K

Report of the Panel on Optical and Infrared Observations from the Ground

EXECUTIVE SUMMARY

We face a future filled with extraordinary opportunities. New ground-based optical and infrared (OIR) observational facilities are central to addressing the most pressing and fundamental questions in astronomy and astrophysics, as assessed by the six science panels of the Astro2020 decadal survey.¹ The importance of some of these questions transcends the boundaries of science: How did we get here? Are we alone? To exploit these opportunities, the United States—which has led the world for decades in ground-based OIR astronomy—must overcome some enormous challenges.

First, to maintain a leadership position in the 2030s and beyond, investments of an unprecedented scale by the National Science Foundation (NSF) in ground-based OIR astronomy will be required. The panel has carefully evaluated the proposal to Astro2020 to create and fund a unified U.S. Extremely Large Telescope Program (U.S. ELTP) that will combine the resources and capabilities of NSF’s National Optical-Infrared Astronomy Research Laboratory (NOIRLab), the Giant Magellan Telescope (GMT), and the Thirty Meter Telescope (TMT). Combined with Key Science Programs facilitated by NOIRLab, this U.S. ELTP will create a system for the broad U.S. community that is fully competitive with, and complementary to, the European Extremely Large Telescope (E-ELT) and one that maximizes synergies with the current U.S. multi-billion-dollar bi-hemispheric system of ground-based astronomical facilities. The programmatic challenges facing the U.S. ELTP are daunting indeed, and it is not at all clear that there will be adequate financial capacity to complete the construction and fund the operations of this two-telescope system. The panel has reached the consensus that the rewards of a successful outcome are high enough for NSF and the other GMT and TMT partners to be given the opportunity to try to achieve this success. The panel believes that a zero-ELT outcome will gravely damage the U.S. astronomy community for decades.

The second challenge is posed by the need to exploit the immense investment that has already been made in the past 30 years to create a powerful and flexible U.S. ground-based OIR system. This is particularly pressing because the scientific payout from a U.S. ELTP is over a decade away. The panel has

___________________

¹ See Appendix A for the overall Astro2020 statement of task, for the set of panel descriptions that define the panels’ tasks, and for additional instructions given to the panels by the steering committee.

reviewed 50 thoughtful white papers from the ground-based OIR community, and has identified a set of thematic areas in which relatively modest investments (e.g., at or below the level of the NSF Mid-scale Research Infrastructure-2 [MSRI-2] program) in existing telescopes could reap major returns during the 2020s. The panel also highlights several opportunities in this medium-scale range to build new special-purpose telescopes or telescope arrays. In some cases, these could be interagency projects (e.g., NSF and the National Aeronautics and Space Administration [NASA], or NSF and the Department of Energy [DOE]). In other cases, they could be in the context of an international partnership. Last, the panel emphasizes the importance of modest strategic investments in technology development and software, and in the further development of the systems-level approach to optimizing the performance of the OIR system in an era of time-domain/multi-messenger astrophysics.

For this plan to succeed, there needs to be a fundamental change in the way in which the federal, state, and private funding sources for ground-based OIR facilities interact. A broad partnership is necessary for the United States to maintain leadership. NSF will need a major boost in the Major Research Equipment and Facilities Construction (MREFC) funding line, a robust MSRI program, and a new model for how operations of new facilities are paid for. If we can accomplish all of this, we can fully reap the extraordinary scientific harvests for decades to come.

K.1 INTRODUCTION

K.1.1 Setting the Stage

It is axiomatic that major scientific discoveries are driven by new technology, and in no field is this clearer than in astronomy. Galileo Galilei did not invent the telescope, but he was the first to use it to observe the sky and record his discoveries. His book The Starry Messenger (1610)² reported on his observations of the Moon, Jupiter, and the Milky Way. These observations revolutionized our understanding of the cosmos and ushered in centuries of discoveries to come based on ever-more-powerful telescopes. The era of astrophysics can be said to have begun roughly a century ago, launched by the construction of large telescopes armed with spectrographs. For many decades, the United States was the unrivaled leader in the construction and utilization of such facilities, from the 100” (arcsec) at Mount Wilson (1917) to the 200” at Mount Palomar (1948). This U.S. leadership was made possible largely through an unmatched level of philanthropy.

Such days are over. Starting in the late 1960s, the level of funding provided by the U.S. federal government and of other nations produced a suite of telescopes that matched the capabilities of the largest private/state-funded facilities. This situation has continued into the current era of very large telescopes, which started to come into operation in 1990s. The importance of the next-generation Extremely Large Telescopes (ELTs) has been recognized for at least 20 years, and indeed, an ELT was the top ground-based recommendation of the 2000 decadal survey. Yet as we survey the landscape today, we see that the scale of investment needed to construct and operate the next generation of ELTs is severely straining the financial model that has served the U.S. astronomical community so well for over a century (which has largely segregated private/state and federally funded telescopes). Initially, two competing ELT projects with major U.S. involvement emerged: the GMT and the TMT. Since the previous decadal survey, New Worlds, New

___________________

² Galilei, Galileo, and Tommaso Baglioni. Sidereus nuncius: Magna longeque admirabilia spectacula pandens, suspiciendaque proponens vnicuique, praesertim verò philosophis atque astronomis, quae à Galileo Galileo, patritio Florentino, Patauini Gymnasij publico mathematico, perspicilli: nuper à se reperiti beneficio sunt obseruata in lunae facie, fixis innumeris, Lacteo Circulo, stellis nebulosis, apprime verò in quatuor planetis circa Iouis stellam disparibus interuallis, atque periodis, celeritate mirabili circumuolutis … atque Medicea sidera nuncupandos decreuit. [Venetijs: Apud Thomam Baglionum, MDCX, 1610] PDF. https://www.loc.gov/item/2010667904.

Horizons³ (Astro2010), there has been an enormous amount of work done to retire the most challenging technical risks to the construction of GMT and TMT. These two projects have now joined forces with NSF’s NOIRLab to propose a unified U.S. ELT program seeking substantial federal funding for, and providing access to, both ELTs by the U.S. community.

In addition, two major new ground-based OIR telescopes have just seen (or soon will see) first light. The 4 m Daniel K. Inouye Solar Telescope (DKIST) is the largest solar telescope in the world, with a focus on understanding the Sun’s explosive behavior. The Vera C. Rubin Observatory (the highest ranked ground-based project in Astro2010) will undertake an astronomical survey, the Legacy Survey of Space and Time (LSST), that will fully open the window of time-domain astronomy. Together, these two facilities cost just over $1 billion in 2020 dollars for construction, and continuing investment will be needed to realize a commensurate return on investment.

