V A

NATIONAL

OPTICAL

ASTRONOMY

OBSERVATORIES

NATIONAL OPTICAL ASTRONOMY OBSERVATORIES

LONG RANGE PLAN

FY 1998 - FY 2002

June 10,1997

LONG RANGE PLAN

TABLE OF CONTENTS

I. EXECUTIVE SUMMARY 1

II. NIGHTTIME ASTRONOMY: THE FACILITIES 2

A. Overview 2

1. The Telescopes 32. Observatories or Observatory? 4

B. The Transition Plan 5

1. SOAR 5

2. WTYN 5

3. Access to the Independent Observatories 64. Gemini 7

5. 2.4-m Telescopes 86. Telescope Closures 87. Management Structure 9

m. NIGHTTIME ASTRONOMY: SCIENTIFIC DIRECTIONS 10

A. The Large-Scale Structure of the Universe 11B. The Formation and Evolution of Galaxies 14

1. Galaxies in Clusters 14

2. Field Galaxies 16

C. The Structure of Nearby Galaxies 17D. Star Formation 18

E. Stellar Populations 20F. The Chemical Evolution of the Galaxy from High-Resolution Spectroscopy 21

IV. NIGHTTIME ASTRONOMY: THE PROGRAM 23

A. NOAO Nighttime Instrumentation 231. Scientific Planning for Instrumentation Development 242. Infrared Instrumentation 26

3. Optical/UV Instrumentation 29B. Telescopes and Upgrades 30

1. Initiation of Adaptive Optics at Kitt Peak 302. KPNO Telescope Improvements 313. Plans for Observing Facilities at CTIO 35

C. Center for High Angular Resolution (CHARA) 39D. US Gemini Program and NOAO Science Operations 40

1. Roles of the USGP 40

2. Gemini Instruments: Construction Phase 42

3. Gemini Instruments: Operations Phase 434. Gemini Instruments: Procurement Plan 44

5. National Access to Gemini and Other Telescopes 45

V. SOLAR ASTRONOMY 47

A. NOAO Solar Strategic Plan 47

B. Science at NSO 53

1. Understanding the Solar Activity Cycle 532. Understanding the Coupling of the Interior and the Surface: Irradiance Variations 573. Understanding the Coupling of the Surface and the Envelope: Transient Events 614. Exploring the Unknown 64

C. Initiatives for Solar Physics 661. NSO's Future Directions Plan 67

2. Image Quality Improvement Program for NSO Telescopes 693. Radiative Inputs of the Sun to Earth (RISE/PSPT) 704. Planning the Flagship Solar Telescope(s) 70

D. Instrumentation 71

1. Active and Adaptive Optics 732. Mark II Correlation Tracker 75

3. Near Infrared Magnetograph 2 754. CCD Cameras 75

5. Infrared Cameras 76

E. Facilities/Infrastructure Upgrades 771. KPVT Optics Upgrade 772. McMath-Pierce Telescope FTS Polarimeter Upgrade 773. Evans Solar Facility Instrument Upgrades 774. KPVT Telescope Control System Upgrade 785. McMath-Pierce Telescope Control System Upgrade 78

VI. NOAO SUPPORTING SERVICE 79

A. Computer Support 791. Save-theBits 79

2. Downtown Tucson Computer Support 803. IRAF 80

4. KPNO Computer Support 815. CTIO Computer Support 826. NSO/SP Computer Support 847. NSO/Tucson Computer Support 85

B. Facilities 86

1. CTIO 87

2. KittPeak 89

3. NOAO Tucson 92

4. NSO 95

VII.PUBLIC INFORMATION AND EDUCATION 100

A. General Trends 100

B. Visitor Center Programs 101C. Educational Outreach Programs 102D. Undergraduate and Graduate Education 103

VTH. BUDGET 104

A. Base Budget 104B. Funded Project 106C. Proposed Initiatives 106

Appendix A: The Albuquerque ProcessAppendix B: History of the GONG ProjectAppendix C: Budget Summary

ii

I. EXECUTIVE SUMMARY

This long range plan for NOAO spans the time interval 1998-2002 and defines the program for acritical phase in the evolution of NOAO. The nighttime program will complete the transition froman earlier generation of telescopes and instruments to an almost completely new suite of facilities.During this same time interval, the solar program will be redefined, and what is at stake is nothingless than the scope and future of groundbased solar physics in this country.

Gemini South is currently scheduled to begin scientific operations in 2002. This milestone willmark the completion of an effort to replace essentially the entire complement of telescopes offeredby NOAO to the community with the exception of the Blanco and Mayall 4-m telescopes, whichwill remain in operation. The construction that remains to be completed are the 4-m SOARtelescope, for which it now appears that funding is available from Brazil, the University of NorthCarolina, and Michigan State University, and the 2.4-m wide-field O/TR imaging telescopes, whichwe believe are essential for supporting Gemini observations. Funding for these two telescopes hasnot yet been identified.

During this same period, we expect to complete several major instruments, including 8K x 8Kmosaic CCD imagers for both CTIO and KPNO; a multi-fiber spectrograph for the Blancotelescope, which will match the capability of the Hydra spectrograph already at WIYN; mediumresolution IR spectrometers for both CTIO and KPNO, with the CTIO instrument being a clone ofthe Gemini IRS and the KPNO spectrograph being a simpler device with more limited capability;wide-field LR imagers for both sites, with one being a single channel device with a lKx IKdetector and the other being the multi-channel SQLTD, upgraded to 500 x 500 arrays; a single highresolution (R = 100,000) spectrometer to be shared by both sites and with Gemini; and possibly avery high throughput single object optical spectrometer. A high resolution IR imager will be sharedbetween CTIO and Gemini South during commissioning. We expect to provide a low-order natural-guide-star adaptive optics system at WIYN.

At the end of this time period, NOAO will be the only organization that offers observing time inboth hemispheres, on telescopes with a range of apertures, and with a broad complement of infraredand optical instrumentation. In order to take advantage of this uniqueness, we must learn how to usethis suite of telescopes in an optimum fashion. How do we match complex observing programs totelescopes in order to maximize the scientific output? How do we inform the community mosteffectively about options for their programs? How do we best support programs that require the useof more than one facility? What kinds of supporting infrastructure are required in order to use thetime on Gemini and the large telescopes at the independent observatories most effectively? Thisproposal outlines some steps that we propose to take to develop answers to these questions and toprovide user support that is unified across all of the facilities accessed through NOAO.

The changes in the nighttime offerings of NOAO, including access to Gemini and some of theindependent observatories through the NSF instrumentation program, represent a qualitative changein what we offer the community. There will be fewer (by about a factor of 2) individual nightsavailable, but the throughput of photons per unit time will in some cases be literally orders ofmagnitude higher than it was two decades earlier. It is almost certainly true that we will supportfewer users than we have in the past, but this higher throughput will enable new types of science,including much larger sample sizes for a range of programs, from the study of stellar activity levelsin nearby open clusters to the analysis of the dynamics of distant clusters of galaxies. The use ofnew observing modes, such as queue scheduling, can potentially increase efficiencies still furtherand enable studies, that have up until now been impossible, of variable objects and targets of

opportunity. During the period covered by this long range plan, we will re-examine the way weschedule the telescopes and experiment with a variety of modes of observing (queue, remote, andservice, in addition to conventional observing with the astronomer present) so that we canunderstand what strategies will best support the diverse science proposed by the large communityserved by NOAO.

NOAO continues to look beyond this five year interval in its long range planning. The crucialquestion is whether major gains in sensitivity and angular resolution beyond what is provided bythe 8- to 10-m class telescopes linked to satellite telescopes for interferometry can be achieved mostcost effectively in space or on the ground. We are working with a variety of groups andcommittees, including most notably ACCORD (the council of directors of major US observatories),to evaluate this issue. We also continue to work on the development of interferometry techniquesthrough our participation in CHARA, which is building an interferometry on Mt. Wilson.

The issues that will be addressed by NSO during the next five years are very different from theissues confronting the nighttime program. It has been nearly forty years since a large-aperture solartelescope was built by NSO. In that same period of time, the nighttime program has built twogenerations of facilities: the Mayall and Blanco telescopes, followed by new technology telescopesin the form of WIYN, SOAR, and Gemini. The GONG project is an excellent example of how newtechniques can yield fundamental advances in our understanding of solar structure. Application ofmodern technologies to the construction of a large-aperture solar telescope for studies of the solardisk and atmosphere can be expected to lead to equally profound advances in our understanding ofthe nature, variability, and origin of solar activity.

During the next five years, the central effort in the solar program will be to define the scientificrequirements for a large-aperture telescope, establish the technical feasibility of building atelescope that will meet those requirements, identify the best possible site, and prepare and submit aproposal for such a telescope to the NSF.

The remainder of the NSO program is well defined and includes the continuation of GONG for atleast one 11-year solar cycle; the preparation of a proposal to upgrade the spatial resolution of theGONG cameras by a factor of 4; the construction by the time of the next solar maximum of SOLIS,which is a suite of instruments designed to support Synoptic Optical Long-term Investigations ofthe Sun; completion of the construction and the operation of the PSPT (Precision Solar PhotometricTelescopes) to study irradiance variations; demonstration and use of a low-order adaptive opticssystem at the NSO/SP Vacuum Tower Telescope; and the continuation of the infrared program atthe McMath-Pierce Telescope with upgraded cameras based on the IK x IK InSb arrays.

II. NIGHTTIME ASTRONOMY: THE FACILITIES

A. Overview

This long range plan covers the time period when both Gemini telescopes will begin scienceoperations, access to the independent observatories will be offered to the community, and theconstruction of SOAR will be completed. We also still hope to initiate the construction of twomoderate aperture telescopes for wide-field imaging, especially in the near-infrared part of thespectrum. This section gives an overview of the program changes and issues that are associatedwith the replacement of nearly all of the facilities offered through NOAO to the usercommunity. Subsequent sections of the long range plan provide information about the detailedimplementation plans.

1. The Telescopes

The following tables compare the nighttime facilities available in each hemisphere at thetime that NOAO was established and those that we expect to be operational in 2002, at theend of the five-year period of time covered by this long range plan. The number inparentheses indicates the fraction of time that will be available to the NOAO community.

Capabilities - North1984

4-m Mayall2.1-m

1.3-m

Coude Feed

0.9-m

0.9-m

Burrell-Schmidt (1/4)

No. of Nights: 2281

2000

8-m Gemini (0.425)4-m Mayall3.5-m WIYN (0.4)2.4-m

9-m HET (0.07)6.5-m MMT (0.07)

No. of Nights: 1082No. of nights x (D)z: 9056 No. of nights x (D)z: 22315

Capabilities - South1984

4-m Blanco

1.5-m

l.O-m(Yale)0.9-m

Curtis-Schmidt (Mich.)

No. of Nights: 1460 (est.)No. of nights x (D)2: 7264

2000

8-m Gemini (0.45)4-m Blanco

4-m SOAR (0.3)2.4-m

No. of Nights: 1004No. ofnights x (D)2: 18925

D = Diameter

These tables highlight the major issue confronting nighttime astronomy: the number oftelescopes being operated is decreasing but the aperture is increasing. More specifically,the number of nights available to be assigned to the user community will decrease by abouta factor of two in the north and somewhat less in the south. However, the number of

photons collected will increase by about a factor of 2.5. If we allow for the increase inthroughput of the instruments (multi-object fiber spectroscopy, mosaics of CCDs, largeformat infrared detectors), the throughput of the telescopes will be literally orders ofmagnitude higher in 2002 than it was two decades earlier. This change in capability shouldnot be taken to imply that we can support more programs and more astronomers with fewertelescopes. Just as more powerful computers enable researchers to undertake more complexcalculations, so too astronomers are taking advantage of improved telescope throughput toincrease sample sizes, spectral resolution, angular sky coverage, etc. Nevertheless, thisqualitative change in the types of observations that are possible does require a reexamination of how we should schedule this powerful suite of telescopes in order to usethem most effectively. What innovative approaches to operation will be required?

2. Observatories or Observatory?

NOAO is the only observatory in the world that offers open, competitive access to abalanced suite of telescopes with a range of apertures and instrumentation in bothhemispheres. If the limited number of facilities we operate are to make a disproportionatecontribution to advances in astronomy, we must take maximum advantage of what wealone can offer. That means we must emphasize a systems approach to schedulingscientific programs. Observers must have the opportunity to optimize their observingstrategies by being able to submit proposals for entire programs that make use of a varietyof telescopes in a variety of locations with optimum observing run lengths on each.

What this means conceptually is that there should be "one-stop shopping" for observingtime. There should be a single interface to the user community—one source of informationabout observing capabilities, one form for observing proposals, and one deadline forproposal submission. Observers should not have to be concerned (at least until, or if, theyhave to buy plane tickets) where the telescopes are located or who operates them. In otherwords, we should become the National Optical Astronomy Observatory—not severaldifferent National Optical Astronomy Observatories offering independent anduncoordinated observing facilities.

This concept is more easily stated than realized. And realizing it will affect every aspect ofthe way we interact with the user community. The diversity of our offerings will be solarge that we will need to re-invent the way we provide information to the community.Now people can request manuals for particular instruments on particular telescopesbecause they already have a good understanding of the best place for a particular type ofobservation. In the future, we may have to construct a "smart system" for helping others todecide where observations should be done; observers should be able to enter a Web pagewith information about what they want to observe (resolution, limiting magnitude, S/N,etc.) and be directed to options for where they might make those observations. We willhave to offer simulators for complex instruments so that observers can optimize observingstrategies in advance. We will have to take advantage of queue scheduling for at leastthose types of observations that require observing conditions significantly better than themedian. In many cases, people may not require whole nights of observing but perhaps onlya few hours of imaging on a 2.4-m class telescope to select objects, several nights forspectroscopy at a 4-m telescope to characterize the selected sample, and a few hours onone of the Gemini telescopes for detailed spectroscopy of the faintest and/or mostinteresting members of the class.

If we view the role of the national observatory as delivering data rather than nights, as isimplicit in some of the alternative approaches to scheduling, then additional questionsarise. What level of investment should we make in archiving? With larger format detectorsand higher throughput, it is more likely than 20 years ago that the utility of a dataset willnot be completely exhausted by the observer for whom it was obtained. The value of theHST archive is already apparent. Gemini does have plans for putting data into an archive;what investment should NOAO make to ensure that the archive is accessible? And what

data from our own telescopes should be easily accessed?

And what about providing datasets and surveys? For example, Gemini can do IRspectroscopy at least five magnitudes fainter than the limiting magnitude of the 2MASSsurvey. Similarly, optical spectroscopy will be possible to faintness limits well beyond the

deepest large area sky surveys currently planned. Should NOAO undertake very deepsurveys over limited parts of the sky and make the data available to the entire community?

It is likely that the efficient use of Gemini will require not only deep surveys but otherpreparatory observations. Should observing time for these preparatory measurements begranted automatically to successful proposers for Gemini time? How much time will berequired for this purpose and what will be the impact of such a policy on other types ofscience that may not depend on the Gemini telescopes?

B. The Transition Plan

The astute reader will notice that most of the telescopes listed as available in 2002 are not yetin operation. And several others now in operation will not be supported by NOAO after the turnof the century. How do we plan to get from here to there? In the following sections, we discussfirst the plans for telescopes operated directly by NOAO, then the support for US participationin the Gemini project, and finally the implementation of access to the independentobservatories.

1. SOAR

The primary initiative within NOAO is the SOAR project, which has as partners CNPq(Brazil), Michigan State University, the University of North Carolina, and NOAO.Observing time will be allocated in proportion to the financial contributions of the partners.The goal is to have these contributions to construction, commissioning, instrumentation,and operations be such that NOAO and Brazil will each receive 30 percent of the observingtime, UNC and MSU will receive 15 percent each, and the remaining 10 percent will beallocated to Chilean astronomers in accord with the agreement under which AURAoperates in Chile. Letters of intent for sufficient funds to complete the SOAR project havebeen signed by the persons in each participating organization who are authorized to commitfunds.

Since August, the project has been conducting preliminary design and costing studies. Thiseffort is being led by Gerald Cecil from UNC, who has been resident in Tucson. Severalinnovative designs have been evaluated, with special emphasis on quantifying the gains tobe achieved from an off-axis low-scattered-light design. Cost estimates have been obtainedfrom vendors for the major components, and project and payment schedules are being puttogether. An interim board has been established with one member from each institution,and Sidney Wolff is serving as chair. The goal for the science working group is to convergeon the science requirements this spring, establish with high probability that a telescope canbe built within a budget that meets these requirements, and obtain approval from theinterim board to proceed with the full project. Tom Sebring, who was the project managerfor the Hobby-Eberly Telescope, has been selected to be project manager for SOAR.

2. WIYN

WIYN has now completed 18 months of successful science operations. The mediandelivered image quality throughout that time has been maintained at 0.8 arcseconds.However, there are a number of issues that must be addressed in order to obtain maximumproductivity from this telescope.

In the south, the SOAR and the Blanco 4-m telescopes are being treated from the outset as atelescope system with complementary and non-duplicated capabilities. Partners will tradeobserving time in order to obtain access to the full range of capabilities for theircommunities. It has become clear that the WIYN partners would like access to observingcapabilities that are already available at the Mayall 4-m telescope, including most notablyinfrared instrumentation and long-slit spectroscopy, while NOAO has very little interest induplicating capabilities it already has. The obvious solution is time trades. In order to makethis possible, we must find ways to assure the partners that, if such trades are allowed, theirproposals for NOAO telescopes will not be downgraded by the TAC on the grounds thatthey already have access through time trades.

A second issue for the WIYN consortium is to improve the operation. During the first twoyears of science operations, a number of concerns have been identified that relate to thesafety of the equipment during operation (hardware and software safety interlocks, etc.) andto the efficiency of observing (for example, automation of the procedure for obtainingwavefronts). The partners have pledged $400,000 over a two year period to makeimprovements.

A third issue for the consortium is to find a mechanism for funding new instrumentationand for building it. An ongoing instrumentation program was not funded as part of theWIYN agreement. Comparison with other large telescopes, such as CFHT, suggests that anannual expenditure of $1M is probably required to keep the telescope competitive. NOAOhas included its share of this budget ($400,000 annually) in its long range plan, but theuniversities have not yet found a way to fund their share. There is also currently no long-term plan for instrumentation agreed to by the consortium. NOAO has recommended thatthe WIYN consortium hold a workshop this spring to develop a strategic plan for WIYN,and the partners have agreed to attend such a meeting in mid-June. The details of this planwill depend on whether or not the issues relating to time-sharing with the 4-m Mayalltelescope can be resolved.

A final issue for WIYN is to assess the effectiveness of the queue observing programcurrently being operated by NOAO. The WIYN queue program requires between 2 and 3staff in addition to what would be required if we were supporting observers who come tothe telescope to make their own observations. The queue program has enabled certain typesof science not possible with conventional scheduling models, most notably themeasurement of supernova light curves in distant galaxies. What we need to do beforedeciding whether to continue the queue program is to develop models to assess thescientific throughput of the telescope with both conventional and queue scheduling andalso to poll those who have received WIYN queue observations to determine the quality ofthe data. An additional problem for this program is the departure to the ESO/VLT of DaveSilva, who has been responsible for both the completion of the telescope and theimplementation of the queue observing program. KPNO does not have experiencedscientific staff currently available to assume the responsibilities held by Silva.

3. Access to the Independent Observatories

The NSF has funded instruments for the Hobby-Eberly Telescope and for the upgraded(6.5-m) MMT. In return, approximately 7 percent of the time on each will be available tothe community for six years. The proposals for this observing time will be received andreviewed by NOAO. Discussions have begun with the independent observatories on how to

implement this program. The goal is to develop an MOU for NSF approval that will addresssuch issues as who is eligible to apply (i.e., what level of experience must observers alreadyhave), who is responsible for helping users understand the instruments well enough toprepare competitive proposals, how manuals and training are to be provided, when timeassignments are to be made in order to meet the scheduling constraints of the independentobservatories, what data reduction software is available for processing the data, what is tobe done about proposals from NOAO users and members of the host institution forsubstantially similar scientific products, and who owns the data (whether it can be placed inan NOAO archive).

4. Gemini

Gemini North is scheduled to achieve first light at the end of 1998 and Gemini South in2000, with operational handover occurring about two years after first light. That means thatboth telescopes should be producing scientific data by the end of this five-year plan. NOAOis responsible for providing US user support for those activities that do not occur at thetelescopes themselves (help with proposal preparation, evaluation of proposals, assistancewith planning observations, data reduction, and access to the Gemini archive). The staffingand support required for these activities have been summarized in earlier long range plansand in the proposals to renew the cooperative agreement, and we are beginning to transferresources from other parts of NOAO, mainly KPNO, to support these activities.

NOAO is also responsible for managing the entire Gemini infrared instrumentationprogram, including those instruments being built outside NOAO, and is itself committed toproviding the Gemini infrared spectrometer, the controllers for the LR arrays, thecharacterization of the LR arrays, and possibly some portions of the hardware/software forthe CCD controllers (proposals are currently under review). The risks to NOAO in buildinginstrumentation for Gemini are substantial. There is the financial cost; the international

agreement clearly states that the funding for instruments is not intended to be high enoughto cover the full costs (direct plus indirect) of the instruments. There is the financial risk;we are to agree to fixed price contracts, with the costs of overruns to be borne by us. Andthere is the impact on other NOAO programs as we direct our most experienced personnelto work on Gemini instruments.

Other issues are related to the acquisition of future Gemini instruments. In particular, thereis the question of how the US is to live up to the international agreement to provideinstrumentation given that Gemini does not expect to fund it fully. Many universities maybe unwilling to sign fixed price contracts and to provide the required matching funds,which are estimated to be approximately one-third the full cost of an instrument. The NSFhas indicated that its agreement to cover these additional costs for the near-infrared imager,which is being built in Hawaii, should not be taken as a precedent but that it will deal withfuture instruments case by case. NOAO has sufficient confidence in its own management—and sufficient commitment to the success of Gemini—to be willing to sign fixed-pricecontracts. The increase in our actual indirect costs as a result of doing Gemini work is small(< $100,000 per year per project). However, NOAO is not in a position either to guaranteethe cost overruns of work performed in a university or to pay their indirect costs.

There will be instruments that Gemini desires to have built in the US where NOAO is not

the best choice as the vendor. How are these contracts to be paid? Guaranteed telescopetime is the most obvious compensation. However, there are relatively few university groups

that are capable of building the complex instrumentation required by Gemini, and most ofthose are already associated with the independent observatories so that telescope time is notattractive.

We have identified no universal solution to this problem but will simply attempt to dealwith it by working with proposers and the NSF on a case by case basis for each instrumentassigned to the US.

A final issue in the instrument program is to develop an effective model for the hand-offfrom the construction team to the operational teams at the sites where the telescopes arelocated. I am not aware of any observatory that has successfully solved the problem ofensuring a graceful transition from construction to maintenance. Traditionally, KPNO andCTIO instruments were built locally, with the construction teams only an hour and a halfaway from the operating telescopes. Therefore, responsibility for maintenance was neverfully transferred. In the new model, where instruments are built in Tucson to be shipped toChile and Hawaii (or in the case of the Arcons, built in Chile for use at Kitt Peak), we mustlearn to manage this transfer effectively.

5. 2.4-m Telescopes

The two telescopes on the list of operational facilities in 2002 for which we have notidentified funding are the 2.4-m class imaging telescopes. It remains NOAO's point of viewthat wide-field imaging, especially in the near-TR, will be essential to select andcharacterize objects for detailed study with the Gemini and other large telescopes.However, in light of the NSF reviews of this proposal, which recommended that weestablish a community consensus on what is required for Gemini support, we plan to host aworkshop to examine this issue broadly. The goals of the workshop will be: (1) to identifythe observations and capabilities required to support effective scientific use of all of thelarge telescopes being built in the US; (2) to inventory the facilities already availablewithin the US community; and (3) to develop a plan for obtaining support for the facilitiesnot already available, including recommending which organization should assume the leadresponsibility for each.

6. Telescope Closures

According to the plan outlined above, several operating telescopes are slated for closure.CTIO has already indicated that it will transfer the Yale 1-m back to Yale, which hasrecently reached agreement with Portugal and Ohio State for joint operation; 10 percent ofthe time will still be available to the US community. KPNO has already closed the 1.3-mtelescope and this summer will issue an Announcement of Opportunity seeking anotheroperating university or consortium. KPNO has also indicated that it will withdraw supportfor the Burrell-Schmidt effective 1 October. Case has arranged to continue operation bybringing in several new partners, and NOAO will leave all of its equipment at the Schmidton long-term loan so that the telescope remains operational. KPNO is exploring thefeasibility of closing the Coude Feed. Because the high resolution mode (R = 200,000) is aunique capability, we are looking at the feasibility of moving the spectrograph optics to the4-m Mayall and feeding them with a fiber.

In all of these closures, the goal is to preserve scientific capability while reducing theoperational support to levels sustainable by the current staff. When the Mosaic CCD

Imager is available at the 0.9-m, for example, it will offer better capability for wide-fieldphotometry because of better sampling of the point spread function than does the Schmidt.While the competition for observing time may become more intense, the intention is not toeliminate any classes of science.

Additional closures of telescopes would be linked (in the optimistic case) to the initiationof operations at the new facilities listed above or (in the pessimistic case) to furtherdecreases in budget.

7. Management Structure

In order to operate this complex suite of facilities as an observing system (effectively as asingle observatory with "one-stop shopping" for the user community), we have recentlyimplemented a change in management structure. This change will result in unifying all ofthe activities relating to support of users before and after their observing runs at all of theobservatories that are accessed through NOAO (Gemini, KPNO, CTIO, and theindependent observatories) under a single office, which may be called the ScienceOperations Division (names are still under discussion). The office will be headed by ToddBoroson, who will also continue to serve as US Gemini Project Scientist. The logic behindthis is that the USGP must perform this support function for the US observers at the Geminitelescopes. There will likely be efficiencies within NOAO if these services are provided forall telescopes by a single office, with minimum duplication of functions within the differentunits (USGP, KPNO, and CTIO). There will certainly be simplifications for the users.

To support this suite of activities, a number of KPNO scientific staff have been assigned tothe USGP (Pilachowski, Gatley, Jannuzi, and Lauer). This office will immediately begin tohandle the receipt and evaluation of KPNO proposals, with the actual scheduling of theselected proposals to be handled by KPNO. It will also handle proposals for theindependent observatories. A Web-based proposal form is being designed. Models of thescience throughput of alternate scheduling models are being developed, and simulations ofthe rules for Gemini queue-scheduling have already led to interesting conclusions abouthow to use those telescopes most effectively.

As part of this re-organization, Richard Green has been named Director of KPNO, and hewill also, continue to serve as head of the instrumentation program, which is responsible forbuilding instruments for KPNO, CTIO, and for Gemini. Again, this combination ofprograms unifies activities along functional lines. All of the engineering and technical staffbased in Tucson (and outside of NSO) report to Green, who can easily handle trades amongprograms and also work toward common hardware and software standards across all ofNOAO.

This current organizational structure is an interim one and takes advantage of existingpersonnel. It also lodges responsibility for advance planning of new approaches toobserving in an organization (the Science Operations Division) that is not currently likelyto be diverted by the daily pressures of operating an observatory. At its recent meeting, theAURA Observatories Council recognized the merits of re-organizing along functional lines,but recommended that the ultimate structure should combine user support across all of theobservatories with the operation of KPNO. In this model, the instrument group would beheaded by a separate director. NOAO will move toward that realignment over the next twoyears as Gemini approaches its operations phase and USGP activities related to the

construction of those telescopes and their initial complement of instrumentation begin towind down.

While scientific staff will be assigned to KPNO, the instrumentation program, or theScience Operations Division as their primary home, responsibilities will still be sharedacross divisions. For example, Tod Lauer (Science Operations) is the imaging scientist forKitt Peak, while Mike Merrill (Kitt Peak) is project scientist for the IR array controllers.Staff will also be reassigned from one division to another from time to time. Jay Elias(CTIO), for example, is currently resident in Tucson as project scientist for the GeminiIRS, while Ron Probst (KPNO) has been temporarily transferred to CTIO to help supportthe IR program. This flexibility in staff assignments is consistent with a more unifiedobservatory. Careful coordination of functional responsibilities will, however, be requiredin order to ensure that staff has time for research.

A further advantage of the recent re-organization of NOAO is that it reduces the number ofdirectors by one. As we slim down all parts of NOAO and realign our priorities to beconsistent with the facilities that we plan to offer in 2002, it is only appropriate that themanagement structure should be streamlined and modified to emphasize our changedpriorities.

III. NIGHTTIME ASTRONOMY: SCIENTIFIC DIRECTIONS

The following sections describe first the scientific context within which we have developed thelong range plan for the nighttime facilities and then the plan for the evolution of NOAO's program.The priority for groundbased IR and optical astronomy for the next decade will be to exploit fullythe capabilities of the facilities that are now on line or will be completed in the next few years. TheGemini telescopes have been designed to take advantage of modern technologies to achieveunprecedented image quality combined with low emissivity in the infrared. These new telescopeswill almost surely increase the demand for observing time on 4-m class telescopes; observers willuse these telescopes to conduct surveys to select objects suitable for detailed study with the Geminitelescopes; to obtain calibration data on nearby bright objects that can be used to interpret limitingobservations made with the Gemini telescopes; and to test new observing protocols, techniques, andinstruments for subsequent use by the Gemini project. In addition, there are some types of keyobservations, such as wide-field imaging in the optical with CCD mosaic arrays and multi-fiberspectroscopy over a wide field, that will not be possible with the initial Gemini implementation.Therefore, a key priority for the nighttime program is to increase access to 3-m to 4-m classtelescopes and to equip those telescopes with state-of-the-art instrumentation. If at all feasible, thismight include construction of new facilities, possibly through a leveraged financial arrangement. Ahighly capable instrumentation group is also needed to support the fabrication of the complexinstrumentation required to realize the full potential of the Gemini telescopes.

It is not at all clear how groundbased optical/IR astronomy will evolve after this generation of 8-mand 10-m telescopes is completed. The AURA Coordinating Council of Observatory Directors(ACCORD) is actively discussing this issue. One obvious extension is the inclusion of adaptiveoptics at varying levels of sophistication on both the 4-m and 8-m to 10-m class facilities.Following this development, it is widely thought that an interferometric array is the most likelycandidate for a major new facility, but many technical issues remain to be resolved before such afacility could be designed. This is the decade of prototype interferometers, and NOAO willcontribute technical and scientific assistance to one of these arrays in order to gain experience withinterferometry, understand the technology, and begin developing community interest in the

10

technique. Only after the prototype arrays begin delivering scientific results will it be possible toset the design parameters for a major array.

In summary, then, the emphasis in the nighttime program is on exploiting fully the facilities wehave, including the Gemini telescopes, by ensuring that they are properly instrumented andeffectively operated, and on exploring some potential new technologies that may provide the basisfor the facilities of the next century. In addition, if overall economies can be realized by replacingolder facilities with more efficient state-of-the-art 3-m to 4-m class telescopes, then suchpossibilities should also be fully explored.

Science at CTIO and KPNO

Approximately 13,500 scientific papers have been published between 1961 and 1996 by visitorswho made use of NOAO facilities and by NOAO staff. In this plan, we summarize keyobservational results in a few particularly active areas of research, and in addition we brieflydescribe how these areas will benefit from use of the current and planned NOAO facilities duringthe next five years.

A. The Large-Scale Structure of the Universe

At the heart of cosmology is the "cosmological principle," which posits that the universe ishomogeneous and isotropic on large scales. Thus if one examines a large enough volume of theuniverse, its average properties and the structures formed by the galaxies within, that volumeshould be no different from those in any other large volume of the universe. The size of thevolume that is required to include a fair sample of the universe is, however, unknown. In recentyears, observations made with telescopes operated by NOAO and by other institutions haveshown that the universe is organized on scales so large that all current theories, in order toexplain how galaxies formed and clustered to develop the structures that we see, must either befundamentally flawed or be in need of significant revision.

The first glimpses of large-scale structure were seen over a decade ago when in 1981 R.Kirshner (Center for Astrophysics) and his collaborators used the KPNO telescopes to identifywhat appeared to be a gigantic void in the direction of the constellation Bootes. In 1986, thissame group used the KPNO facilities to confirm the existence of this void, which occupies avolume of over one million cubic megaparsecs and is roughly spherical in shape, with adiameter of about 120 Mpc. Also in 1986, V. de Lapparent, M. Geller, and J. Huchra (allCenter for Astrophysics) published results from the extended Center for Astrophysics redshiftsurvey, which showed galaxies residing on the surfaces of contiguous bubble-like structureswhose diameters are typically 25 Mpc with a maximum of 50 Mpc. A similar picture of thedistribution of galaxies in space emerged from a 21-cm survey of 2,700 galaxies published in1986 by M. Haynes (Cornell U.) and R. Giovanelli (Arecibo Obs.). Data on even largerstructures has emerged from the work of D. Koo (Lick Obs.) and his collaborators. Using pencilbeam surveys of the north and south Galactic poles, these investigators have sampled thedistribution of galaxies over an unprecedented scale of 2000 Mpc, which is far deeper than anyprevious surveys. They find evidence for large-scale features in the galaxy distribution out tothe limits of the survey. Work is continuing on this project using many different telescopes, andthe existence of walls and voids on scales of 100 Mpc has recently been reported by Bellangerand de Lapparant (Institute d'Astrophysique) using the ESO telescopes. In short, so farobservers continue to find galaxies organized into larger and larger structures as they surveylarger portions of the universe.

11

Equally exciting has been the discovery of streaming motions of galaxies on an extremely largescale. These motions result from the dumpiness of matter in the universe and provide separateevidence that matter in the universe is organized on extremely large scales. In the late 1970s,M. Aaronson (U. of Arizona), J. Huchra, and then-KPNO staff member J. Mould pioneered theinfrared Tully-Fisher technique for determining distances to spiral galaxies. In 1986, Aaronsonand collaborators employed this method on a sample of ten northern hemisphere clusters,observed mostly at KPNO and Arecibo, and obtained a positive detection of the motion of thelocal group toward the 3K dipole anisotropy. In the same study, Aaronson, et al. concluded thatthe rms peculiar velocities of clusters must be smaller than 500 km/sec. This picture of arelatively quiescent local Universe was very short-lived, however, and in 1987 a group of seveninvestigators (A. Dressier, D. Lynden-Bell, S. Faber, D. Burstein, R. Davies, R. Terlevich, andG. Wegner) announced the discovery of streaming motions of galaxies. Their analysis ofspectroscopic and photometric data of 400 elliptical galaxies, obtained with telescopes atKPNO, CTIO, and other sites, revealed a net streaming motion of galaxies over a region about100 Mpc in size, with a velocity at the Sun of roughly 570 km/s over and above the uniformHubble flow. The source of this bulk motion was attributed by Dressier and collaborators to theexistence of a "Great Attractor," located in the direction of the constellations of Hydra andCentaurus, at a distance of 20-30 Mpc and with a mass of about 5 x 1016 solar masses.

Observations of galaxies from the southern hemisphere are thus of crucial importance inassessing the all-sky nature of peculiar velocities. Unfortunately, the largest radio telescope inthe southern hemisphere, the Parkes 64-m, has only 1/12 the collection area of Arecibo, and sodoes not generally provide high quality data for galaxies with velocities greater than 5000 km/s.Using an imaging Fabry-Perot interferometer on the CTIO 4-m telescope, however, R.Schommer (CTIO), T. Williams (Rutgers U.), G. Bothun (U. of Oregon), and J. Mould (Mt.Stromlo Obs.) developed an optical Tully-Fisher method, which has been used to determinedistances to galaxies with velocities as great as 10,000 km/s. This group has published ananalysis of peculiar velocities for a sample of 48 late-type spiral galaxies that are located in thevicinity of the Great Attractor. These data confirm the existence of positive velocity residuals,with amplitudes of 500-2000 km/s, in the Hydra-Centaurus region, but a symmetric infallpattern centered on the Great Attractor is not seen. These results suggested that theidentification of the source of the observed streaming motions was as yet not well determinedand that the possibility of large peculiar motions on scales of 100 Mpc remains.

