very common in national laboratories in this transition period). Without reorganization, these problems will only become worse in the Gemini era.

The most successful cases of instrumentation development at NOAO can be traced to good teamwork. Examples are the teams that developed the Hydra multifiber spectrograph and the infrared cameras. The IGP engineering group operates very effectively in this way. NOAO might do well to emulate the IGP's most successful teams in all the NOAO engineering programs.

NOAO should consider contract engineering firms as an alternative source of engineering support to replace a fraction of its present engineering and technical staff. Supplemental engineering talent could be brought in as needed for Gemini instruments, for example, as those instruments will be bid in an open competition and will bring their own funding.

In any case, an engineering and scientific core must exist within NOAO to, at a minimum, sustain the telescopes, the control systems, and instruments, and to help set specifications and see that they are met for facility-class instruments. Access to engineering time is crucial, whether the instrument is built inside or outside NOAO.

NOAO should concentrate resources for in-house instrument development to build on its current strengths, with a focus on detectors, controllers, and fibers. Telescopes need large formats in the optical and infrared, especially given the clear needs for wide-field imaging. NOAO should play a role in Gemini instrumentation development.

Both KPNO and CTIO should, whenever appropriate, build their instruments in collaboration with outside groups.

Looking toward the future, and to maximize efficiency, NOAO should actively explore time trading and dedicated facility instrument collaborations with private observatories that have new-technology telescopes. In the best scenario, time trading could result in a net savings for the NSF, better science, and reduced operations and maintenance costs.

NOAO has performed an extremely important service in the development and maintenance of the IRAF image data analysis software system, which has become the most widely used international standard for astronomical data analysis. However, IRAF was written in a fashion that makes it difficult for outside groups to contribute original code; the result is a product that is too dependent on the programming staff in Tucson. The IRAF development did not take full advantage of the very considerable software expertise outside Tucson. The panel encourages NOAO to consider the development of the next generation of data analysis software, but this time to develop a more open system with stronger community participation in the project.

Finally, in order to ensure that astronomers who win time on NOAO telescopes have a minimum level of support necessary to carry out their proposed science, the panel recommends that NSF give NOAO the responsibility and the necessary funds to support travel, lodging, and publication costs of observers who win time at NOAO facilities but lack other sources of support.

The panel repeats its main recommendations for the future role of NOAO as the Gemini era approaches. These recommendations are appropriate no matter what the future budgets may be.

• U.S. user interface
• Technical support for observing
• Liaison with IGP
• Performance optimization of Gemini telescopes
• Support for Gemini instrumentation development

• Support for visitors at CTIO telescopes and Gemini South
• Performance optimization and operation of Gemini South and telescopes on La Serena and Cerro Pachon
• Development of some instruments
• Development of a new-technology 3- to 4- meter- class telescope

• Support of visitors at KPNO telescopes
• Performance optimization and operation of several telescopes, especially WIYN
• Development of some instruments

• NOAO administrative headquarters
• Community user interface and service, including wide-band links
• Coordination of north-south and Gemini-KPNO-CTIO observing programs
• Targeted facility instrument production

There may be a way, however, for all U.S. astronomers to retain access to a broad spectrum of observing options even as NOAO becomes more narrowly focused. For a possible means to achieve this, we turn to the independent observatories, discussed in the following section.

The National Academies Home Page | NAP Home Page | Report Home Page