The past decade has also seen investment in new capabilities for existing U.S. ground-based OIR telescopes. The twin national 4-m telescopes have been converted to wide-field survey machines with the Dark Energy Camera (Blanco), and the Dark Energy Spectroscopic Instrument (Mayall) made possible by an NSF/DOE partnership. The 3.5-m WIYN⁴ telescope is now equipped with a state-of-art spectrograph (NN-EXPLORE Exoplanet Investigations with Doppler Spectroscopy [NEID]) for precision radial velocity measurements of exoplanets, thanks to an NSF/NASA partnership. The development and deployment of multi-object and integral-field spectrographs, and of advanced adaptive optics (AO) and high-contrast imaging systems has made these technologies powerful tools on the largest telescopes. Pioneering work on optical/near-IR interferometers has opened up a whole new terrain mapped with high angular resolution.

In this report, the panel assesses the investments needed in the next decade to ensure that the United States remains at the forefront in addressing the most important and enduring scientific questions (many of which can only be answered with the next generation of ground-based OIR observatories—see Section K.2 below). In the case of the ELTs (discussed in Section K.3), the cost will be considerable, with the investment needed now. The scientific payoff—while immense—will be more than a decade away. The panel argues in Section K.4 that additional investment is essential so that, in the meantime, we can fully realize the potential of the powerful and flexible OIR system constructed at great cost over the past few decades.

K.1.2 Input to the Panel

The analyses reported below are based on extensive input to the panel. The panel individually discussed 50 white papers from the community. These were used primarily to define the topical areas described in Section K.4 and guided the panel’s thinking in these areas. The panel had two presentations by NSF: one in which they provided an overall budgetary framework and entreated the panel to think boldly and a second in which they clarified the nature of the cross-divisional support in NSF for solar astronomy and the role of Astro2020 in providing guidance in this area.

The panel received and analyzed the more detailed Requests for Information Version 1 (RFI 1’s) from the GMT, TMT, and NSF’s NOIRLab for the U.S. ELT Program, Center for High Resolution Astronomy (CHARA) (OIR Interferometer), Coronal Solar Magnetism Observatory (COSMO) (solar observatory), and the Mauna Kea Spectroscopic Explorer (massively multiplexed spectrograph). The panel received and analyzed the less detailed RFI 2’s from the Fiber-Optic Broadband Optical Spectrograph (FOBOS) and MegaMapper

___________________

³ National Research Council, 2010, New Worlds, New Horizons in Astronomy and Astrophysics, The National Academies Press, Washington, DC.

⁴ University of Wisconsin–Madison (W), Indiana University (I), Yale University (Y), and the National Optical Astronomy Observatories (N).

massively multiplexed spectrographs, the Navy Precision Optical Interferometer (NPOI) and Magdalena Ridge Observatory Interferometer (MROI) OIR interferometers, and the Liger and wide-field adaptive optics (WFAO) next-generation AO systems. In the case of the U.S. ELT program, the panel received full Technical, Risk, and Cost Evaluation (TRACE) analyses (see Appendix O). The panel went through three iterations of questions and answers with the U.S. ELT program based on the RFI 1’s and initial TRACE results. The panel also had a face-to-face meeting with the U.S. ELT program leads and key personnel (from Association of Universities for Research in Astronomy (AURA), GMT, NOIRLab, and TMT). The panel did not request a TRACE for the other projects with RFI 1 or 2, because the panel did not deem them as being high-priority candidates for NSF MREFC-funding.

K.2 THE SCIENCE FRONTIER

Before considering the different proposed programmatic elements in the future U.S. ground-based OIR system, the panel would like to set the stage by highlighting four areas at the science frontier for which these facilities would play the most essential role. These are drawn from the reports of the science panels, representing this panel’s synthesis of their analyses.

K.2.1 Exoplanets and Astrobiology

Is the solar system a cosmic rarity or a galactic commonplace? How do Earth-like planets form, and what determines whether they are habitable? Is there life on other worlds? The National Academies report Exoplanet Science Strategy⁵ and the Panel on Exoplanets, Astrobiology, and the Solar System have identified two ground-based capabilities in optical and infrared astronomy that are essential to realize the great opportunity in exoplanets and astrobiology that is open before us.

First, the GMT and TMT will open an unprecedented discovery space in the study of planet formation, mature gas giants, and even terrestrial worlds. The unprecedented contrast and angular resolution of the GMT and TMT will enable profound advances in imaging and spectroscopy of entire planetary systems, over a wide range of stellar and planetary masses, semi-major axes, and wavelengths, including both reflected light and thermal emission. For low-mass stars, the reduced glare of the central star may permit the GMT and TMT to image temperate rocky worlds. These facilities will also allow the detection of newly formed planets in their natal disks, providing the ground truth for the time scale of planet formation and permitting studies of the dynamical interaction between disks and planets. With the high spatial resolution of the GMT and TMT, researchers will finally be able to search the inner parts of planet-forming systems. The unprecedented light-gathering capability of the GMT and TMT, coupled with high-resolution optical and infrared spectrographs, will be powerful tools for studying the atmospheres of transiting and non-transiting close-in planets. For the closest and least massive stars, these observatories may detect molecular oxygen in the atmospheres of transiting terrestrial planets orbiting within the habitable zone of their stars, landmark discoveries with implications far beyond astronomy and even science.

Second, as mass is the most fundamental property of a planet, the astronomical community needs to develop the ability to measure the radial velocities of stars to a precision sufficient to measure the masses of their attendant temperate terrestrial exoplanets. These mass measurements are essential both for determining the planet’s bulk composition (and, by inference, its formation history) and for interpreting studies of the planetary atmosphere, since the atmospheric scale height depends on a combination of surface

___________________

⁵ National Academies of Sciences, Engineering, and Medicine, 2018, Exoplanet Science Strategy, The National Academies Press, Washington, DC.

gravity, temperature, and mean molecular weight. Exoplanet Science Strategy found that “Radial velocity measurements are currently limited by variations in the stellar photosphere, instrumental stability and calibration, and spectral contamination from telluric lines.”⁶ Hence, progress will require a coordinated national initiative in extreme precision radial velocities that includes observers, instrument builders, stellar astrophysicists, heliophysicists, and statisticians. Although the bulk of this effort will involve smaller ground-based telescopes, the photon gathering capability of the GMT and TMT will also play a role, enabling photon-noise-limited Doppler precision of several centimeters per second on time scales of a few minutes, allowing researchers to disentangle the signals from the stellar photosphere and orbital motion.

The grand opportunities now afforded by exoplanets cannot be addressed with a single telescope: the optimal targets for some of the most demanding exoplanet studies will be rare and demand access to the entire celestial sphere. In addition, the technology roadmap to enable the full exoplanet science potential of GMT and TMT will be realized only by leveraging the existing network of U.S. centers and laboratories and current 8- to 10-m class facilities.