This possibility may well be confirmed by some recent results on large scale structure obtainedby T. Lauer (KPNO) and M. Postman (STScI). Using facilities at both CTIO and KPNO, theseobservers have determined that the brightest galaxies in clusters of galaxies can be used asreliable standard candles because the luminosity measured within a fixed metric radius of 10kpc is nearly constant, with a dispersion of 0.24 magnitude. This leads to an error of only 17%when these objects are used as distance indicators. With this technique they have derivedredshifts for 119 clusters of galaxies and have determined the velocity of the Local Group withrespect to all Abell clusters within 15,000 km/s redshift. The objective was to determinewhether the motion of the Local Group relative to this very distant set of clusters is the same asthe motion of the Local Group relative to the microwave background. If so, then the largestscale anisotropic streaming would be confined to scales smaller than that sampled by theclusters. Lauer and Postman do not find this convergence. A straightforward interpretation oftheir results is that the entire volume enclosed by the cluster sample is itself streaming at avelocity of about 700 km/s relative to the microwave background. If this result is confirmed, itimplies streaming and inhomogeneities on a scale much larger than previously observed.

12

All of these large-scale features in the galaxy population pose severe challenges to currentcosmological models. The difficulties lie in the inability of the models to produce such largestructures by purely gravitational means. The hot dark matter models use the hot particles tosuppress excessive small-scale fluctuations, but in conventional cosmologies with Q. = 1, theydo not reproduce the galaxy-galaxy correlation function, and they have difficulty producing thegiant mass fluctuations required to explain the large-scale streaming motions. Cold dark mattermodels cannot produce the large structures, and they yield too much small-scale structureunless ad hoc assumptions such as biased galaxy formation are introduced. If the most recentresults of Lauer and Postman are confirmed, then virtually all current cosmological models,including hybrid models and those with ad hoc fixes, are in serious jeopardy. Non-gravitationaltheories using shock waves to initiate galaxy formation may successfully account for theobserved structure, but they require as yet unobserved explosive events of enormous energy,about 10 5ergs per event, which is equivalent to the energy radiated by 10,000 galaxies overthe age of the Universe.

An essential element in any study of large scale structure and cosmology is the issue of thedistance scale. The determination of this fundamental yardstick has had a colorful andcontroversial history, and the elements of controversy remain to this day. Arising from workcarried out with NOAO telescopes is a growing hope that agreement may finally be near on thevalue of the Hubble constant. Two recent techniques for distance determination that haveshown significant promise are the use of planetary nebula luminosity functions (PNLF) andsurface brightness fluctuations (SBF). Both of these techniques claim accuracy in distancedeterminations of about 10% out to distances of 20 Mpc. Development of both methods hasrelied on the use of NOAO facilities, and recently a collaborative effort among the proponentsof these methods sought to test the similarity of their results. G. Jacoby (KPNO), R. Ciardullo(Penn State U.), and J. Tonry (Massachusetts Inst, of Tech.) compared direct observationalresults for a sample of 15 galaxies. This test shows that the two methods yield scales which areoffset only by 0.07 mag, and this offset can be attributed entirely to uncertainties associatedwith measurements of extinction toward the calibration galaxies M31 and M32. The verystrong consistency between the two methods, one using radiative heating and cooling ofgaseous nebulae and the other depending on our understanding of the evolution of stars on thegiant and asymptotic giant branches, provides a high level of confidence in the distancesderived. These methods provide a relatively large value of about 80km s"1 Mpc"1 for the Hubbleconstant. This value poses yet another challenge for conventional closed universe cosmologieswhen the ages of the oldest stars are compared with the predicted age of the Universe. Supportfor value of the Hubble constant only about 10 percent lower than this has now been providedby observations of Cepheids with the repaired Hubble Space Telescope. Work by the CTIOgroup, by Perlmutter and his colleagues, and by Kirshner and his collaborators on the use oflight curves from type la supernovae can, when coupled with Cepheid data, provide values forboth Ho and qo- Early results from this technique are yielding values for H0 of around 65-70km/s/Mpc. Intensive work continues in this area, and the queue-scheduling, good image quality,and large aperture of the WIYN make it particularly well-suited for obtaining light curves ofsupernovae in distant galaxies.

Future programs investigating the large scale structure of the Universe will push out to higherredshifts. Early results by Steidel and others using Keck indicate that large scale structure(bubbles and voids) is already present by redshift 3, thereby creating severe problems formodels that assume critical density. A thorough investigation of the evolution of large scale

13

structure with cosmic epoch requires the execution of wide-field surveys that can identify, anddetermine the distribution of, galaxies and QSOs at various times. Such deep wide-field surveysare ideal for the 4-m and WIYN telescopes at NOAO. Further into the future, the largecollecting area of 8-m telescopes and the upgraded MMT, when coupled with the speed offiber-fed multi-object spectrometers, will result in an improvement of more than two orders ofmagnitude over current NOAO facilities. This capability will allow the analysis of faint quasarsand clusters of galaxies at redshifts between 1 and 3, which is an essential period in the study ofthe evolution of large-scale structure. In addition, the reddest galaxies at redshifts greater than 3have no spectral features in the optical region and must be studied exclusively in the infrared. Itis these most distant and possibly evolved galaxies that will provide constraints on basiccosmological parameters and on the age of the Universe through use of the Gemini telescopes.

B. The Formation and Evolution of Galaxies

The question of how galaxies form and evolve is one of the fundamental problems in modernastronomy. The current view is that galaxy formation is a complex, dynamic process in whichsmall objects first become gravitationally unstable in the early universe, cluster, and theneventually merge to form a larger galaxy-mass object. The subsequent evolution of this galaxy-mass object is governed by the aging of its stellar populations, variations in its star formationrate, and its merger history. These processes determine the morphological, dynamical, andspectral properties of the final product: the present day galaxies. Although the theoreticalinvestigation of this problem, which began with Sir James Jeans in 1928, has now reached alevel of considerable sophistication where predictions can be made of galactic properties (mass,luminosity, spectra) as a function of time, it is only recently that the observational datanecessary to confront the theoretical studies have become available.

1. Galaxies in Clusters

One of the first indications that the effects of galaxy evolution are directly observable wasrevealed nearly 18 years ago in a study by H. Butcher (Kapteyn Obs.) and G. Oemler (YaleU.). Using KPNO telescopes in an observing program spanning several years, Butcher andOemler discovered that although nearby rich clusters like Coma tend to be populated withold, quiescent elliptical galaxies, more distant clusters (observed at earlier epochs) containan increasing fraction of blue galaxies, which are undergoing significant amounts of starformation. The clusters surveyed extended out to a redshift of approximately 0.4, whichcorresponds to looking back nearly one-third of the age of the Universe (we adopt H0 = 50km/s/Mpc, n = 0.2). This result, which has since been confirmed spectroscopically byDressier (Carnegie), Gunn (Princeton), and their collaborators, indicates a rapid evolutionof galaxies in rich cluster environments. High resolution optical imaging of the redshift 0.3-0.4 clusters with HST shows that the blue star-forming galaxies tend to be disk systems,with the majority showing distorted morphologies and signs of interactions, suggesting thatgalaxy interactions may drive the star formation and the morphological evolution ofgalaxies in clusters.

Less is known about the evolution of the reddest galaxies in clusters; although there isevidence that some of them (the so-called E+A galaxies) may have undergone recentepisodes of star formation, there is also some evidence that the red elliptical galaxypopulation evolves fairly passively from redshifts greater than 2 (when our Universe wasless than one-quarter of its present age) to the present. For example, the study by

14

S.A. Stanford (IGPP/LLNL) and his collaborators using the NOAO telescopes foundevidence for only mild evolution in the red elliptical galaxy population in clusters betweenredshifts around 0.4 and the present. The questions of the age and evolution history of thereddest galaxy population (the 'red envelope') is of great importance, since these objectsrepresent the oldest stellar conglomerates; age-dating them can provide the epoch when thebulk of their stars were formed and, for the objects in the most distant clusters, can place alower limit on the age of the Universe.

Clues to a method for finding distant clusters of galaxies may be found in the systematicstudy of the environments of quasars out to redshifts of ~ 0.7 by H. Yee (U. of Toronto),R.F. Green (KPNO) and E. Ellingson (U. of Colorado). Using telescopes at NOAO andSteward Observatory and the Canada-France-Hawaii Telescope, these investigatorsestablished that both radio-loud and radio-quiet quasars tend to be situated in regions ofhigher-than-average galaxy density, and in particular, that higher redshift radio-loudquasars are found in much richer environments than their lower redshift counterparts.Similar evidence was also found by J.A. Tyson (Bell Labs.), who used the CTIO andKPNO telescopes to observe quasars at two different epochs (redshift ranges 0.1-0.5 and1.0-1.5). Tyson found that the more distant quasars had ten times more galaxies aroundthem than did the nearby sample. Similarly, T. Heckman (Johns Hopkins U.) and hiscollaborators have used NOAO telescopes to determine that galaxies associated with radiosources often reside in regions of higher galaxy density than do similar galaxies that areradio-quiet. These results have the important implication that quasar and radio galaxyevolution may be a strong function of environment, and also that distant clusters can befound from searches around distant quasars and radio galaxies. Using KPNO telescopes,M. Dickinson (STScI) has identified a few rich clusters at redshifts around 1 (looking backmore than half the age of our Universe) by using high redshift radio galaxies as signposts ofrich environments. In one redshift 1.2 cluster, Dickinson has discovered a population ofred, elliptical galaxies, which have colors similar to present-day ellipticals, suggesting thatthe old population in clusters may be very old, indeed. Dickinson, A. Dey (KPNO), andH. Spinrad (UC, Berkeley) are now following up the KPNO imaging study by obtainingoptical spectroscopy of the cluster members using the Keck Telescope. A search for normalgalaxies in the vicinity of very distant quasars (at redshifts around 3; when the Universewas less than a fifth of its present age) has been carried out by C.C. Steidel (Caltech),D. Hamilton (Heidelberg), and their collaborators using the NOAO 4-m telescopes. Theirimaging survey identified several faint candidate star-forming galaxies on the basis of theircolors. Steidel and his collaborators have spectroscopically confirmed the redshifts and thestar-burst nature of these galaxies using the Keck Telescope.

What about the primeval galaxies: systems undergoing their first epoch of star formation?Over the last decade, several searches have been carried out for these elusive objects withnull result. The best candidates may still be the highest redshift radio galaxies discussedabove, but even the most distant ones known (observed when the Universe was less than 10percent of its present age) show signs of metal enrichment over a large volume. Extremelylow-metallicity, dynamically young, star-forming galaxies are yet to be discovered. Sincethe forming galaxies may be faint, novel methods that make use of gravitationalamplification by foreground clusters or of the absorption imprints they leave on the light ofbackground quasars may be the only recourse. Gravitational lens arcs were first discoveredat KPNO by R. Lynds (KPNO) and V. Petrosian (Stanford) and are now thought to beimages of distant galaxies produced by the foreground cluster which acts as a gravitationallens. Several such lens systems are being studied by various groups in an attempt to both

15

determine the mass distribution in clusters and deconvolve the lensed images to catch aglimpse of the distant, young galaxies.

2. Field Galaxies

The modern study of field galaxy evolution was pioneered by J.A. Tyson (LucentTechnologies) and P. Seitzer (U. of Michigan), who conducted deep CCD surveys of thesky using the CTIO and KPNO 4-m telescopes. They found a vast population of faint, bluegalaxies whose number counts appeared to be in excess of that predicted by simple "no-evolution" models, implying that this population of galaxies evolves very rapidly. Recentdetailed studies (high resolution imaging with HST and groundbased spectroscopy)strongly suggest that the faint blue galaxies are fairly low mass dwarf galaxies that happento be undergoing a burst of star formation at low redshifts (between 0.5 and 1.0). At presentit is unclear why the dwarf galaxies appear to undergo delayed formation or starburstactivity, or how this is a function of environment. In contrast, studies of the red fieldgalaxies appear to show little evolution out to redshifts of around 1. This has beendemonstrated by D. Hamilton using the NOAO telescopes, and most recently in the deepCanada-France-Hawaii Redshift Survey conducted by a team led by S. Lilly (U. ofToronto). These results are also supported by the HST Medium-Deep Survey, which findsthat the bulk of the bulge-dominated field galaxy population is in place by about half theage of the Universe and then evolves only passively to the present time, in contrast to thedisk galaxies which appear to show signs of interactions and mergers. Spectroscopy of theMDS survey galaxies for the purpose of redshift determination is now proceeding at NOAOfacilities. More evidence that galaxies of higher mass undergo more slow evolution isfound in a study by Steidel and Dickinson, who used the KPNO and Michigan-Dartmouth-MIT telescopes to study the population of galaxies responsible for causing MgU absorptionlines in QSO spectra. They found that this unbiased sample of luminous field galaxiesexhibits little evidence for evolution in their photometric properties or space densities outto redshifts of around 1.

The culmination of the road paved by the groundbased CCD surveys is the Hubble DeepField, the deepest image of the sky ever obtained. KPNO telescopes are already being usedin follow-up studies: a team led by Dickinson (STScI) obtained deep infrared images of thefield (currently impossible with HST) and made them available for general study.Investigations of the faintest galaxies in the HDF will be crucial to our understanding of theearliest epochs of galaxy formation. The existing data, however, already demonstrate thatthe relationship between morphological type, galaxy mass, and evolutionary history in bothfield and cluster populations appears to be complex and at present is not well understood.At least at the low and intermediate mass end of the galaxy mass function, severalunanswered questions remain.

Future programs in galaxy formation and evolution will fully utilize all of the plannedobserving capabilities at NOAO through the end of the current millennium. Thecombination of deep optical and infrared imaging surveys on 2-m and 4-m class telescopeswith sub-arcsecond resolution using the imaging arrays under development at NOAO,together with follow-up optical and infrared spectroscopic studies available with theGemini telescopes, will provide a combined observing capability of immense power. Thisworld-class facility will permit the deepest and most thorough look yet at the epoch ofgalaxy formation and early evolution. Photometry and spectroscopy of distant galaxies inthe infrared is critical, since most of the well understood optical features are shifted into

16

this wavelength regime at high redshifts. The importance of wide-field imaging surveys,especially in the infrared, cannot be overemphasized. The use of 4-m class telescopes forsuch surveys will be a necessary condition for the efficient and productive use of theGemini telescopes. In addition, targeted narrow band imaging surveys with 4-m telescopescan be used to search for high redshift protogalaxies and protoclusters in the redshiftedLyman a band or in the redshifted Ha band in the infrared. The spectroscopic power of the8-m telescopes will not only address galaxy evolution at early epochs, it will also allowstudy of the dynamics of galaxy clusters at redshifts greater than one, thus gaining access tothe epoch of cluster formation, and it will in addition permit a first view of the internalkinematics of individual galaxies at high redshifts.

C. The Structure of Nearby Galaxies

The investigation of galaxy formation and evolution would be incomplete without a completeunderstanding of the physical processes that govern the large scale properties of nearbygalaxies. The stellar populations, global and local star formation rates and efficiencies, gascontent, and kinematic properties of the nearby ellipticals and spirals are being studied byvarious teams. Nearby massive star-forming regions like 30 Dor have been explored at bothoptical and UV wavelengths using groundbased telescopes, IUE, and HST. In addition, galacticscale winds driven by large scale starbursts have been discovered in several edge-on spirals byHeckman and his collaborators using NOAO telescopes. The importance of recent mergers inthe late evolution of ellipticals has been studied by D. Silva (KPNO) and G. Bothun (U ofOregon), who find that recent bursts of star formation in ellipticals account for less than tenpercent of the light, suggesting that the nearby ellipticals are predominantly old systems. Otherinvestigators are searching for low surface brightness galaxies, which may be quiescent orfailed massive disk galaxies. These investigations (and others) of nearby galaxies will outlinethe parameter space occupied by the end products of galaxy evolution, and in addition they willallow us to study in detail the physical processes that play a role in galaxy evolution.

A low redshift population of galaxies that may be undergoing dramatic evolution and is oftentouted as protogalaxies are the ultraluminous galaxies detected by the Infrared AstronomicalSatellite (ERAS). These objects emit 10 to 100 times the luminosity of a normal spiral galaxy,with the emission mostly in the far infrared region, and are found to have very disturbedmorphologies indicative of mergers between two massive systems. The presence of copiousamounts of dust and molecular gas has now been established through millimeter andsubmillimeter wave studies, but the origin of their prodigious luminosity is still not understood.It is thought that these galaxies may be undergoing violent global star formation or harboring aproto-quasar, but more observations will be required to define these objects properly. Othercandidates for galaxies in the earliest stages of formation are the extremely hydrogen-rich andstar-poor objects found nearby, such as the hydrogen cloud discovered by Giovanelli andHaynes. This elongated cloud seems to offer support for the idea that disks of galaxies can formslowly throughout the history of the Universe.

Thus, recent observations suggest that galaxy formation was far from coeval, that galaxies mayhave been forming throughout the age of the Universe, and that many, but not all, haveundergone dramatic evolutionary changes in that time. In many cases, those galaxies that showstrong evidence for evolution are found to exist in special circumstances: in clusters ofgalaxies, near quasi-stellar objects, or associated with strong radio sources. It is by no meansclear which of these processes is relevant or dominant, and much further work needs to be

17

done. Future NOAO facilities can play a prominent role in these investigations. For example,the 8-m Gemini telescopes using spectroscopy in combination with adaptive optics will permitexamination of the metallicities and kinematics of nearby galaxies out to an R magnitude ofabout 24. Such fundamental investigations of the internal structure and star formation history ofnearby galaxies could in principle consume all available Gemini observing time over a periodof five years.

D. Star Formation

Many aspects of the process of star formation remain poorly understood. The formation andevolution of the parent clouds, their fragmentation and collapse, the role of angular momentumand magnetic fields, the establishment of the initial mass function, and the evolution ofprotostars and very young stars are all areas of active investigation. NOAO facilities have beenused to observe regions of star formation in our own and in nearby galaxies, and the advent oftwo-dimensional detector arrays that operate in the infrared is stimulating even more observingprograms relevant to this topic. Moreover, these arrays are particularly well-suited for use inmulti-wavelength, multi-observatory programs that address key questions in the area of starformation. In particular, they will complement observations by HST, SDRTF, SOFIA, andmillimeter and sub-millimeter telescopes and arrays.

It has become well established that the collapse of a protostellar cloud to form a new star isaccompanied by an outflow of mass from the central region, and several programs have usedNOAO facilities to investigate the nature of these outflows and their implications for the starformation process. The outflow is usually anisotropic, often bipolar, and sometimes very highlycollimated. The origin of this outflow and the mechanism for its collimation have beenexamined by several groups. In some cases, the outflow is seen to be collimated to within 100AU of the surface of the young stellar object, but it is still unclear whether the wind isintrinsically bipolar or is collimated by material around the star. Evidence that circumstellardisks are found around virtually all young stellar objects has been obtained by S. Strom (U. ofMassachusetts) and his collaborators, arguing for an external collimation mechanism.Moreover, these observers also find the inner portion of the disks to be cleared away over aperiod of about 10 million years, leaving a small remnant of material that may eventuallyaggregate into planets. Understanding this very common outflow phenomenon would providevaluable information about the role of angular momentum in the star formation process, aboutthe efficiency of stellar collapse and the role of circumstellar material, and about the conditionsat the surface of the young star itself.

The Herbig-Haro objects (emission line nebulae thought to originate from the interaction ofmatter ejected from a young star with the interstellar medium) also hold clues to the starformation process. Recent work by J. Bally (U. of Colorado) and his collaborators now make itclear that these objects are associated with very highly collimated and narrow outflows from ayoung stellar object. In many cases several HH objects are associated with one outflowingstream, and Bally and his co-workers have found spectacular symmetric outflows extending inopposite directions from young stellar objects. In some cases these supersonic double jetsextend more than a parsec and contain many HH objects. The relation of these outflows to themore massive, slower, and less collimated molecular outflows in which they are oftenimbedded is not yet clear, nor is the mechanism known that produces the high degree ofcollimation. Once the relationships among these outflow phenomena are understood, we willhave a much clearer picture of the early evolution of protostars and of star formation itself.

Studies of angular momentum in evolving protostars shed further light on the role of disks instar formation. For example, J. Stauffer (Smithsonian Astrophysical Obs.) has obtainedrotational velocities for stars in the Pleiades cluster. He finds that nearly half of the starsobserved have rotational velocities that are much higher than expected. This indicates that atleast some stars retain a major portion of the angular momentum of the protostellar cloudduring collapse. Further insight into this phenomenon has come from a study by S. Edwards (U.of Massachusetts) using optical and IR data obtained at KPNO. A sample of 34 T Tauri starswith and without accretion disks was examined for rotational periods; those stars with accretiondisks were found to have lower rotational velocities than those without. These results suggestthat the disk may be a sink of angular momentum, presumably through some kind of magneticfield link between the disk and the star. Once the disk is lost, stars can spin up as they contractdue simply to the conservation of angular momentum. The most rapidly rotating stars on themain sequence may be those that, for whatever reason, lost their disks during the early stages oftheir collapse.

Productive areas for future investigation include the local phenomena of protostars, massoutflow, circumstellar shells, and accretion disks, along with such global aspects as the initialmass function and the variation of star formation with changes in metallicity, turbulence, andmagnetic fields. A major impetus for new and continuing programs in star formation has comefrom the availability of two-dimensional infrared array detectors. Already these arrays havedetected circumstellar disks of molecular gas around protostellar objects, and they have alsorevealed rich groups of very young stars whose existence was heretofore only assumed. P.Hartigan and his collaborators at the Center for Astrophysics have used a two-dimensionalarray to determine that molecular cooling is a principal energy loss mechanism in thedeceleration of outflows from Herbig-Haro objects. In another study using the four colorinfrared imager (SQILD), E. Lada (U. of Maryland) and C. Lada (Center for Astrophysics) havestudied the star formation history of the young cluster IC 348. These new results show that starformation in this object did not take place in a single burst (as seems to be the case in the Orionclusters) but instead extended over a period of 5 to 7 million years. About 20% of the clustermembers show an infrared excess indicative of a proto-planetary disk. The differences shownin star formation history for this cluster when compared with Orion are indicative of the manyfactors that influence the complex phenomenon of star formation. These programs are thebeginning of what is seen as an era of major advances in the understanding of star formation,not only in our own Galaxy but eventually in other galaxies such as the Magellanic Clouds andmembers of the Local Group. In the Galaxy itself, the problem of low mass star formation andthe definition of the initial mass function (IMF) for low-mass stars can now be addressed.Studies of the upper end of the IMF, where many of the more massive stars have beenheretofore obscured by the dust associated with their birthing process, have also becomepossible. It will be feasible to identify and study stars below the critical mass for nuclearignition and to examine accretion disks around protostars and young stellar objects. The arrayswill also make possible the search for particle disks around main sequence stars, these being thepresumed progenitors of planetary systems.

Complementary to large scale surveys are detailed studies of individual star-forming regions.KPNO Astronomer Phil Massey has studied the formation of massive stars in many nearbygalaxies, and his intensive examination of these objects has provided new insights into theupper end of the mass function and its dependence on both local and global factors. Forexample, Massey and his co-workers, in studying the Magellanic Clouds from CTIO, havefound that the most massive field stars are as massive as those found in large OB complexes,but that the slope of the IMF for field stars is much steeper than that for the OB complexes.

19

More recently Massey and his collaborators have found circumstellar disks around intermediatemass stars in Ml6. The observations of these short-lived disks, as well as observations of theyoung stars themselves, will permit definitive tests of current ideas on mass loss and angularmomentum transfer in the earliest stages of stellar evolution.

Future facilities at NOAO will ensure active and forefront programs in this area through theend of the century. In addition to the infrared imaging of specific regions described above, thedeep, wide-field optical and infrared surveys used for extragalactic programs will also yieldvaluable new data about the populations of brown dwarfs and protostars as well as sampling thestellar mass function at the elusive but critical faint low mass end. Four-meter class telescopeswith sub-arcsecond seeing can now be used to examine specific protostellar objects todetermine in detail the disk kinematics. The advent of the Gemini telescopes will permit thesekinematic studies to proceed to fainter objects and to probe much closer to the young star itself.Use of the infrared regime will allow studies of protoplanetary disks and provide new highresolution data about the cores of molecular clouds where the youngest protostars are formed.

E. Stellar Populations

The study of stellar populations not only provides information about stellar evolution, it alsosheds light on such diverse topics as star formation, the chemical and dynamical evolution ofthe Galaxy, the calibration of the distance scale, and the age of the Universe. One area of stellarpopulations that has been especially fertile in recent years is the chronology of the formation ofthe Galactic halo, as evidenced by globular clusters and field stars. A number of studies havepresented findings that are not accounted for by the traditional model for the formation of theGalactic halo via a rapid, monolithic collapse.

One major development in the area of stellar populations is the methodology for datingglobular clusters and the resulting age range in the cluster system, and inter alia the Galactichalo. Independently, groups at DAO and Yale proposed overlaying the main sequence andsubgiant regions of the color-magnitude diagrams of globular clusters of similar metallicity andderiving the age difference rather than the absolute age. Age differences can be derived withminimal reliance on theoretical isochrones. Such studies revealed a significant age range amongglobular clusters of intermediate metallicity (~ 3 Gyr). This favored the Searle-Zinn accretionpicture for the formation of the Galactic halo over the rapid collapse model. The majority of thecolor-magnitude diagrams on which this work relies were constructed using CCD images fromthe CTIO 4-m telescope. In particular, the cluster pair with the most convincing age difference,NGC 288 and NGC 362, was studied differentially at CTIO by M. Bolte (UC, Santa Cruz), andindependently by B. Green (Steward Obs.) and J. Norris (Mt. Stromlo Obs.). Another clusterthat figures prominently as a young halo cluster, Pal 12, was observed with the CTIO 4-m by P.Stetson (DAO) and collaborators.

Another observation that has had a strong influence on scenarios for the formation of theGalactic halo is the discovery of the Sagittarius dwarf galaxy. Found serendipitously during asurvey for Galactic bulge stars, Sagittarius is located only 16 kpc from the Galactic center andshows distortions that are likely indicative of its being disrupted by the tidal forces of theGalaxy. Sagittarius may thus be a system where the accretion of a dwarf galaxy by the halo is inprogress. The field population of Sagittarius and M54, the globular cluster that lies at its center,have been studied by Sarajedini (KPNO) and Layden (then at CTIO) using the CTIO 0.9-m.These authors investigated the metallicity distribution in Sagittarius and found both metal-richand metal-poor components. The metallicity-luminosity relation for dwarf galaxies then

20

suggests a much larger integrated magnitude for Sagittarius than is measured, suggesting thatits interaction(s) with the Milky Way have stripped away a significant fraction of its originalstars. M54 and Sgr have the same distance modulus and radial velocity within the uncertainties,and thus M54 may represent the nucleus of Sagittarius.

Another approach to understanding the formation of the halo is mapping the space velocity andmetallicity of stars in selected areas. Majewski (U. of Virginia), Munn (U. of Chicago), andHawley (Michigan State U.) are carrying out such a program in SA57 at the north Galacticpole. The first step in their study is a deep absolute proper motion survey using multi-epochKPNO 4-m plates. They use these data to isolate a relatively pure sample of halo stars byselecting only objects more than 5.5 kpc above the Galactic plane. Spectra are then obtainedwith Hydra, yielding radial velocities and metal abundances. This survey has revealed largescale streaming motions in the Galactic halo. Thus the halo is not dynamically mixed, butcontains a significant fraction of stars with membership in correlated stellar streams. Thissuggests that the accretion of satellite systems likely played a significant role in the formationof the Galactic halo.

The above examples illustrate the broad range of stellar populations programs being carried outwith NOAO facilities. For many of these problems, spectroscopy of more stars, intrinsicallyfainter stars, and more distant stars is needed. Scientific programs in the area of spectroscopy ofstars in clusters and stars found in selected area surveys benefit especially from multi-objectcapability, which allows observations of many individual stars simultaneously. Both CTIO andKPNO have multi-object fiber-fed spectrographs in operation, and the NOAO instrumentationprogram for the next five years plans to enhance this capability. In addition, many programsrequire precise photometry for samples of stars over a significant number of square degrees.The deployment of the NOAO 8K x 8K CCD Mosaic Imager will be an asset to such scientificendeavors, especially when used in conjunction with the planned wide-field optical andinfrared surveys.

F. The Chemical Evolution of the Galaxy from High-Resolution Spectroscopy

Spectroscopy of stars in our Galaxy is providing additional constraints on the formation of theGalaxy, and by implication, on the formation of all spiral galaxies. K. Gilroy and C. Sneden (U.of Texas), C. Pilachowski (KPNO), and J. Cowan (U. of Oklahoma) used the KPNO 4-mechelle spectrograph to investigate the abundances of very-heavy (Z > 30) elements in metal-poor halo stars. Such elements may be formed via slow neutron captures (the s-process) on Fe-peak seeds during the late quiescent stages of the evolution of low to intermediate mass stars orby rapid neutron captures (the r-process) during the supernova explosions of high mass stars.Their results suggest that for stars more metal-poor than about l/100th of the solarcomposition, evidence only for r-process nucleosynthesis could be found in the abundances ofthese elements. The s-process contributions to the observed abundances seemed to appear onlyat higher metallicities. This allows an estimate to be made for the timescale of the formation ofthe very metal-poor galactic halo: about 10 million years.

The Gilroy, et al. study raised the possibility of large star to star scatter in the abundances ofvery heavy elements. Such a scatter, if confirmed, could set important limits on the mixingtimescales of the early galactic halo. At what overall metallicity does one see evidence for"local" nucleosynthesis events in the then-young Galaxy? To answer that, T. Armandroff, C.Pilachowski (KPNO), D. Burris and J. Cowan (U. of Oklahoma), and C. Sneden (U. of Texas)currently are analyzing the very heavy element abundances of a sample of metal-poor field halo

21

stars, using spectra again acquired with the KPNO 4-m echelle spectrograph. The focus will beon the balance of r- and s-process contributions to the very heavy element abundances and onstar to star element scatter, both as functions of stellar overall metallicity. Such questions canbe cleanly addressed only with large sample sizes.

In another study of chemical evolution, Beers, Preston, and Shectman have produced a catalogof ultra metal-poor stars (those with metal abundances less than l/1000th of the Sun's). Out ofa high resolution study of some of these stars came the discovery of a star with a hugeoverabundance (10-40 times more abundant than the ordinary metals in this star) of very-heavyelements. C. Sneden and B. Armosky (U. of Texas), A. McWilliam and G. Preston (CarnegieObs.), and J. Cowan and D. Burns (U. of Oklahoma) performed the most detailed very heavyelement abundance investigation ever attempted in a metal-poor star, using CTIO 4-m echellespectra. Abundances for 20 elements with Z > 30 were determined, including 5 elements (Tb,Ho, Tm, Hf, and Os) never before detected in metal-poor stars. An excellent match to a scaledset of solar system r-process abundance fractions was found; no other nucleosynthesis modelcan be made to fit the observed abundances. The very-heavy elements in this star must havebeen made in prior supernova explosions. And the huge overabundance of all very-heavyelements argues that only one supernova was the culprit: the nucleosynthesis event was a localone. Finally, the clean detection of thorium in this star sets an approximate age for itssupernova progenitor: about 15 billion years.

The Beers, et al. survey of ultra metal-poor stars has been used for only a few abundancestudies to date. Even so, hints of more abundance peculiarities in individual stars have beenslowly appearing in the literature. Abundances of ultra metal-poor halo stars provide thecleanest "laboratories" for testing explosive nucleosynthesis theories. Thus, detailed abundancestudies of other very metal-poor stars will be carried out in the future with 8-m class telescopes;these stars are generally found only in the more distant galactic halo regions.

Another important site for studying chemical evolution in the galaxy lies in globular clusters.Abundance variations have been known to occur in the stars residing in these clusters for morethan 30 years, but only now are we reaching any understanding of the causes. It is now clearthat the CNO element group as well as Na, Al, and now Mg, have large and mostly correlatedvariations among giant stars of many clusters. The average abundances of each of theseelements also vary from cluster to cluster. All of these abundance variations are probably due toa complex set of several (often coupled) proton capture chains that operate in the hydrogenburning shells of low mass, low metallicity stars as they evolve up the first ascent red giantbranch. These abundance variations in many globular cluster giants imply envelope-shellmixing that extends far beyond the expectations of ordinary "first dredge up" from the CN-cycle. The CNO reactions deplete C and O and enhance N. At the same time, Na manufacturedfrom (unobservable) Ne and Mg is reduced as it is turned into Al.

An effective study of the C-N-O-Na-Mg-Al abundance variations requires the acquisition ofhigh resolution spectra of many stars in several globular clusters. A series of papers by R. Kraft(Lick Obs.), C. Sneden (U.of Texas), and E. Langer (Colorado Coll.) used Lick 3-m Hamiltonechelle spectra to study samples (~ 10) of red-giant-branch tip stars in six clusters. But eventhose sample sizes do not allow coverage of stars in different evolutionary states of theseclusters, so these investigators have joined with C. Pilachowski (KPNO) to acquire spectra ofnot ~ 10 but ~ 100 evolved stars in several clusters, employing the WIYN Hydra multi-objectfiber positioner and bench spectrograph. Their study of 130 M13 giants shows that Na and Mg

22

abundances vary by large amounts from star to star in Ml3, and that the variations occur in thefaintest stars on the red giant branch.

For Na in Ml3, there is a very large star to star abundance scatter among the lower luminosityred giant branch stars. The Na abundances are nearly always large among stars of the red giantbranch tip. But the Na abundances are smaller in the (presumably more evolved) asymptoticgiant branch stars. These Na differences as a function of evolutionary state obviously argue forevolutionary changes in envelope compositions of elements that are affected by proton capturesynthesis chains. These changes require deep mixing of substantial amounts of envelopematerial through the region of the hydrogen-burning interior shells of the stars that are observed(as opposed to acquisition of the abundance peculiarities at stellar birth). The giants withhighest Na contents probably also have enhanced envelope He contents.

All of these observations of M13 thus imply that our understanding of the interior structure anddynamics of these stars is incomplete, and that more sophisticated theoretical modeling isrequired, especially concerning the role of convection. These insights into the interiornucleosynthesis in giants of Ml3 could only have been accomplished with the multi-fiberHydra system. The next step is a parallel study of M3, a cluster that has a metallicitycomparable to Ml3 but whose red giant tip stars show milder star to star abundance variations.In addition similar observations are needed of the more metal-poor cluster pair M92 and M15in order to investigate how the stellar envelope mixing efficiency changes with clustermetallicity. This and similar studies in the area of chemical evolution would benefit in a trulysignificant way from the use of a multi-object high resolution spectrometer coupled to an 8-mtelescope. With the current generation of telescopes, it is possible only to study the evolution ofelement abundances in stars no fainter than those of the cluster horizontal branches. Clearly thefinal story will not be written in this field until large samples of lower luminosity giants andsubgiants can be subjected to the same type of spectroscopic scrutiny as we can now give thebrighter cluster stars.

IV. NIGHTTIME ASTRONOMY: THE PROGRAM

A. NOAO Nighttime Instrumentation

Milestones for Instrument Completion:1998 Hydra for CTIO

SQIID Upgrade

1999 Gemini Near-IR SpectrographCCD Mosaic 2

COB Upgrade

2000 Wide-field IR ImagerWIYN Adaptive Optics ImagerPhoenix Upgrade

2001 Near-IR Spectrograph CloneHigh-Throughput Optical Spectrograph

2002 KPNO IR SpectrographSOAR Instrumentation

23

1. Scientific Planning for Instrumentation Development

NOAO must provide the national community of users with facilities that enable them tocarry out observations competitive on an international level. That charge alone is sufficientto compel a substantial ongoing investment in new and upgraded focal planeinstrumentation. Advances in detector performance, delivered image quality, and in userexpectation all drive the development of new capabilities. The international competitionalso forces an extremely focused investment. Large instruments produced recently byNOAO have required some 20 FTE (work-years) of effort and the order of $500K incapital. For comparison, the submillimeter instrument SCUBA, produced and currently incommissioning by the Royal Observatory Edinburgh, required 75 FTE and over $1 millionin capital.