K.2.2 The Fundamental Physics of the Universe

Over the past century, some of the most revolutionary discoveries in fundamental physics have come from OIR astronomy: from the confirmation of Einstein’s prediction of the trajectory of light passing close to the Sun and Hubble’s discovery of the expansion of the universe to the discovery of dark matter and of an accelerating universe (likely owing to dark energy). In this vein, there are a host of exciting paths to follow over the next decade and beyond. Here, the panel focuses on cases in which ground-based OIR observations are essential.

Tests of general relativity (GR). The galactic center is an ideal laboratory for testing GR in the context of a supermassive black hole (BH). In the coming decade, it will be possible to use the next generation of AO to precisely measure the orbits of stars close to Sag A*. Such measurements will test both the Einstein equivalence principle and the form of the Kerr metric around a BH, allowing a search for a massive graviton or additional interactions (like scalar fields). By using ELTs to go 5 magnitudes fainter than the confusion limit today, there will likely be multiple stars with periods as short as 1 to 2 years, providing even more stringent tests of GR. With the improved sample and astrometric and radial velocity precision enabled by ELTs, it will be possible to detect GR effects that have no analog in classical dynamics, such as the precession of the periapse and the Lense-Thirring or frame dragging effect, which is owing to the spin of the BH.⁷

The nature of dark matter. Models of dark matter span an enormous range in particle mass and in the degree to which the dark matter particles self-interact. The combination of light-gathering power and high angular resolution will enable the new ELTs to test these models. By measuring precise radial velocities and proper motions, it will be possible to construct maps of the 3D stellar orbits at dwarf galaxy centers to map the radial distribution of dark matter to thereby determine whether self-interacting dark matter is required. Similarly, precise maps of gravitationally lensed background sources utilizing the sensitivity and spatial resolution of an ELT will make it possible to determine the dark halo mass function at low masses, which depends on the nature of dark matter.

___________________

⁶ National Academies of Sciences, Engineering, and Medicine, 2018, Exoplanet Science Strategy, The National Academies Press, Washington, DC., p. 96.

⁷ T. Do and A. Ghez, 2019, “Envisioning the Next Decade of Galactic Center Science: A Laboratory for the Study of the Physics and Astrophysics of Supermassive Black Holes,” white paper submitted to Astro2020: Decadal Survey on Astronomy and Astrophysics, https://arxiv.org/pdf/1903.05293.pdf.

Tests of the expansion history of the universe. These will not only constrain the nature of dark energy but can also look for signs of “new physics,” which is hinted at by the so-called H₀ tension (an apparent difference in the Hubble constant based on local scales versus the cosmic microwave background). Ground-based OIR astronomy will play a crucial role in this. Key programs include improved measurements of the evolution of the large-scale structure of the universe (made using the next generation of massively multiplexed spectrographs) and improved and refined measurement of the distance scale using new standard candles, standard sirens (electromagnetic counterparts to gravitational wave events), and standard clocks (time delay measurements of lensed Type Ia supernovae). These techniques really require the sensitivity and spatial resolution provided by ELTs to achieve their full potential. ELTs are required to reach 1 percent precision H₀ measurements with multiple techniques. With ELTs, tip of the red giant branch distances will reach deep into the Hubble flow to ~100 Mpc, enabling large numbers of direct measurements, while geometric eclipsing binary distances will be possible for all Local Group galaxies anchoring these distances. Gravitational lensing systematics, both the mass distribution of quasar hosts and the line-of-sight mass distribution, can be minimized with deep and high-resolution Integral Field Units observations (IFU). Standard siren host galaxies are likely to be faint; ELTs will be needed for spectroscopic follow-up to obtain redshifts.

Understanding the physics of cosmic inflation. Maps of the large-scale structure of the universe made possible with galaxy redshift surveys undertaken with a future generation of massive-multiplexed wide-field OIR spectrographs are crucial here. For example, a clear detection of non-Gaussianity in the primordial fluctuations would rule out the simple “single field” inflation scenario. The detection of departures from a power-law form of the primordial power spectrum would offer the opportunity to explore the history of the primordial seeds of large-scale structure during their production in the era of inflation. These require sampling significantly larger volumes with a much larger number of galaxies than can be done with existing or nearly complete facilities.

K.2.3 Galaxy Evolution

In the next decade and beyond, ground-based OIR facilities will remain key to unraveling the fundamental processes that drive galaxy evolution. Future capabilities will transform our view of galaxies, their constituent components, and their vital link with the circumgalactic and intergalactic medium (CGM/IGM) via baryon cycling. Specific areas of progress are captured by the themes identified by the Panel on Galaxies: (1) the evolution of the IGM and the sources of radiation during the first billion years from cosmic dawn throughout the epoch of reionization; (2) the gas, metals, and dust flows into, through, and out of galaxies at all epochs; (3) the formation of supermassive black holes and the coupling of their growth with that of their host galaxies; (4) the impact of the assembly history of galaxies and dark matter (DM) halos on their observable properties; and (5) the discovery and investigation of the first galaxies.

The complex interplay between baryons and DM, and the richness of physical processes involved, call for multi-scale multi-wavelength approaches probing the full ecosystems across time. Ground-based OIR systems play a pivotal role by accessing a wealth of key spectral tracers of stars, gas, and metals. Transformative advances rely on progress along two synergistic axes: spatially and spectrally detailed information and multiplexing. Key enabling capabilities are (1) R ~5,000–50,000 spectroscopy; (2) resolution in the infrared down to 0.01–0.02” (roughly 100 pc); (3) dense/full spectroscopic coverage up to ~2 arcmin scales (roughly 1 Mpc at z >1); and (4) 10–100× higher sensitivity than existing facilities.

Detailed information is vital to unveil physical processes at the level of individual stars and their surrounding nebulae in the Local Group, of individual 100 pc-scale star-forming complexes and globular cluster progenitors at z >1, and of the first galaxies at z >10 with compact <<1 kpc sizes. The characterization of their dynamical, chemical, and thermal state requires velocity resolution <10–100 km/s. Achieving

the corresponding angular resolution of 10–20 mas and spectral resolution of several 10³ to a few 10⁴, with the necessary sensitivity for such faint sources, will only be possible with ground-based ELTs, coupled with high-performance AO systems. With 5–10 times sharper views, ELTs+AO will map z >1 galaxies in similar detail as we have now for galaxies a mere 100 Mpc away—a drastic advance reaching fundamental sub-galactic components. The ELTs will thus uniquely enable major breakthroughs in galaxy evolution.

Multiplexing is essential to build up sample statistics and to cover wide contiguous areas. This is needed for (1) a complete census and detailed characterization of ~10⁸ stars and the low-mass satellites that encode the Milky Way’s chemo-dynamical history and the baryon-DM interplay, (2) to connect directly individual galaxies and their surrounding CGM/IGM, and (3) to map the distribution and properties of the first generations of galaxies holding keys to the history and topology of reionization. Ongoing projects and future concepts for massively multiplexed spectrographs—be they multi-object or integral field instruments—will deliver the necessary 10–100× larger samples. Coupled with high sensitivity and angular resolution, they will reach unexplored low mass/luminosity regimes that are very sensitive to the physics of star and galaxy formation, overcome crowding in nearby studies, and greatly boost observational efficiency for distant sources.

K.2.4 A Stellar Renaissance

Over the past 15 years, the solar and stellar astrophysics community has ignited a scientific renaissance through a remarkable investment in facilities for global synoptic and high-resolution solar observations, extreme-precision stellar radial velocity, and milliarcsecond imaging of nearby stars, as well as ultrawidefield surveys focused on high-precision position measurements, high-cadence time-domain measurements, medium-resolution optical-IR spectroscopy, and multi-band precision photometry. Ground-based OIR facilities have played, and will continue to play, the central role.

A crowning achievement for ground-based solar astronomy is the recent completion of NSF’s 4 m DKIST, whose unprecedented spatial resolution and spectropolarimetric sensitivity will open a new window on the magnetohydrodynamic phenomena that affect convective motions and drive the storage and release of magnetic energy (reconnection). These, in turn, result in sunspots, flares and eruptive events, solar energetic particles, coronal heating, and perhaps most fundamentally, the solar cycle, all of which have impact on our home planet through space weather. Solar measurements made from the ground (DKIST, Global Oscillations Network Group (GONG), Synoptic Optical Long-Term Investigations of the Sun (SOLIS), Goode Solar Telescope (GST), and other observatories, as well as the planned COSMO, next generation Ground-based solar Observing Network (ngGONG), and DKIST-II) and from space (Solar Dynamics Observatory (SDO), Solar and Heliospheric Observatory (SOHO), Solar Orbiter [with ESA], Parker Solar Probe, and the in-development Solar EUV High-Throughput Spectroscopic Telescope [with the Japan Aerospace Exploration Agency]) mean that the coming decade will provide significant new insight into physical processes that apply to all stars. High time-cadence observations by NASA’s Kepler mission have opened an analogous window on the interiors of distant stars via asteroseismology and photometry, while high spatial-resolution observations with ground-based optical interferometers such as the Center for High Resolution Astronomy (CHARA) and the Very Large Telescope Interferometer (VLTI) are producing astonishing insights into the surface phenomena and asymmetries of nearby stars.

Meanwhile, the combination of ground- and space-based ultra-wide-field imaging, spectroscopic, and high-precision astrometric surveys (e.g., Sloan Digital Sky Survey [SDSS]-IV Apogee, ESA’s Gaia, DOE/NSF Dark Energy Survey) are enabling a giant leap forward in researchers’ ability to disentangle galactic structure and more precisely constrain the dynamical and chemical evolution of the Milky Way over its entire lifetime, as well as providing new, more precise measurements of fundamental stellar properties and life

cycle stages across the HR diagram. Indeed, such large surveys have been critical in detecting examples of statistically rare, short-lived stellar types.

The renaissance in solar and stellar astrophysics will continue and accelerate in the decade ahead thanks to OIR ground-based facilities that have just started operation or will begin soon, such as NSF/DKIST, DOE/NSF/Rubin Observatory, DOE/NSF/DESI-Stellar, NASA/NSF/NEID, and ESO/4MOST. High-time cadence information will be added for large stellar samples by NASA/Transiting Exoplanet Survey Satellite (TESS) and ESA/Planetary Transits and Oscillations of stars (PLATO). Many of the brightest stars in those samples will be observable by ground-based optical interferometers and high-spectral resolution spectrometers and spectropolarimeters to open a new window on the connection between stellar surface and interior phenomena. Looking further into the future, the spatial resolution of ELTs will isolate single stars below the main-sequence turnoff, map the orbits of tight binaries, and enable proper motion selection in distant clusters in currently inaccessible environments throughout the Local Volume. The ELTs will also be useful to follow up the detections of candidates for Pop III stars in the halo of the Milky Way and in other Local Group galaxies.

Affiliating and effectively using these highly multidimensional peta-scale data sets will require new approaches to managing and analyzing positional, photometric, and spectroscopic information, likely incorporating database and machine-learning techniques developed outside astronomy. Astronomers have always worked closely with mechanical, optical, electrical, and computer engineers. In the decades ahead, astronomers will also embrace working with data scientists in ways not broadly appreciated even 10 years ago.

K.3 THE U.S. EXTREMELY LARGE TELESCOPE PROGRAM

K.3.1 Introduction

We stand at a watershed moment. A new generation of Extremely Large Telescopes is essential to answering the most important questions in astronomy and astrophysics in the 2030s and the decades that follow. It seems clear that without a major federal investment, the two ELT programs with U.S. stakeholders, the Giant Magellan Telescope (GMT) and the Thirty-Meter Telescope (TMT), will both fail, severely damaging U.S. astronomy for decades.

At this pivotal time, the two ELT projects have joined forces with NSF’s NOIRLab to propose a radically new concept to Astro2020: a unified U.S. ELT program that will provide the broad U.S. astronomy community with access to both GMT and TMT in exchange for a substantial federal investment in the ELTs and with an interface to this broad community provided by NOIRLab. The panel then considered two stark choices: (1) Endorse an unprecedented level of NSF investment and a revolutionary new “business model” to capitalize on this investment or (2) cede U.S. leadership in the frontier science enabled by ELTs for decades to come. The panel argues below in favor of the first choice on the basis of the review of a proposed U.S. ELT program uniting GMT, TMT, and NOIRLab.

K.3.2 The Science Case

In Section K.2, the panel described four broad areas at the scientific frontier where the next generation of ground-based OIR facilities are most essential and explicitly identified there the need for ELTs, from the search for life on distant planets, to the births and lives of galaxies, to the fundamental physics of the cosmos.

Here, the panel reports on its analysis of the reports of the science panels to identify which of the specific questions and discovery areas (DAs) will need ELTs. For each of these, the panel has evaluated whether the ELTs are essential (18 cases), useful (8 cases), or not needed (4 cases) to answer the questions posed. The essential items are listed in Table K.1, which makes it clear that the science case for ELTs is

TABLE K.1 List of Essential Needs for ELTs in Science Frontier Panel Reports

extraordinarily broad and deep (and likely unmatched by any single other new ground- or space-based observatory currently under consideration).

K.3.3 The Components of the U.S. ELT Program

The U.S. ELTP as proposed to Astro2020 is made up of three elements: GMT, TMT, and NOIRLab.