A focused program of investment in instrumentation therefore demands that we play to thestrengths of the telescopes and sites, exploit new detector and optical technology tomaximum advantage, and manage existing resources with extreme diligence.

The existing facilities at CTIO and KPNO will provide observations that support andcomplement observations with the Gemini telescopes. We intend to exploit the one-degreefields of view of the 4-m telescopes for deep optical imaging and fiber-coupledspectroscopy. The new technology telescopes, WIYN and SOAR, will deliver imagequality nearly undegraded from that provided by the atmosphere. These telescopes cansupport rapid image compensation systems through upgrades to adaptive optics andspectroscopy with higher angular resolution than offered by the older 4-m telescopes. Theproblems of star formation and evolution of galaxies at intermediate and high redshiftsdemand supporting data in the near-ER. Both 4-m telescopes and SOAR will be equippedwith imagers and spectrographs to tackle these challenging IR investigations. The long-term goal for the plan period is to construct a 2.4-m telescope for each site. Thosetelescopes will provide the imaging surveys in the optical and near-infrared required toisolate samples suitable for study with 8-m and 10-m class telescopes.

Rapid advances in both optical and near-ER detector technology have created opportunitiesand strong user demands for new instrumentation. NOAO will continue to exploit therevolution in near-ER indium antimonide arrays that it has helped create by deploying asuite of new and upgraded instruments designed around IK x IK and larger formats. Thesewill include multi-color and wide-field imagers, and two new slit spectrographs. Thedevelopment of 2K x 4K, three-side buttable CCDs, driven in part by NOAO and itsGemini partners, enables very large format optical imaging. Mosaic imagers with 8Ksquare format will be deployed for the prime foci of both 4-m telescopes, and mini-mosaicswill be created for WIYN and SOAR. NOAO is developing new array controller hardwareand software for Gemini that will be adapted for NOAO instruments as well. The Geminitelescopes will take on the proposals based on multi-object optical spectroscopy of thefaintest objects. The niche for the national 4-m telescopes will be fiber-coupled wide-fieldspectroscopy, and very high efficiency limited-field beam-fed spectroscopy. Newholographic grating technology holds the promise of offering very high efficiencies for newinstruments for SOAR and at Kitt Peak.

Are the instrumentation resources being tightly managed in a manner satisfactory to bothsites? The Instrument Projects Advisory Committee (EPAC) provides scientificprioritization to the Instrument Projects Group. Its current members are Taft Armandroff

24

(NOAO-Opt. Instr.), Todd Boroson (USGP), Dave DeYoung (KPNO), Jay Elias (GNERSProject), Richard Green (NOAO-Chair), Brooke Gregory (CTIO), and Bob Schommer(CTIO). EPAC meets with the instrumentation engineering managers once a month toreview priorities, schedules, and budgets. They develop the scientific content of the long-range plan, on the basis of input from the users through the Users' Committee and personalcontact, the WIYN and SOAR partners, and the Gemini advisory structure through theUSGP. Every instrument under development has an instrument scientist from the NOAOscientific staff. We believe that this arrangement is essential for successful development—each instrument must have an intellectual champion to see that the project meets itsscientific performance goals.

EPAC provides a venue where the interests of each site are fairly represented. Concurrentwith the establishment of EPAC, we put into place a revised system of project management.It placed greater emphasis on initial design and planning and on overall accountability andresource tracking. Phoenix, the high-resolution near-infrared spectrograph, represented atransition case. The fabrication was managed quite closely, but the commissioning efforttook longer than anticipated, in large part because of the problems associated withdeploying the first ALADDIN arrays. Phoenix is now in successful operation on thetelescope.

The CCD Mosaic imager enters shared risk use in the Spring semester of 1997. It is anexcellent example of the current project approach. The imager hardware design phase wasextensive, with strict external review. The hardware was completed on schedule andfunctioned fully in every major aspect on its first engineering run at the telescope.Innovative design aspects included an articulated filter transport and a pneumatically drivenblade shutter for high-precision exposure timing over the large array format. Themultiplexed Arcon controller was produced by CTIO in this fully collaborative effort. NeilGaughan, the Tucson EPG Program Engineer, worked with the CTIO technical team todevelop project planning tools and methods for resource tracking.

The specific instruments included in the current plan, with start and completion dates withtheir total direct costs, are listed in Table 1. To estimate a true cost for comparisons withcosts of instruments built in universities, an average ETS labor cost rate has been applied tothe manpower estimate. For a more direct comparison with any particular institution, anappropriate overhead rate must be applied as well.

25

Table 1

Major InstrumentsProject

Scientist

Start

Date

CompletionDate

ManpowerMM

Capital($K)

Labor

($K)

Total

Cost

($K)High Throughput Spectrograph Armandroff 1998 2000 125 171 725 896

Mosaic II Armandroff 1997 1999 136 139 789 928

HYDRA for CTIO Barden 1997 1999 102 160 592 752

SOAR Instrumentation Green 2000 2003 164 262 951 1213

CTIO Med Res Spectrometer Elias 1999 2002 276 665 1601 2266

New Start Green 2002 2004 107 300 621 921

SQIID Upgrade Merrill 1997 1998 50 80 290 370

WFOV IR Imager Merrill 1997 2000 107 260 621 881

KPNO IR Spectrometer Joyce 2001 2003 135 350 783 1133

IR Lab Array Controller Merrill 1997 1998 25 100 145 245

The instrumentation group will also be responsible for the development of arrays for bothoptical and infrared astronomy and for the associated control electronics. In addition, thestaff responsible for instrumentation upgrades and for maintenance of instruments anddetectors at Kitt Peak have been assigned to the instrumentation group. The total staffingfor upgrades and improvements at KPNO and CTIO is the same. Table 2 shows the yearlycosts of the four components of the instrumentation group program: operations andmaintenance of already completed equipment now in use primarily at KPNO; upgrades toKPNO instruments; R&D of detectors; and major instruments.

2. Infrared Instrumentation

A comprehensive approach to the development, upgrading, and deployment of infraredinstrumentation for the nighttime telescopes has been defined by CTIO and KPNO. Thelong-term goal is to provide substantial imaging and spectroscopic capabilities for bothsites, based on the anticipated production of large-format InSb arrays. The steps to that goalinclude upgrades of existing instruments and development of new spectrographs for bothsites. A collaboration of Hughes Santa Barbara Research, NOAO, and the US NavalObservatory (USNO), known as the ALADDIN project, has been formed to provide the1024 x 1024 InSb arrays necessary for the next generation of IR instruments and upgrades.

At the present time, the ER instruments deployed at CTIO are an ER Spectrometer (256 x256 InSb) and an ER Imager (256 x 256 HgCdTe), as well as the Cryogenic Optical Bench(512 x 512 ALADDIN InSb). The ER instruments at KPNO are Phoenix, the high resolutionnear-ER spectrograph with a 512 x 1024 ALADDIN InSb array, CRSP (256 x 256 InSb),EREM (256 x 256 HgCdTe), and the Ohio State/NOAO Imaging Spectrograph, a sharedinstrument with a 512 x 1024 ALADDIN InSb array.

By the end of the plan period, most of these instruments will either be replaced or upgradedwith better detectors so that both existing observatory sites will have state-of-the-artimaging and spectroscopic capabilities. CTIO will have the clone of the Gemini Near-ERSpectrograph to share between SOAR and Gemini South, and will support imaging with theupgraded SQIID (simultaneous four-color imager). The Cryogenic Optical Bench will be in

26

IPAC 5 YEAR PLAN

FY-1998 FY-1999 FY-2000 FY-2001 FY-2002 TOTAL

Manpower Capital Manpower Capital Manpower Capital Manpower Capital Manpower Capital Manpower CapitalPROJECT PI (Months) ($K> (Months) ($K) (Months) ($K) (Months) ($K) (Months) ($K) (Months) ($K)

Operations and MaintenanceInfrared Merrill 26 50 26 50 26 50 26 50 26 50 130 250

O/UV Armandroff 21 20 21 20 21 20 20 20 20 20 103 100

subtotal 47 70 47 70 47 70 46 70 46 70 233 350

KP UpgradesArcon Upgrades 18.5 7 4 2 22.5 9

KPNO Support 15.5 15 10 10 10 60.5 0

Image Improvements Jacoby 20 10 30 50 10

4m Coude Spectrograph Barden 10 20 20 50 0

Gold Cam II Jannuzi 10 5 10 5

WIYN Instrument Upgrades Jacoby 16 20 41 60 41 60 10 108 140

subtotal 60 32 60 62 61 60 60 10 60 0 301 164

R&D

IRR&D Gatley 33 48 33 49 33 49 33 49 33 49 165 244

O/UV CCD Development Armandroff 27 20 27 20 27 20 27 20 27 20 135 100

New Technology (fibers) Barden 2 1 2 1 2 1 2 1 2 1 10 5

Detector Purchases Green 300 300 300 300 300 0 1500

subtotal 62 369 62 370 62 370 62 370 62 370 310 1849

Instrument ProjectsCCD Mosaic Imager Boroson 4 5 4 5

High Throughput Spectrograph Armandroff 7 1 58 90 60 80 125 171

Mosaic II Armandroff 60 60 4 59 64 119

Mosaic II Controllers 60 0 60

HYDRA for CTIO Barden 45 29 3 9 48 38

SOAR Instrumentation Green 9 55 62 60 102 124 164

CTIO Med Res Spectrometer Elias 53 75 76 335 50 175 40 219 585

CTIO MRS Controller 47 80 47 80

New Start Green 47 47 0

Phoenix Spectrometer Hinkle 20 40 20 40

COB Upgrade Probst 20 40 20 40

SQIID Upgrade Merrill 30 37 30 37

WFOV IR Imager Merrill 10 58 260 30 98 260

KPNO IR Spectrometer Joyce 55 60 60 175 115 235

Aladdin Controller Electronics Merrill 15 125 10 25 10 12 35 162

Gemini IR Array Controllers Merrill 20 20 0

IR Lab Array Controller Merrill 5 5 0

Solar IR Array Controller Merrill 10 10 0

OSU Instruments Green 235 0 235

subtotal 206 552 206 558 205 467 207 377 207 277 1031 2231

Total IPG Labor/Capital Required 375 1023 375 1060 375 967 375 827 375 717 1875 4594

IPG Budget 375 417 375 417 375 417 375 417 375 417 1875 2085

Capital Recovery From KPNO 32 62 60 10 0 164

Capital Recovery From Gemini 224 281 190 100 795

Capital Recovery From Solar 50 50

TOTAL 375 723 375 760 375 667 375 527 375 417 1875 3094

use to commission the SOAR telescope. Kitt Peak will have a wide field of view near-ERimager for the Mayall 4-m and the new 2.4-m telescopes, and a spectrograph produced inpartnership with Ohio State University. A new spectroscopic capability will be underdevelopment for the Mayall or for WIYN, and near deployment by the end of the planperiod. This complement of five advanced facility ER instruments has been selected in orderto provide powerful and balanced capabilities to users of our nationally supported facilitiesin both hemispheres.

In addition to instruments for CTIO and KPNO the ER instrument team is responsible forthe production of the near-infrared arrays and controllers for the two initial Gemini near-ERinstruments, the imager and spectrograph. The arrays are based on the ALADDINdevelopment program. NOAO's responsibility is to manage Gemini's foundry production atSBRC. The development of the new array controller presents both hardware and softwarechallenges, and is strongly schedule driven. The first controller will be delivered to theimager project at the University of Hawaii prior to the start of FY 1998. The secondcontroller will be provided for the spectrograph in FY 1998. The Gemini controller designwill be used as the basis for all the planned NOAO near-ER instruments, allowinguniformity in support of Gemini operations in the South and both NOAO sites.

The major instrument under production for Gemini is the Near-ER Spectrograph (GNERS).This project is the largest ER instrument ever undertaken by NOAO. The dewar will be 2meters in length, and the instrument weighs some 2000 kg. It will provide long-slitcapabilities with a range of dispersions through selectable gratings, and two pixel scales,covering the wavelength region from 1 to 5 microns with four cameras addressing a single1024 square ALADDEN-type InSb detector. Fabrication begins in early FY 1998, withdelivery to the Mauna Kea site planned for late calendar 1999.

Spectroscopy at CTIO will be carried out on the new long-slit and cross-dispersed ERSpectrograph to be built later in the plan period. The CTIO spectrograph will be acustomized version of the GNERS. The compatibility of the CTIO spectrograph withGemini protocols will allow it to be shared with Gemini South, and it has the strongendorsement of the Gemini ER Instrument Science Working Group. This ERS will beaccommodated on the new SOAR telescope, since its weight and volume exceed thecapacity of the Blanco 4-m. The Cryogenic Optical Bench serves as a high-resolutionimager, with 0.1" pixels matched to the image quality of the Blanco 4-m with tip/tiltcorrection. COB will be upgraded (with partial support from Gemini) to an InSb detectorwith more than one working quadrant and a Gemini-style controller in FY 1999. NOAO hasarranged with Gemini that COB will be available for somewhat more than one year as thecommissioning imager for Gemini South, starting in mid calendar 2000. At that point,CTIO will require a replacement imager for its users.

Three new infrared instruments will be developed for imaging at CTIO and imaging andspectroscopy at KPNO. The scientific staff is currently exploring concept designs andformulating the details of the integrated plan. The development path presented in Table 2 isthe best projection for the second half of the plan period, but may be modified on furtherconsideration. A high priority for the users, as expressed through the Users' Committee, iswide field near-ER imaging. The first realization of that capability will be in the upgrade ofSQEED, the four-color imager. The 256 square PtSi arrays will be upgraded to 512 squareInSb arrays with a customized NOAO/Gemini Controller. That instrument will be availablefor shared risk user observations on Kitt Peak in fall semester 1998.

28

The next instrument may be produced in collaboration with the astronomy department ofthe Ohio State University. They may be able to modify their existing dewar/optical benchdesign to accommodate spectroscopy with a resolution near 5000 and a slit widthappropriate to the Mayall 4-m image scale. If that is the case, the cost recovery forproducing instruments for Gemini will be applied to this collaboration. NOAO wouldsupply the ALADDIN detector and probably the controller. Deployment would be plannedfor fall semester 1999.

To meet the needs for wider field imaging, the major instrument produced within the EPGwill be an imager that addresses a large-format near-ER detector, probably 2K x 2K. Thatimager will be used on both the Mayall 4-m and the new 2.4-m imaging telescope. It islisted in Table 2 as WFOV ER imager, and is scheduled to be completed in FY 2000. Thedeployment of that imager is part of the integrated plan in supporting both observatoriesand Gemini. At that time, SQEED can be sent to CTIO to support near-ER imaging programswhile COB is being used to commission Gemini South.

The last ER instrument in queue is listed as a KPNO ER spectrometer, planned to be startedin FY 2001. That instrument will address one of two needs: a more substantial approach tomoderate resolution spectroscopy from 1 to 5 microns, or the provision of near-ERrecording capability for the multi-fiber output of the WIYN Hydra positioner. The latterconfiguration would be unique in covering a large area of sky spectroscopically atwavelengths between 1 and 2 microns. The surface densities of objects discovered in the2MASS and Sloan Digital Sky Surveys will be an appropriate match for such aninstrument.

3. Optical/UV Instrumentation

The driving principle of the O/UV program is to develop capabilities for CTIO and KPNOthat will complement the strengths of optical observing on the Gemini telescopes. Currentplans concentrate on exploiting the wide-field of the 4-m telescopes. Individual instrumentsare described below.

Hydra CTIO. The next project in the queue is to rebuild the Hydra positioner for the CTIO4-m telescope. In the area of wide-field multi-fiber spectroscopy, the existing 4-mtelescopes can and should complement the capabilities of the Gemini telescopes. Thepresent multi-fiber instrument at CTIO is Argus, which was designed almost a decade ago.It is capable of observing 24 objects at a time, which puts it at a factor of four disadvantagecompared with more modern instruments such as KPNO's Hydra. The new multi-fibersystem is being built in Tucson and is based largely on the design of Hydra as it wasconverted from use at the Mayall to use at the WIYN. New motor controllers, a newgripper, and new fiber cables are being produced for the system for CTIO, along with a newwide-field corrector with atmospheric dispersion compensating prisms. This instrumentwill be located at the f/8 Ritchey-Chretien focus and will take advantage of the excellentimage quality by using fibers of smaller diameters than those of Argus. Delivery and firstcommissioning activities are planned for October 1998. The new motors and motorcontrollers will provide a much more rapid positioning time than is currently achieved withHydra on WIYN. The new technology will then be retrofitted as an improvement to WIYNHydra, as a WIYN instrument upgrade.

29

Mosaic Imager 2. The major instrument currently in commissioning is the CCD mosaicimager, to be deployed at KPNO, then cloned for CTIO. EPG has produced an imager with8096 x 8096 format that has an active imaging area of over 12-cm on a side. The Mosaiccurrently contains eight 2K x 4K three-side buttable CCDs from SETe, but they areengineering grade. Scientific grade CCDs have been ordered for two Mosaic imagers,through a consortium purchase with Carnegie Observatories. The controller is amultiplexed quadruple version of the ARCON, developed and produced at CTIO. The firstMosaic Imager will be made available for shared risk observing on KPNO telescopes inJune. The second Mosaic is scheduled for completion in FY 1999, with deployment atCTIO in the first half of the calendar year.

Adaptive Optics Imager. An extremely high priority for the entire WIYN partnership is theimplementation of a low-order, natural guide star adaptive optics imager for the WIYNtelescope. A general description of the capability is given in the KPNO section of the plan.NOAO's participation in the project is listed in Table 2 under WIYN instrument upgrades.

New spectrographs. Long-term studies are currently underway to define a new generationspectrograph for the 4-m telescopes. Scientific performance tradeoffs are being investigatedto identify the most effective combination of field of view, spectral dispersion, andwavelength coverage. A key goal is to use the new generation of large-format CCDs withsmaller slits and adequate pixel sampling, in order to exploit the expected improvement indelivered image quality. In addition, the use of holographic volume phase gratings mayallow a significant increase in throughput. Brazil has expressed significant interest incollaborating with NOAO in exploring such a design for use on SOAR. The spectrographthat results from these considerations is planned as the next major instrument to follow theWIYN imager upgrade to a mini-Mosaic of two 2K x 4K SETe CCDs.

4-m Coude Spectrograph. To maintain the capability for very high resolution stellarspectroscopy, KPNO must find a way to continue use of the coude spectrograph, currentlyat the 2.1-m/Feed. To reduce operations stress because of the long-term reduction inpersonnel, the Feed telescope itself is planned for closure. To open a significantly largervolume of space accessible for observation, fainter limiting magnitudes could be achievedby moving the coude spectrograph to the 4-m. It could be fed by a simple insertable fiber,to achieve high stability as well as high resolution. A new image sheer must be designed torealize high signal to noise with respect to the narrow physical slit required for the highestresolutions. This project is currently under study, and will be implemented through theportion of the IPG that is committed to support of KPNO instrument upgrades.

B. Telescopes and Upgrades

1. Initiation of Adaptive Optics at Kitt Peak

In order to select the approach to be taken in providing adaptive optics systems for existingtelescopes, one must consider three very different field-of-view domains with differenttechnical requirements and different scientific applications: 1) diffraction-limited imagesover a few arcseconds achieved by high-order corrections; 2) sub-half arcsecond imagesover 3-5 arcminutes achieved by relatively low-order corrections; and 3) uncorrectedimages over as wide a field-of-view as the telescope-detector combination delivers. This

30

third capability, imaging over wide fields-of-view with detectors of the highest possiblequantum efficiency, has traditionally been one of the strengths of NOAO/KPNO, and thenew 8K x 8K CCD mosaic imager will keep the observatory at the cutting edge withrespect to this type of observation.

The detailed specifications for an AO system are currently under discussion with theWIYN partners. However, NOAO believes that the most useful scientific advance would beto achieve 0.3 arcsec images in the near-ER and sub-half arcsec images in the blue overfields of view of 3-5 arcminutes, with the AO system then being coupled to imagers andpossibly spectrometers. Such an AO system would be designed to make partial correctionsfor wavefront errors introduced by the atmosphere, focus variations of a few tens of Hz,high-speed guiding errors, and residual telescope aberrations. Initially, we would like todevelop the AO system for the WIYN telescope, which now produces the finest images onKitt Peak, but such a system would be useful at the Mayall 4-m as well.

The likely success of this approach has already been demonstrated by several recentexperiments. Tests at the KPNO 2.1-m telescope employing an active secondary and at theKPNO 4-m using real-time shift-and-add confirmed that tilt correction alone can achievediffraction-limited imaging in the near infrared. This work led to the deployment of a fasttip-tilt ER secondary at the CTIO Blanco 4-m telescope, which is currently in thecommissioning process. Recent tests of a simple tip/tilt mirror system provided by SantaBarbara Instrument Group demonstrated improvements in the Strehl ratio in the R-band atWIYN of about 60 percent with coherence at the 85 percent level over 80 arcsec. Thefrequency of image motion power is mostly below 10 Hz, requiring a system bandwidthgreater than 100 Hz for full compensation. The "HR Cam" and PUEO instruments at CFHTconfirm the effectiveness of low-order correction and have proven to be very user-friendlyinstruments that have yielded significant scientific results.

There are no current plans for installation at Kitt Peak of a high order system, which wouldrequire laser guide stars. Recent experience has somewhat dashed expectations because ofthe complexity of high-order AO systems, the limited scientific results achieved to date,and the high cost of operation. The US is, however, likely to be assigned the responsibilityfor delivering laser guide star adaptive optics to Gemini South. Through its management ofthe procurement process, NOAO will have the opportunity to assess what is available in thecommunity and determine whether and how to proceed with such a system at one or moreof its existing telescopes.

2. KPNO Telescope Improvements

Telescope improvements projects have the dual goals of enhancing performance, especiallyimage quality, and reducing maintenance costs.

Milestones for Telescope Operations

1998 Complete ventilation of 4-m Mayall domeVentilate volume above 4-m primary mirrorRe-submit proposal for 2.4-m telescopesWithdraw from joint operating agreement for Burrell-Schmidt

31

1999 Encode f/8 secondary motion at 4-mComplete control of remaining thermal sources at 4-mComplete wavefront analysis systemInstall active primary mirror support at 4-m if requiredInstall mirror heating system at 4-m if feasibleComplete WIYN improvements program

2000 Complete 2.4-m telescopeWithdraw NOAO support for 2.1-m and 0.9-m after 2.4-m completedTransfer coude spectrograph to 4-mInstall low order natural guide star AO system at WIYN

2001 Install new telescope control system at 4-m

2002 Consolidate telescope operations in single control center

a. Mayall 4-m

The primary goal of the work planned for the Mayall 4-m telescope is to achieve sitelimited image quality as much of the time as possible with this 25-year-old telescope.Smaller average image size makes it possible to detect fainter objects, to observe equalbrightness objects in shorter exposure times, and to achieve better angular resolution.There are two reasons to believe that consistently better image quality is attainable.First, WIYN has convincingly demonstrated that Kitt Peak is a much betterastronomical site than previously thought. Second, the Blanco 4-m at CTIO, which hasthe same design as the Mayall, now routinely averages sub-arcsecond images.

A major project to improve image quality at the Mayall was initiated three years ago,and will require an additional three years to complete. The goal is to achieve a mediandelivered image quality of 0.9 arcsec. (This should be compared with the WIYN whichhas a median delivered image quality of 0.8 arcsec, with 25% of the time better than0.7 arcsec and 10% better than 0.6 arcsec.) We already know that excellent imaging ispossible because 0.6-0.7 arcsec images have been seen when the conditions at the 4-m,and especially the thermal characteristics of the mirror and enclosure, were just right.

The next major tasks will be ventilation of the dome (installation to begin in summer1997 and to be completed in FY 1998), ventilation of the volume immediately abovethe mirror (FY 1998), control of the remaining major thermal sources (FY 1998-FY 1999), encoding of the f/8 secondary motion, and completion of the wavefrontanalysis system (FY 1998-FY 1999). We will also investigate the necessity of installingactive primary mirror supports. (If required, this work will be done in FY 1998 or FY1999.) We continue to follow the development of the mirror heating system by Gemini,and, if found feasible for the 4-m mirror, a similar system will be installed in FY 1998or FY 1999.

b. WIYN3.5-m

NOAO's goal for WIYN is to ensure that it continues to achieve high image quality andreliable operation. Our experience during the first year of science operations (FY 1996)

32

identified areas where the safety of the telescope and of personnel working in thetelescope environment could be improved and where the performance of the telescopeand observing efficiency could be enhanced. The four partners of the WIYNObservatory have approved a two-year special assessment totaling $400K to addressthese issues.

c. 2.1-m

No work is planned at the 2.1-m pending approval of construction of a new 2.4-m.

d. 0.9-m

The lifetime of the 0.9-m will determine whether any improvements are made to it. If itappears that this telescope will remain in operation for more than three additionalyears, as it may have to if the 2.4-m is not built in a timely fashion, a new telescopecontrol system will be required in order to improve reliability and lower maintenancecosts. The current system makes use of FORTH and very old DEC computers. It islikely that we would purchase a commercial TCS rather than attempt to build one in-house.

e. Coude Feed

No improvements are planned for this telescope. We are in the early stages ofexamining the feasibility of moving the optics of the coude spectrograph to the 4-mtelescope to enable spectroscopy at resolution greater than 200,000 to fainter limitingmagnitudes than are now possible.

f. Projects impacting on scientific operations at all telescopes

• Modern software architecture Telescope Control System for the Mayall 4-m andother telescopes in operation.

The TCS at the Mayall 4-m is a derivative of the original KPNO FORTH-basedtelescope control code and has neither the modern message passing system of theWIYN TCS or the data-base TCS of Gemini. The TCS that is developed for thenew 2.4-m telescopes will be based on a software design which includes many ofthe aspects of the Gemini telescopes. This design will be ported to the 4-m andother telescopes in operation on the mountain so that we maintain compatibility ofTCS throughout the mountain for ease of maintenance and upgrades. Theanticipated time scale is to initiate planning and design in FY 1999, withcompletion in FY 2001.

• Infrastructure to support the next generation of instrumentation control.

The next generation of instruments will use a control philosophy based on theGemini design with control computers located immediately adjacent to theinstrument for high-speed access. This design impacts on our instrumentinfrastructure in three ways:

33

1) Necessity to provide a cooling source to draw off approximately 1KW of heat(present solution is to provide—for example, at the 4-m—cold glycol to theCassegrain cage).

2) High-speed, high-bandwidth networking capability.

3) Epics layer to the TCS software for compatibility with Gemini-styleinstrumentation.

Design of these improvements will begin in FY 1998 and will be completed in timefor deployment of these new instruments (FY 1999 at the earliest).

Replace telescope guider electronics.

The leaky memory guide electronics—integrators with time-weighted history tofeed correction pulses to the telescope servo control—are based on commercialcircuit-boards which are no longer available and cannot be readily repaired at thecomponent level. The leaky at the 4-m will be replaced in FY 1998 with acommercial frame-grabber unit similar to that at WIYN. Other operating telescopeswill be reviewed at that time and their leaky memories will be replaced withidentical units depending on the projected lifetime of the telescope.

Update data computers to multi-processor units.

The generation of computers used currently for data acquisition were installed in1991. While we have upgraded them with increased memory and disk space and, insome cases, faster processors, we must provide significantly increased computingcapacity for the Mosaic CCD imager and other future high data-rate instruments. Inaddition, we must provide computers at the telescope which are at least as powerfulas astronomers have on their desktops at their home institutions. Current computertechnology uses multi-processors to provide increased computing power. Becausedevice drivers would have to be re-written for the operating systems provided withthe multi-processor machines, this upgrade is not a simple replacement.Investigation of the scope of this improvement will begin in FY 1998 withimplementation in the years to follow.

The year 2000.

The year 2000 has unknown consequences to the computer systems on themountain. So that operations do not grind to a halt on midnight of 1 January 2000or put us back to 1900, we plan in FY 1998 to determine all of the centurydependencies in our software and to make changes in early FY 1999 before themillennium arrives.

Investigation of consolidation of all control rooms and technical shops in onebuilding.

A modern multi-telescope observatory where telescopes are designed for remoteoperation could have a single control facility to consolidate operations and

34

technical staff with the astronomers as they observe and reduce their observations.As the older telescopes on Kitt Peak are closed and replaced with modern facilities,such consolidation may be feasible on Kitt Peak. In conjunction with the design ofthe new 2.4-m we plan to investigate this operations philosophy. We mustunderstand what this style would entail for the Mayall 4-m and WEYN. Designstudy will proceed in concert with that of the 2.4-m (FY 1998 and beyond).

3. Plans for Observing Facilities at CTIO

Introduction

CTIO in the next century will consist of a complementary system of larger telescopesalongside individual smaller facilities, the latter being set up and financed by US universitygroups and others. Gemini South and these smaller facilities will contribute to, and benefitfrom, the overall AURA/NSF infrastructure in Chile. The inventory of telescopes isexpected to be:

Gemini South, 8-m

SOAR 4-m

Blanco 4-m

O/ER survey telescope 2.4-mSmaller support facilities for synoptic use, initial surveys, etc.

Milestones for Telescope Operations

1998 Complete conceptual design study for SOAR

1999 Complete 2MASS southern survey

2000 First light Gemini South

2001 Install 2.4-m telescope

2000 Gemini science operations beginFirst light for SOAR

a. Gemini South 8-m Telescope

CTIO will give its highest priority to supporting the run-up to full Gemini operations.This will include the signing of an MOU, which will govern the way in which Geminiwill operate in Chile and define how the existing infrastructure will be shared.

In the first years of the long range plan period, CTIO expects to provide, on a cost-reimbursable basis:

• Logistics and administrative support for completion of the Gemini South buildingand dome.

35

• Assistance with the transfer of AURA Chilean engineering and technical staff fromCTIO to Hawaii to work on commissioning Gemini North, followed by a return towork in Chile.

• Scientific, technical, and logistical assistance with characterization of theatmosphere over Cerro Pachon.

• Publishing SCEDAR measurements made using the 1.5-m telescope on CerroTololo.

• Representation of AURA (including Gemini) in Chile.

• Support for installation, testing, and commissioning of the Gemini SouthTelescope.

• Scientific and technical assistance with implementation of Gemini'sCommunications Plan for Chile.

• Support for operations of Gemini South.

b. SOAR 4-m Telescope

CTIO is currently working with partners in NOAO/Tucson, Brazil, the University ofNorth Carolina, and Michigan State University to complete the definition phase of theSOAR project. Here again, CTIO staff are involved in many activities similar to thoselisted above for Gemini; in addition, CTIO will itself operate the SOAR telescope forthe consortium. More information about SOAR is given in Section II.B.l.

c. The Blanco 4-m Telescope

At the beginning of the planning period, this telescope will still be the largest operatingroutinely in the Southern Hemisphere. By the end of the period, at least sevensignificantly larger telescopes will be in operation in Chile alone.

The project to improve the image quality at the Blanco telescope has been essentiallycompleted. During the initial part of the period covered by the current long range plan,the emphasis at CTIO will be to complete the installation and commissioning/recommissioning of the various instruments being built in Tucson for use on the BlancoTelescope, in particular the MOSAIC 8K x 8K CCD imager, the Hydra multifiberspectrograph, and the Phoenix high-resolution ER spectrometer.

The first two instruments will fully exploit the wide-field imaging capabilities of theBlanco telescope and permit some deep survey work in preparation for the arrival ofGemini South.

d. The 2.4-m O/IR and Optical Survey Telescope(s)

Gemini South will easily be able to carry out spectroscopy to magnitude K = 20. The2MASS survey is based on an all-sky survey with 1.3-m telescopes (one of which is on

36

Cerro Tololo), but will reach only to magnitude K = 14.5. The need for a surveycapability in both hemispheres appropriately matched to Gemini's needs is obvious.Given the special ER capabilities of Gemini and the effort by ESO to provide southernhemisphere optical survey capabilities for the VLT, we feel it is wise to start first withthe installation of the ER survey telescope. This can perhaps be followed by anextremely wide-field optical survey telescope in Chile. We expect to firm up our long-range plans for telescopes supporting Gemini following national and internationalworkshops on this subject later this year.

e. The Older, Smaller Telescopes

The smaller telescopes currently in operation will be replaced by the 2.4-m surveytelescope(s). Because CTIO offers unique access to the southern hemisphere for mostof the US community, we are endeavoring to find creative ways to operate thosetelescopes that the NSF can no longer afford to support. We have already taken steps tolower operations costs; all telescopes smaller than 1.5-m on Cerro Tololo are operatednow with only one, fixed instrument. We have also sold observing time; fifteen percentof the 0.9-m is used for MACHO follow-up studies, financed through the MACHOprogram. This funding allowed the performance of the 0.9-m to be substantiallyimproved.

As another example of these ad hoc, creative initiatives, the 1-m telescope (which wasclosed to general users in January 1997 and is currently being upgraded with fundsfrom consortium members) will be operated for the first three years of the long rangeplan period by a consortium of Yale University (30%), Ohio State University (30%),The University of Lisbon, Portugal (30%), and NOAO (10%). This telescope will beoperated in service mode, exclusively for synoptic programs—a new venture inoperations strategy at CTIO. The small access retained for NOAO will provide mostNOAO users with their only opportunity for long-term studies of variability in thesouthern hemisphere.

f. New Small Telescopes

CTIO does not itself intend to build new small telescopes. However, there is increasingpressure from the community for access to such facilities in the southern hemisphere.CTIO is therefore responding to this demand by offering to facilitate, at cost, theinstallation and operation of smaller telescopes by qualified user consortia led by USuniversities and other institutions. Some examples follow:

• 2MASS

By the beginning of the period covered by this long range plan, the 2MASS all-skyER survey will already be well underway. It is intended to provide a census ofgalaxies and stars in the J, H, and K bands, complete (at high galactic latitudes)down to K = 14.3. The 2MASS telescope on Cerro Tololo complements a northernequivalent on Mount Hopkins in Arizona. Survey observations will be completedby 1999. Options for subsequent use of the telescope will be explored once themain survey gets underway.

37

• GONG

This successful, six-station, worldwide robotic network of telescopes to probe theinterior of the Sun is described in Section V.C.I.b. GONG will probably beupgraded and operated throughout a full solar cycle. Impact on the rest of CTIOoperations is likely to remain very light.

• USNO Astrometric Survey

This survey is expected to be fully underway at the start of the report period and tocontinue until mid-1999.

The United States Naval Observatory's CCD South Astrometric Catalog is toproduce a high density catalog of reference stars in the Southern Hemisphere withaccuracies of about 20 milliarcseconds at epoch for the magnitude range of about 7to 14 and extending to 16 magnitude with about 70 mas mean error.

The plan is for a two-fold overlap of exposures, preliminary reduction with Tychostars, direct link to Hipparcos stars, and link to extragalactic sources. A globalblock adjustment will be done to develop the final catalog.

The USNO catalog will have about 2000 stars per square degree, positions betterthan Tycho for 9th magnitude and fainter at epoch. The catalog will bridge themagnitude gap between Hipparcos and faint photographic surveys. It will providereference stars for faint observations. It will give reference stars for ER surveys(such as the 2MASS survey) at the same epoch. Improved proper motions will laterbe derived for all stars in the catalog when combined with other observations atother epochs.

The USNO twin 8-inch astrograph will be housed in an existing small dome on thesummit platform of Cerro Tololo.