The primary mirror of the GMT consists of seven 8.4 m segments, for a total diameter of 24.5 m. It is a two-mirror aplanatic Gregorian telescope yielding a 25 arcmin field of view (FOV) at f/8 with ground-layer AO. Alignment and phasing are achieved with adaptive secondary mirrors. Laser tomography AO utilizes six sodium laser beacons and edge sensors. With this, diffraction-limited images of 10 mas at 1 μm can be achieved over a 20–30 arcsec FOV. The first generation of instruments includes the GMT Multi-Object Astronomy and Cosmology Spectrograph (GMACS), the GMT Integral Field Spectrograph (GMTIFS), the GMT Consortium Large Earth Finder (G-CLEF), and the GMT Near-IR Spectrograph (GMTNIRS). The GMT will be located at the Las Campanas Observatory in Chile. The majority of the GMT partners are U.S.

institutions, with international partners in Australia, Brazil, and Korea. The site has been excavated, the site water and electrical distribution upgrades have been completed, and the sixth mirror has been cast.

The primary mirror of the TMT consists of 492 1.44-m mirror segments, for a total diameter of 30 m. It is a three-mirror Ritchey-Chretien telescope delivering a 20 arcmin FOV. Ground-layer AO can improve image quality over a 2 arcmin FOV. Six sodium lasers and 3 tip/tilt natural guide stars will be employed. With this Narrow Field Infrared Adaptive Optics System (NFIRAOS), diffraction-limited images of 8 mas will be achieved over a 15–30 arcsec FOV. Instruments will be deployed at the Nasmyth platforms. The first-generation instruments include the Infrared Imaging Spectrometer (IRIS), the Wide-Field Optical Spectrometer (WFOS), and the Multi-Object Diffraction-Limited High-Resolution Infrared Spectrograph (MODHIS). The TMT will either be sited at the Mauna Kea Observatory in Hawaii (MKO) or at Roque de los Muchachos Observatory in the Canary Islands (ORM). The majority of the TMT partners are international, with the participation of institutions in the United States, Canada, China, India, and Japan.

NSF’s NOIRLab is the U.S. national center for ground-based, nighttime optical and infrared astronomy. It operates five programs—Cerro Tololo Inter-American Observatory (CTIO), the Community Science and Data Center (CSDC), Gemini Observatory, Kitt Peak National Observatory (KPNO) and the Vera C. Rubin Observatory (Rubin Observatory). Its roles in the U.S. ELTP will be to (1) provide access to a bi-hemispheric ELT system; (2) enable and support large-scale, systematic, collaborative research (Key Science Programs—see Section K.3.10, below); (3) provide user support, broaden participation in TMT/GMT science, and foster research inclusivity; and (4) engage and represent the whole U.S. community in GMT and TMT governance, scientific planning, and instrumentation development.

K.3.4 Technical Risks

Since Astro2010, both the GMT and TMT projects have devoted significant resources to identify and retire the major technical risks. It is the panel’s assessment that—for the most part—these are now both technically mature projects that are mastering the implementation of new technology. However, there are still issues to be resolved. The TRACE analysis by the Aerospace Corporation gave both projects a medium rating in technical risk. The largest concerns for both projects were the active and adaptive optics systems.

For GMT, the most important issue is the complexity of 7-segment adaptive secondary mirrors, since phasing and alignment to desired specifications across four modes is required. The on-axis guide star and laser guide star modes were found to be the most challenging. Quantitatively, the adaptive 7-segment secondary mirror has 675 actuators per segment. Challenges with this system could lead to schedule delays and may limit desired scientific performance across select observing modes. The laser tomography AO mode depends on six sodium lasers and edge sensing to <15 nm root mean square (RMS). Its design requires significant development.

For TMT, the biggest concern is that the NFIRAOS AO instrument is complex, with multiple sodium lasers, wavefront sensors, deformable mirrors, and downstream imagers and spectrographs. It is a single point failure whose performance directly impacts two of three first-generation science instruments. It still requires significant development across multiple partners. The primary mirror integration and wavefront control system for the 492 segments involves 1,476 actuators and 2,772 edge sensors, and will also be challenging. Problems with this system could lead to delays.

K.3.5 Construction Costs and Funding Model

In the panel’s view, the most serious risk to the successful construction for both the GMT and the TMT projects is financial rather than technical. The cost estimates provided by the projects are considerably

higher than those previously understood by the astronomy community and will severely strain the financial capacity of even a full partnership between the projects and NSF.

The GMT project now estimates a total construction cost of $2 billion in real year (RY) dollars. Of this amount, 20 percent has been spent to date. An additional 10 percent has been committed by current partners. The project plans for the remaining 70 percent to come from NSF (40 percent), additional (uncommitted) funds from existing partners (15 percent), and funds from unidentified new partners (15 percent). To state it differently, even with all the funds expended and committed, and with an $800 million commitment from NSF, there is still a shortfall of roughly $600 million.

The TMT project now estimates a total construction cost of $2.65 billion in RY dollars. Of this amount, 18 percent has been spent to date. An additional 41 percent has been committed by the current partners. The project plans for the remaining 41 percent to come from NSF (30 percent) and additional (uncommitted) funds from current partners (11 percent). In this case, with all the committed funds and full funding from NSF, there is still a shortfall of $310 million.

The construction costs estimated in the TRACE reports are about 20–30 percent higher than the project estimates ($2.6 billion for GMT and $3.1 billion for TMT). They are largely the result of conservative assumptions made by the TRACE analysis as to risk. All of these numbers are summarized in Table K.2.

K.3.6 Programmatic Risks

The largest single programmatic risk to each project derives directly from the funding issues summarized above. Both projects need significant additional new funding beyond the planned request from NSF. Both projects believe that the combination of the imprimatur of a top ranking in the decadal survey, followed by the full financial involvement of the U.S. federal government, would make it possible to secure additional resources from existing partners and possibly from new partners.

Before discussing the two projects, it is important to emphasize that they are based on very different funding models, each with different potential risk factors. The majority of the GMT partners are U.S. universities (plus the Carnegie Institution for Science). The model for the contribution and allocation of financial resources is strongly cash-based. In contrast, the majority of the TMT partners are international entities funded by their respective national governments. The TMT funding model is largely based on significant in-kind contributions from these partners in terms of completion of assigned technical work packages.

The GMT project’s estimate of the additional funds needed in addition to the funds currently committed and the funds requested from NSF is $600 million. If the construction cost for the TRACE is adopted, this increases to $1.2 billion. A further risk factor for GMT is the relative immaturity of estimates for cost-to-go.

TABLE K.2 Summary of Construction Costs and Funds, in Millions of Dollars and Real Years

Only 16 percent of these are based on signed contracts or detailed bids, with the other 84 percent being based on “rough-order-of-magnitude” or project estimates. This leads to additional uncertainty in cost and schedule.

The TMT project’s corresponding estimate of these additional new funds is $310 million. If the TRACE cost estimate is adopted, this becomes $760 million. A further potential risk factor is TMT’s reliance on international partners. This is based on a model of low-cost in-kind critical components and sub-systems. While TMT has agreements in place that a given international partner would shoulder any additional costs associated with the delivery of their work packages, this model has not been tested under extreme circumstances. If this funding model is successful, the TRACE delta could be significantly decreased.

An additional potential programmatic risk for TMT is posed by the uncertainty in its choice of site. Based on the documents presented by TMT, which were analyzed by the panel and in the TRACE report, a timely decision to build TMT on ORM would not lead to an increase in cost or a delayed schedule compared to MKO. Moreover, the panel has reviewed the relevant metrics on site quality and finds that—while MKO is the superior site—the ORM site is acceptable. The largest impact would be in the thermal infrared and in the ultraviolet near the atmospheric cutoff. Despite this assessment, the choice of a site still poses a significant programmatic risk because it could adversely affect the partnership.

The TRACE analysis—based primarily on the considerations above—gave both projects a medium-high programmatic risk. The TRACE evaluation flagged the schedules as being too aggressive. GMT plans for 12 years, including Laser Tomography Adaptive Optics (LTAO) commissioning, while the TRACE estimate was 13 years. TMT plans for 10 years, while the TRACE estimate was 13 years. The panel notes that there is better agreement with risk-adjusted schedules from the projects: 13.7 years for GMT and 11.2 years for TMT. Both stated that their costing included the risk-adjusted schedules.

K.3.7 Life Cycle Costs

As is discussed further in Sections K.3.8 and K.3.9, the panel regards it as essential that a plan is developed to ensure that adequate funding is in place through a combination of federal and project funds to operate the U.S. ELTs with high efficiency, and to continue to provide them with state-of-the-art instruments throughout their scientifically productive lifetimes.

The two projects have developed bottom-up estimates for operations of these facilities and have budgeted for partial funding of future instruments. The panel is concerned that the estimated operations and instrumentation budgets are too lean, for both GMT and TMT. GMT has budgeted $30.6 million per year for operations and $10 million per year for new instruments. The total annual amount of $40.6 million represents about 2.1 percent of construction costs (all numbers being in 2020 $). TMT has an annual budget of $33.3 million for operations and $13.7 million for new instruments. The total annual budget of $47 million represents 2 percent of construction costs. In comparison, the corresponding fractions are 4 percent for the E-ELT and 5 percent for Rubin Observatory and Atacama Large Millimeter/submillimeter Array (ALMA).

It is the panel’s assessment that the currently proposed operations budgets are too low to support the type of highly efficient, flexible mode that will be needed to capitalize fully on the financial investment. Likewise, the budget for future instrumentation will not be adequate to ensure a continuing line of new, state-of-the-art instruments.

K.3.8 Consideration of the U.S. ELT Program

In this section, the panel considers the case for the proposal to Astro2020 by the U.S. ELTP for federal investment that would unite the GMT, TMT, and NOIRLab. The panel first discusses the question of whether this should be done and then asks whether it can be done, from a financial point of view.

The panel emphasizes here that a single proposal was received from the U.S. ELTP (not separate GMT and TMT proposals). The panel has therefore considered this proposal for investment in a two-ELT system. In Section K.3.9, the panel addresses what would happen if NSF and the two ELT projects cannot firmly commit to the total financial resources required.

K.3.8.1 The Strategic Case for the U.S.-ELTP

The panel believes that the scientific case for a U.S. ELT program is strong and compelling. The technical risks, while challenging, are manageable. The greatest threat to these projects is financial in nature and could potentially be retired through an NSF investment, if that is part of a full and robust financial plan with the other partners. Given the status of the GMT and TMT projects, and the ongoing construction of the E-ELT, the time to invest is now.

Here, the panel summarizes the strategic considerations that informed its thinking. These considerations form the core of the argument to Congress for federal support at such a substantial level.

Ensuring U.S. Leadership: The construction of the E-ELT is well under way. Absent a federal investment, it seems clear that the U.S. community will cede leadership in what has been the backbone area of observational astrophysics for over a century. In this context, a federal investment only makes sense if the program it enables is at a level that effectively competes with and complements the E-ELT capabilities over a long lifetime of discovery.

The E-ELT (39-m aperture) is larger than either the GMT (24.5 m) or TMT (30 m). A U.S. federal investment in either GMT or TMT on its own will not achieve parity. In particular, the sizes of the professional astronomical communities in the United States and Europe are similar, so a fraction of a single ELT would underserve the U.S. community.

There are a number of ways in which the sum of the GMT and TMT can offer important advantages over the E-ELT, some of which are described below. Here the panel emphasizes two points. The first is that the United States has a powerful legacy of astronomical discoveries using a bi-hemispheric OIR system of telescope. Unlike Europe, this bi-hemispheric approach has enabled the U.S. community to undertake major surveys or follow up on high-stakes targets in any hemisphere required. The second is that GMT and TMT have respective unvignetted fields of view that are six and four times larger than the E-ELT in terms of solid angle, making them more efficient for undertaking surveys. The panel notes that the U.S. community has extensive experience in large astronomical surveys. This wide-field, bi-hemispheric capability makes the U.S. ELTP complementary to the larger-aperture E-ELT.

Capitalizing on Investments and Assets: The strongest argument in favor of a bi-hemispheric U.S. ELTP is the importance of being able to fully exploit the synergy made possible by the investments already made in bi-hemispheric facilities. No matter how powerful the ELTs may be, they are part of a system of ground-based facilities that have been constructed at great cost (over $3 billion). These facilities are needed to unlock the scientific power of the ELTs, and vice versa. These facilities are (by hemisphere, in alphabetical order):

• North: ARC 3.5, Gemini North, High-Altitude Water Cherenkov Observatory (HAWC), Hobby-Eberly Telescope (HET), Keck I and II, Large Binocular Telescope Observatory (LBTO), Mayall (DESI), Multiple Mirror Telescope (MMT), Palomar 5 m, Panoramic Survey Telescope and Rapid Response System (Pan-STARRS) I and II, Palomar Transient Factory (PTF), Sloan Telescope (SDSS N), Subaru (Hyper Suprime-Cam [HSC] and Prime Focus Spectrograph [PFS]), Submillimeter Array (SMA), Very Large Array (VLA), WIYN (NEID)
• South: ALMA, Auger Observatory, Blanco (Dark Energy Camera [DECam]), DuPont (SDSS-S), Gemini S, IceCube, Magellan I and II, Rubin Observatory, South African Large Telescope (SALT), Southern Astrophysical Research Telescope (SOAR), South Pole Telescope (SPT) (cosmic microwave background/Sunyaev-Zel’dovich [CMB/SZ])

In addition to these ground-based facilities, space-based facilities are inherently all-sky and require bi-hemispheric ground-based facilities to fully exploit their power. Last, there are fields on the sky that are only accessible to either a northern- or southern-hemisphere ELT and represent either an enormous existing investment in space-based and/or ground-based observations or are astronomically unique, including:

• North: Extended Groth Strip (EGS), Euclid NEP, Great Observatories Origins Deep Survey (GOODS)-N, Kepler, M 31, SDSS (main), TESS-N
• South: GOODS-S and Euclid DF-S, Euclid DF-Fornax, Large Magellanic Cloud (LMC), Roman Space Telescope’s (ST) High Latitude Survey (HLS), SDSS (S), Small Magellanic Cloud (SMC), TESS-S

There are also very rare and important sources to be discovered that may be visible from only one hemisphere, and there are monitoring programs where the longitudinal coverage afforded by two ELTs would be valuable.