• Robotic Networks for Research and Education

As part of an overall effort to improve access to the southern hemisphere skies,CTIO is in very preliminary conversations with two groups that are proposing toset up non-profit global networks of small, optical, robotic telescopes. The GNATconsortium wishes to set up 1-m-class telescopes, while Remote Telescopes, Inc. isplanning a 0.5-m-class network. There would be some advantages to the US usercommunity in having photometric calibration measurements taken each night onCerro Pachon and Cerro Tololo. Such networks are also likely to prove a valuableresource for education and public outreach (see Section VEE). Once the first US-based telescope in each network is functioning successfully, CTIO is prepared toconsider hosting the first overseas telescope station in each network, at cost.

38

g. New Large Telescopes

CTIO continues very preliminary conversations with the Hobby Eberley Consortiummembers, with a view, if funding can be identified, to hosting a similar, or largerfacility in Chile.

C. Center for High Angular Resolution (CHARA)

It is widely believed that the next major initiative in groundbased astronomy may be in the areaof optical interferometry. Several prototype systems, developed in the university and DODcommunities, have demonstrated the capability of optical interferometry to deliver high angularresolution. Optical interferometry was ranked highly by the Astronomy and AstrophysicsSurvey Committee (AASC), and two moderate scale interferometric facilities are now fundedand under development in the US.

The NOAO staff has a long record of participation in the development of high angularresolution techniques, including visible and infrared wavelength speckle interferometry, doubleFourier interferometry, and other more recent multi-telescope interferometry. In recent years,NOAO staff members and long-term visitors have collaborated with astronomers at theUniversity of Wyoming, University of Paris, and the Harvard-Smithsonian Center forAstrophysics in design, construction, and operation of several prototype interferometer systems,including the Infrared Michelson Array (ERMA) and the Infrared Optical Telescope Array(IOTA). The FLUOR program, carried out at Kitt Peak with NOAO support, resulted in thefirst interferometric linkage of telescopes with optical fibers. More recently, the classical andfiber infrared beam combiner systems which NOAO developed for ERMA and FLUOR, havebeen moved to IOTA. The fiber system has recently obtained the most accurate stellar angulardiameters yet recorded with optical interferometry, and the system is now in regular use forscientific observations.

During FY 1996, NOAO completed the second year of a contract with the Georgia StateUniversity Center for High Angular Resolution (CHARA) to support the design of thetelescopes for the CHARA interferometric array. The design was completed in FY 1996 and thefive 1-m aperture telescopes are now under construction. The first will be delivered in August1997 and the last should be complete in early 1998. The contract has been extended to includework on additional areas including the telescope enclosures and the optical delay system.

The AASC Interferometry Panel working papers recommended work in this decade leading tothe development of a plan for a large interferometric optical array, with construction possiblystarting towards the beginning of the next decade. With several prototype and permanentinterferometric facilities operational or in progress, discussion in the interferometry communityis now turning toward the possibility of preparing a plan for a major facility. NOAO will beprepared to join in these discussions, to support the planning process, and to collaborate in theconstruction and operation of a major interferometric array.

39

D. US Gemini Program and NOAO Science Operations

Milestones for USGP

September 1997

October 1997

June 1998

December 1998

Conduct workshop to define infrastructure required to optimize scienceproductivity of Gemini telescopes

Unify proposal submission process for KPNO and CTIO

Deliver Gemini IR controller to University of Hawaii

Deliver Gemini CCD controllers to optical instruments

Schedule open access nights at upgraded MMT

First light on Gemini North

February 1999 Schedule open access nights at HET

January 2000 Deliver near-IR spectrograph to Gemini North

March 2000 Science operations begin at Gemini North

June 2000 First light on Gemini South

June 2001 Science operations begin at Gemini South

1. Roles of the USGP

The US Gemini Program (USGP), which was established in 1993, is the fourth and newestdivision of NOAO. Its initial role was to serve as the interface between the international

Gemini Project and the US community. Recently, this division has begun to take on the roleof providing the interface to users of all of the NOAO nighttime telescopes.

During the period of time covered by this long range plan, the Gemini telescopes will makethe transition from construction to operation, and the activities of the USGP will evolve aswell. In the construction phase, the primary role of the USGP and the national projectoffices in the other partner countries is to assist in the management of work packageswithin their countries. Specifically, the USGP has taken on the responsibility formanagement oversight of the entire infrared instrumentation program for the internationalGemini Project, including work being carried out at both NOAO and other institutions. TheUSGP is also managing the effort to provide CCDs and CCD controllers for the Geminioptical instruments. Ln addition, the USGP is responsible for obtaining input on scientificrequirements and goals from the US community, and disseminating information about theproject. The USGP supports design reviews and other project activities, identifies USrepresentatives for international Gemini committees and working groups, and otherwiseassists the project.

40

In the operations phase, the national project offices will retain the above roles but also willbe responsible for supporting the national access to the Gemini telescopes. In generalterms, the USGP will be required to support all activities before and after the observingrunitself. We have decided that the US user community will be better served if we unify theinterface to all telescopes to which NOAO provides access. The USGP will therefore alsosupport national access to KPNO, CTIO, the independent observatories that supply nationalaccess time in exchange for NSF instrumentation funding, as well as to the Geminitelescopes. Specific responsibilities include the following:

• Distribute information to the US community about all national access facilities andinstrumentation.

• Supply or coordinate expert advice and support for observing proposal preparation.

• Solicit observing proposals and perform or coordinate initial technical evaluation.

• Evaluate proposals through Time Allocation Committees (peer-review) and organizerepresentative participation in TACs run by the non-NOAO organizations providingaccess to the US community.

• Establish and support a Users' Committee structure that provides effective feedback toNOAO on the entire suite of facilities to which national access is provided, andorganize participation or input to the non-NOAO organizations providing this access.

• Supply data reduction software and data reduction support to the US community for allinstruments (NOAO, Gemini, independent observatories) for which packages in theIRAF environment are desirable.

• Provide or coordinate support for observing programs through instrument specialists.This support will likely be provided from the NOAO-wide pool of astronomers andscientists.

• Establish and maintain whatever type of data archive is cost-effective. Provide orfacilitate US community access to this database.

• Investigate the feasibility and desirability of operating remote observing stations andsupport this type of access.

• Provide assistance to the community for queue-scheduled observations.

• Carry out public relations and outreach activities aimed at the astronomical community.

The following are additional responsibilities that are uniquely aimed at support of theGemini facilities:

• Coordinate and manage the US-allocated Gemini instruments. This includes supportingcommunity participation in defining new instrumental capabilities.

41

• Coordinate the maintenance support for systems provided by the US to Gemini.

2. Gemini Instruments: Construction Phase

The international Gemini agreement is designed to ensure that all the partner countriesmaintain full intellectual participation in the project. In general terms, this goal has beenachieved by assigning specific responsibilities to each of the partner countries, as describedabove. Active participation in the definition and implementation of the instrumentationprogram is a particularly important component of the responsibilities assigned to thepartner countries. According to the terms of the international agreement, work on thedesign and fabrication of the instruments will be allocated to the partner countriesapproximately in proportion to their financial participation in the project.

In accordance with the international agreement, responsibility for approximately half of theinitial complement of instruments has been allocated to the US. These US-allocatedinstruments are:

• Near-IR (1-to 5-micron) Imager• Near-IR Spectrometer• Mid-ER (8- to 30-micron) Imager• Optical detector arrays and array controllers• Near-IR Arrays• Near-IR Array Controllers

Near IR Imager. This instrument was allocated by the NSF to the University of Hawaii. Acontract between Gemini (AURA) and the University of Hawaii for design and fabricationhas been negotiated. The USGP has assumed responsibility for an assessment ofperformance against schedule and budget and is making regular reports to the internationalGemini project.

Near-IR Spectrometer. This instrument was the subject of an open competition, based onan RFP. NOAO won this competition, and refinement of the proposal design is now inprogress. The preliminary design review was held in October 1996.

Mid-IR Imager. Interest in designing and building this instrument has been expressed by anumber of university groups. The USGP ran a competition and selected two groups forpartial funding of conceptual design efforts. These and two unfunded design studies weresubmitted for review. A group will be selected through a second competition to complete adetailed design and fabricate the instrument. This approach maximizes communityinvolvement and assures Gemini of the best possible instrument.

Optical Detector Arrays and Array Controllers. This work package consists of threeparts, the acquisition of CCDs, the design/fabrication of CCD controllers, and theintegration of these two items into the cameras of the Gemini optical spectrographs. TheUSGP organized an international consortium of CCD experts to study various approachestoward acquiring the 2048 x 4096 CCDs. Based on the evaluation of proposals from threepotential vendors by this consortium and another committee of Gemini instrumentscientists, EEV was chosen to supply the optical detectors.

42

For the controllers, the USGP was faced with a situation similar to that of the Near-ERspectrometer, in that both NOAO and outside groups wished to be considered as a source.Thus, this evaluation was the first to follow the process defined at the AURA-sponsoredworkshop in Albuquerque in March 1995. Appendix A describes in some detail how thiswas undertaken.

Based on the evaluation from the external committee, which is part of the AlbuquerqueProcess, the USGP plans to adopt the following approach for providing CCD controllers toGemini:

• The required hardware will be purchased from SDSU. This choice follows directly therecommendation of the evaluation committee.

• The Royal Observatories will be asked to provide the low-level software (where theirexpertise is most relevant). NOAO will be responsible for the high-level software,which will be a modification of the software already being written for the ERcontrollers. This approach is consistent with the evaluation committee's identificationof the strong points of the various software efforts.

• NOAO will assume overall management responsibility for integrating the software andthe hardware. NOAO has already been assigned the responsibility for integrating thedetectors with the controllers.

Near-IR Arrays. The Gemini near-ER instruments are designed around 1024 x 1024 LnSbarrays. The only such device is the ALADDIN array, currently being developed at SBRCunder a contract funded jointly by NOAO and USNO. The USGP has initiated a foundryrun at SBRC aimed at fabricating a sufficient number of arrays to assure two high-qualitydetectors for the Gemini imager and spectrograph. NOAO will characterize the arrays.

Near-IR Array Controllers. NOAO and Gemini requirements for controllers are similarbecause both observatories plan to use 1024 x 1024 InSb arrays. Because of the advantagesof commonality, including the possibility of sharing ER instruments built by NOAO for itsown telescopes, the international Gemini project indicated that it wishes to use the samearray controllers as NOAO. The two builders of near-ER instruments (the University ofHawaii and NOAO) have agreed that the NOAO controllers will meet their requirements.An assessment by the USGP of other alternatives for the ER array controllers indicated thatno other single vendor could provide an end-to-end system that meets the Geminispecifications, and development of a new controller is inconsistent with the Geminischedule. Therefore, NOAO is building array controllers for the two Gemini Near-ERinstruments. In order to facilitate the integration of the controller with the Near-ER imagerat the University of Hawaii, NOAO will integrate the controller with an engineering-gradeALADDIN array in a test dewar and ship this assembly to Hawaii.

3. Gemini Instruments: Operations Phase

During the construction phase of the Gemini project, instrumentation planning hasemphasized providing a suite of basic instruments for spectroscopy in both the optical andinfrared and for infrared imaging. With the completion of this initial complement, thereshould be more opportunities to provide innovative and special purpose instruments. There

43

must also be an evolution of the basic capabilities in order to ensure that the Geminifacilities remain competitive with other large telescopes. In order to encourage input fromthe US community to the ongoing process of Gemini instrument planning, the USGPorganized a workshop to discuss priorities of future capabilities. This workshop includedparticipation by representatives of most of the other large telescope facilities. The format ofthe workshop was to split the 25 participants into four discipline-based subgroups, each ofwhich discussed the major scientific questions likely to be attacked with large telescopesover the next ten years. Programs that could answer these questions were translated intoastronomical capabilities and these were all discussed to identify areas of commonality orunique importance. This US workshop fed into an international workshop with the sameformat organized by the international Gemini project. It is planned to hold similarworkshops to review and extend the ongoing instrumentation program every two years.

The new high-priority instruments identified by these discussions include a laser guide staradaptive optics system for Gemini South; two Near-ER multi-object spectrographs, oneoptimized for a small AO-corrected field, and one for as wide a field as can be easilyaccommodated; and a Near-IR imager/coronagraph, optimized for use with the AO system.

4. Gemini Instruments: Procurement Plan

As the first generation of Gemini instruments are well into final design or fabrication, it hasbecome obvious that some substantial issues should be resolved concerning the process thatthe international project uses to procure future instruments for Gemini. In particular, it isagreed that there should be a uniform and consistent method of establishing the constraintson instruments, the performance, the cost, and the schedule. In order to address this andother procurement issues and to converge on an actual program for the ongoinginstrumentation, Gemini recently convened the first Gemini instrumentation forum. Thismeeting included the national project managers and national project scientists as well as theGemini personnel associated with the instrumentation program. The recommendations forprocurement that came out of this forum include:

• The procurement process will begin with a very rough estimate of the cost for aninstrument with a given set of scientific capabilities, generated by the Gemini project orby one or more of the national project offices. These estimates and the assumptionsused will be presented to and discussed by the forum.

National project office representatives will poll their communities to determine interestin providing any of the instruments currently on the table. They will then advocatethose interests to the forum. There will also be opportunities for the national projectoffices to propose to the forum existing instruments that could be shared with Geminias well as programs aimed at instrument-oriented technology development.

One or more conceptual designs for each instrument will be commissioned and at leastpartially funded by Gemini. Selection of the groups to perform this design work will bethrough an open, informal discussion involving all interested parties, in the presence ofthe instrumentation forum members, supplemented by unbiased experts from thepartner communities. Additional groups may elect to develop unfunded conceptualdesigns in parallel with the selected groups.

44

• Following presentation of the conceptual design studies, the instrumentation forum willdecide which, if any, of the groups it would like to fund to complete the detailed designand fabrication. The desire is to negotiate one or two fixed-price phases, based on theconceptual design work already done.

• Over the long run, instrument allocation should approximate the relative financialparticipation of the partners. However, it is recognized that in the short-term,departures from this will be necessary. It is also recognized that collaborative effortsare desirable both because the smaller partners may not have the national infrastructureallowing them to build entire instruments and because of the intellectual benefits ofsuch collaborations.

It can be seen that the process proposed by the Gemini instrumentation forum is in manyways quite similar to that implemented by the USGP following the AURA-sponsoredworkshop held in Albuquerque in March 1995 (see Appendix A). The two significantdifferences from current Gemini procedures are that instruments are not first assigned tocountries and that the members of the forum itself have put themselves in the role ofrecommending the selection from among the interested groups. The potential involvementof all countries is seen as desirable in order not to force the assignment of an instrument toa particular partner on the basis of equity as opposed to technical expertise. Therecommendation of the forum members will be forwarded to the Gemini Board.

The USGP's role in this new instrument assignment and procurement process will be muchthe same as if instruments were assigned to the US at the start. The USGP will solicit inputand expressions of interest from the community for each of the instruments that theinstrumentation forum considers. Assistance in the development of cost estimates will berequested from groups with relevant experience. The USGP may run pre-forum discussions(particularly in cases where both NOAO and outside groups have expressed interest) toidentify the few best groups and help them refine their presentations to maximize theiropportunities. After instrument design and fabrication work is assigned, the USGP willmaintain its management role for Gemini. It may also be desirable for the USGP throughNOAO to act as an integrator in order to maximize participation by smaller groups in thecommunity, or to supply special pieces of Gemini instruments, such as the software, incases where other existing groups do not have that expertise.

5. National Access to Gemini and Other Telescopes

Just as the design and construction of the Gemini telescopes rely on assistance from thenational project offices, the Gemini operations plan assumes that the partner countries willbe actively involved in providing user support. Given this requirement, it seems most usefulfor planning purposes to consider the US share of the international Gemini project assimply the newest of the NOAO telescopes. If NOAO integrates the US support of Geminiwith the analogous parts of the KPNO and CTIO operations, and adds the requirement tosupport access to the independent observatories who will provide national access inexchange for NSF instrumentation funding, NOAO can carry out this responsibility bymaking effective and appropriate use of existing NOAO resources and at the same timeminimize the impact on support of the existing telescopes. Accordingly, the USGP is nowmoving toward supporting all the pre- and post-observing run activities for all the nighttimetelescopes to which NOAO provides access. A few Gemini-specific responsibilities

45

(primarily concerning the ongoing Gemini instrumentation) remain within this NOAO-widescience operations division.

The operations-phase activities were summarized in the introduction to this section and aredetailed below.

Instrumentation Program. The USGP will be expected to continue to manage thedevelopment of those Gemini instruments that are assigned to the US. Through the Geminiinstrumentation forum and the Gemini science advisory structure, the requirements andestimated costs for elements of the ongoing instrumentation program will be defined. TheUSGP will be responsible for identifying US interests in designing and fabricating theapproved instruments.

The Proposal/TAC Process. The USGP will develop an efficient and uniform method bywhich the community can propose to use any of the facilities to which NOAO suppliesaccess. The process will facilitate the approval and execution of scientific programs, ratherthan differentiating among different telescopes. The information in submitted and approvedproposals will transfer into the various databases by which the different facilities managetheir operation, and into a database that allows NOAO to track its success at facilitating thecommunity's science. The USGP will also integrate the TAC process for all thesetelescopes, so that uniform, consistent, and appropriate criteria are used to ensure that thebest scientific programs are executed successfully. The USGP will explore means toincrease the efficiency of the entire process, including investigation of automaticscheduling of the NOAO telescopes.

Scientific Support. The USGP will assemble and unify the information that proposers andobservers need to use the facilities effectively. This includes making general informationabout telescopes, instruments, and techniques available electronically, providing softwaresuch as exposure time estimators, and documentation such as instrument manuals. It alsoincludes the provision or coordination of pre- and post-observing run support byastronomers who can answer questions about specific programs, instruments, telescopes, ordata reduction capabilities.

Data Reduction. The USGP will ensure that software is available to reduce data from

instruments to which NOAO supports access by the community. This will involve theproduction of ERAF applications in some cases (such as the Gemini instruments) and willinvolve the distribution of software produced by outside groups in other cases (such as theindependent observatories). The effort to be put into new software should be appropriate tothe expected demand.

Alternate Modes of Observing. The efficient use of telescope facilities of the future maywell require a change in the way that astronomers obtain their observational data. Withspace-based and radio groundbased observatories we have seen the advent and the initialdevelopment of practical remote observing and queue scheduling. These types of observingand scheduling may be essential elements of the national facilities in order to optimize boththe science and the operational efficiency of those facilities. The international Geminiproject science and management teams have already established that the Gemini telescopeswill be queue scheduled at least 50% of the time. The Hobby-Eberly Telescope, to whichthe community will get some access, will be entirely queue-scheduled. The USGP will

46

explore alternate modes of observing 1) in order to ensure that the US community can getmaximum benefit from those programs that are not scheduled in the classical way, and 2) tounderstand whether the scientific and operational benefits are substantial enough to warrantimplementing such modes for additional facilities. Such implementation might include aremote observing station with high-bandwidth links to Cerro Tololo or the Geminitelescopes, with eventual support for remote observing from the observer's home site.

Archiving. Although Gemini maintains the requirement that instruments write data into anarchive, no support is provided for access by the astronomers of the partner countrycommunities. Similarly, KPNO currently writes all data frames into a save-the-bits archive,but no access is provided other than to recover data that is lost or damaged. The USGP willinvestigate the feasibility, the cost, and the anticipated benefits of supporting general accessto the data obtained on the various telescopes. If such access seems warranted and feasible,USGP will develop a plan for providing it. As an associated issue, it may be desirable toidentify certain datasets as being of general scientific interest and to support acquisition of,and access to, these datasets.

Scientific Outreach. The USGP has a specific responsibility to act as the liaison betweenthe international Gemini project and the US community. It satisfies that obligation byproviding information about the project and the facility capabilities through newsletters,displays, and talks, by convening workshops on special topics relevant to Gemini and ofinterest to the US community, and by soliciting input, both through its advisory committee,the US SAC, and informally. In its new, broadened role, the USGP will have responsibilityfor NOAO scientific outreach. The USGP will carry out a set of activities similar to thoseundertaken in support of Gemini. Information will be broadcast through newsletters,displays, talks, and electronic media. Sessions at AAS meetings and workshops will beorganized so that the community can provide input to guide the NOAO program. Users'committees will meet annually to discuss issues of current concern.

SOLAR ASTRONOMY

A. NOAO Solar Strategic Plan

Humanity has always viewed the Sun as special, a reliable source of the energy required for itsexistence. The large luminosity changes related to the evolution of the Sun as a star occur ontime scales of billions of years, much longer than the time frame associated with thedevelopment of life on our planet. Other events, possibly catastrophic, will probably terminateterrestrial life before this evolution of the Sun will force its extermination billions of years fromnow.

Ever since the first telescopic observations of sunspots by Galileo, and especially since thediscovery by Schwabe in 1843 of the 11-year cyclic variation in their number, many havesought variations on Earth that can be related to the changing Sun. These variations, if presentat all, are small compared to the changes due to orbital and geological effects. Major ice ageswith durations of 20,000 years and amplitudes of 5 to 10 degrees C appear to be the result ofthe latter, although there is now strong evidence that "little ice ages," which are less severe(few degrees C) but are of shorter duration and have a more rapid onset (< 100 years), arerelated to major changes in the solar activity cycle itself. Total solar irradiation changesobserved in the last activity cycle amount to 0.1 - 0.2%, probably too small to have affected the

47

weather on Earth but large enough to have a severe effect on upper atmospheric chemistry andcause variations in atmospheric ozone as large as man-made effects, which are of major currentconcern. The solar cycle is accompanied by large, often rapid, changes in ultraviolet and X-rayirradiation that, together with major changes in the solar particle flux in the solar wind, have asevere effect on our magnetosphere and on what is referred to as "space weather." The moretechnically sophisticated civilization becomes, the more it becomes subject to these solarvariability related changes.

Not only is solar variability of major interest as an astrophysical phenomenon, butunderstanding the origins of solar activity is essential to our future vitality. Only byunderstanding the origins of solar variability can we develop the tools needed to predict thatvariability. NSO's strategic plan aims to both fill this need and to satisfy our curiosityconcerning the origins of solar variability. It aims at achieving these goals by a combinationofboth in-house research and broad community participation.

The following text describes the NSO strategic plan in more detail. (Elsewhere this plan hasbeen referred to as the Future Directions Plan or FDP.) It makes use of a major recentbreakthrough in solar observing capability: the ability to image the solar interior regionsinaccessible to conventional telescopes. The rapid instrumental and interpretive advances inhelioseismology in the last decade allowed the National Solar Observatory to deploy in1994/1995 a "telescope"—the Global Oscillation Network Group (GONG)—that can image thesolar interior by means of sound waves over at least a few solar cycles. This "interiortelescope" will, for the first time, allow us to observe the physical conditions responsible forthe solar cycle and the short-term variations within it. Coupling its observations with existingoptical and X-ray telescope capabilities and with in-situ observations in the Earth'senvironment will allow us to develop models of the complete variable Sun-Earth system withmajor predictive capability. GONG is complemented by other helioseismology experiments,especially by the Solar Oscillations Imager (SOI) experiment on the Solar and HeliosphericObservatory (SOHO) satellite. Although the SOHO interior telescope is scheduled to operatefor less than a solar cycle, it has the advantage over GONG of higher spatial resolution,allowing observations closer to the solar surface. The comparison of the GONG andSOI/SOHO data over the few years of simultaneous observations will be invaluable to establishthe quality and reliability of the observations.

The proposed program for the understanding and prediction of solar variability is divided intothe following four research goals.

Goal 1: Understanding the solar activity cycle.

As seen at the solar surface and in the solar envelope, the solar cycle is a cycle of magneticfields. Sunspots result from strong magnetic fields at the solar surface inhibiting convection,which is the main form of energy transport in the Sun's outer layers. Magnetic fields couple thesolar surface to the solar envelope and the solar wind. They are in some, as yet poorlyunderstood way, basic to the heating of the million degree solar corona and to the generation ofshort-term, transient solar phenomena such as solar flares and coronal mass ejections.

How does the solar magnetic cycle arise? Because the cycles of the northern and southern solarhemispheres are in phase even after many millions of solar cycles, it results clearly from aglobal solar mechanism. We now believe that large-scale motions in the 200,000 km thick outerconvection zone of the Sun, combined with a dynamo mechanism, result in the periodic

48

amplification of the internal magnetic fields which then, as they penetrate the solar surface,result in the observed solar activity phenomena. Although a number of models exist based onthis concept, they are clearly inadequate primarily because of the lack of knowledge of theactual conditions in the convection zone. They are based on extrapolations downward fromobservations of the solar surface. Limited helioseismic observations of the solar interior show

inconsistencies in these flow/magnetic field models. The same observations have, however,also clearly demonstrated that with helioseismology it will be possible to observe the flows andstrong magnetic fields well below the solar convection zone. Also, as time progresses, changesof the solar surface waves are clearly seen to be related to the changes in the progressingcurrent solar cycle. We therefore have full confidence that extended observations with GONG,covering one or more solar cycles, will result in vastly improved and perhaps definitive modelsof the solar cycle.

We will therefore pursue the following program to accomplish that goal:

• Operate GONG for One or Two Solar Activity Cycles

This includes both the operation of GONG and the development of the mathematicalinversion techniques required to obtain the sharpest and most accurate images of thetemperature, velocity, and magnetic fields in the solar convection zone. Extended operationbeyond the approved three years and ongoing refinement of the imaging techniques areneeded to satisfy our goal. An absolute minimum operation duration is one full sunspotcycle of about 11 years. The magnetic polarity of the sunspot cycle reverses every cycle,resulting in the so-called 22-year "Hale cycle." It is desirable to continue GONG for a fullHale cycle.

• Continue and Refine Ongoing Observations of Solar Surface Flows and Magnetism

The National Solar Observatory has already accumulated archival observations of flowsand magnetic fields on the solar surface based on Doppler shifts and the Zeeman effect andcovering almost two sunspot cycles. Coupling GONG interior observations with ongoingsurface observations will allow us to: understand the coupling of the interior and surfaceproperties, and based on this understanding, extrapolate models of the solar cyclebackwards in time. NSO submitted in FY 1996 a proposal to NSF for a renewal of ourexisting synoptic facilities by replacing them with a fully autonomously operated, modernfacility called SOLIS. If funded, it will be ideal for carrying out this program.

• Develop a Computer Model of the Solar Cycle

We anticipate rapid advances in the dynamo and magnetohydrodynamic theories from thestimulus of the helioseismological observations. Based on these it will be possible todevelop powerful computer models of the solar convection zone and of the solar magneticcycle. In many ways these models will be similar to models of the terrestrial climate. Theymay, however, share similar limitations in their ability to predict solar activity.

• Test of Solar Cycle Model on Sun-Like Stars

We can test a model for only a few solar cycles out of the hundreds of millions of cycles• that have and will occur in the life of the Sun. This snapshot in time may very well not be

49

typical for the behavior of the variable solar cycle. The study of activity cycles on Sun-likestars will give us snapshots at very different, randomly selected times in the long period ofcyclic activity.

Goal 2: Understanding the coupling of the solar interior to the surface and the predictionof solar irradiance changes.

When observed at the solar surface, the solar magnetic field is almost fully concentrated in fluxtubes, which range in size from 100,000 km to 200 km or smaller and with strengths rangingfrom 3500 gauss to 1400 gauss. The few larger flux tubes cause sunspots; the numerous smallerflux tubes, or magnetic elements, appear to be the basic building block of most of the surfacemagnetic fields outside the sunspots. The larger flux tubes are powerful enough to dominateand therefore modify the convection zone near the solar surface; the smaller flux tubes, on theother hand, are jostled about by surface convective cells known as "granulation." The dynamicsof these flux tubes are undoubtedly the major ingredient in the transport of the solarnon-thermal energy, which heats the corona, causes solar transients, and ultimately drives thesolar wind. All manifestations of the interior activity cycle are ultimately revealed to theobserver by surface magnetic fields and their properties. It is essential to investigate whetherthe predictions made by the solar cycle model are confirmed by these surface magnetic fields.

We propose the following program to accomplish that goal:

• Make High Quality Observations of the Outer Layers of the Solar Convection Zone

To couple the interior observations and model of solar activity to the observable solarphotosphere it is necessary to zoom in with the helioseismological techniques on theuppermost layers of the solar convection zone representing the outer few percent of theSun. Considering the very fine scale structure of the observed solar surface magnetic fields,it will be necessary to obtain the very highest angular resolution possible. An upgrade ofGONG from a 256 x 256 array imaging system to a 1024 x 1024 system will be proposedto NSF to accomplish this.

• Relate the Changes in the Outer Convection Zone to the those of the Solar Surface

High quality, quantitative observations of the solar surface temperature, flows, andmagnetic field strength can be obtained with existing NSO telescopes but will requireupgrading the existing instrumentation or the construction of new instruments. SOLIS willbe central to this. We also want especially to capitalize on the increasing availability ofhigh-quality infrared array detectors in observing the magnetic field, which is the maindriver of solar activity in these layers.

• Determine the Origin of the Total Solar Irradiance Changes

Space observations tell us that the solar irradiance changed over the present solar cycle byas much as 0.15%. We don't know whether this is typical, although stellar observationssuggest that it is lower than average by a factor of 2 to 3. Anyway, it is likely that theobserved long-term variations in the solar cycles can result in larger irradiance variations,which are the likely cause of the "little ice ages." These irradiance changes could havethree causes: (1) the dark sunspots, (2) the bright faculae that accompany the small

50

•

magnetic elements, and (3) surface temperature changes reflecting interior temperaturechanges rather than solar surface phenomena like the first two contributors. The two-stationNSO network of Precision Solar Photometric Telescopes (PSPT), which is part of the RISEprogram (Radiative Inputs of the Sun to Earth) program, will be used together with thesolar cycle model to pinpoint the origin of the irradiance changes and to forecast andpostcast these changes, which are of major interest to the global change program.

Develop a Model of the Solar Surface Magnetic Field Behavior

It should be possible to develop models of the typical response of the magnetic flux tubesand their physical conditions to interior conditions. These models are needed as input toour third goal given below: understanding the coupling of the solar surface behavior to thesolar envelope.

Test this Model at Very High Angular Resolution with Surface MagneticField Observations

From existing observations we know that surface magnetic fields show very largevariations on spatial scales that are poorly resolved with present solar telescopes. Largegroundbased telescopes equipped with adaptive optics are needed to make these tests. Inthe meantime, the rapid advances in active optics and image selection techniques at the SacPeak Vacuum Tower Telescope are allowing diffraction-limited imaging at 1/6 arcsec.

Goal 3: Understanding the coupling of the solar surface to the solar envelope and theprediction of solar transient events

Flows control the magnetic fields over most of the solar photosphere and below, but above itthe situation dramatically reverses. The photosphere represents not only a boundary layer foralmost all of the solar radiation, it also is the layer where convection ends and where plasmamotions begin to be controlled by magnetic fields. Above it, major departures from hydrostaticequilibrium and spatial homogeneity appear. The temperature starts rising upwards, first in thesolar chromosphere which, at temperatures of 10,000 to 100,000 degrees, becomes the sourceof the highly variable solar near- and far-ultraviolet radiation. This radiation has major effectson the properties of the Earth's upper atmosphere, such as its temperature and chemicalcomposition (for example, ozone); these in turn affect mankind by changes in UV radiationreaching the Earth's surface and by the decay of satellite orbits.

Even higher, in the solar corona, the temperatures reach one to three million degrees, and insolar transient events like flares, 10 to 30 million degrees, comparable to the 16 million-degreesolar core. In the corona the magnetic field geometry dominates everything. Changes in thisgeometry result from flows in the solar interior but are observable at the photospheric levels.The solar wind is strongly modulated by the magnetic geometry; instabilities occurring in thisgeometry and topology lead to a number of fast, transient phenomena, which can havecatastrophic effects on our existence on Earth. Solar flares, probably the result of rapidreconnections of magnetic fields, and coronal mass ejections are among the best known typesof these events. Related magnetic storms affect power and communications distributionsystems. Energetic X-rays and particle radiation pose a threat to artificial satellites and mannedspace flight; rapid related changes in the near- and far-ultraviolet radiation cause transients in

51

the Earth's thermospheric temperature and density structure, resulting in loss of satellitetracking capability.

We propose the following program to address this goal:

• Determine the Mechanisms that Control the Geometry and Topology of the CoronalMagnetic Fields

Required is a combined program of theoretical and observational research focusing both onlarge- and small-scale aspects of this topology. A vigorous program in theoretical researchis needed to take the observed and predicted photospheric magnetic field patterns anddetermine from them the magnetic field topology in the corona. Our expectations are that aprogram of good, high resolution observations of the vector magnetic fields and currentpattern in the photosphere, coupled with these geometry and topology models, which thenare tested for validity on X-ray images and infrared coronagrams, will result in a muchimproved predictive capability for this topology.

• Research the Mechanisms Responsible for Chromospheric and Coronal Heating

Notwithstanding many years of research, these mechanisms still elude us. The dynamicaction of photospheric convection on the small photospheric magnetic elements is a likelysource of this heating. Higher angular resolution observations of photospheric (vector)magnetic fields and of photospheric and chromospheric structure are needed. These requirea large-aperture solar telescope equipped with adaptive optics and sophisticated optical andinfrared instrumentation.

• Develop a Model for Coronal Transients and Solar Flares

The large-scale magnetic field geometry and topology in the corona changes as a result ofthe flows in the solar photosphere and below. Occasional instabilities will occur whichresult in rapid magnetic field reconnections. These reconnections are seen in X-ray images.The resulting solar storms cause large changes in the interplanetary magnetic fields andmassive coronal mass ejections. Solar flares are also believed to be one of theconsequences. The resulting effects on space weather and on Earth can be catastrophic.With the mechanisms underlying the coronal magnetic field topology and heating known, itshould be possible to develop a model with predictive capability for these transients.

Goal 4: The exploration of the unknown

In developing this program for the understanding and forecasting of solar variability, one isstruck by the fact that most of the knowledge and the tools on which we base our planning arethe result of work by astrophysicists who were driven in their research, not by well-definedprogrammatic goals as outlined in this plan, but by curiosity and by the desire to innovate.Without the results of this fundamental research, we would not have the theoretical andobservational tools available to define and carry out the solar variability program developedhere. The creation of the main tool that will be used in the program proposed here forunderstanding the solar cycle, helioseismology, is a very good example.

52

We therefore consider the exploration of the unknown to be an essential component of ourproposed research. The "practical return" of truly innovative, exploring programs in solarphysics, leading to understanding and predictive capabilities, has always been very high. Anexample of such a program—and one that is central to NSO's future plans—is the explorationof the properties of the Sun at infrared wavelengths. We have as yet no knowledge of theinfrared spectrum of the solar corona. Exploring it will in all probability lead to improved toolsfor measuring coronal conditions and magnetic fields, leading to improved models of, forexample, coronal mass ejections and their effect on Earth. No spectrum has as yet beenobtained at infrared wavelengths of the high density, high temperature plasma present in solarflares. Its detection is likely to lead to new insights in the understanding of solar transients andin plasma physics. Another example is the exploration of the small scale structure of solarmagnetic fields. No one has so far adequately resolved the "magnetic elements" (smaller than100 km) that make up most of the solar surface magnetic fields. Coronal heating and energybuildup are likely to occur on very small scales, below what can be resolved now. Largetelescopes and clever techniques to remove the effects of atmospheric image disturbances willbe needed to figure out if and how the magnetic fields "do their work" on the solar envelope.NSO is presently developing a concept for a large solar optical/ER telescope which will replaceits current facilities and which will allow it to engage in cutting edge research at the beginningof the 21st century.

The NSO strategic plan covers the next 11-year solar cycle starting in about 1998. Its goal isvery ambitious and requires participation of a broader segment of the solar community than justNSO itself. The time between now and the beginning of the next solar cycle will be used togenerate this broadened plan, including probably participation not only by the US community,but also by the worldwide community, using both groundbased and space-based facilities.Broad community participation and focusing NSO's efforts on the requirements of this planwill also provide the strategic tools needed in case a downsizing of NSO cannot be avoided.