Risk Mitigation: There are serious programmatic risks for each of the ELTs, leading to multiple ways in which either could fail. By pursuing the as-proposed U.S. ELTP, the risk of a zero-ELT outcome is reduced.

Platform for Innovation: A two-telescope system provides a broader platform for innovation. First, two telescopes provide more nights of observations to the community, making a broader scientific program possible. Second, while the first generation of instruments on GMT and TMT have overlapping capabilities, this need not be the case in the future. The U.S. ELTP could decide on a strategic plan through which future generations of instruments are complementary and access is shared across the partners. This would add to the advantage that the smaller plate scales for GMT and TMT mean that instruments will be less expensive compared to those on the E-ELT, making it possible to achieve greater diversity in capabilities for a fixed investment.

NOIRLab as a “Force-Multiplier”: A crucial strategic consideration in the panel’s view is the use of NOIRLab as a “force-multiplier” in terms of return to the community. The panel agrees with their plan (described in Section K.3.10) to develop a set of large strategic community-driven Key Science Programs on a scale that would greatly boost the impact of the U.S. ELTP. In fact, the panel believes that the true power of this approach really requires that the non-U.S.-federal partners also contribute observing time to these programs and are fully engaged with the broader U.S. community in terms of selecting, designing, and executing these programs.

Trans-Pacific Partnership and Worldwide ELT System: The successful formation of the U.S. ELTP would join together the United States in a scientific partnership spanning the Pacific Ocean, from North America (Canada and the United States) and South America (Chile), to Australia, East Asia (China, Japan, and Korea), and South Asia (India). This partnership would be able to forge collaborations with the European ELT from a position of strength and in so doing form a powerful worldwide ELT system that fully exploits the complementary strengths of these facilities for decades to come.

K.3.8.2 Can This Be Done?

As discussed in Section K.3.5, the intention of the U.S. ELTP is to request a total of $1.6 billion RY dollars funded by NSF’s MREFC line. These funds would be split evenly between GMT and TMT. Can this be accommodated within NSF’s budget?

NSF presented an aspirational MREFC budget for the coming decade that represents a very significant increase over the mean annual budget over the past decade. The U.S. ELT program has also presented a baseline for when the funds from NSF would be needed. The comparison between this request and the notional NSF MREFC budget is shown in Figure K.1.

From this, two conclusions are obvious. Yes, the U.S. ELTP can fit within the NSF guidance. However, the MREFC line pertains to the entire NSF program (not just astronomy), and over the 5 fiscal years starting in 2023, the U.S. ELTP by itself would require 70 percent of the total MREFC budget. In other words, NSF might have the capacity to provide the requested funds, but only if the U.S. ELTP was the number one major NSF construction project until late in the decade.

The panel has expressed its concerns in Section K.3.7 that the proposed funding level for operations and new instruments may be insufficient to realize a commensurate return on the enormous investment in construction. Even the NSF share of these budgets as proposed by the projects is problematic, since the current NSF model is one in which operations costs of observatories built using MREFC funds are borne by the Division for Astronomical Sciences from its existing budget. This cannot work in this context. The panel was briefed by NSF on this issue and told that changes are being considered. Nevertheless, this remains a major concern of the panel.

K.3.9 Conditions for an NSF Investment

The panel believes that maximum flexibility on the part of NSF, in the terms of the final arrangements it negotiates with the U.S. ELTP, will lead to the best outcome. However, in order for NSF to agree to provide the requested funds to the U.S. ELTP, the panel believes that there are critical conditions to be met. These are of two kinds. The first are programmatic in nature, and the panel assumes that they would be met through the due diligence of NSF. The second concerns the critical role played by the NOIRLab on behalf of the U.S. community.

K.3.9.1 Programmatic Conditions

The following programmatic considerations would be preconditions for NSF investment in the U.S. ELTP: First, the partners in GMT and TMT present a full financial plan that has all the resources to complete construction, including adequate contingency. Second, NSF caps its investment at the requested amounts

of $800 million (RY) for each project. Third, the share of observing time secured by NSF is in direct proportion to its investment in fixed-year dollars (and hence, larger than the notional 25 percent if $1.6 billion is allocated). Last, the current U.S. ELT partners and NSF need a firm and realistic plan for life cycle costs for at least the first decade (covering both operations and new instruments). This plan would be rigorously reviewed by NSF to determine whether the proposed budgets represent the best trade-off between cost and scientific return. The NSF portion of the operations budget would explicitly include a program of competitively selected grants to enable research using data obtained as part of the Key Science Programs (Section K.3.10) undertaken with the ELTs.

If it is not possible for all parties to commit to the full financial requirements, NSF could then examine whether it is possible to achieve a full financial commitment for one telescope, and decide whether to proceed with that plan to realize a U.S.-ELTP involving NOIRLab.

K.3.9.2 The Role of NOIRLab

The second set of conditions deals with the role that NOIRLab would play in making the U.S. community a full scientific partner in the U.S. ELT program and ensuring that the return on the federal investment to this community is maximized. The U.S. federal government, through the NSF investment, will be far and away the largest single stakeholder in each of these telescopes. Thus, a strong U.S. leadership position would be enabled by empowering NOIRLab to act on behalf of the full U.S. community.

The federal investment in the U.S. ELTP needs to accomplish more than simply providing much of the missing financial resources needed to construct and operate these observatories. This investment can be leveraged as part of the negotiations to create the right framework: the NSF resources can be used in part to provide the “glue” to form a true partnership. NSF can ensure this by empowering NOIRLab to play a central role in the partnership. More specifically, the panel envisages that NOIRLab would be charged to:

1. Provide access to a bi-hemispheric U.S. ELT system.
2. Engage and represent the whole U.S. community in GMT and TMT governance and scientific planning.
3. Provide user support, broaden participation in GMT and TMT science, and foster research inclusivity.
4. Assist the community in developing Key Science Programs (see Section K.3.10 below), and solicit the active involvement of the entire GMT/TMT U.S. and international partnership in these, including contribution of their observing time.
5. Develop the capability for queue scheduling applied to all GMT/TMT observations to optimize the way in which data are taken and maximize their immediate and legacy scientific value.
6. Facilitate plans for new instruments (including a process of open competition for building these instruments by members of the U.S. community when federal funds are involved).
7. Ensure that well-calibrated and well-characterized data and resulting data products are in an easily used archive containing all TMT/GMT data.