B. Science at NSO

The study of our nearest star continues to yield important results about its structure andevolution that are widely applicable to astrophysics and solar-terrestrial physics. As describedin the previous section, NSO intends to organize its research effort over the next fifteen yearsaround four broad goals connected with understanding solar variability. The current scientificprogram is presented here under those four headings. However, no simple categorization cancomfortably embrace the full range of science carried by NSO staff and visitors.

1. Understanding the Solar Activity Cycle

Local Helioseismology

NSO has played a key leadership role in the emerging field of local helioseismology. In thisfield, the solar oscillations are analyzed in a variety of ways to infer conditions in localizedvolumes, rather than globally averaged over the entire Sun. NSO staff members have beeninstrumental in developing four major methods of local helioseismology: p-modeabsorption, acoustic holography, time-distance, and ring diagrams.

Observations obtained at the Vacuum Telescope on Kitt Peak have shown that sunspots andactive regions absorb a significant fraction of incident p-mode acoustic wave energy. Thishas provided a new diagnostic tool for understanding the structure and dynamics of

53

magnetic regions. Data obtained both with GONG, and at the South Pole in an NSO-NASA-Bartol collaboration, have been used to measure wave scattering properties ofsunspots. The scattering and absorption of acoustic waves by sunspots offers the prospectof studying the subsurface structure and evolution of sunspots and active regions by a formof acoustic holography. Indeed, analysis carried out on the South Pole data as well asHigh-/ Helioseismograph (HLH) data may have already detected the presence of subsurfacemagnetic fields before their appearance at the surface. Acoustic power maps from SouthPole observations show enhanced power in the form of "halos" surrounding active regionsas well as occasional "finger"-like features, which extend outward from active regions.Theoretical work has been carried out to understand the physical mechanism responsiblefor the absorption and to explain the observed scattering properties of active regions bywave and geometric acoustics.

Time-distance helioseismology, developed at NSO, provides another excellent tool forlocal helioseismology. Directly analagous to classical terrestrial seismology, this methodmeasures the time it takes a wave to travel over a certain distance. It is now known that

sunspots produce a measurable signature in time-distance plots, and a secondary reflectionof the acoustic waves from a layer in the solar atmosphere is detectable. Theoretical workdone at NSO has reproduced the time-distance observations and studied the forwardproblem of deciphering the subsurface structure of sunspots through time-distancehelioseismology.

A three-dimensional analysis of oscillation data has been developed in which the signatureof the oscillations is a set of trumpet-shaped surfaces in the kx-kv-co volume. When thesurfaces are sliced at a constant temporal frequency (co), a set of rings appears. The centralpositions of these rings are proportional to an average over depth of the horizontalcomponents of the velocity of the solar plasma. By analyzing rings from data obtained atdifferent heliocentric positions, three-dimensional maps of the horizontal flows can bedetermined. The ring-diagram technique has now been applied to observations obtainedwith the Sacramento Peak Vacuum Tower, the Kitt Peak HLH, GONG, the Mt. Wilson 60-

Foot Tower, the Tawainese Oscillation Network (TON), and the MDI/SOI instrument onSOHO. In all cases, the results indicate the presence of a horizontal flow field whosedirection spirals with depth at a given location. This pattern could be a consequence of a setof oppositely-directed shear layers associated with the ionization zones in the outerconvection zone, and is consistent with the production of helicity in magnetic fieldsobserved at the surface. The magnitudes of the flows are also consistent with the velocitiesof thermal winds that could be the consequence of large-scale thermal variations alsoobserved as changes in the shape of the solar limb. Synoptic observations obtained with theHLH over the course of a solar cycle are now being analyzed to study the temporalevolution of the flows in the outer solar convection zone, as well as flows associated with

magnetic helicity and active longitudes. The possibility that the formation of active regionsis presaged by a recognizable flow pattern in the convection zone holds out the hope that along-range forecast for specific solar activity may eventually be available. Theoretical workis also underway to develop methods to infer sub-surface magnetic fields from the ringdiagrams. Here, the shapes of the ring (rather than the positions) are altered by the presenceof magnetic field.

Using this arsenal of analysis techniques, NSO plans to interpret local helioseismicobservations over the next five years, during the declining, minimum, and ascending phases

54

of the solar activity cycle. The observations will be obtained from GONG, theSpectromagnetograph and the HLH at the Kitt Peak Vacuum Telescope, the UniversalBirefringent Filter at the Sac Peak Vacuum Telescope, and through the NSO-NASA-BartolSouth Pole project.

Large Scale Fields

Helioseismology is helping to map the internal flow fields that are central to the solarcycle, but the interaction of these flows with (as yet undetected) strong magnetic flux tubesremains obscure. Three-dimensional simulations of the rise of magnetic flux tubes from thebase of the convection zone are used to understand the formation of sunspots, plages andephemeral regions. The properties of these active regions—such as rotation, meridionalcirculation, growth, and decay—can be determined by studying the dynamics of flux tubes.This should shed light on the underlying mechanism that drives the solar cycle.

The newly-commissioned spectromagnetograph at the Kitt Peak Vacuum Telescope is nowproviding vital data for the study of the solar cycle. Full-disk digital magnetograms,Dopplergrams, and intensity maps in the A.10830 line of helium are now produced daily anddistributed to the solar-terrestrial community. These data, along with the 20 years of nearlydaily full-disk magnetograms acquired with the now-retired 512-channel magnetograph, arecurrently being used to study the emergence patterns of active regions over a size scaleextending from the smallest ephemeral regions to large complexes of activity. These studieshave already led to important and new information about the solar cycle. We now know, forexample, that the new solar cycle begins with the emergence of reverse-polarity, high-latitude ephemeral regions approximately two years before the first appearance of a new-cycle sunspot region and three years before sunspot minimum. The diachronic observationsat NSO and the research investigations using these data by the NSO staff and visitors willyield further clues and constraints to the modeling of the solar cycle.

Sunspots are perhaps the most obvious manifestation of solar activity. A long-term programis underway at NSO, in collaboration with colleagues at the Indian Institute of Astrophysicsin Bangalore, India, to study the daily positions and sizes of sunspots. The photographicarchives at Mount Wilson and Kodaikanal (Lndia), which date back to the early 1900s, havebeen measured for this information, and these datasets continue to be analyzed for detailedregularities in sunspot motions. These results will bear on the problem of the solar cycle.

A new observational tool for mapping large-scale horizontal flows in the photosphere hasbeen developed at NSO: local correlation tracking of granules for many hours per day. Thegranules, which live for 10-15 minutes, serve as tracers of persistent flows that interactwith active region magnetic fields. Ln combination with other tools, this technique allowsclose study of the forces that build flare energy.

The GONG network is providing coarse full-disk maps (16-arcsec resolution) of the surfacemagnetic field and Doppler velocity every 20 minutes for three years beginning in 1995.From these data the large-scale surface horizontal flows (e.g., differential rotation, torsionaloscillation, meridional circulation) can be derived, as well as non-axisymmetric flows atleast as large as supergranulation. Simultaneous observations from the Solar OscillationsImager aboard SOHO will complement the GONG observations, at 4-arcsec resolution, andwill be used to track the fine-scale flows (vortices, sources, and sinks smaller than

55

supergranules). In addition, if sufficient funds are available, we plan to operate a three-station network for at least a full 11-year solar cycle. NSO considers this extension ofGONG to be its highest scientific priorityfor new initiatives.

Solar-Stellar Physics

The principal objectives of the NSO and NOAO program in solar-stellar physics are to:(1) advance solar physics by testing theories developed in a solar context throughobservations of stars that, in total, represent a range in physical parameter spaceunavailable using the Sun alone; and (2) advance stellar physics through the application ofthe methods and results of solar physics to the study of solar-type phenomena observed inother stars.

The McMath-Pierce solar-stellar program has made important contributions in the generalarea of the magnetic field properties of solar-type stars, thus yielding insights on the natureof stellar dynamos. These include the delineation of magnetic field characteristics such asstrength and filling factor, and their empirical dependence on stellar properties that includerotation, fractional convection zone depth, and photospheric gas pressure. The results serveas critical constraints for the development of dynamo theory. This effort relied on highresolution spectroscopic techniques often carried out in a synoptic mode that produced dataon time scales ranging from that of the stellar rotation period to actual magnetic cycleperiods.

Within the NSO, these goals had been observationally addressed through a program of highresolution synoptic spectroscopy carried out by a resident observer on behalf of thecommunity at the McMath-Pierce telescope from approximately 1984-1995. In FY 1995,budgetary stringencies led to the closure of this program at the McMath-Pierce.

Ln FY 1996, the NSO was able to renew nighttime operations at the McMath-Piercetelescope using contributed funding from principal investigators in the solar-stellar andplanetary communities. A vigorous (but closed) program of nighttime work is now in effectat the McMath-Pierce facility based on these outside contributions and science programsthat also passed review by the NSO Telescope Allocation Committee (TAC). The currentPis include: K. Strassmeier (University of Vienna), J. Eaton (Tennessee State University),F. Roessler, F. Scherb, and R. Reynolds (University of Wisconsin), A. E. Potter(NASA/JSC), T. Morgan (Southwest Research Institute), and D. Deming (NASA/GSFC).The science programs range from Doppler imaging of active, spotted stars to observationsof molecular features in comet Hale-Bopp.

The instrumentation includes the cross-dispersion upgrade that was installed (but notcommissioned for routine use) in 1995. The NSO looks forward to the eventual receiptfrom the NOAO instrumentation program of a large-format CCD that will replace thecurrent TI 800 x 800 CCD that has been in use since 1987. The continuation of nighttimeoperations at the McMath-Pierce telescope is contingent upon the availability of outside,contributed funds on a year-to-year basis.

Even though nighttime operations at the McMath-Pierce have been curtailed, a forefrontresearch effort in solar-stellar astrophysics is ongoing at the NSO. This effort reliesincreasingly on the utilization of NOAO nighttime telescopes and instrumentation. Acollaboration of KPNO and NSO scientific staff in SONG—Stellar Oscillations Network

56

Group—is exploring techniques for the detection of p-mode oscillations in solar-type stars.Experiments are being conducted at both the McMath-Pierce telescope (solar data) and theCoude Feed (stellar data) on Kitt Peak. In addition, the WIYN telescope with its multi-object spectrograph (HYDRA) is being utilized to investigate the chromospheric and cycleproperties of the numerous solar-type stars in M67. The many "Suns" in this solar-age andmetallicity cluster are counterparts of our own Sun, thus serving as representative examplesof the Sun at all possible, or potentially possible, phases of its activity cycle. Such datawould require centuries to obtain if we were to rely on observations of the Sun alone. BothNSO and KPNO staff will utilize the new cryogenic echelle PHOENIX—for sensitive, highresolution observations of infrared spectral diagnostics of magnetic fields in solar-typestars. The infrared is a particularly powerful region for magnetic field studies because ofthe wavelength sensitivity of Zeeman splitting. We also note that KPNO and NSO staff areinvestigating the establishment of a bench-mounted, fiber-fed spectrograph at the Mayall4-m telescope on Kitt Peak for high resolution (as high as R = 300,000), long-term studies.Finally, the NSO maintains contact with developments in the area of small telescopenetworks, such as GNAT (the Global Network of Automated Telescopes). Photometric datafrom small telescopes is a potentially valuable augmentation to spectroscopic studies thatinclude, for example, Doppler imaging and the joint behavior of magnetic field propertiesand luminosity variability on short and long time scales in solar-type stars.

2. Understanding the Coupling of the Interior and the Surface: Irradiance Variations

Overview

Measuring and understanding the Sun's variable outputs poses one of the most importantproblems in solar research. Not only is this problem linked to fundamental questionsconcerning the structure of the interior of the Sun and to the unknown mechanisms of theactivity cycle, but it also has an impact on climate and atmospheric chemistry. In recentyears much has been learned about variations in the solar radiative output, but much is stilllacking in our knowledge of this important field.

Investigators now believe there are at least three, possibly four, contributors to solarirradiance variations: active regions (sunspots and plage), strong magnetic networkelements, weak magnetic network elements, and possibly the centers of supergranular cells.The relative importance of these components as a function of time in the short- and long-term variations in solar irradiance is not well established.

In a cool star like the Sun, energy is transported by convection in a thick shell that lies justbelow the visible surface, the photosphere. Convective motions also generate and amplifysolar magnetic fields, which rise buoyantly to the photosphere and expand into the outersolar atmosphere. The interaction of these motions and fields is responsible for all thedetailed structure that we observe (sunspots, coronal streamers, active regions) as well asthe great variety of non-thermal phenomena that comprise solar activity. Convectivemotions also generate sound waves, gravity waves and MHD waves, which can propagateoutward and heat the solar atmosphere.

The interaction of magnetic fields and convective motions ("magneto-convection") occurson a broad range of temporal scales (seconds to months) and spatial scales (arcseconds tothe solar radius). Such phenomena as solar flares, the solar wind, and many other solar

57

cycle manifestations are products of this interaction. This broad subject forms one of themajor research areas in solar physics and is a central concern in NSO's long range plan.

Over the last two decades, NSO instruments have acquired a set of observations thatprovides a strong foundation for studies of the evolution of the surface features on timescales of minutes to that of a solar cycle and on spatial scales of around 2 arcsec to the fulldisk (the Sun as a star). These data are now being used to study with moderate spatialresolution the detailed spatial and temporal comparison of photospheric magnetic fields andthe corresponding absorption and emission structures observed in the photosphere andchromosphere. Work at NSO now involves the separation of several different componentsof magnetic activity (e.g., active regions, decaying active regions, and network), usingmasks derived from the NSO full-disk magnetograms, to determine the contributions ofthese different components to variations in solar irradiance and how these vary with time.

The continuation of synoptic observations at NSO in the future, especially when SOLIS isoperational, will play a central role in developing an understanding of the physicalmechanisms that drive solar activity and the resulting variations in solar irradiance, and indetermining the effect of solar variability on the Earth's environment.

The diachronic magnetogram data from the KPVT are an important existing tool forunderstanding how the irradiance signal is distributed across the surface of the Sun. Suchgroundbased data may help to settle a long-standing controversy about whether allirradiance variations arise from localized magnetically associated features or whether othermore global changes are involved. Since April 1992, additional thermodynamicinformation, strictly co-spatial and co-temporal with the magnetograms, has been obtained.The analysis of both archival and current data by NSO scientists (including NASA andSPRC staff) will be a major effort during the next decade.

RISE/PSPT

High-precision measurements of continuum images yield important data on the diminutionof the total irradiance by sunspots as well as the enhancement of this irradiance by faculae.Similar measurements in the core of the Ca II K line yield equally important data on thedistribution and strength of plages and network in the chromosphere. These bright plagesand weaker network contribute directly to the Sun's output in the UV and EUV regions ofthe spectrum. In large part these measurements describe the effects of emergence andevolution of magnetic flux in the solar photosphere on the Sun's irradiance. NSO, incollaboration with a consortium of scientists, has begun the RISE program (RadiativeInputs from the Sun to the Earth) to increase our knowledge in this important area. RISEhas attracted—and depends on—separately-identified, incremental funding from the NSF.NSO is currently building two identical instruments (Precision Solar PhotometricTelescopes), capable of 0.1% photometry of the continuum irradiance with a one-hourcadence. The instruments are designed around a 15-cm objective, a 2K x 2K pixel CCD,and fast frame-selection electronics. Full-disk surface photometry with 0.1% per pixelaccuracy with a one hour cadence will be obtained. In addition to their role in solving thesolar irradiance puzzle, the PSPT data will satisfy a long-standing need for precisionphotometry in many other areas of solar physics.

58

Convection

Two convective eddy sizes have been recognized for decades: the granulation, with typicalsizes of 1500 km (2 arcsec), and the supergranulation, with sizes around 30000 km andlifetimes of around a day. Only recently, advanced observing and analysis techniques(developed at NSO and at Lockheed Palo Alto Research Labs) have confirmed theexistence of a third scale—the mesogranulation, with typical sizes of 7000 km. Theorists inthe US and Denmark are engaged in modeling the dynamic behavior of the turbulent,compressible, radiating flows that are involved in stellar convection, and they useobservations obtained at NSO and La Palma to guide their models. Certain numericalstudies of convection predict the existence of convective cells on the size of roughly200,000 km. The search for these giant cells using the high-resolution magnetograms fromthe KPVT have shown that even if they exist they probably do not show up at thephotosphere. These magnetograms, taken since 1975, have been studied extensively toanalyze the surface flow fields, like meridional circulation, torsional oscillations, rotation,diffusion of fields, and the variation of these parameters over the solar cycle.

During 1994, USAF scientists resident at NSO obtained a long, continuous granulationsequence at the Swedish Observatory in La Palma with sub-arcsecond resolutionthroughout. These data will help to improve kinematic models, which already give realisticrepresentations of the time-evolution of solar granulation. In addition, three-dimensionalsimulations of solar convection, based on the hydrodynamic equations, have now reached adegree of realism that should allow quantitative comparison between theory andobservations.

NSO scientists plan to collaborate with scientists involved in the SOI instrument aboard theSOHO satellite in a three-year study of convection. The SOI will be capable of observingthe evolution of granulation and its interaction with small-scale magnetic fields for severaldays, and with uniform spatial resolution of 1.2 arcseconds. The VTT at Sunspot is capableof sub-arcsecond resolution over a larger field of view for several hours, but only onfavorable occasions. These two facilities thus complement each other and will yield newinformation on convection at these scales.

The interaction of convection and differential rotation in the outer layers of the convectionzone generates a variety of vortical and shearing motions of intermediate spatial scales (10arcseconds), which are relevant to the evolution of active regions, but are only nowbeginning to be studied with the new tool of local correlation. Two scientists at NSO havepioneered in this work and shown both its feasibility and scientific potential for the studyof flare energy buildup. Ln the quiet Sun, such motions must participate in the redistributionof magnetic flux over the surface. NSO plans further studies of this important process.

Recent observations at NSO have revealed a previously undetected feature of stellarconvective motions: the existence of rising and falling "plumes" of gas, distributedrandomly over the photosphere and occupying a small fraction of the surface. These plumesare well-known in terrestrial hydrodynamic systems (e.g., the ocean and the atmosphere),and have recently appeared in some solar simulations. New techniques for analyzing full-disk velocity movies are being developed at NSO to reveal this complex circulation.

Convection in the Sun has two important side effects. First, the turbulent motions producesonic noise, which propagates into the overlying atmosphere, produces shock waves, and

59

heats the low chromosphere. Gravity waves are also predicted above the convection zone,but their wavelengths are too short to be detected directly, except for their role inbroadening photospheric line profiles. Recent work at NSO is directed at estimating theenergy flux in such gravity waves from high spatial resolution observations of granulation.Second, convective motions transport magnetic fields that emerge from the solar interior. Inthe photosphere, these fields have kilogauss strength and occupy only a tiny fraction (10"4)of the surface area, mainly between granules.

Hydrodynamic simulations of convection predict the existence of high-speed vortices thatlie between emerging granules and that have diameters of 100 km at most. These vorticesmay play a crucial role in twisting the footpoints of coronal magnetic fields and thusheating the solar corona.

In order to study convective motions and magnetic fields on such small scales, NSO isdeveloping several complementary approaches to enhancing image quality. An adaptiveoptics system for the Vacuum Tower Telescope at Sunspot is being developed inpartnership with the US Air Force. This system is intended to correct the VTT image fordistortions produced in the Earth's atmosphere ("seeing") and to deliver a diffraction-limited image (0.17 arcsec) for many hours a day. The program is expected to deliver a 20element system in two to three years. Algorithms for image restoration and speckle arebeing applied and improved. Image selection techniques now allow diffraction-limitedsingle images. Two-dimensional spectroscopy at video rates is now available and canfreeze the solar image in moments of good seeing. In partnership with the KiepenheuerInstitute and Johns Hopkins University, NSO has developed a correlation tracker, which isnow available at the VTT at Sunspot. These developments are expected to culminate in asuperb facility for imaging and spectroscopy at 0.15 arcsec (100 km) resolution.

Small-Scale Magnetic Fields

In the quiet Sun, magnetic fields cluster at the borders of supergranule cells. Small bipolesconstantly emerge inside the cells and drift to the borders, where they disappear or mergeand reconnect with opposite polarities. This continuous dynamic process is basic to thedispersion of magnetic flux from active regions to high latitudes and ultimately controls theevolution of the large-scale fields, which impose structure on the corona. This process willcontinue to be a focus of studies by NSO scientists and visitors during the next five years.

The interaction of small-scale magnetic fields in the quiet Sun sometimes produces He I10830 dark points and X-ray bright points. Such events are most often associated with theremoval of magnetic flux and the disappearance of small-scale (10-30 arcsec) bipolarmagnetic structures. Research is now being conducted, in a collaborative study of X-raybright points observed with the Yohkoh Soft X-ray Telescope, of He I 10830 dark pointsand the underlying photospheric magnetic field in order to understand the conditions andprocesses leading to the chromospheric and coronal response to changes in thephotospheric magnetic field.

The Advanced Stokes Polarimeter, a collaborative effort by the High Altitude Observatoryand NSO, has been operating at the VTT at Sunspot for several years. It provides high-precision measurements of small-scale vector magnetic fields. During the next years, it willbe used for studies of the structure of sunspots, the dynamics of spicules, MHD wavephenomena, and many other research programs.

60

The image-quality enhancement techniques described above (adaptive optics, correlationtracking, image selection and restoration) are vital for studying fundamental magneticprocesses on their intrinsic physical scales. Another important component of NSO's effortis the collaboration with space-based observations from the SOHO, TRACE (TransitionRegion and Coronal Explorer), and Solar-B missions. SOHO, launched in 1995, is a majorEuropean/US spacecraft that brings a powerful array of instruments to bear on the structureof the solar corona and its connections both inward to the photosphere and outward to theheliosphere and the Earth. TRACE (1997-1998) is a NASA mission that will overlap andaugment SOHO by exploring the connections between fine-scale magnetic fields and hotplasma structures. Solar-B is a Japanese mission expected to fly around 2001. It will featurevisible and soft X-ray imaging with high angular resolution.

The infrared portion of the solar spectrum offers important advantages for magnetic fieldmeasurements (Section 4). To exploit these advantages, NSO (with support from NASA)has built a Near Infrared Magnetograph (NEM), now available to staff and visiting scientistsat the McMath-Pierce telescope. The NEM has enabled us to study for the first time thespatial distribution of kilogauss flux tubes in solar plages. A second-generation NEM-2 willbe completed (also with NASA support) in 1996. The VTT at Sunspot transmits infraredradiation to 2.5 microns, a range that includes the Zeeman-sensitive iron line at1.56 microns. Recent studies by NSO scientists at the VTT have demonstrated the power ofcombining the high-resolution of the telescope with the sensitivity of the infrared line.NSO's infrared magnetometers will become even more powerful over the next five years asinfrared array technology advances.

Weak Magnetic Fields

The solar photosphere is penetrated by a turbulent magnetic field. The largest of thesemagnetic field elements are probably the so-called intra-network magnetic fields, whosemorphology has been studied in the last two decades. As turbulent fields may be ofconsiderable importance for the solar cycle, the physical parameters of these fields (fieldstrength, temperature, etc.) need to be determined. The Zurich Imaging Stokes Polarimeter(ZEMPOL I) has been used at the McMath-Pierce facility to record spectra in circular andlinear polarization of these weak fields. Using the Zeeman effect, it appears that thestrength of these fields is less than 500 Gauss. Investigations of the depolarizing propertiesof the Hanle effect suggest the field strength is even weaker, on the order of 50 Gauss. Thetemperature is significantly cooler than in plages and the network at the level of lineformation.

The role of these weak magnetic fields in the solar cycle is still unexplored because of thedifficulty in observing them; however, theoretical studies at NSO and elsewhere haveemphasized that they may play a key role in the magnetic cycle.

3. Understanding the Coupling of the Surface and the Envelope: Transient Events

Solar active regions are composed of low-lying magnetic arches and arcades in the vicinityof sunspots. These fields have their roots in the photosphere and are distorted by a varietyof convective motions and horizontal shearing flows. The free energy stored in the field bythese motions can ultimately discharge in a violent solar flare, with major terrestrial effects.One of the major goals of solar physics is to understand, in detail, how this process of

61

energy storage and release occurs. The problem has many facets, ranging from the study ofsub-arcsecond field footpoints to the study of persistent large-scale horizontal flows. NSOis bringing on-line new tools for this major effort.

A new instrument for high spatial and temporal resolution imaging of solar active regionsand flares in the 10830 A line of He I has been developed jointly by NSO andNASA/GSFC for use at the KPVT. The instrument is built around a conventional but

cleverly packaged five-element Lyot filter. Ferroelectric liquid crystals modulate bandpassresponse and polarization at video rates. Subtracting accumulated observations in threepossible pairs of filter transmission modes provides active-region imagery of 10830 Aequivalent width, line-of-sight velocity, or longitudinal magnetic field with time cadence ofa few seconds. Nearly two decades of full-disk observations in the 10830 A line at theKPVT show a wide range of solar phenomena such as coronal holes, filament channelsassociated with coronal mass ejections, long-duration flares, and analogues to X-ray brightpoints, which are otherwise viewed on the disk only from spacecraft. The filtergraph willprovide a new view of transient phenomena in solar active regions with temporal andspatial resolution that cannot be obtained from the spectromagnetograph, where dataimages are built by scanning the telescope's solar image across a long entrance slit. Thenew data will be particularly timely for comparisons with observations from the SOHOspacecraft (to be launched in the summer of 1995) which, in combination with recentdevelopments in modeling the formation of neutral helium, promise to resolve longstanding controversies on how the formation of the line depends upon external EUVradiation from the corona and low-temperature transition region.

The solar corona is coupled to the photosphere by its magnetic field. According to currentideas, the power necessary to balance coronal radiative losses (and losses in the solar wind)originates in convective motions in and below the photosphere. A major goal in solarresearch is to understand how convection heats and shapes the corona via the magneticfield.

The earlier idea of acoustic heating by waves propagating from the photosphere has nowbeen substantially discarded since propagation of these waves is inefficient and observedupper limits are insufficient to heat the corona. Processes related to the electric currentdissipation or MHD wave generation and damping are now receiving more attention, butobservations of small-scale dynamical phenomena are needed to support and offer furtherdevelopment of these ideas.

The horizontal convective motions and interactions of the small-scale magnetic fields in thephotosphere twist and braid the magnetic field loops that extend into the corona. Accordingto a plausible theory, the complex coronal field develops current sheets, in which freemagnetic energy is released to heat the corona. One of the prime objectives forobservational solar astronomers in the coming five years is to test this idea. We mightexpect to observe transient, high-temperature fluctuations in the active-region corona,accompanied by turbulent high-speed flows. The initial volumes of these transient"nanoflares," may be very small, however, and difficult to detect. Vector magnetic fieldmeasurements in the photosphere, at sub-arcsecond spatial resolution, and fast imagingspectroscopy of the solar corona will be required. NSO is developing the equipment neededto pursue these crucial investigations.

62

Flow and condensation of the coronal plasma are known to exist above active regions andaround prominences. Flows are particularly impressive in the case of post-flare loops,where rapid evolution is seen. Recent studies with the NSO coronagraph provide evidencefor the reconnection of post-flare loops with the release of small quantities of energy. If thisis a common phenomenon in the quiescent corona, as two staff scientists suggest, it maygive support to the nanoflare heating process. Further studies of this important questionwill be performed.

High speed solar wind streams originate in coronal "holes" (regions of depleted density).Themechanisms thataccelerate coronal plasma to speeds as high as 800 km s"1 arenot wellunderstood. Some form of MHD wave, generated low in the corona by convective motionsin the photosphere, may be responsible. To date, little convincing evidence for such waveshas been found. Recently, however, the source of coronal heating and wind acceleration hasreceived attention by two of NSO's visiting scientists. They have used the 40-cmcoronagraph at the Evans facility to measure the height-dependence of the two strongestforbidden coronal lines and find linewidth changes consistent with the deposition of Alfvenwave heating. They expect to continue these studies to sort out the various complicationsduring the next few years.

Small-scale explosive or impulsive events are believed to produce fast ejections thatpropagate along the open magnetic field lines of a corona hole, producing the fast solarwind. Until recently, no detailed observations existed of the ejection process. At the 1991total eclipse of the Sun, such an event was detected for the first time with the 3.6-mCanada-France-Hawaii Telescope on Mauna Kea. A team of NSO and French astronomersdiscovered the event. Further searches for such minute ejections will continue.

Of more practical relevance are the large-scale eruptions called coronal mass ejections(CMEs). These occur once or twice per day at all phases of the solar cycle and involve theejection, by magnetic forces, of perhaps 1015 grams of plasma into interplanetary space.When such ejections interact with the Earth's magnetic field, violent geomagnetic stormscan occur, sometimes with devastating effects on power grids and communications. NSO,in collaboration with the High Altitude Observatory, will continue to investigate the originof such events and the instability of the solar corona, with the ultimate aim of being able topredict them.

The high-resolution NSO magnetograms and He I 10830 spectroheliograms provideimportant data related to the study of CMEs. The KPVT observations provide informationabout magnetic field configurations and the evolution of the magnetic field that eventuallyleads to the occurrence of a CME.

One important type of event, the two-ribbon event, has been studied using the daily, as wellas higher time-resolution, He I 10830 spectroheliograms. These events are the large-scale,long-duration end of a wide spectrum of flares and are more easily seen in He I 10830, as apair of dark ribbons, than in H-alpha. Most of these events are located outside activeregions, though the strongest occur in active regions, and all follow the eruption of afilament or of a filament channel. Because of their high association with CMEs (96%), ourability to identify these events provides an important tool in locating the probableoccurrence of CMEs that are likely to affect the Earth. The transient formation orexpansion of adjacent coronal holes is observed with many of these events, indicating asignificant alteration of the magnetic field configuration in the vicinity of these flares.

63

Collaborative studies of the two-ribbon events seen in He I 10830 with X-ray images fromYohkoh are currently being undertaken to investigate coronal activity and changes inmagnetic connections and configurations inferred from these data.

It is clear that better coronal and chromospheric observations are needed to attack many ofthe major problems connected with the coupling of the surface and the envelope. Most ofthe physics seems to reside in small volumes. To carry out detailed spectroscopic studies athigh temporal and spatial resolution, observers will certainly need more photon flux.Conventional coronagraphs have been limited in aperture by the difficulty of making near-perfect lenses. A new era in coronal studies has been opened by the advent of super-polished mirrors. Now meter-class coronagraphs are feasible, providing more photons infully achromatic images, which can be used to investigate the fundamental physicalprocesses. NSO, in collaboration with the Institute of Astrophysics in Paris, is engaged inbuilding such mirror coronagraphs, with USAF funding. Two prototypes have beencompleted successfully, and a third 55-cm instrument is under construction. Whencompleted in a year or two, it will produce a fast series of monochromatic images in theprincipal optical spectrum lines of the corona. The spatial resolution of the 55-cm apertureinstrument will be limited only by the seeing and should be sufficient to permit studies ofcoronal fine structure at 2 arcsec scales.

The temperature and density structure of the solar corona varies systematically throughoutthe solar cycle as its magnetic topology changes. NSO scientists are using new ER detectorsto examine several forbidden line ratios that are sensitive to electron density, at differenttemperature regimes. This diagnostic work is important for an understanding of thecoupling of the solar surface and interior.

4. Exploring the Unknown

The Infrared Frontier

Infrared observations offer unique advantages for probing the magnetic and thermalstructure of the solar atmosphere. For example, because the ratio of the Zeeman splitting ofa spectral line to its Doppler linewidth increases linearly with wavelength, it is possible inthe infrared to measure the intrinsic strength and orientation of solar magnetic fields with asensitivity that is difficult or impossible to achieve in the visible region. The discovery atNSO of Zeeman-sensitive emission lines near 12 microns and a sensitive Fe I absorptionline near 1565 nm has made this possibility a reality.

Results from the Near Infrared Magnetograph (NEM) have demonstrated the existence ofspatially coherent and correlated variations of field strength and area filling factor in plageregions. A collaborative program with NASA Goddard Space Flight Center has beeninitiated to study the three-dimensional variation of field strength within individualmagnetic structures by making simultaneous and co-spatial measurements at 1565 nm (baseof the photosphere) and 12 microns (top of the photosphere). A secondary goal of NEM is tostudy sunspot magnetic fields. It has been difficult to measure the variation of field strengthacross a sunspot because the umbra is so dark compared to the photosphere (which createsproblems with stray light and dynamic range). At 1565 nm, however, the intensity contrastbetween the umbra and the quiet photosphere is only about a factor of two; moreover, thecontinuum opacity is simple and well understood, so it is straightforward to estimate thetemperature at the level of continuum formation. NEM observations of about a dozen spots

64

have so far revealed a characteristic nonlinear relation between field strength andbrightness temperature that may place new constraints on magnetohydrostatic sunspotmodels. A modification to NEM that will greatly enhance its sensitivity to weak magneticfields has been tested at the McMath-Pierce and shows great promise for use with a large-format ALADDIN array camera.

NEM-2, a Fabry-Perot-based imaging magnetograph now undergoing engineering tests, willyield improved measurements of extended targets such as sunspots and active regions.

Infrared observations at wavelengths shorter than 2.3 microns are being pursued at the SacPeak VTT and ESF, including 1565-nm magnetometry with high angular resolution andimaging spectropolarimetry of the He I 1083-nm line. The 1083-nm line is also studied atthe Kitt Peak Vacuum Telescope, where the spectromagnetograph recently discoveredevidence for chromospheric outflows in coronal holes.

Other major diagnostic resources of the infrared spectrum include molecular band systemsand the infrared continuum. For example, the vibration-rotation bands of carbon monoxide(CO) at 2.3 and 4.7 microns are a sensitive thermometer that is being used to probetemperature inhomogeneities in the upper photosphere and to test the validity of widely-used atmospheric models. The published results of this program are substantial andgrowing. Also, the CO lines show prominent intensity oscillations that can be used to studythe penetration of solar p-modes into the upper atmosphere. The 1.6 microns continuumarises deeper in the solar atmosphere than any other observable wavelength and provides aunique window on magnetic fields and convection below the photosphere. Because theinfrared continuum approximates the Rayleigh-Jeans portion of a blackbody curve (withintensity directly proportional to temperature), it is possible to study temperature variationsboth vertically and laterally.

The infrared spectrum of the solar corona is largely unknown. The first survey was doneduring the total eclipse of 1970 from a high-altitude aircraft, in the range 1-3 microns. Evenless is known about the ER spectrum of flares. Predictions of the coronal spectrum out to 30microns are scanty at best. New ER technology is slowly changing this situation. Sensitivemeasurements can be made well into the ER where unknown and potentially useful coronalspectral lines may exist, including diagnostics for temperature, density, and magnetic field.NSO/SP has initiated a program to survey the ER coronal spectrum, which is beginning toyield promising results. A team was sent to observe the total solar eclipse of November1994 in Chile and returned with exciting data in the 1-1.5 microns region. A coordinatedground- and aircraft-based experiment for the 1998 eclipse is currently under development.

The McMath-Pierce Telescope is particularly well-suited to infrared observations on thesolar disk beyond 2.3 microns. Its all-reflecting optics provide high transmission, lowscattering, and low instrumental polarization at all infrared wavelengths accessible togroundbased observation. Its large aperture is a major advantage both for potential angularresolution (0.25 arcsec at 1.6 microns) and for light-gathering power (the number of solarphotons per Doppler linewidth is some 20 times smaller at 10 microns than at 0.5 micron).The scientific potential and emerging technical capabilities have stimulated broad interestin an enhanced solar infrared facility. A comprehensive approach to pushing back theinfrared frontier is embodied in NSO's plans for a "flagship" large-aperture visible/ERtelescope.

65

C. Initiatives for Solar Physics

Milestones

1998

• Complete phase C (final) Design of SOLIS• Start construction of SOLIS

• Start construction of high resolution camera for GONG• Construction of components for 20 Zernike adaptive optics system• Deploy second Precision Solar Photometric Telescope (PSPT) to Mauna Loa• Construct third PSPT. Deploy on Sac Peak• Complete feasibility study of CLEAR• Develop community consensus on science requirements and priorities of Flagship Solar

Telescope• Set resulting design parameters for Flagship Solar Telescope• Continue Flagship Solar Telescope site testing

1999

• Continue construction of SOLIS

• Complete construction of and install GONG high resolution camera• Integrate 20 Zernike AO system in Sac Peak VTT. Start tests• Operate RISE/PSPT network• Possible redeploy Sac Peak PSPT to Learmonth, Australia• Select site for Flagship Solar Telescope• Complete phase B study of Flagship Solar Telescope• Prepare proposal for Flagship Solar Telescope• Explore national and international partnership(s) for Flagship Solar Telescope• Review of Flagship Solar Telescope by NAS/NRC decade astronomy study

2000

• Complete SOLIS construction• First light SOLIS• Start observations with GONG high resolution camera• Regular observation of 20 Zernike AO system• Go-ahead for construction of Flagship Solar Telescope• Phase C study of Flagship Solar Telescope

2001

• Regular SOLIS operations• Propose for continued operation of RISE/PSPT network• Complete phase C study of Flagship Solar Telescope. Start phase D (construction)

2002

• Continue construction of Flagship Solar Telescope

2006

• GONG observations of first eleven-year solar cycle completed

66

2007

• GONG data reduction first eleven-year solar cycle completed

1. NSO's Future Directions Plan (FDP)

NSO's long range plan and most of NSO's initiatives are directly related to its FutureDirections Plan (FDP) (Section V.A.), which aims at the understanding and prediction ofsolar variability using observations both of the solar interior with GONG and of the visiblesolar atmosphere initially using NSO's existing telescopes, upgraded and equipped withadvanced instrumentation but ultimately SOLIS and the large, flagship solar telescope. TheStrategic Plan, initially viewed as the 10-20 year plan for NSO recommended by the 1993NSO/NOAO Observatories' Visiting Committee, is likely to evolve as a national, or eveninternational, plan for solar research in the next solar cycle.

Two major components of the FDP are now being proposed to NSF: SOLIS and theextension and upgrade of GONG. These are described in the following.

a. SOLIS

SOLIS (Synoptic Optical Long-term Investigations of the Sun) is a project to makeoptical measurements of processes on the Sun whose study requires sustainedobservations over a long time period. The primary scientific motivation for SOLIS is totest and improve our understanding of how and why stars like the Sun exhibit activity.The NSO has been making observations of this type for more than twenty years,observations that have contributed much to our present understanding of solar activity.However, these heavily used and productive synoptic facilities are aging and need to bereplaced to keep pace with major advances in the study of the solar interior and itsouter atmosphere, and to take advantage of the greatly improved quality and quantity ofobservational results that are enabled by modern detectors and data analysis techniques.SOLIS consists of four instruments that will replace and improve upon existing NSOfacilities. These instruments are a vector spectro-magnetograph to make high sensitivityfull disk measurements of the Sun's magnetic field, a full disk patrol to obtain highaccuracy images of solar activity in numerous photospheric and chromosphericspectrum lines, a solar coronal imager to measure the intensity and Doppler shift of fivecoronal emission lines above the solar limb, and a solar spectrometer to measure theshape and intensity of solar spectrum lines integrated across the solar disk. A $6.72M,three-year proposal has been submitted to the NSF. The project has been favorablyreviewed and is awaiting funding by the NSF. A scientific advisory committee of non-NSO experts has been formed to advise NSO how best to carry out its mandate toprovide state-of-the-art synoptic observations for the benefit of the nation's researchcommunity. If funding is provided, the project will commence later in 1997 and shouldbe finished in time for the next solar maximum expected in mid-2000. An operationallifetime of 25 years is planned.

b. Global Oscillation Network Group (GONG) Project

The Global Oscillation Network Group (GONG) is an international project studying theinternal structure and dynamics of the closest star by measuring resonating waves thatpenetrate throughout the solar interior—a technique known as helioseismology. To

67

overcome the limitations imposed by the day-night cycle at a single observatory, GONGhas built and deployed a six-station network of sensitive and stable solar velocitymappers located around the Earth that is obtaining nearly continuous observations ofthe "five-minute" oscillations, as well as direct measurements of the "steady" motionsof the solar surface itself. GONG has also established a distributed data reduction and

analysis system to facilitate coordinated analysis of these data by the community. Theprimary analysis is being carried out by half a dozen science teams, each focusing on aspecific category of problems. Membership in these teams is open to all qualifiedresearchers; there are currently 163 members, representing 70 different institutions.NSO is carrying out the GONG project in close collaboration with the community.

A history of the GONG project is given in Appendix B.

The full GONG network became operational in October 1995, and data from three ormore sites was obtained starting in May 1995. The instruments' performance has metour objectives both operationally and scientifically.

Analysis of the data and its distribution to the community have been able to keep upwith the data flow, and refinement of the algorithms is proceeding now that the networkdata is available.

The community published the first scientific results in a special issue of Science in lateSpring 1996, and the initial results were presented to the broad scientific community ata joint meeting with the American Astronomical Society and its Solar Physics Divisionin June 1996.

It has now been established, beyond a shadow of a doubt, that the solar resonantoscillations, which GONG was undertaken to study, vary significantly over the solarcycle. These solar activity-related frequency variations contribute both noise and signal.On one hand, the influence of the variations must be removed to achieve the ultimatepotential of the solar interior inversions. On the other, these variations represent one ofthe largest potential advances for understanding the time-variable Sun. We plan tomaintain the GONG network for a period comparable to the solar cycle, to achieve itsfull potential and to advance our understanding of the mechanism of the solarvariability.

We have undertaken a design study to evaluate the possibility of increasing the spatialresolution of the system by a factor of four. This appears to be achievable by changingonly the camera and its associated electronics, without any significant optical ormechanical modifications. We believe that deploying such a detector upgrade, on thetimescale of three years, would be a cost-effective attack on the next phase of our studyof the solar interior and a critical component in a systematic study of the variation ofthe solar interior over the solar cycle. This development will also provide continuoussurface velocity and magnetic field measurements, which will support other importantscientific objectives. NSO will continue to involve the scientific community in all of thetechnical and scientific decisions in the development of the replacement detectorsystem.

The cost of continued operation of the GONG network is included within the currentNOAO budget. Support for the new cameras will be requested from the NSF through a

68

special proposal. The cost of these two activities is shown separately in the followingtable. The increase in operating costs after the new camera is completed reflects that therate of data flow will increase by a factor of 16. However, we believe that withimprovements in the price/performance of modern computing systems, we can keep theincrease in annual operating costs to $200K. The increase is largely in the area ofstorage media and in processing. Some additional operating costs will be incurred in thesparing and maintenance of the increased on-site image storage electronics.

Camera Upgrade

Base BudgetNew Cameras

Annual Total

Fiscal Year

1998 1999 2000 2001 2002

1,850

402

2,252

1,850

835

2,685

1,850

574

2,424

2,000

2,000

2,000

2,000

2. Image Quality Improvement Program for NSO Telescopes

Major progress has been made in the last few years in improving the image quality withNSO's highest angular resolution telescope, the Sac Peak Vacuum Tower telescope. This isa result of a systematic program aimed at achieving diffraction-limited imaging with thattelescope. This program has successfully completed a number of components:

(a) Improving the optical quality of the telescope's entrance window by active thermalcontrol.

(b) Improving the optical quality of the alt-azimuth flat mirrors.

(c) Installing a rapid tip/tilt system relying on correlation tracking to locate the solar imagewith the required accuracy.

(d) Applying image selection techniques in combination with the correlation tracker toselect the moments of best atmospheric seeing for the observations.

(e) Developing a wavefront sensing system and applying it initially with a low temporalbandwidth to an adaptive mirror to correct the residual telescope aberrations (so-called"active optics").

(f) Speeding this system up, using parallel processing, to achieve a full-up adaptive opticssystem capable of correcting atmospheric seeing.

At this moment all steps of this program have been successfully implemented except for there-figuring of the altitude flat mirror and the speeding up to achieve full adaptive opticsperformance. Doing the latter is one of our highest priority programs for the next year.Given sufficient funding, we expect a full-up 20 Zernike adaptive optics system to befunctioning in two years (see section V.D.I.).

Because of the improvements already achieved, NSO has now obtained the highest angularresolution images of the Sun ever achieved.

69

Adaptive Optics is of course a central component of our plans for a large O/ER solarfacility, so that this program is essential technology development towards that telescope.

This program is supported as part of the NSO annual budget with the significant additionalfinancial and personnel support by the USAF Phillips Lab branch residing at Sac Peak.

3. Radiative Inputs of the Sun to Earth (RISE/PSPT)

Measuring and understanding the Sun's variability constitutes one of the most importantproblems in solar and solar-terrestrial research. Proposed in 1990, RISE is a systematicprogram of research to increase our knowledge of solar irradiance variations and theireffect on global change.

An important part of the overall RISE plan calls for the design, construction, deployment,and operation of precision solar photometric telescopes (PSPTs) that will work together toprovide a reliable daily (and higher cadence, as weather permits) record of solar surface-brightness variations. These telescopes will define state-of-the-art precision forgroundbased solar differential photometry. A proposal by NSO and the RISE ScientificAdvisory Committee to construct two PSPTs was accepted by the NSF (ATM) in 1992.The scientific objectives of the RISE program call for this PSPT network to be operationalfor at least three years.

Each PSPT will be a small, low-scattered-light refracting telescope consisting of a 15-cmdoublet, a fast guider, a 0.3-nm Call K-line filter (and several additional filters forcontinuum observations at other wavelengths), and a 2048 x 2048 CCD camera. Theenclosed telescope package will be equatorially mounted and computer controlled. Eachinstrument will achieve photometric precision approaching 0.1% per pixel across the full2048 x 2048 pixel array in continuum wavelength bands extending from the green to nearER wavelengths. Image data of this quality will elucidate the cause of the small irradiancevariations that are seen in satellite observations of integrated sunlight. The dailyobservations and occasional campaigns of higher-cadence time-series data will allow newstudies of the transport properties of the upper convection zone in the presence ofphotospheric magnetic fields.

The prototype instrument was completed in FY 1995. It was purchased by the RomeObservatory and is being operated now on an informal basis. A second telescope, fundedby the Rome Observatory will also be constructed this fiscal year. Deployment andoperations of the two NSO stations (in addition to the prototype which will be relocated atSac Peak) are expected to begin in FY 1997 and FY 1998.

The PSPT is jointly funded by the MPS/AST and GEO/ATM division of NSF.

4. Planning the Flagship Solar Telescope(s)

NSO is planning a major solar observing facility with the goal of having a phase A/Bdefinition in hand to present to the next astronomy decade review. At that time we will becelebrating the 28th, 31st, 39th, and 46,h anniversaries of NSO's large facilities! They stillhave a lot of life in them, but if we are to solve many of the outstanding questions in solarphysics and open up new areas in the observational parameter space so essential for new

70

discoveries, NSO will need a state-of-the-art, broadly capable solar facility. The contrastwith nighttime astronomy, which will see the construction of about 15 modern 8-m classtelescopes in the 1990s, is startling.

The international proposal for the Large Earth-Based Solar Telescope (LEST) was aserious attempt to achieve a modern high-angular-resolution solar facility for observationsat visible and near-infrared wavelengths. Unfortunately, it was rejected by the NSF, and theUS is no longer an active participant in that endeavor. A phase A study was completed ofan upgrade of the McMath-Pierce facility to a 4-m aperture telescope (known as the "BigMc"). This would greatly enhance the light gathering power of that telescope, as well as itsangular resolution at far-infrared wavelengths. Higher sensitivity for both solar andsolar-stellar observations would result. The McMath-Pierce facility is unique in being theonly sizable solar telescope in the world capable of imaging at all wavelengths out to thefar-infrared (35 microns). A third concept for a 4-meter aperture, all wavelength, lowscattered light solar telescope is currently in a phase A study. The acronym for thistelescope CLEAR, which stands for Coronagraph and Low Emissivity AstronomicalReflector, is somewhat misleading since its other primary characteristics are high angularresolution, broad wavelength coverage, and high polarimetric accuracy. This technical andbudgetary feasibility study is nearing completion. Parallel with this study, NSO is engaginga dialogue with the solar physics community to explore the scientific opportunities whichsuch a telescope presents. On the basis of these, priorities will be set and choices will bemade on the characteristics of the telescope, especially angular resolution/diameter,wavelength coverage, polarimetric characteristics, and coronagraphic/scattered lightproperties.

Parallel to the planning for this large O/ER solar facility, NSO is involved in a number ofrelevant technology development programs: (1) development of adaptive optics for solarapplications (see section V.D.I.); (2) testing of reflecting coronagraphs (USAF-supportedprogram resulting in two completed prototypes and another one underway); and(3) evaluation of the observing qualities at potential solar observing sites. The latterprogram emphasizes seeing testing. Evaluation of daytime seeing conditions are underwayfor a number of existing observing sites (Big Bear, Kitt Peak, La Palma, Mauna Loa, andSac Peak). First results show that under appropriate observing conditions seeing can be asgood at daytime observatories as it is at the best nighttime observatories (like Mauna Kea),lasting most of the day.

D. Instrumentation

Milestones for the Active and Adaptive Optics Program:

FY 1997: Finish active optics experiments.Detailed design and costing of low order AO.

FY 1998: Develop SH-WFS for low order AO. Procure reconstructor.

FY 1999: System integration [SH-WFS, reconstructor (Phillips Lab), XineticsDM] and tests at VTT/SP.

The instrumentation program given in this section and the planned upgrades in the next sectionassume level funding over the next five years. Some of the programs are therefore very limited

71

^1to

FY-1998 FY-1999 FY-2000 FY-2001 FY-2002 Total

NSF Labor] Non Payroll NSF Labor Non Payroll NSF Labor Non Payroll NSF Labor Non Payroll NSF Labor Non Payroll NSF Labor Non Payroll

(months) ($K) (months) ($K) (months) ($K) (months) ($K) (months) ($K) (months) ($K)

Infrared ProgramNear Infrared/Magnetograph 2* Rabin 5 7 5 7

Large-format array and controller Rabin 12 110 9 48 8 35 29 193

8-12 um array* Rabin 3 10 3 10 16 40 18 40 40 100

Telescope ImprovementKPVT Telescope Control System I Harvey 25 15 25 15

Refurbish CSIRO Filter Keller 3 2 3 2

McMath-Pierce Spectrograph Cont Keller 8 10 8 10

McMath-Pierce Telescope Control Keller/Harvey 40 30 39 25 39 40 118 95

Image Quality ImprovementMcMath-Pierce Seeing Keller 1 2 1 2

McMath-Pierce Guider Keller 11 11

Correlation Tracker Rabin/Rimmele 13 58 13 58

Detector Improvement

Fast CCD Imager Keller 6 20 6 20

Spectral Hole-Burning Device+ Keller 9 56 6 56 22 56 37 168

Imaging FTS* Hill 4 10 25 50 29 60

Subtotal 65 146 65 146 65 146 65 146 65 146 325 730

]'Supported in part by NSO partners+Requires additional incremental funding (NSF,USAF,NASA,etc.)

H65a;n"

in scope. Where more rapid progress is desired, like the ones related to image qualityimprovements, initiatives are indicated in the previous section. Table 3 summarizes theinstrumentation and upgrades program.

1. Active and Adaptive Optics

Observations of highest resolution are extremely important to address fundamentalscientific questions in solar astronomy [e.g., coronal heating, the physics of fundamentalmagnetic elements (flux tubes), and magnetoconvection]. Substantial progress could bemade in these areas if existing telescopes like the VTT/SP were equipped with an adaptiveoptics (AO) system providing diffraction-limited resolution over an extended period oftime. It is also widely acknowledged that any new large aperture, groundbased telescopewill need AO to be an effective tool for high resolution solar astronomy.

NSO/SP is developing an AO system to be deployed as a user system at the VTT/SP. AnAO system consists of three major components: a wavefront sensor that measures thewavefront aberrations introduced by atmospheric turbulence, a reconstructor that convertsthe wavefront measurements into drive signals for the third component, and the deformablemirror, which compensates the wavefront errors. Deformable mirrors and reconstructors arenow commercially available or can be cloned from nighttime AO systems.

The critical challenge facing the Adaptive Optics program at NSO/SP is developing a solarwavefront sensor. Solar wavefront sensing is difficult because point source sensing targetsare never available everywhere on the solar disk. Most nighttime AO systems use Shack-Hartmann wavefront sensors (SH-WFS), in which a point source target (natural or laserbeacon) is imaged through an array of subapertures. The displacement of each subapertureimage with respect to a reference is proportional to the wavefront slope within thatsubaperture. Displacements can be measured by simple center-of-gravity algorithms orquad-cell devices. However, tractable sensing targets such as small sunspot or pores arescarce on the Sun, and laser beacons are not bright enough to provide practical substitutes.In general, solar wavefront sensing has to be performed using the granulation, an extended,low contrast and evolving structure, as its sensing target.

Two wavefront sensing approaches are currently under consideration at NSO/SP: 1) amodified Shack-Hartmann wavefront sensor that employs cross-correlation techniques tomeasure the displacements of granulation images formed by a subaperture array; and 2) aspatial filtering approach that uses a mask placed in an image plane to modify the light in away that makes wavefront gradients visible as intensity fluctuations in a pupil plane. Thelatter technique can be understood as a generalization of the classic Foucault knife edgetest, in which the "mask" is a straight edge placed over image of a point source. Typicallythe partial derivatives of the target structure are mapped in binary form on a transmissiveliquid crystal screen which is placed in an image plane. A small, fast CCD records thewavefront information in a pupil plane.

There are advantages and disadvantages to both approaches. Shack-Hartmann wavefrontsensors are widely used in nighttime AO, and are comparatively well understood. The low(order 2%) contrast of the granulation image formed by a small subaperture is not aproblem. The Mark II correlation tracker recently completed at Sac Peak has successfullydemonstrated its ability to stabilize the granulation image formed by a 7-cm subaperture—

73

i.e., a single channel of a correlating solar SH-WFS has been implemented in hardware. Togain a more detailed understanding of the SH-WFS approach, we have also recordedsimultaneous SH-WFS data and granulation images. The measured instantaneouswavefronts are being used to deconvolve a time sequence of granulation images.Preliminary results show that the SH-WFS provides phase information with the requiredaccuracy.

We also have used a SH-WFS to test the concept of an active optics system operating at theVTT/SP. The active optics system is intended to correct aberrations in the VTT optics andalso serves as a stepping stone towards a full AO system. Using a 127-element LCD-wavefront corrector provided by the Phillips Lab/Albuquerque, we were able to close theloop and correct the optical aberrations present in the VTT/SP, improving the Strehl ratiofrom 0.28 to 0.9. A Xinetics 97 actuator deformable mirror (DM) was purchased usingyear-end funds provided by the Air Force Phillips Lab and NSO. The 97-channel DM,which in contrast to the LCD-wavefront corrector has sufficient bandwidth to be used in an

atmospheric compensation system, was implemented in the active optics system and testedat the VTT. The loop was closed and the capability of the system to measure and correctoptical aberrations introduced, for example by the entrance window of the VTT, wassuccessfully demonstrated.

In principle, the "LCD Knife-Edge" approach offers an elegant and inexpensive solution towavefront sensing, since much of processing is performed optically rather than digitally.Wavefront slopes are directly coded as pupil plane intensity variations by the appropriatemask. No subapertures are involved. Theoretical studies and simulations performed severalyears ago at NSO/SP by J-M. Conan and R. Radick did suggest potential problems withsensitivity and resolution when granulation is used as target structure. Engineering issues,like noise due to the refresh cycle of the LCD and the imperfect contrast of the LCD mask,pose further practical problems. Ln December 1996, observations made were aimed atproofing the actual "LCD Knife-Edge" approach using an active matrix LCD computerdisplay as the focal plane mask. Various target structures (sunspots, pores, and granulation)were observed, and various approaches for generating mask on the LCD were explored.The principal result emerging from this experiment is that the "LCD Knife-Edge" approachencounters increasingly severe S/N problems as the contrast of the target structuredecreases. With a sunspot as the sensing target, we typically achieved a S/N of 3.5. A smallpore delivered a S/N of about 2. Granulation, however, produced a S/N of unity or less,even under good seeing conditions. We are currently exploring ways to improve the S/N;unless this can be done, the correlating SH-WFS may be the only practicable approach tosolar wavefront sensing.

A SH-WFS for a solar AO system operating on granulation requires the implementation ofhigh-speed cross-correlation algorithms. Although the hardware needed to do this is nowcommercially available, the bandwidth requirements point to a massively parallel system.We estimate the cost of a full-up AO system with 80-100channels to be of order $1M+. Atthis point, we are confident enough to contemplate proposing to the NSF and othersponsors implementation of a full-up AO system. In the meantime, we will pursue theimplementation of a low order AO system as the logical next stepping stone toward thefull-up AO. A low order system would significantly improve the capabilities of VTT andproduce new science and thus support the proposal for the full-up AO system.

74

Provided the low order system has sufficient bandwidth (> 100 Hz) to correct on the orderof 20 Zernike modes, the Strehl ratio will be:

~ 0.6 for r0 = 15 cm (good-excellent seeing);~ 0.5 for r0= 10 cm (good seeing);-0.1 for r0 = 5 cm (medium-bad seeing).

This means the system will deliver satisfactory results only in good seeing conditions in thevisible. However, in the near infrared the system will deliver diffraction-limited imagery atthe VTT/SP with even unexceptional seeing.

2. Mark II Correlation Tracker

The development of the Mark El correlation tracker has been completed, and the system isnow operational at the VTT/SP. Compared to the Mark I CT, the Mark El CT has improvedperformance, in particular a higher servo bandwidth (60 Hz) and a capability to track lowcontrast granulation images during moments of bad seeing. We are considering cloning theCT system to provide tip/tilt correction at the McMath telescope. Also the Big Bear SolarObservatory has expressed interest in getting a copy of the Mark II CT.

3. Near Infrared Magnetograph 2

NEM-2 is a near-infrared imaging magnetograph based on a rapidly tunable Fabry-Perotetalon and an infrared array camera. It will complement rather than replace the existingspectrograph-based NEM instrument. NEM-2 exploits the highly Zeeman-sensitive 1565 nmline of Fe I (g - 3) to provide true field strengths in sunspots and plages. The line iscompletely split in kilogauss fields, so the three polarized Stokes components arecomparable in magnitude, an advantage for vector magnetometry. At the expense ofsomewhat lower spectral resolution, NEM-2 will have an advantage over NEM in studyingspatial structure and evolution because the whole spatial field is imaged simultaneouslyrather than built up by scanning the image.

NEM-2 is undergoing engineering tests and will be commissioned during FY 1997.

4. CCD Cameras

For over a decade NSO/SP has been relying on the Multiple Diode Array (MDA) system asits main data acquisition system. The MDA system was custom developed at NSO/SP.Although state-of-the-art at the time, this RCA CCD-based system is now outdated.NSO/SP is upgrading its CCD detectors with the goal of replacing the MDAs by state-of-the-art, large format, fast-readout CCD systems. Where possible, commercially availablesystems will be implemented. Ln-house development efforts are kept to the necessaryminimum.

New CCD systems currently available to the users are:

• Thomson IK x IK, fast read-out (10 Mpix/sec), 10-bit resolution, 4-5 frames/sec.

75

• Kodak Megaplus 1300 x 1024 pixels, fast read-out (lOMhz), 8-bit resolution,5 frames/sec.

• XEDAR 2K x 2K (developed by XEDAR ENC. for the RISE project). Fast read-out[8 Mpix/sec (multiplexed mode), 16Mpix/sec max, 4-quadrant], 2 frames/sec, highdynamic range (30e read noise, 12-bit resolution), low dark current [thermoelectriccooling (-30C)], MPP mode, Lumigen coated.

Features of the new CCD systems include frame selection and on-the-fly co-adding ofimages. Each CCD detector has its own data acquisition, i.e., parallel datastream frommultiple cameras.

Driven by user requests, in the near future we plan to add the following CCD systems:

• one to two additional XEDAR 2K x 2K cameras.

• IK x IK blue sensitive CCD (40-50% QE @ Ca-K) for K-line.• two to three 1024 x 300 spectroscopic CCDs, for Horizontal and Echelle spectrograph

observations.

5. Infrared Cameras

To realize their full potential for infrared solar astronomy, NSO's major telescopes must bemated to modern infrared cameras and data systems. NSO will participate in an NOAO-wide initiative to incorporate 512x512 or larger infrared arrays in astronomicalinstruments.

In typical solar applications, low-noise infrared arrays saturate in a fraction of a second. Tobuild up the signal-to-noise ratios that will often be required, it is necessary to averagemany frames of data at each spectral/spatial location. Thus, NSO needs data systems withreal time processing. There is a similar requirement in nighttime work, so the effort is againNOAO-wide.

Because much of the infrared work at the McMath-Pierce Telescope is at wavelengthslonger than 2.5 microns, NSO/KP has asked NOAO to supply one or more ALADDIN (1-5microns) arrays from the joint NOAO/USNO development effort, and NSO/KP hasreserved internal funds for an NOAO-developed ALADDIN array controller.

At NSO/SP, ER observing accounts for approximately 25% of all allocated telescope time atthe ESF and VTT. It is necessary to develop a better infrastructure to satisfy the needs ofthe unique aspects of its ER capabilities. Resources required are a liquid crystal polarizer,leak detector system, new detectors, and a tunable filter. These will provide the basic toolsfor coronal spectroscopy and high-resolution magnetometry out to 2.5 microns. Our longrange goals include the development of new ER arrays based on HgCdTe materials.Continued progress in upgrading the existing near-ER camera systems with larger formatarray detectors will be needed. The current SP/ER program has been limited by theavailability of our shared cameras. The HgCdTe arrays were jointly developed with MSUand the Wyoming Infrared Observatory and are available to the NSO/SP user communityfor only a fraction of the year.

76

E. Facilities/Infrastructure Upgrades

1. KPVT Optics Upgrade

The NSO Kitt Peak Vacuum Telescope is one of the largest solar telescopes in the world. Ithas a window of 86-cm aperture and an image forming mirror of 70-cm aperture. Thetelescope was built in 1973 to provide seeing-limited synoptic observations of the Sun'smagnetic field and chromosphere. The focal plane instrumentation was built with onearcsecond sampling. Twenty-three years of experience have shown that the image quality isoften limited by this coarse sampling rather than by the intrinsic seeing on Kitt Peak. Newinstrumentation either built recently or planned will permit finer sampling. Experience hasalso shown that the optics of the telescope are frequently a limiting factor in getting highangular resolution. Measurements indicate that several of the optical components are notfigured to contemporary quality standards. Additionally, the image forming mirror ismounted in such a way that it vibrates at about 30 Hz with an amplitude that may reach onearcsecond. Remounting this mirror and refiguring the worst optical elements will allow theKPVT to produce significantly better image resolution. This will not only improve thequality of data for existing research projects, but will also enable new projects that are notcurrently practical. In this way, the facility will be able to continue to make majorcontributions to solar physics as emphasis on high angular resolution continues to increase.This project is proposed to be jointly supported by NOAO and NASA.

2. McMath-Pierce Telescope FTS Polarimeter Upgrade

The NSO Fourier Transform Spectrometer located at the McMath-Pierce Telescope is aunique and very powerful instrument for studies of the solar system. It was installed in1975, and a few years later a polarimeter was built as an accessory to obtain solar spectra incircular and one state of linear polarization. This polarimeter has been very successful andmuch of what we know about solar magnetic fields can be attributed to this instrument.More could be learned with some modest upgrades. First is a mechanical change to permitcalibration optics to be reproducibly placed in the beam. Some electronic changes wouldproduce a significantly higher signal-to-noise ratio. A change to the modulator assemblywould improve beam pointing stability. Adding a movable half-wave plate would allow allstates of polarization to be measured efficiently. A final, fairly major upgrade would be toadd a zinc selenide modulator and suitable detectors to allow polarizations to be measuredlongward of the current 2.5 microns limit.

3. Evans Solar Facility Instrument Upgrades

The new ESF control system is in place. Many of the instruments were upgraded in theearly 1980's, but there was not close follow up to maximize their performance. The newcontrol system has uncovered some deficiencies, which are straightforward to correct. Theimprovements would complete the automation started in the 1980s and would improve thedata quality.

The automation will allow the observers to make quality observations with a minimalamount of daily setup and interaction. This is especially important because of the reducedobserving staff size. The data quality would improve because, e.g., settable parameterswould always be set the same way, and instrument stability would be improved. Also,maintenance would be easier and less time consuming on an upgraded system rather than

77

the old systems. Improvements already have been implemented since the new controlsystem has been in place, and we are anxious to finish the improvements.

Anticipated upgrades in FY 1998 - FY 2000 are:

Replace the relay racks with solid state relays.Upgrade the coronagraph aperture select, coronal lens select, dustcap, and fan controlto allow for computer control.Upgrade the air lines and air pressure system.Add coronagraph fast mirror guider.Add sky brightness monitor.Add coelostat guider.Add linear arrays on the Spectroheliograph to replace film.Add linear array detector for the LSG.Automate the Littrow Spectrograph.

4. KPVT Telescope Control System Upgrade

The goal of this project is to upgrade the 23-year old control and guiding system of theNSO/KP Vacuum Telescope. The maintenance of the guider and control system hasbecome difficult. Many of the parts are obsolete or no longer available. The guider does notfunction properly in some operational modes or in light clouds. Recent magnetographcomparisons indicate that spurious image motion is a serious problem for overall magneticcalibration and they highlight the importance of accurate polarimetry. These problems arebeing addressed by a multi-year, phased upgrade project supported by both NOAO andNASA. The first elements of the upgrade project have been completed with the installationof new stepper motor controllers and a new drive gear reducer for the large mirror used toscan the solar image. A new mounting for another mirror is about to be installed. Thismount will allow it to be moved rapidly to correct deleterious image motion. The phases ofthe project which are yet to be completed are described next.

The fast image motion compensation will be completed in 1997 by replacing the detectorson the existing limb guider translation stage by ones which operate much more rapidly andhave no moving parts. These will provide both an error signal to feed to the movable mirrorand also much improved guiding performance under less-than-ideal observing conditions.

A major part of the upgrade is to replace the ancient PDP 11/73 and CAMAC controlsystem. The design of this upgrade is complete and involves the use of a distributed controland command system. The design uses well-established hardware and software methodsproven in the field of industrial control, and is similar in many ways to the systemdeveloped for the GONG project. It breaks ground for the proposed SOLIS facility in itsability to be accessed remotely and in managing large amounts of data efficiently. It will beeasily maintainable and has flexibility to allow new capabilities to be added easily.

5. McMath-Pierce Telescope Control System Upgrade

The McMath-Pierce is the largest solar telescope in the world. Upgrades to the telescopecontrol system are needed to ensure that the facility remains competitive and maintainable.

78

The proposed upgrades will take advantage of the engineering experience from theupgrades to the KPNO 2.1-m and 4-m telescopes, as well as the NSO KPVT and VTT.

The proposed control system will allow accurate pointing, tracking, guiding, and scanningof the main and East Auxiliary telescopes at any coordinate on the solar disk. Thiscapability is particularly needed for collaborative efforts with other ground- and space-based solar telescopes.

The computer will be a VME-based system with VX-Works and an Ethernet interface. Thesame computer system will be used to replace the obsolete FORTH-based TelescopeControl Program and to bring the dual-grating (infrared/visible) main spectrograph, whichalready incorporates accurate absolute encoding, under flexible computer control.

VI. NOAO SUPPORTING SERVICE

A. Computer Support

Milestones

Dec. 1997: Release of Open IRAF, phase 1Oct. 1998: Completion of ATM link between La Serena and Cerro TololoDec. 1998: Release of Open IRAF, phase 2Oct. 1999: Extension of ATM link to Cerro Pachon

Dec. 1999: Release of Open IRAF, phase 3Jan. 2000: 100 Mbps Fast Ethernet extended to every NOAO-Tucson scientist's

desktopOct. 2001: Establishment of ATM link between CTIO and Tucson

1. Save-the-Bits

The Save-the-Bits Project provides a mechanism to archive all data from the NOAOnighttime telescopes. The system has been in use at KPNO for several years, and wasrecently implemented at CTIO as well. Data from the McMath nighttime program and fromthe WIYN Observatory have also been routinely archived as part of Save-the-Bits. TheWIYN Telescope Board recently approved and funded a CDROM-based upgrade that willserve as the primary data distribution mechanism for the telescope, in addition to itsarchival functions. Save-the-Bits has also been adopted by other observatories, includingKeck, Lick, and the CFHT.

At NOAO, Save-the-Bits data are stored on Exabyte tapes. Two copies of all data arewritten, and stored at different locations. The Save-the-Bits archive provides the primaryback-up for observers' data. The volume of NOAO data archived is now 2 terabytes, andmore than 1 million science images have been saved. In addition to providing new copies ofdata lost to media failures, the Save-the-Bits Archive is also used to collect data obtainedby different observers monitoring particular objects. For example, an astronomer in Hawaiiis conducting a long-term campaign to study comets and asteroids. Save-the-Bits is alsoused for special, intensive observing programs such as the SONG observations of Procyon.

In the long term, an archive's real value is only achieved when its data are accessible fornew scientific programs. It is the goal of Save-the-Bits to adapt existing and successful user

79

interfaces such as those demonstrated by STScI and by the CADC to make the NOAO dataavailable to the community. Data from the existing NOAO nighttime programs will providea path for the development of NOAO services for the archive of Gemini data for the UScommunity.

2. Downtown Tucson Computer Support

The computer facilities run by CCS in the Tucson office complex serve three general needsfor NOAO-Tucson: data reduction and analysis for the scientific staff and visitors, generalcomputing for all staff members, and ERAF development and support. Our distributedcomputing strategy for Tucson implements a combination of central, shared facilities, and avariety of desktop facilities including workstations, X-terminals, and personal computers.Computing systems are linked with a set of Ethernets transmitted by wire in Tucson, opticalfiber on Kitt Peak, and leased-line between Tucson and Kitt Peak. (Work stations forindividual staff members are funded by the division or unit to which they are assigned.)

In FY 1999 we will continue an extensive improvement program to our central computingfacilities, replacing older high-maintenance systems with modern low-maintenance ones,achieving concurrently a major upgrade of capabilities. In subsequent years, we willcontinue our program of upgrades for increased performance with reduced maintenance andoperating costs. In general, the lifetime of a computer or peripheral is four years.

Thus, in FY 1998 we plan to replace our general timesharing machine, Orion, and ourprinter control machine, Solitaire. In FY 1999 and FY 2000 we will replace the server forthe Science Workstation Network, Gemini. In FY 2000 and FY 2001 we will restart the

cycle and replace our "fast" machine, Ursa (which was replaced in FY 1996) and inFY 2001 we will replace our ERAF development machine, Tucana, which we plan toreplace in FY 1997.

We also plan to continue our ongoing projects to restructure the network infrastructure inthe downtown NOAO office complex to provide multiple 100 Mbps network connectionsin each office, tie all these connections into an efficient building network, provide effectivefacilities for staff members to utilize our computer systems from home or while traveling,and provide adequate connectivity to the worldwide internet.

A staff of approximately 5 FTE maintains the downtown computer facilities (approximately100 machines including desktop workstations and approximately 200 personal computers),supports the network, and provides support to users (both staff and visitors).

3. IRAF

ERAF (the NOAO Image Reduction and Analysis Facility) is a portable software systemused for astronomical data reduction and analysis, general image processing and graphics,and astronomical software development. The ERAF software, first released over a decadeago, is now in constant use within the NOAO observatories, at over a thousand sites in theworld astronomical community, and within the NASA astrophysics community.

Approximately 6 FTE NSF-supported programmers work on ERAF systems software andscientific applications and provide support to the ERAF community. Continuing effort isrequired to keep ERAF current with the ever-changing world of computing and to

80

accommodate the ever-changing data acquisition, reduction, and analysis needs of NOAO.(Current projects include improved software for large CCD and ER arrays such as theNOAO mosaic-CCD imager and an efficient and useful real-time display system for usewith the mosaic and other instruments.)

The ERAF group has begun a large new effort, supported with funds from NASA'sAstrophysics Data Program, in collaboration with STScI, SAO, and EUVE. Called "OpenIRAF," the goals of this effort are to:

• make it easier for non-experts to develop applications in the IRAF environment.• make it easier to use pieces of IRAF independently, outside of the IRAF system.• make it possible to use pieces of IRAF with non-ERAF software.• extend support for additional, popular image formats.• make it easier for others to enhance or extend the IRAF system software.

The plans for acquisition of new computer systems are summarized below in Table 4. Onlythe funds for the NOAO/Tucson portion of the program are funded out of CCS. The otherportions are funded from the divisional budgets. In addition, individual workstations arepurchased as required.

Table 4

NOAO/Tucson Schedule for Major Computer Capital Expenditures

Item 1998 1999 2000 2001 2002 2003

Central Facilities 65K 65K 65K 65K 65K 65K

Downtown Network

Infrastructure 35K 35K 35K 35K 35K 35K

Total $100K $100K $100K $100K $100K $100K

4. KPNO Computer Support

KPNO will modify and improve the telescope control systems at its major facilities duringthe course of this long range plan. Adoption of the Gemini software TCS where plausiblemay save substantial software effort. The hardware to support this code is primarily a VMEenclosure with a CPU, WWV interface, and some hardware to connect to the observatoryhardware.

In the realm of instrument control, the adoption of Gemini standards and the constructionand testing of actual Gemini instruments will drive changes at the major KPNO facilities.The same VME enclosure and hardware as for the TCS is needed. Improvements to thedome LAN will be needed, and the LAN must be brought to the telescope. Additionally weneed cooled enclosures on the telescope for the VME hardware.

During the next five years, the internet bandwidth will surely increase, as will the demandson the data line from Tucson to Kitt Peak. As technology permits, we will increase thebandwidth of this link beyond the current level of two Tl lines (1.5 Mbps each) and alsoimprove the bandwidth and robustness of the Kitt Peak mountain network. The overallbandwidth between Kitt Peak and Tucson is limited by the costs and capabilities of thetribal utility company.

Table 5

KPNO Schedule for Major Computer Capital Expenditures

Item 1998 1999 2000 2001 2002 2003

Telescope Control 25K 25K 25 K 25K 25K 25K

Instrument Control 20K 20K 20K 20K 20K 20K

Mountain IRAF workstations

and upgrades 45K 45K 45K 45K 45K 45K

Computer Spares 10K 10K 10K 10K 10K 10K

Mountain Network

Infrastructure 15K 15K 15K 15K 15K 15K

High Speed Tucson-KP Link 40K 40K 20K 20K 20K 20K

Downtown Staff Computing 60K 60K 60K 60K 60K 60K

Archiving 25K 25K 25K 25K 25K 25K

Total $240K $240K $220K $220K $220K $220K

5. CTIO Computer Support

The CTIO computer plan covers both the Cerro Tololo and La Serena sites. The bulk of theexpenditures each year will go towards maintaining and gradually upgrading the individualworkstations (currently 55) in our network of Sun computers, which are used for dataacquisition, data reduction, and laboratory development work. To keep pace with theincreasing size of detectors and to avoid falling too far behind the state-of-the-art incomputer hardware, we need to replace or substantially upgrade our machines onapproximately a four-year cycle at an annual cost of at least $70K. During FY 1996 severalolder workstations used by scientific staff and by members of the CTIO programminggroup were replaced with new machines. In FY 1997-1998 there will be a substantialupgrade of the data acquisition computers at the 4-m telescope and of the downtownreduction facilities in order to support the CCD mosaic imagers (BTC in service now, andthe NOAO mosaic scheduled for deployment at CTIO in mid-1998).

Until recently all computers on Cerro Tololo were connected to a single 10 Mbps Ethernetbackbone. During FY 1996 we began the process of partitioning the mountain network intoseparate sub-nets. All the computers at the 4-m telescope were connected via a smart hub toform a local network coupled to the backbone through the hub. During FY 1997 a similarsub-net will be created at the 1.5-m telescope. We will also split the La Serena network, byestablishing a separate sub-net for the scientific workstations. At the same time the 4-mtelescope and scientific sub-nets will be upgraded to 100 Mbps Ethernet.

The networks in La Serena and on Cerro Tololo are connected via a commercial microwave

link with backup and maintenance provided by the contractor. At present, one of the fouravailable El channels (2 Mbps) is used for a network connection and another for voicelines. During FY 1998, we plan to install ATM routers in La Serena and on Cerro Tololoand begin to use ATM on the link between the two sites. The computer networks in bothlocations will be connected to the switches along with the existing telephone plants. ISDNboards will be installed in both telephone plants along with ATM trunk line equipment.

82

Changing the link structure in this way incurs a capital cost of approximately $70K for thetwo ATM switches required. This will immediately increase the maximum speed of datatransfer between La Serena and Cerro Tololo from the current 2 Mbps (El) to 8 Mbps (E2)at no additional recurring cost. This, combined with the installation of boards supportingISDN lines to external equipment will permit videoconferencing and direct computerconnections at up to 128Kbaud to be made between the two sites and to all ISDN-compatible locations worldwide, such as Tucson and Hawaii.

During FY 1999, we anticipate that the Cerro Tololo ATM switch will be connected to aplanned link to Cerro Pachon to meet the needs of Gemini and SOAR and perhaps also toLas Campanas. An upgrade of the speed of the CT-LS connection, to at least E3 (34 Mbps)will be necessary in order to support the resulting increase in traffic. The scalability andflexibility of ATM make network changes of this nature straightforward, relatively lowcost, and non-traumatic. Cost sharing arrangements will have to be worked out between thevarious users (CTIO, Gemini, SOAR, etc.). The budget shown only includes an estimate ofCTIO's share of these costs.

The system connectivity to the outside world is currently throttled by a 56Kbps satellitelink to the US, supplied by NASA. A higher speed connection is a prerequisite forremote/queue scheduled observing and will be essential before the turn of the century.Fortunately, a transition to the use of terrestrial fiber links for telecommunications is takingplace in Chile. Fiber connections already exist between Santiago and La Serena and to theUSA via Argentina and Brazil. Direct connections from Chile to the US west coast arebeing installed. As a result it is likely that a higher speed connection between CTIO and theUS will be commercially available, at an affordable cost, within the next few years. TheATM switch in La Serena can be connected directly to such a terrestrial fiber optic link,once it becomes available, to provide a high bandwidth connection to the US Internet.However, it must be recognized that rental and traffic costs for this connection will need tobe paid from CTIO's budget, while the present satellite connection is paid for by NASA.The budget shows these increased recurring costs starting in FY 1999. Again, we willinvestigate cost sharing arrangements with the other observatories in order to obtain thenecessary bandwidth in the most cost-effective way.

During FY 1996 a 64Kbps terrestrial link between La Serena and Cerro Calan Observatoryin Santiago (and thence to the Chilean Internet) was implemented in partnership withUniversidad de Chile. This allows for more efficient communication between CTIO and the

Chilean astronomical community and the other observatories in Chile. It also serves as abackup internet connection.

83

Table 6

CTIO Schedule for Major Computer Expenditures

Item 1998 1999 2000 2001 2002 2003

Distributed Computing:Spares/Maintenance 35K 35K 35K 35K 35K 35K

Upgrades 95K 95K 70K 70K 70K 70K

Network Infrastructure (capital costs):La Serena - Tololo Link

Local Networks

70K

10K

10K

10K 20K 20K 30K 30K

Network Recurring Costs:Chile -US Link

La Serena - Tololo Link

10K

15K

20K

15K

30K

30K

35K

30K

40K

30K 30K

Total $235K $185K $185K $190K $205 K $165K

6. NSO/SP Computer Support

The computer facilities at NSO/SP consist of a campus-style layout with five main areas;Main Lab (ML) building, Evans Solar Facility (ESF) telescope building, Hill Top (HT)telescope building, the Vacuum Tower Telescope (VTT) building, and the administrativeoffices building. Each building is connected to our central Ethernet switch by a lOBaseFfiber optic link, and is wired or will be wired for lOBaseT (cat5 twisted pair).

The HT, ESF, and VTT computer systems are mainly used for telescope control and datacollection with limited data analysis. The ML facility is used for data reduction, analysis,and general computing by both visiting scientists and local staff.

A staff of 1 FTE maintains the NSO/SP computer facilities (approximately 140 nodesincluding 40 workstation/servers, 17 X-terminals, 20 personal computers), supports thenetwork, and provides support to users (both staff and visitors).

Our main objectives over the next five years are to:

• upgrade our old, inefficient, and high maintenance computer systems with newhardware.

• upgrade our network towards a 100 Mbps plus network. This will include upgradingour Ethernet switch to a high bandwidth switch and high bandwidth links to eachbuilding, finishing the upgrade of all buildings to 100BaseT wiring, and adding smallerswitches and hubs to segment the network load at each building.

• purchase a DLT 7000 tape unit for the ML server. Allow data collected at the VTT tobe read at the ML.

• provide an optical disk system for data storage and data archiving.

• move toward a centrally administered network by purchasing software that will allowthe 1 FTE to manage the network from a single site.

• hardware/software upgrades.

84

Table 7

NSO/SP Schedule of Major Computer Capital Expenditures

(Fiscal Year)

Item 1999 2000 2001 2002 2003

Computer Supplies 25K 25K 25K 25 K 25K

SW Maintenance 10K 10K 10K 10K 10K

HW Maintenance/Upgrades 25K 25K 25 K 25K 25K

New Purchases 2K 10K 10K 10K 10K

DLT 7000 Tape Drive 8K

Total* $70K $70K $70K $70K $70K

* The above chart assumes level funding over the next five years.

7. NSO/Tucson Computer Support

NSO/Tucson will address five areas:

• On Kitt Peak, a continuation of a major modernization of data acquisition, datareduction, and telescope control at the Solar Complex: Currently, the TelescopeControl System (TCS) of the Kitt Peak Vacuum Telescope is being upgraded to takeadvantage of the latest developments in real-time operating systems and databasemanagement. This, in conjunction with a completely revamped guiding system, willgreatly improve the productivity of the telescope. The lessons learned from this projectwill then be applied either to SOLIS (if funded), or to the McMath-Pierce telescope,completing the modernization of the NSO/KP observing facilities.

• In Tucson, an ongoing program to provide and upgrade scientific workstations orX-terminals and associated peripheral devices: The ever-increasing size of solardatasets and the further development of sophisticated and numerically intensiveanalysis algorithms creates a perpetual need for additional resources, particularly datastorage space, I/O transfer rates, and CPU speed. These demands will be met by acontinual migration towards higher-speed multi-processor SPARC stations, largercapacity disk drives, denser off-line storage media (such as the 7/14-GB Exabyte drivesor Digital Linear Tape drives), and perhaps optical storage devices. NSO already hasboth Hyper- and Ultra-SPARC workstations in-house and will continue to upgradeolder equipment as funds allow. In addition, better access to commercial softwarepackages has been provided by the installation of EDL on several machines. This trendwill most likely continue in the future.

• On-line archival storage for major datasets such as FTS spectra, full-disk synopticmagnetograms and spectroheliograms, and the High-l Helioseismometer: Thearchiving of the entire set of transformed FTS spectra onto CD-ROM is complete. SolarFTS Spectral atlases are on-line and heavily used by the scientific community. Recentmagnetograms and He I 1083-nm Spectroheliograms are available to the usercommunity via anonymous FTP and the World Wide Web. A 300-CD (200 GB) CD-ROM jukebox has been installed on Argo (NSO/Tucson's data server) and nowcontains 57 disks. The Kitt Peak Magnetograms and the untransformed raw FTS dataare being migrated onto CD-ROMS. Ln about one to two years, the new generation of

85

Digital Video Disk (DVD) technology will be widely available, providing as much as13 GB of storage on a single 5.25-inch CD-ROM. This increase of a factor of 20 overcurrent capacity will allow NSO to place its entire archive on-line. A user interface tothe digital library is now being developed using funds from the NSF Space Weatherprogram.

• Stable access to programming resources sufficient to allow the planning andexecution of major software packages: Although this situation was greatly improvedwith the addition of three assigned programmers to the NSO staff, the subsequent lossof one position to budget cuts eroded the capability of NSO to develop new software.Programming time continues to be the limiting resource in NSO/Tucson projectplanning.

• Increased presence on, and use of, the internet: The NSO Web pages are continuallyupdated. The user interface to the search engine for the digital library is Web-based.NSO/Tucson is also contributing to the development of Web-based educationalmaterials on solar topics. Finally, NSO/Tucson has successfully experimented with theuse of the internet for limited scientific meetings. Potentially, conferencing over theinternet could result in a substantial reduction in travel costs.

The following schedule shows that NSO/Tucson derives the great majority of its fundingfor computer-related capital expenditures from outside grants and contributions from its"partner" agencies.

Table 8

NSO/Tucson Schedule for Major Computer Capital Expenditures

Item 1998 1999 2000 2001 2002

Distributed Computing 20K 45K 5K 15K 20K

Equipment Upgrades 26K 24K 34K 25K 25 K

Application Software 6K 9K 3K 10K 10K

Archives 15K 3K 13K 20K 25K

Total $67K $81K $55K $70K $80K

Non-AURA Percentage 85% 88% 82% 85% 88%

B. Facilities

Table 9

Facilities Maintenance Requirements

Maintenance, Construction and Safety Summary 1998-2002

CTIO $660K

KPNO $807K

NSO/Kitt Peak $245K

Tucson $472K

NSO/Sacramento Peak $604K

Total $2788K

86

1. Cerro Tololo

Milestones:

Projects (FY 1998) Estimated Cost

1. 4-m telescope: Fire protection system $ 40K2. Vehicle fleet renovation 20K

3. Main access road to Tololo 75K

Total $ 135K

Projects (FY 19991 Estimated Cost

1. 4-m telescope: Repainting $ 20K2. Vehicle fleet renovation 20K

3. Main access road to Tololo 75K

Total $115K

Projects (FY 2000) Estimated Cost

1. Water system, Tololo (pipeline) $ 25K2. Vehicle fleet renovation 20K

3. Main access road to Tololo 75K

Total $ 120K

Projects (FY 2001) Estimated Cost

1. Power House $ 25 K

2. Vehicle fleet renovation 20K

3. Main access road to Tololo 75K

Total $ 120K

Projects (FY 2002) Estimated Cost

1. Bus Replacement $ 150K2. Vehicle fleet renovation 20K

Total $ 170K

With the recent closings of the Lowell and Yale telescopes, CTIO now operates four nighttime telescopes: the 4-m, 1.5-m, 0.9-m, and the Univ. of Michigan Curtis Schmidt. By theyear 2002, we expect to be participating in the commissioning of the SOAR 4-m telescope,most likely on Cerro Pachon. In addition to the routine maintenance of the physical plantson Tololo and La Serena, we propose to carry out the following major maintenance andconstruction activities:

a. Maintenance

Fire Protection:

A fire occurred in the 4-m building in 1995 which, if not detected in time, could haveproduced serious damage to the facility. The fire occurred in an area that is out of sightof normal operations personnel but not covered by fire detection devices. Serious fireshave occurred recently in astronomical facilities on Mauna Kea, one with fatalconsequences. We have recently installed a significantly improved system of fire

87

detection at the 4-m, but our top infrastructure priority at CTIO is to improvesignificantly the level of fire protection at the 4-m.

Tololo Water System:The San Carlos water line (CTIO's only water supply) has been in continuous operationfor over 20 years and is nearing the end of its expected 25-year useful life. A gradualearly replacement must be planned of those sections of pipe that are worn by use or bypitting due to the problems caused by iron-eating bacteria (innocuous to human health)that badly affected the line during the first few years of use. This problem has beendealt with so far by plugging the pipe or removing small stretches where solderingwould not be possible because of the extent of the damage.

The funds have been found to acquire a stock of pipe tubing. It is necessary now to hirecontract personnel to install it.

4-m Telescope Repainting:The dome and building must be repainted every five years. Delaying this kind ofmaintenance activity leads to higher costs later.

Vehicle Fleet Renovation:

CTIO maintains a vehicle fleet for transportation of both personnel and goods,especially between La Serena and Cerro Tololo. Heavy equipment is needed tomaintain the road and to support access to Cerro Pachon. The vehicle fleet has beenreplaced only sporadically through special allocations from the NOAO Director,especially including the bus to transport workers to the mountain; the steady statebudget for CTIO is inadequate to support regular replacement. Failure to replacevehicles in a timely manner is a financially inefficient policy that adds to maintenancecosts in subsequent years.

Cerro Tololo Power House:

During the past four years, CTIO has been carrying out a plan to upgrade the TololoPower House. The next phase of work involves replacement of obsolete switching andcontrol gear, which has been used for well over 25 years.

Bus Replacement:The CTIO bus which transports day workers between La Serena and Cerro Tololoevery workday will be ten years old in 2002. Careful maintenance should allow the busto keep operating until then, but it should be replaced at that point. We will, of course,examine leasing or rental options, but currently in Chile it is more economical topurchase a new bus.

b. Construction

Cerro Tololo Road Improvement:The greatest safety problem at CTIO continues to be accidents on the access road toTololo. Last year there were several near misses; in one earlier case a car driven by atourist with five family members rolled over on a narrow stretch of the road andstopped just short of falling down into a deep gulch. Miraculously, no one was badlyhurt.

Ln recent years CTIO has installed 483 linear meters of guard rails in sectors of thegreatest potential danger. In order to continue offering greater protection to peopleusing the road, an estimate of a further 10,200 m of guard rail are needed, at a cost ofUS 27.50/linear meter ( 44,155 per mile, total 300 K).

2. Kitt Peak

Milestones:

Projects (FY 1998)

1. Telecommunications systems upgrades2. Fuel tank monitoring3. Energy management4. Propane system5. Bus replacement6. General building repairs

Total

Projects (FY 1999)

1. Water system2. Energy management3. Propane system4. Replacement of water, sewer, and telephone lines5. 4-m aluminizing tank enclosure6. Road repairs7. General building repairs

Total

Projects (FY 2000)

1. 4-m dome and building repairs2. Replacement of water, sewer, and telephone lines3. Energy management4. General building repairs

Total

Projects (FY 2001)

1. 4-m dome and building repairs2. Replacement of water, sewer, and telephone lines3. ADA compliance for dormitory4. Energy management5. General building repairs

Total

Projects (FY 2002)

1. 4-m dome and building repairs2. Energy management3. Improvements to administration building4. General building repairs5. Septic system

Total $ 138K

89

Estimated Cost

$ 20K

7K

10K

5K

120K

18K

$ 180K

Estimated Cost

$ 30K

10K

5K

20K

40K

60K

18K

$ 183K

Estimated Cost

$ 70K

60K

10K

18K

$ 158K

Estimated Cost

$ 50K

40K

30K

10K

18K

$ 148K

Estimated Cost

$ 20K

10K

60K

18K

30K

KPNO mountain-based facilities will continue to be downsized over the next few years.However, the nighttime facilities, with only four telescopes being currently operated full-time by NOAO, are near the projected minimum complement of three telescopes. Furthermajor reductions can occur only if the NSO facilities are closed, as they will be if theflagship facility described elsewhere in this plan is funded. However, the savings will berequired to operate the new solar site.

For the next two years, the highest priority for both maintenance and improvements on KittPeak will be the WIYN telescope. Over this two-year period, $400K (40 percent fromNOAO) has been earmarked by the consortium members to bring the WIYN telescope up toits optimum level of performance and reliability. As a result, some maintenance work onfacilities and other telescopes will be deferred to future years.

a. Maintenance

The list of projects which we feel are essential to the successful operation of KPNOover the next five years is:

Water System:Water for use on Kitt Peak is obtained primarily from runoff collected in catchmentbasins. These basins have leaks and must be repaired. Adequate supplies of water mustbe maintained not only for human consumption, but for fire fighting as well. In dryyears, runoff is inadequate to meet these requirements and water must be hauled to thesummit for processing. The problem is compounded by leakage. As a conservationmeasure, existing high water use fixtures are being replaced with no, or low, water usefixtures to minimize usage.

4-m Dome and Building Repairs:The 4-m dome and building were last painted in 1992. The exterior of the facilityshould be repainted no later than the year 2000. In addition, the dome and shutter sealsare worn and should be replaced. The 4-m enclosure consumes roughly one-quarter ofall power used on Kitt Peak. We are currently studying why this energy use is so highand reviewing the best method of including the 4-m telescope in our mountain-wideenergy management program.

Replacement of Water, Sewer, and Telephone Lines:Most of the underground utilities were installed in the early 1960's, and are well pasttheir expected lifetimes. The main sewer lines were replaced in FY 1995. At that time,it was decided to defer replacement of the small feeder lines. We have experiencedleakage with several of the main water lines this past year. When warmer weatherpermits we will closely examine the condition of these lines and replace them asneeded. Usable telephonecable pairs remain in short supply, especially on the east sideof the summit.

Road Repairs:The roadways at the summit of Kitt Peak, as opposed to the access road, areobservatory responsibility. These roads are subject to weather damage and mustperiodically be resealed. We must chip-seal these roads no later than 1998.

90

Telecommunications System Upgrades:Although AT&T System 75 telephone switches for Kitt Peak and NOAO Tucson werepurchased in 1984, we have been successful in keeping them state-of-the-art technologythrough occasional upgrades. To extend our switchboard coverage, we will purchasenew software that ties the two switches together and allow them to operate as one. Thisimprovement will allow us to extend our Tucson headquarters voice mail system to KittPeak.

We will continue to work with our telephone service provider, the Tohono O'OdhamUtility Authority (TOUA) in their plan to implement greatly increased bandwidthbetween Kitt Peak and NOAO Tucson, through a fiber-optic network.

Bus Replacement:A bus provides the most cost-effective means of transporting staff to Kitt Peak eachworkday. The bus was last replaced in 1987 and has accumulated roughly 290,000miles to date. Recent repairs have extended the useful life of the bus. However, itshould be replaced no later than 1998. The option of lease rather than purchase is beingconsidered.

Septic System:A study of the mountain-wide septic system in 1993 identified numerous problemareas. More frequent removal of sludge and a second grease trap for the kitchen appearto have improved the condition of our main leaching field. Replacement tanks may berequired in several areas within this five-year period.

Propane System:Many of the underground lines connecting our propane tanks with structures on KittPeak were installed 25-30 years ago. In order to assure safe operation, we will pressuretest and visually inspect the propane distribution system. Pipes, tanks, and fixtures willbe replaced as required.

Energy Management:With the demonstrated success of our energy management program for the NOAOTucson headquarters building, we now wish to proceed with the same strategy on KittPeak. Over the next five years we will begin to make energy improvements, startingwith the shortest time and largest payback items first.

General Building Repairs:Including the telescope enclosures, KPNO has in excess of 130K s.f. of buildings. Wehave developed a plan to repair, paint, and roof all facilities on a five-to-seven-yearrotational basis. This program can be carried out within the current budget.

b. Construction

4-m Aluminizing Tank Enclosure:The 4-m aluminizing facility is the only large coating chamber that is made routinelyavailable to astronomy-related organizations. More than half the mirrors coated eachyear belong to organizations other than NOAO. The tank is built into the ground floorof the 4-m building within a large open space. It is very difficult to keep this area cleanenough to prepare mirrors for new coatings. Over the next two years, we plan to

91

enclose the tank and mirror preparation area so that the environment approximates thatof a semi-clean room.

ADA Compliance for Dormitory:Presently, there are no handicapped-accessible dormitory facilities to accommodatedisabled workers or observers. We plan to modify a portion of one of our dormitories tobring the Observatory into compliance with the Americans with Disabilities Act.

Improvements to Administration Building:The functions of some of our buildings have changed over the years. FacilitiesOperations is reviewing the use and condition of the various support buildings withcentralization and cost savings in mind. One building that will play a prominent role infuture operations is the Administration Building. Proposed modifications may include aremote observing area, which could eventually serve as the main control room for allnighttime telescopes; office space for WIYN staff members and observers; the additionof telecommunications/video capability to assist with remote engineering evaluationand repair; and improved office space for observing technicians and programming staff.

Fuel Tank Monitoring:To meet EPA standards, by the end of 1998 we will need to install automatic fuel tankmonitoring equipment on all underground fuel tanks on Kitt Peak.

3. NOAO Tucson

Milestones:

Projects (FY 1998)

1. Tucson/Kitt Peak telecommunications

2. Energy management control system3. Fuel tank monitoring4. Replace shuttle vehicles5. Alternate fuel storage facilities

Total

Projects (FY 1999)

1. Correction of headquarters building (HVAC)2. Fire/security alarm system3. Asphalt parking lot replacement4. Energy management control system5. Replace shuttle vehicles

Total

Projects (FY 2000)

1. Power distribution

2. Correction of headquarters building (HVAC)3. Energy management control system4. Replace shuttle vehicles

Total $ 80K

92

Estimated Cost

$ 30K

10K

7K

30K

10K

$ 87K

Estimated Cost

$ 10K

35K

35K

10K

20K

$ 110K

Estimated Cost

$ 40K

10K

10K

20K

Projects (FY 2001)

1. Water pipe replacement2. Correction of headquarters building (HVAC)3. Energy management control system4. Replace shuttle vehicles

Total

Projects (FY 2002)

1. Repaint exterior of headquarters building2. Energy management control system3. Replace shuttle vehicles4. ADA building upgrades

Total $115K

NOAO remains an obstacle to future University of Arizona campus development. However,during this five-year period neither organization is likely to be able to identify sufficientfunds to facilitate the relocation of the NOAO headquarters to another nearby location.Therefore, NOAO plans to continue to make investments designed to lower the cost ofoperation of the existing building. At the same time, we recognize that portions of theheadquarters building are now more than 35 years old. Accordingly we must continue amethodical replacement of likely to fail building components.

a. Maintenance

Energy Management Control System:The Central Facilities Operations staff set a goal of lowering the Tucson facilitieselectric utility costs to $200K/year by December 1996. We missed that goal by $19K,but still intend to meet it sometime over the next two years. Since these savings are notone-time events, but instead are realized annually, our energy management programremains at the top of our list of maintenance improvements. Additional savingsmeasures remain to be implemented, and our long-term goal is to complete all of themwithin this five-year period.

Power Distribution:

Electrical protection circuits now in use cause all power to be shut down whenever weexperience even brief interruptions. The power cannot be restored without manualintervention. Also, in portions of the building, the routing of power lines needs to beimproved. We have correlated power line electromagnetic interference with computermonitor instability. The availability of our energy management controller makes itpossible to deal with the protection circuit problem. The extent of effort required forthe power lines remains to be determined.

Tucson/Kitt Peak Telecommunications:

Our plan to upgrade voice/data communications capability between Tucson and KittPeak was drastically altered recently when the Tohono O'Odham Utility Authority(TOUA) finally, after years of lobbying on our part, reduced our single T-l telephoneservice cost from $3749/month to $l,575/month, or roughly a 58% decrease. Inaddition, the TOUA offered to rent us two more T-l lines for $500/month each. Weimplemented the rental of a second T-l line, and are studying the need for a third. This

93

Estimated Cost

$ 20K

10K

10K

40K

$ 80K

Estimated Cost

$ 35K

10K

20K

50K

reduction in cost was very timely as we were beginning to experience a slowdown inour data flow because of insufficient bandwidth.

As a result of the price reduction by the TOUA, along with the progress they aremaking on a plan to provide fiber-optic capability to Kitt Peak, we have placed our planto purchase and install a microwave link between Kitt Peak and Tucson on hold. At theold T-l line rental rate, the microwave equipment had a payback time of four to fiveyears. At the new rates, we can afford to wait to see whether the fiber link is installedand whether we can use it at a reasonable price.

Asphalt Parking Lot Replacement:The asphalt surface of our main parking lot will need to be replaced over the next five-year period. The surface is no longer responding to the sealing procedure used toprevent deterioration. The old material will need to be dug out and replaced with newasphalt. The other parking areas and service yards will continue to be sealed asrequired.

Water Pipe Replacement:With portions of the headquarters building more than 35 years old, we have beenconcerned about the condition of the water pipes. During a recent investigation of acooling problem, a broken hot water pipe (used to heat the building) was discovered.Due to the construction of the building, the repairs are likely to be both extensive andexpensive. We expect that during this five-year period, replacement of portions of thepotable water pipes will be required as well.

Repaint Exterior of Headquarters Building:Within the next five years, the stucco exterior surface of the headquarters building willneed to be crack sealed and painted.

Complete Correction of Headquarters Building (HVAC) Problems:The headquarters building main boiler was replaced in 1997, but additional work willbe required on the HVAC system (replacements and upgrades to the air distribution andpiping equipment).

Replace Shuttle Vehicles:A shuttle fleet is utilized for transportation of staff and visitors to/from Kitt Peak andfor travel to out-of-Tucson destinations. To maintain safe highway vehicles, we have aprogram in place for annual replacement of high mileage vehicles. We replace, on theaverage, three vehicles per year.

b. Construction

Alternate Fuel Storage Facilities:The new EPA alternative fuel vehicle AFV regulations may prove extremely costly toNOAO. The downtown fleet is small enough so that we are not affected by the newregulations. However, the initial ruling by the NSF is that the fleet size is determinedby the combination of the number of vehicles located both in Tucson and assignedpermanently to the mountain, where obviously the mileage is very low and air pollutionis not a problem. Unless we receive a waiver, in 1998 50% of our new vehiclespurchased must meet AFV standards. In 1999, the requirement increases to 75% and

94

remains at that level. The AFV options for transportation between the mountain anddowntown are propane- and methanol-fueled vehicles. These vehicles cost more topurchase, more to operate, and will require the installation of additional fuel storagefacilities in Tucson and possibly on Kitt Peak as well.

Fuel Tank Monitoring:To meet EPA standards, by the end of 1998 we will need to install automatic fuel tankmonitoring equipment on our Central Facilities underground fuel tank.

Fire/Security Alarm System:In FY 1996, we replaced the antiquated NOAO headquarters computer room fire alarmsystem. The main building fire alarm system is also obsolete and in need ofreplacement. The recent break-in and computer theft at the Central AdministrationServices building further demonstrate how vulnerable we are even with guard serviceprotection. To improve security in the buildings outside the main headquartersbuilding, we will install both fire and burglar alarms.

ADA Building Upgrades:The Tucson headquarters building is not in full compliance with Americans withDisabilities Act (ADA) requirements. NOAO has completed initial ADA modificationsto the restrooms as well as to a side entrance. Access to all levels in the building,however, has not yet been provided.

4. NSO

NSO's telescope buildings at both observing sites, and the entire infrastructure atSacramento Peak, date mainly from the 1950's and 1960's. These structures do not meetcontemporary standards for energy efficiency, ease of use, or maintainability. Past andcontinuing budget constraints have precluded an adequate preventative maintenanceprogram: for at least 10 years, "maintenance" has mainly consisted of repairing or replacingequipment that fails.

The following lists itemize what is required to upgrade NSO facilities to the point thatregular maintenance and upgrades over the past 10-15 years would have provided. Such aninvestment would allow the maintenance program at both sites to become proactive andensure the future viability of the infrastructure.

NSO will maintain its facilities with an eye toward the future NSO sketched in the SolarStrategic Plan (section V.A). As that vision is realized, some of the maintenance categorieslisted below will become unnecessary. However, until key components of the plan becomereal, NSO must and will protect the public's investment in the facilities under ourcustodianship.

95

a. NSO/Sacramento Peak

Milestones:

Projects (FY 1998) Estimated Cost

1. Painting - telescope buildings $ 25,0002. Power line replacement 20,000

3. Cloudcroft facility/RCA building reroof 75.000

Total $120,000

Projects (FY 1999) Estimated Cost

1. Painting - commercial buildings $ 25,0002. Staff vehicle replacement 20,000

3. Lighting upgrades 32,000

Total $ 77,000

Projects (FY 2000) Estimated Cost

1. Painting - commercial buildings $ 25,0002. Lighting upgrades 32,000

3. Storage facility 40,000

Total $ 97,000

Projects (FY 2001) Estimated Cost

1. ADA compliance $ 50,0002. Underground utility replacement 35,000

3. Housing upgrades 25.000

Total $110,000

Projects FY 2002 Estimated Cost

1. Road repair/resurfacing $175,0002. Housing upgrades 25,000

Total $200,000

Road Repair/ResurfaceRoads and driveways throughout the housing area are in serious need of repair. Harshweather conditions and minimal preventive maintenance in past years have led toserious deterioration. A complete removal of existing paving and replacement will berequired.

Commercial and Residential PaintingAll the redwood houses, the Vacuum Tower Telescope, Evans Solar Facility dome andseveral maintenance buildings on the site require painting. The redwood housing waslast painted in 1978, many of the maintenance buildings and telescopes prior to that.The paint is peeling on almost all redwood houses with deterioration beginning aroundfoundations. Continued neglect will result in higher repair costs when undertaken.

Cloudcroft Facility Building Reroof (RCA)NASA recently spent over $175K renovating the main telescope building at theCloudcroft Telescope Facility. Other buildings, in particular the RCA building, requireimmediate roof repair to prevent extensive water damage during the upcoming rainy

96

season. Age and deterioration of this building will soon make it unusable. Access tooccupants and visitors has been restricted.

Lighting UpgradesLighting upgrades in all commercial buildings to modern fluorescent fixtures withcurrent switching would result in savings in electricity usage and improve the quality oflighting in all areas.

ADA ComplianceCurrently, none of our buildings are in ADA compliance. To assist a disabled summerstudent who was on site a couple of years ago, we renovated a restroom in the MainLab. We need to modify access to several buildings, and renovate more restrooms andan apartment in the VOQ to allow handicap access.

StorageCurrent storage requirements are being met with surplus quonset huts constructed inthe 1950s. Due to their age and condition, these facilities are not weatherproof, norheated.

Underground Storage Tank ComplianceTwo petroleum underground storage tanks will require upgrade to maintain EPAcompliance by the end of 1998. Upgrades will include automatic leak detection,spill/overfill protection and cathodic protection. An option that is being explored isreplacement instead of upgrade.

Housing UpgradesThe housing at the Observatory was constructed in the 1950's and 1960's. Very littlehas been done to improve or upgrade them since installation. Many have the originalplumbing, electrical fixtures, piping, wiring, floor tile, and cabinets.

Vehicle/Equipment ReplacementThe current fleet of vehicles and equipment is very old and consists of unreliableequipment and vehicles. Many pieces of equipment are 1960's vintage with highmileage or hours. Current staff vehicles are over 10 years old with some over 100Kmiles each. The advanced age of the fleet increases maintenance costs dramatically.

Power Line ReplacementPower lines in the housing area were installed in the 1950's and are beginning to showserious deterioration. Many of the installations are sub-standard.

97

b. NSO Kitt Peak

Milestones:

Projects (FY 1998) Estimated Cost

1. McMath-Pierce electrical upgrade $30,0002. McMath-Pierce painting 15,0003. Reroofing and sealing 5,0004. McMath-Pierce energy management 5,0005. McMath-Pierce fall protection 5.000

Total $60,000

Projects (FY 1999) Estimated Cost

1. McMath-Pierce electrical upgrade $30,0002. Reroofing and sealing 5,0003. McMath-Pierce energy management 5,0004. McMath-Pierce fall protection 5,000

5. KPVT vacuum line cold-trap 2.500

Total $47,500

Projects (FY 2000) Estimated Cost

1. McMath-Pierce electrical upgrade $30,0002. Reroofing and sealing 5,0003. McMath-Pierce energy management 5,000

4. Extend KPVT loading dock 2,500

Total $42,500

Projects (FY 2001) Estimated Cost

1. McMath-Pierce electrical upgrade $30,0002. Reroofing and sealing 5,0003. McMath-Pierce energy management 5.000

Total $40,000

Projects (FY 2002) Estimated Cost

1. McMath-Pierce electrical upgrade $30,0002. McMath-Pierce painting 15,0003. Reroofing and sealing 5,0004. McMath-Pierce energy management 5,000

Total $55,000

In addition to the structural maintenance items summarized above and described below, theaging Telescope Control Systems (TCS) at the McMath-Pierce and Vacuum Telescopesconstitute a serious long-term maintenance issue. The current 20-plus-year-old controlsystems are increasingly difficult to maintain, resulting in increased downtime. Thereplacement of these systems with modern, lower-maintenance hardware and software is ahigh priority.

As described in the instrumentation section of this document, the design phase for theKPVT TCS upgrade was finalized in FY 1995*, and implementation of phase 1 (upgrade ofthe #2 mirror drive assemblies) was completed in early FY 1996. Phases 2 and 3 (image

98

jitter compensation and guider upgrade) will be completed in fiscal years 1997 and 1998.Upgrades to the McMath-Pierce telescope control computer hardware and software willfollow.

The cost of these TCS upgrades, including labor, will exceed $500K but is not included inthe summary above.

McMath-Pierce Electrical Upgrade:Sections of the wiring and electrical components within the main observing room andthe telescope structure as a whole require replacement and upgrade to ensure that thefacility remain operational and maintainable.

McMath-Pierce Interior Painting:The exterior of the McMath-Pierce was painted in 1991 and has held up reasonablywell-some touch up is scheduled for this summer. Maintenance of the interior surfacesand caulking, however, is long overdue. The section of the interior windscreen justbelow the top of the pier requires extensive work.

Reroofing and Sealing:Due to deferred maintenance, there now exist several rainwater leaks in the McMath-Pierce complex. The two leaks in the main observing room ceiling and FTS stairwellhave become serious and require attention.

McMath-Pierce Energy Management:Improve energy management within the McMath-Pierce complex. Currently the lower-tunnel cooling and fans are run open loop. When in operation, the cooling system andfans run continuously even if weather conditions change. Upgrading the system to"closed-loop" operation by the addition of a smart controller and sensors will result insignificant savings. Also, additional modest savings may be gained by changes in thegeneral building lighting.

McMath-Pierce Fall Protection:

Access to the east and west auxiliary heliostat drives is difficult and can, under certaincircumstances, present a safety hazard. A redesign of the accessway and rails shouldeliminate the hazard. Access has been restricted.

KPVT Vacuum Line Cold Trap:Oil backstreaming from the vacuum pump into the Kitt Peak Vacuum Telescope'svacuum chamber has been a continuing problem. Adding a large cold trap to thevacuum line should eliminate the problem.

Extend KPVT Loading Dock:The Kitt Peak vacuum telescope loading dock does not extend sufficiently far out fromunder the roof overhang. This has resulted in damage to the roof by cranes unloadingequipment onto the dock. Extending the dock will eliminate the problem.

99

VII. PUBLIC INFORMATION AND EDUCATION

A. General Trends

NOAO must reach several audiences with its programs in outreach and public education. Firstis the professional community, and given the major changes now in progress in both the solarand nighttime programs, it is essential that we work closely with the community so that theyhave opportunity for input, can gain a good understanding of why changes are being made, andalso have the information necessary to design experiments that take maximum advantage ofnew observing capabilities. Improving communications with the users is one of the primarygoals of the reorganization of the nighttime program, which placed user support of all of thefacilities accessed through NOAO under a single Director.

In addition to communications with professional colleagues, NOAO devotes approximately1.5% of its budget to public outreach and educational programs. We have identified three typesof activities that reach three different types of audiences. The first audience is the educationcommunity. The second is the general public. And the third is the media, including the pressand television. Initially, we have chosen to emphasize the program in education and outreach tothe general public, especially in the communities where we have facilities. We are trying toidentify resources that would allow us to be more effective in terms of interactions with themedia involving press releases and in using the Web to make images and other informationavailable to the general public in an attractive format.

The recent initiatives in Educational Outreach and Kitt Peak Visitor Center operations havematured, allowing both the diversification and strengthening of outreach programs at NOAOdespite the static fiscal limit. We have become more informative, responsive, and responsible tothe general public and educational community through increased efficiency, inter-observatorycooperation, and the use of alternate sources of funding.

Alternate revenue sources obtained in the last two years include:

• NASA IDEA Grants (3)

• the Arizona/NASA Space Grant Consortium• Education Supplements to NASA science proposals (2)• NSF Teacher Enhancement Program within the Directorate for Education and Human

Resources ($500K)• Astronomical Society of the Pacific as NOAO became a Project ASTRO expansion site

($87K)• increased monies from the public, in response to increased accessibility to NOAO facilities

through the Kitt Peak Visitor Center.

An example of inter-observatory cooperation is the Southwestern Consortium of Observatoriesfor Public Education (SCOPE), a consortium of several observatories located in the southwestwhich was organized in the last year. The consortium includes the National SolarObservatory/Sacramento Peak, Apache Point Observatory, McDonald Observatory, Very LargeArray/National Radio Astronomy Observatory, Whipple Observatory, Lowell Observatory, andKitt Peak National Observatory. SCOPE is working on the development of promotionalmaterials and plans to coordinate visitor center displays. SCOPE recently served as an efficientand effective distribution channel for educational materials developed at NOAO.

100

The overall NOAO reorganization will integrate our various outreach programs and resourcesto present a unified message to our many audiences (professional, educational, general public,and internal staff). The WWW pages of NOAO (http://www.noao.edu) describe our programsand facilities in an increasingly accessible manner to all audiences. The NOAO ImageCollection is a resource common to all outreach programs and efforts are underway to increaseits accessibility with less staff intervention. There will be a Web-based pictorial indexcontaining thumbnail sketches and brief captions of every available image, as well asdownloadable digital versions at different resolutions. Once this database is complete, theNOAO Image Collection will be more widely available, more accurately captioned, and morecurrent. This increased efficiency will make the collection more readily available toprofessional scientists and textbook authors. In addition, the collection can be more easilymarketed by the planned Image Publishing division of the Kitt Peak Visitor Center.

NOAO will continue to take a conservative approach to issuing scientific press releases,restricting them to subject matter that is either scientifically significant or of broad generalinterest. Major newspapers and astronomical publications such as Sky and Telescope aretargeted. However, the frequency of more general press interactions has been significantlyincreased. NOAO is actively promoting public outreach efforts at the Visitor Center and seeksto develop a strong working relationship with the local and national media.

B. Visitor Center Programs

The Kitt Peak Visitor Center, attracting upwards of 50,000 visitors per year, is the revenue-based arm of the Public Information Office. An aggressive business plan has been incorporatedto introduce new public services and create additional revenue centers to support NOAO publicoutreach activities.

Retail space renovations, combined with improved product offerings and profit margins, willallow for increased sales revenue over previous fiscal totals. A new Image Publishing divisionwill reproduce and market NOAO images in the form of postcards, notecards, posters, and slidepackages. Additionally, profits will be extended by creating a wholesale division and salescatalog to market NOAO products to museums, bookstores, planetariums, and throughprofessional societies such as the Astronomical Society of the Pacific.

A fee-based Nightly Observing Program for the public introduces participants to stargazing andbasic telescope usage. A limited group of 20 individuals per night are given a tour of the nightsky beginning with a constellation orientation, followed by an introduction to stellarclassifications and evolution using the 16-inch Visitor Center telescope, and concluding withplanet, nebula, and galaxy viewing using binoculars and portable telescopes. A boxed dinner isprovided by the Kitt Peak Dining Hall staff, generating a revenue center for the Kitt Peakmountain operations division. A fee-based Advanced Observing Program will give amateurastronomers the experience of observing on-site at the world's largest optical astronomyobservatory using state-of-the-art equipment and including access to, and instruction in, CCDimaging equipment.

A fee-based automated tour program in the form of recorded information played over individualheadphones will relieve staffing pressures and overcome language barriers while creating anadditional revenue center. The scripted presentation will insure the correct dissemination ofinformation and provide an additional service for the public at Kitt Peak.

101

A Kitt Peak Docent Program has been developed to train volunteers as tour guides or to givespecialized science demonstrations and lectures at the Kitt Peak Visitor Center. Docents willattend an introductory six-week training program on NOAO history and policy and receivecontinuing education throughout the year. In addition, new exhibits and visitor gallery displaysare being installed to enhance the Kitt Peak experience for the public.

Corporate fundraising campaigns will be developed, along with a "Friends of Kitt Peak"program, to establish ongoing sources of revenue for the expansion and improvement of theKitt Peak Visitor Center and public outreach programs through the NOAO Public InformationOffice.

The Sunspot Astronomy and Visitor Center will open this summer and be used as a springboardfor the education and outreach program at Sac Peak. The center will include educationaldisplays and a large meeting room for workshops and colloquia.

C. Educational Outreach Programs

The NOAO K-12 Educational Outreach Program has gone from zero to substantial in the pasttwo years. The program has reached a stable configuration balancing three approaches: directclassroom involvement, instructional materials development, and teacher training in the "Use ofAstronomy in Research Based Science Education." Staffing includes 1 FTE on the NOAOpayroll and one-half to one additional employee funded through alternate sources. The NOAOOutreach Advisory Board, a group of thirteen local middle and high school science teachersknowledgeable about NOAO and the science done here, contribute considerable advice andoccasional labor on a volunteer basis. The overall role of the Educational Outreach Office is to

make the science and scientists of NOAO accessible to the K-12 educational community in auseable way. We see ourselves as a resource to educators as they strive to implement the NRCNational Science Education Standards. We are also a resource to the NOAO staff as theybecome more responsive and responsible to the general public and educational communities.

The three main program areas: direct classroom involvement, instructional materialsdevelopment, and teacher enhancement are further described in the following paragraphs.

1. Project ASTRO, developed by the Astronomical Society of the Pacific, expanded to Tucsonin the fall of 1996 with NOAO as the lead institution. An active and cooperative coalitionof astronomy resources contributes to ASTRO's success in Tucson. The local coalitionincludes representatives from the University of Arizona, Tucson Amateur AstronomyAssociation, Tucson Unified School District, Pima Community College, and others. ProjectASTRO forms ongoing partnerships between astronomers and fourth to ninth gradeteachers in Tucson area schools and community organizations. We expect to form 25partnerships annually through Project ASTRO, with teachers and astronomers receivingtraining, materials, and support throughout the year. ASTRO funds (from the NSF InformalScience Education Program, to the ASP, to NOAO) provide three years of support for ahalf-time Project Coordinator, after which time NOAO is expected to continue the programwithout this source of funding.

2. Instructional Materials in various formats have been developed based on topics of currentastronomical research taking place within the Observatories. Most materials are availableon the NOAO EO Home Page on the WWW as well as in printed formats. We expect to

102

I »

continue with materials development as resources allow. These materials have been wellreceived by teachers, are relatively easy to obtain outside funding for, integrate thedevelopment, production, and distribution of resources available within NOAO outreach,and are fun and satisfying to produce.

Some examples:

• "Visitors from Space - Comets!" is an eight-page color brochure on the topic of comets,Hale-Bopp in particular, developed with the NOAO Outreach Advisory Board to bringHot Topics in Astronomy into the Classroom. Funded by a 1996 NASA EDEA Grant,our goal was to bring timely and accurate science information into the classroom tosupplement newspaper and television headlines. The brochures contained colorpictures, activities, interviews, articles, and pointers to additional sources ofinformation including the internet. Three hundred classroom sets were distributedprimarily in the Southwest, in part through SCOPE.

• "Build Your Own Comet" and "Solar Music - Helioseismology" are two brochuresdeveloped as part of a 1995 NASA EDEA Grant, titled "Active Learning Exercises inPlanetary and Solar Astronomy for K-3 Students."

• "Frequently Asked Questions About Being an Astronomer" contains answers (fromNOAO staff) to questions from students around the country. This brochure provides anefficient way to respond to common questions and will be used as a template for asecond brochure: "Frequently Asked Questions About Astronomy" based on questionsasked ASTRO astronomers as they venture into the classroom.

3. "Use of Astronomy in Research Based Science Education" is a four-year program offeringresearch experience to middle and high school teachers during summer workshops atNOAO beginning in 1997. Participants will experience how science is practiced anddevelop strategies for paralleling those techniques in their classrooms. Teachers areexpected to increase their knowledge of astronomy and the scientific process, their use ofcomputer-based technologies including image processing and use of the internet, theirclassroom use of inquiry-based teaching and learning, and their awareness of careeropportunities in science and technology. The research experience extends to the classroomduring the academic year with datasets in preselected areas of astronomical researchcontinuously being provided over the internet. Local mentors, drawn from the national userbase of NOAO facilities on Kitt Peak, will provide support to participating teachers in theirareas after the summer workshop. NOAO Education Officer Suzanne Jacoby, Sahuaro HighSchool teacher Jeff Lockwood, and UA Astronomer Don McCarthy are co-investigators onthis Teacher Enhancement project, funded through the National Science Foundation'sDirectorate for Education and Human Resources.

D. Undergraduate and Graduate Education

NOAO will continue its role of host to undergraduate summer students, funded by the NSFResearch Experience for Undergraduates (REU) program. The annual budget totals $103K andfunds approximately six summer positions at both KPNO and NSO, and four at CTIO eachyear. Students work closely with a mentor on the NOAO scientific staff throughout thesummer.

103

The NSF has also provided $10K per year for undergraduates, not limited to science majors, toaccompany faculty advisors during observing runs on Kitt Peak. This program has beenenthusiastically endorsed by participants and is expected to continue.

Through its research facilities, NOAO plays a major role in graduate education. About 120graduate students visit KPNO and CTIO each year. In addition, KPNO has recently been ableto offer a graduate intern program where students spend approximately one month in residencesupporting visiting astronomers in return for room and board. Three student interns took part inthis program and found it a useful way to learn about the use and calibration of astronomicalinstrumentation. We will continue to make this opportunity available.

VIII. BUDGET

The budget summary for this long range plan is given in Appendix C. The budget is presented inthree parts: the base budget, which funds the ongoing telescope operations, instrumentationprogram, and scientific staff; new projects for which funding is already available or appears likelyto become available based on proposals already submitted; and program elements that require newfunding. The assumptions used to construct the budget are described below.

A. Base Budget

The staffing levels and investment in instrumentation required to support the core programwere justified in the proposal submitted to the NSF to renew the cooperative agreement underwhich AURA manages NOAO and will be reviewed in the program plan for FY 1998. Thesedocuments make the argument through benchmarking that the current level of staffing isrequired to operate the program as defined and the long range plan budget assumes that thislevel will be maintained throughout the time period covered by this long range plan.

In order to maintain the quality and productivity of the program, AURA must continue to paysalaries that are sufficiently competitive to enable us to recruit and retain people with the skillsrequired for the ambitious and technically challenging program that we have laid out. This willrequire at a minimum that non-technical salaries keep pace with inflation. During the past year,a severe shortage of engineers and programmers has developed, salaries nationwide fortechnical personnel are ramping up sharply, and NOAO's own salaries are no longercompetitive as demonstrated by the fact that we have been unsuccessful in recruiting new staffat salaries that are consistent with our current scales. It is also the case that our scientific staff

salaries remain about 15 percent below those of AURA member institutions. We believe thatsalary freezes are not an option even in a low inflation environment.

Therefore, in modeling the budgets for this long range plan, we have matched the budget levelin FY 1998 to the request submitted by the President to Congress, and for FY 1999-2002 wehave inflated the US-based funding at the rate of three percent per year for both payroll andnon-payroll. In addition, we have incremented technical salaries by $250K above this rate ofinflation in FY 1999 relative to FY 1998. As noted below, the only salary increases in FY 1998will be targeted toward engineering and programming staff. Engineers recently hired away fromus have received increases in salary of 40 percent or more, and so a major adjustment incompensation is mandatory.

104

Inflation in Chile continues to outpace inflation in the US, and salaries of Chilean employeesare, according to union agreement, indexed against Chilean inflation and adjusted quarterly.We have therefore assumed that Chilean peso salaries will increase at 7 percent per year asdenominated in dollars, which is consistent with recent experience. We have assumed that non-payroll expenditures in Chile will increase at the US rate of 3 percent per year, which may beslightly optimistic.

The budgets for the long range plan show a shift of resources away from KPNO and into theUSGP, which will be responsible for implementing the "one-stop shopping" approach tohandling applications for observing time and supporting the community in their use of Geminiand the telescopes at the independent observatories, as well as the facilities at KPNO andCTIO. We have also assigned to the USGP those internal functions that serve the scientific staffof more than one NOAO division. Accordingly, in FY 1997, four scientific staff (Gatley,Pilachowski, Jannuzi, and Lauer) were transferred from KPNO to the USGP; this transfer waseffective for the last three quarters of the fiscal year. Central Computer Services, whichsupports the scientific computing system and network access, and the ERAF group have alsobeen transferred to the USGP. In FY 1998, Dave DeYoung will be transferred to the USGP,along with the library and the photo lab, which he oversees. All of the scientific positions in theUSGP will be funded for the full year in FY 1998. These changes account for the substantialincrease in funding for the USGP from FY 1997 to FY 1998.

All major instrumentation for CTIO and KPNO is now being built in Tucson, as is that portionof the instrumentation for Gemini that is being developed at NOAO. Four scientific staff aredevoting nearly all of their service to the instrumentation program and in FY 1998 will bebudgeted within that program. The staff are Taft Armandroff, who is currently acting head ofthe entire instrumentation program and is also in charge of the optical program; Sam Barden,who is project scientist for Hydra/CTIO and for the Gemini CCD controllers; Ken Hinkle, whois project scientist for Phoenix; and Mike Merrill, who is project scientist for the ER arraycontrollers for both NOAO and Gemini and also for the Aladdin array program. The transfer ofthese individuals accounts for the increase in the instrumentation budget between FY 1997 andFY 1998. Jay Elias, who is project scientist for the Gemini near infrared spectrometer, wasbudgeted within the instrumentation program in FY 1997 and will continue to serve this role inFY 1998.

These transfers of the scientific staff account for the bulk of the decrease in the KPNO budget.

The FY 1997 figures include the carryover for the previous year. However, we have excludedthis carryover in determining the FY 1998 budget. The apparent decrease in some of the FY1998 budgets relative to FY 1997 is due to the subtraction of the carryover. However, therequest level of $28.23M, which represents a 3 percent increase in budget overall, isinsufficient to maintain level of effort. We have inflated the management fee by three percent.The bulk of the remaining increase in budget for FY 1998 relative to FY 1997 is required tocover increased costs in Chile and the loss in indirect income that will result when the

international Gemini staff leaves Tucson. We have also allocated $250K to increasing thesalaries of the engineering staff, and have distributed these funds across all the units inproportion to the size of the technical staff. The actual salary adjustments will be based onperformance evaluations, which will be made later this fiscal year. All other programs will befrozen in FY 1998 at their current dollar levels, thereby necessitating either reductions in staff(and programs) or salary freezes; for the reasons described above, a broad-based salary freezemay be impossible. Detailed planning for FY 1998 is in progress, and the program plan will

105

' *

describe in specific terms the programmatic consequences of a budget of $28.23M. Theamounts for the management fee are presented only as a placeholder. These projected amountsare estimates of what is required to provide for Corporate Office operating expenses; the feeactually charged will be negotiated between the NSF and the Corporate Office on an annualbasis.

We assume that the core program will continue to fund the operation of GONG for a full solarcycle.

B. Funded Projects

Three projects are listed that are either already funded or that are expected to be funded in thevery near future. The first is SOAR. In an earlier year, NOAO was given $2M towardrestructuring the observatory. Approximately $400K was spent to accelerate the GONG project,thereby shortening the construction period by one year and saving approximately $1.7M.Another $400K, approximately, was spent to cover severance costs of staff who were eitherlaid off or offered early retirement. We have indicated to the NSF that we plan to provide $2Mtoward the construction of SOAR, pending the development of agreements that are approved bythe NSF for the construction and operations phases of the project. Of this total, $1.2M willcome from the restructuring money. The remainder will come from the instrumentationprogram; specifically, the near-infrared spectrometer for the southern hemisphere that isincluded in the instrumentation plan will be proposed to the SOAR partnership to make up theremainder of NOAO's contribution. The exact credit to be given to NOAO for the spectrometerwill be a matter of negotiation but will certainly be less than the total cost of construction sincethe spectrometer will be available to SOAR only half time and will be used at Gemini the restof the time.

The construction phase of the RISE project will be completed in FY 1998. The cost ofconstruction is divided between the Astronomy and Atmospheric Sciences divisions of theNSF.

We are optimistic that funds for SOLIS will be provided according to the schedule listed.

C. Proposed Initiatives

We have identified a number of new programs or significant components of the existingprogram that are not currently funded but that are critical to the overall operation of NOAOfacilities. The first is the upgrade of the GONG cameras to 1000 x 1000 CCD arrays, therebyquadrupling the angular resolution. The second is the operation of the RISE telescopes;operation will begin in FY 1998 and the plan is to continue observations through a full 11-yearsolar cycle. Third, the instrumentation budget at its current level does not allow for thepurchase of any new detectors. We expect to obtain many of our IR detectors as third andfourth choices of foundry runs paid for by other astronomy groups in return for characterizingtheir arrays. However, this process is unlikely to yield any 1000 x 1000 InSb arrays for NOAO,and some purchases are likely to be required. We will also have to purchase individual CCDsfor new optical instruments. The CCDs for the two mosaic imagers have already beenpurchased, but given the large number of CCDs that we operate, we can expect a few failuresand will have to make some replacements over the next five years. For example, two CCDs atCTIO failed this past year.

106

Facilities maintenance continues to be underfunded. A steady state expenditure of $500K peryear is required to keep all facilities at all sites in good repair. Alternatively, a one-timemaintenance initiative of $2.5M would address current problems, and another increment wouldnot be required for about five years. Most university observatories, as well as at the ERTF, fundmajor repairs and upgrades to their facilities through such episodic initiatives.

Finally, as noted above, the single biggest problem the observatory currently faces is attractingand retaining highly qualified technical staff. We are undertaking a variety of challenging tasks,including particularly the construction of Gemini instrumentation, and these projects cannot becompleted without skilled staff. As noted above, the budget requested for FY 1998 does notallow for any pay increases for US hires without a reduction in force. What we would like to dois target a pay increase toward engineers and programmers, and we estimate that $500K issufficient to raise average salaries by 20 percent. We would, of course, target the increasetoward critical personnel rather than making the increase across the board, and we propose tomake this increase over a two year period of time.

107

Appendix A

The Albuquerque Process

NOAO, in consultation with AURA, will convene an advisory committee, consisting of representativesfrom institutions interested in building the instrument, representatives of AURA, NSF, Gemini, andNOAO, and scientists with relevant expertise...

The committee will consider proposals and presentations from interested groups. The committee mayrecommend to NOAO for transmittal to NSF. The procedures that the Albuquerque workshoprecommended for Gemini instruments for which proposals are likely to be received from both NOAO andoutside groups are as follows:

• a rank-ordered list of proposals• a single proposal• a collaboration among two or more proposers• a formal competition

A more formal proposal from the selected group will be subsequently reviewed by a scientific andtechnical committee for concluding formal arrangements for producing the instrument.

The CCD controller work package was the first to which the Albuquerque Process was applied, and inworking out the procedures for doing so, the USGP was guided by the above precepts and also by thegoal of obtaining for Gemini a system that provides the best performance at the lowest cost and risk. Thesteps that were followed included:

A plan describing how the procurement would be undertaken was written and submitted to AURA andthe NSF for review.

An Announcement of Opportunity was sent to all potential CCD controller vendors that had indicatedinterest in Gemini work in the past. In addition, the availability of the AO was advertised in theCommerce Business Daily.

At the request of the international Gemini project office (IGPO), the AO was also sent to all potentialvendors identified by the national Gemini Project offices in the partner countries.

The AO specified that the potential vendors should be able to supply a complete controller system,including hardware and software. The reason for adopting this approach was the concern that if groupswere allowed to propose arbitrarily small pieces of the required complete system, the availableinformation would not allow the committee to determine what pieces would fit together and how theywould perform as a system. Therefore, potential proposers were required to develop teamingarrangements in advance, and the USGP supported several groups in searches for teaming partners byissuing announcements on relevant e-mail exploders.

The AO stated the budget and schedule, and potential vendors were asked to indicate that theyunderstood these constraints and that the work proposed would be consistent with these limits.

A total of seven groups responded to the AO with letters of intent to participate in the evaluationmeeting. Of these, four subsequently dropped out.

Sidney Wolff and Richard Reed recused themselves from involvement with the preparation of the NOAOproposal so that they could provide support for the USGP evaluation effort.

Working with the IGPO, the USGP identified a candidate evaluation committee, and this membershiplist, along with logistical information about the meeting of this committee, was forwarded to AURA andthe NSF, both of whom were invited to send representatives and comment on the process and committeemembership.

Each potential vendor was asked to submit written material that answered a set of questions designed toprovide information about their expertise and capabilities. These questions were reviewed and approvedby the evaluation committee.

Additional information about Gemini interfaces and requirements was distributed to the seven groupsthat had submitted letters of intent.

Each group was asked to submit comments on the budget and schedule, and to raise any concerns theyhad about the contract form in a covering letter that accompanied the written answers to the questions.These comments, as well as the submitted written material, were distributed to the committee, with the

exception that all budgetary information was removed. Thus, the committee had no knowledge of the costassociated with any of the proposals.

A meeting of the review committee was held in Tucson in January 1997. Three proposing groupsattended that meeting. They were NOAO; a team consisting of San Diego State University (SDSU) andthe UK Royal Observatories, and a team consisting of SDSU and the Dominion AstrophysicalObservatory.

The order of the presentation was determined by numbers drawn from a hat. Each group was allowed 30minutes to make a presentation, followed by 30 minutes of questioning by the committee. The committeeasked essentially the same questions of each group, with allowance for follow-up questions to clarifyspecific points.

Each of the proposing groups was allowed to hear all of the presentations and to submit questions onthose presentations to the chair of the committee. In keeping with the Albuquerque recommendations, themeeting was open and all groups that had expressed any interest were invited to attend. After the formalpresentations and questions, a second round of questioning was held. The committee then met in closedsession the rest of the afternoon and the following morning to consider and discuss the proposals. Theproposers were given the opportunity to request a private session with the committee to discuss issues ofa proprietary nature, but none chose to do so.

Based on the evaluation from the committee, the USGP plans to adopt the following approach forproviding CCD controllers to Gemini:

1. The required hardware will be purchased from SDSU. This choice follows directly therecommendation of the evaluation committee.

2. The Royal Observatories will be asked to provide the low-level software (where their expertise ismost relevant). NOAO will be responsible for the high-level software, which will be a modificationof the software already being written for the ER controllers. This approach is consistent with theevaluation committee's identification of the strong points of the various software efforts.

3. NOAO will assume overall management responsibility for integrating the software and thehardware. NOAO has already been assigned the responsibility for integrating the detectors with thecontrollers.

This course of action is consistent with the Albuquerque process, which specifically allows for theforming of a collaboration of two or more proposers. In the case of the CCD controllers, this approachcombines the strengths of the SDSU, Royal Observatories, and NOAO proposals as judged by theevaluation committee. This is also the approach preferred by the IGPO.

As expected at the outset, the implementation of the Albuquerque workshop process has required somecomplex details to be worked out. However, all the participants have expressed the belief that theirarguments were fairly heard and that they felt comfortable with the process. The committee indicated thatthey believed that they received sufficient information to make a clear recommendation. The USGPbelieves that this approach has worked exactly the way it was designed to in that it led to a betterunderstanding of the procurement choices available to Gemini and to a better overall system than wouldhave been possible with a standard closed competition.

111

' *

Appendix B

History of the GONG Project

1. Background

In the 1980's, a flood of exciting new results demonstrated that helioseismology couldaddress a growing number of crucial questions; e.g., stellar internal rotation and oblateness,the neutrino deficit, and the efficiency of convection. At the same time, the limitations ofobservations from a single site, including the relatively brief campaigns at the South Pole,were becoming more and more apparent, and pressure to achieve long, continuousobservations was mounting. In April 1984, NSO sponsored a workshop to explore thescientific and technical issues driving the future of helioseismology. As a result of theinterest and support manifested at the workshop, NSO staff, in collaboration with thescientific community, established GONG to coordinate efforts towards the design,construction, and utilization of a network of stations, distributed in longitude and dedicatedto obtaining a uniform set of data. A proposal was generated and submitted to NSF in late1984 by NSO, on behalf of the entire scientific community.

In FY 1985, an examination of worldwide climatic data was undertaken to identifypotential sites. Simulations that allowed for equipment failures and weather history fromsuitably located observatories indicated that a minimum of six sites, spaced roughly equallyin longitude, would be required to achieve the design objective—a nearly unbroken datastream. A robust, automated, digital sunshine monitor for a survey of candidate sites wasdeveloped. Alternative technologies for making the solar velocity measurements of therequisite precision were investigated, and the nature of the overall investigation began totake shape.

In FY 1986, an interim technology development program led to the design of a variable-path-length Michelson interferometer, with a narrow prefilter isolating a single solar line,to provide a stable and sensitive velocity measurement. Early in the year, the site surveystarted operation, eventually growing to 14 potential locations. Results from the site surveyshowed that six sites are capable of observing duty cycles in excess of 90 percent. The finalsites are Big Bear Solar Observatory, Mauna Loa Solar Observatory, Learmonth SolarObservatory, Udaipur Solar Observatory, Observatorio del Teide, and Cerro Tololo Enter-American Observatory.

The project got started in earnest in FY 1987 and produced first light on a full-scaleprototype instrument in March 1990. Major peer reviews of the GONG instrument,reduction and analysis software system, and the overall project were conducted during thewinter of 1990. Subsequent peer reviews of the Data Management and Analysis Center(DMAC) plan and the network production and deployment plan were conducted in 1992and 1993. The DMAC moved into its new quarters in the AURA building across the streetfrom NOAO headquarters in late 1992.

The first field instrument was temporarily "deployed" at the Tucson integration site at theUniversity of Arizona's Campbell Research Farm in September 1994. This station,ultimately deployed at the Big Bear Solar Observatory, was used as the comparisonstandard for the other five instruments as they were operated and certified prior to

disassembly and shipment. Engineering data from a three-site "mini-network" startedflowing in May 1995. The first data from the six-site network were obtained in earlyOctober 1995.

The data reduction and analysis system is a major component of the program and wasdeveloped concurrently. Even before the installation of the network, data acquired duringthe validation of the prototype provided the basis for important research. The design andconstruction of the necessary software tools required the efforts of the project staff as wellas many other members of the solar physics community. It reached operational capabilityalong with the full network.

Coordination with the scientific community continues to be assured by regular consultationwith a five-member Scientific Advisory Committee, a six-member DMAC Users'Committee, a newsletter, and an annual GONG scientific meeting.

2. The Instrument

The five-minute oscillation is a subtle effect. Individual modes exhibit velocities of less

than 20 cm/s, while the sum of all the modes is only a few hundred m/s. The design goalwas to obtain measurements limited only by the Sun's "random" surface motions. Thismeant developing six stable instruments capable of making imaged velocity measurementswith a precision of significantly better than 1 m/s.

The basic idea is to isolate a single solar absorption line and determine its precise(Doppler-shifted) wavelength. The instrument chosen for this task is called a FourierTachometer. Based on a Michelson interferometer, it processes the light from all parts ofthe solar disk simultaneously to produce a velocity-sensitive image of the Sun. This ispicked up by a 256 x 243 pixel solid-state detector and stored in a data acquisitioncomputer's memory.

At the conclusion of each 60-second acquisition cycle, three intensity images, differing inmodulation phase by 120 degrees are produced. These can be summed, differenced, anddivided to produce a single velocity image, as well as intensity and line-strength images.The images are stored on a magnetic-tape cartridge, along with other information, forsubsequent reduction at the central data management and analysis facility.

The instrument receives light from an external feed which is automatic; it turns itself oneach day, locates and tracks the Sun, continues to track using an ephemeris during cloudyperiods, and supervises and reports its environment and operational status. The entire fieldinstrument resides in an environmentally controlled shelter building, which houses theFourier Tachometer, control and acquisition computers, and data recording equipment.

3. Data Management and Analysis

The challenges in the GONG data reduction and analysis are presented by two factors: (1) amonumental volume of data; and (2) a long sequence of complex computing tasks. Eachstation in the network produces at least 200 megabytes of data every day. The wholenetwork generates a gigabyte per day, seven days per week. The field data will exceed oneterabyte every three years. The reduction process itself is far from trivial. Each individual64K-pixel frame from each station must be adjusted, pixel by pixel, for a variety of

• I

instrumental, photometric, and geometric effects. The Doppler effect of the known motionsof the Earth and the Sun also must be removed. As many as three adjacent GONG stationsmay observe simultaneously for periods of several hours. These data must be merged into asingle stream available minute by minute.

Once these preliminary tasks are complete, several more computationally intensivereductions are performed. For example, the decomposition of the image data into timeseries of spherical harmonic coefficients, and their subsequent reduction to frequencyspectra, are standard processes. Finally, the raw field data, the ultimate reduced data sets,and several intermediate stages are placed in long-term storage in computer-based archives.Scientific analysis of the data generally proceeds from these archived data sets. An offsitebackup is also maintained.

Thus, the central GONG computing facility features a powerful network of computers andmass-storage devices. This system both reduces the data and provides for its distribution tothe community.

ill

BASE BUDGET

Cerro Tololo Inter-American Observatory

Scientific Staff and Support

Operations and MaintenanceInstrumentation

Kitt Peak National Observatory

Scientific Staff and SupportOperations and Maintenance

National Solar Observatory

Scientific Staff and SupportOperations and MaintenanceInstrumentation

USAF and NASA Support

U.S. Gemini Project OfficeCentral Computer ServicesImage Reduction and Analysis Facility

CTIO/KPNO Instrumentation

Global Oscillations Network Group

Central Offices

Director's Office

Indirect Cost and Miscellaneous Credits

Central Administrative Services

Central Facilities Operations

Central Engineering and Technical Services

Public Information and Education

REU and Gemini Fellowships

Management Fee

Total Base Budget

SOAR

GONG Upgrade

SOLIS

RISE Construction

RISE Operations

Detector Procurement

Facilities Maintenance

Total Budget Request

NATIONAL OPTICAL ASTRONOMY OBSERVATORIES

FY 1998 - 2002 Long Range Plan

Budget Summary

(Amounts in Thousands)

FY 1997(a) FY 1998 FY 1999 FY 2000 FY 2001 FY 2002

1,797

4,858

611

7,265

1,820

5,085

517

7,421

1,888

5,366

547

7,801

1,960

5,665

563

8,188

2,035

5,982

580

8,597

2,113

6,319

598

9,030

1,703

4,558

6,262

1,154

4,109

5,264

1,189

4,284

5,473

1,225

4,413

5,637

1,262

4,545

5,806

1,299

4,681

5,981

1,430

2,169

865

(567)3,896

1,388

1,930

762

(567)

3,513

1,430

1,987

810

(533)

3,694

1,473

2,047

834

(602)

3,752

1,517

2,109

859

(620)

3,865

1,562

2,172

885

(638)

3,981

866

646

379

1,267

637

357

1,313

656

368

1,352

676

379

1,393

696

390

1,435

717

402>

PIS33

2,727 3,145 3,389 3,491 3,596 3,704

M

2,166 1,850 1,850 1,850 2,000 2,000 X

613 605 624 642 662 681

(466) (235) (242) (249) (257) (264)

1,547 1,487 1,532 1,578 1,625 1,674

1,266 1,343 1,391 1,433 1,476 1,520

632 675 695 716 737 760

207 193 199 205 211 218

143 141 145 150 154 159

550 567 583 601 619 638

28,698

1,210

3,000

279

300

33,487

28,230 29,472 30,401 31,571 32,633

402 835 574

1,000 1,000 1,000

180

152 157 161 166 171

300 300 300 300 300

582 532 498 498 678

30,846 32,296 32,934 32,535 33,782

(a) FY1997 Program Plan Revision I (preliminary) including carryover of $2,487,660.