K.3.10 The Key Science Programs and the Need for U.S. ELT Science Centers

The panel believes that the Key Science Programs are a critical component of the U.S. ELTP. These are community-driven ELT programs that are being facilitated with the help of the NOIRLab. They are based on the principle of creating a powerful scientific legacy through systematic investment in large-scale, transformative research.

These are projects on scales difficult to realize within shares of any of the current GMT/TMT partners and are envisaged as utilizing at least half of the U.S. ELTP observing time. As emphasized above, the panel

believes the true value of the programs will be realized only if all of the ELT partners participate in them, with not only scientific contributions but also observing time.

These are planned to have broad, inclusive scientist participation via open collaboration models, harness resources of a diverse research community, spread scientific benefits widely through the community, and produce data products with high archival reuse value.

While the panel finds these to be an exciting concept, it worries that there is no current funding mechanism in the United States that can support them, nor are they part of the current U.S. ELTP operating budget as presented to the panel. Nonetheless, they are critical to supporting the development of the most effective science program from the U.S. ELTP. Funds for Science Centers would enable collaborations of people and institutions to carry out leadership science and create data, data products, and knowledge infrastructure that would enable multiple generations of scientific usage from large coherent data sets. The creation of a funding mechanism to create Science Centers may potentially be a way for NSF to partner with private foundations, a partnership of increasing interest at NSF.⁸

The national benefit of such Science Centers associated with the U.S. ELTP Key Programs would be manyfold. They would provide the data and knowledge infrastructure necessary to carry out these large, long-term, multi-partner programs. They would produce myriad scientific discoveries derived from these highly leveraged, large coherent data sets, drive the hardware development and produce software tools for enabling science, and create a team of people capable of exciting and effective public engagement and stewardship of the major funding investments made within the U.S. ELT partnership. The panel believes that without these Science Centers, the revolutionary potential of a U.S. ELTP will not be fully realized.

K.3.11 Summary

The coming generation of ELTs will be scientifically essential for decades to come. They will be necessary to address the majority of the questions posed, and discovery areas identified, by the Science Frontier Panels. The European Southern Observatory is well-along in the construction of the E-ELT with an aperture of 39 m. It follows then that without a U.S. response, the United States will be ceding leadership in astronomy (and not just ground-based OIR astronomy) for at least a generation. It is also important to emphasize that—historically—large ground-based OIR telescopes remain highly productive for 50 to 70 years. The cost of the U.S. ELTP is high, but it can be “amortized” over many decades.

There are currently two ELTs with significant U.S. private and state participation, the Giant Magellan Telescope (GMT) and the Thirty Meter Telescope (TMT). It seems clear that neither project can be successfully completed without an unprecedented level of federal support from NSF. GMT and TMT, in concert with NSF’s NOIRLab, have therefore proposed to Astro2020 to create an integrated U.S. ELT Program that will provide access to both GMT and TMT in exchange for federal investment of $800 million in each of these two telescopes and also in the necessary NOIRLab infrastructure.

In the panel’s view, in order to be fully competitive with the E-ELT, to ensure full synergy with the extensive system of U.S. bi-hemispheric ground-based astronomical facilities, and to strengthen the case for this level of financial support (which will take the enthusiastic support of Congress), NSF will need to work with all three U.S. ELTP elements to create the proposed system of two telescopes and implement a bold and visionary new model that forms a true partnership with U.S. community engagement (enabled by the NOIRLab). NSF will be by far the largest single partner in this program and can therefore play an appropriate leadership role.

___________________

⁸ See, for example, Mathematical Sciences Research Institute, “Partnerships: A Workshop on Collaborations Between the NSF/MPS and Private Foundations,” https://www.msri.org/workshops/785.

The path ahead is very difficult. There are serious risks associated with both ELTs, and several ways each could fail. The panel believes the engagement of NSF with both projects helps mitigate the risk of no U.S. ELTP.

For NSF to proceed with this plan, it would need to ensure that each of the other GMT and TMT current and future partners commit the financial and technical resources to fully complete construction and provide the funds needed for operations and a line of new instrumentation. The panel believes that the two-telescope U.S. ELTP is needed to maintain U.S. leadership, and that it would be disastrous if no U.S. ELTP is realized.

K.4 A BALANCED PROGRAM FOR THE NEXT DECADE AND BEYOND

K.4.1 Introduction

As exciting and extraordinarily important as the U.S. ELTP would be, the scientific payoff to the community is over a decade away. Moreover, while the multi-billion-dollar U.S. ELTP price tag may seem daunting, it is crucial to recognize the comparable investment made in existing OIR ground-based observatories, starting with the two Keck telescopes in the 1990s. For telescopes with U.S. stakeholders, this amount is roughly $2.4 billion in 2020 dollars (just in initial construction costs). The DKIST (which has just seen first light), and the Vera Rubin Observatory (operations starting in 2023) together account for about $1.0 billion of this total.

At great cost, we have built a powerful and flexible system of ground-based OIR facilities. How do we exploit this investment during the next decade and beyond? In this section of report, the panel describes what it believes to be the most essential components of this system. The primary goal is to highlight how relatively modest investment either in, or in support of, the existing (or soon-to-exist) ground-based OIR telescopes will pay major dividends during the decade to come. In some cases, the panel also describes possible new telescopes or telescope arrays that could either plausibly be fit into the NSF Mid-Scale Research-2 (MSRI-2) program or funded by NASA. The panel does not consider any of these facilities as rising to the level requiring the MREFC.

K.4.2 Adaptive Optics/High-Contrast Imaging

Adaptive optics (AO) is one of the transformational technologies that have driven breakthroughs in many areas of astronomy and astrophysics in the past two decades, by enabling ground-based telescopes to image at their diffraction-limit. AO development has led, notably, to the discovery of the supermassive black hole at the galactic center, the first images and spectra of exoplanets, and direct evidence for the existence of dark sub-halos as predicted by Cold Dark Matter models. AO is a key component for the success of DKIST.

Many current and new instruments at existing OIR facilities are equipped with AO capabilities. AO is intrinsically embedded in the design and operations of ELTs. Thus, the central role of AO instrumentation and the importance of further development are rapidly growing, with novel concepts pushing toward wider areas, higher performance, and extended wavelength coverage.

• Wide-field AO (ground layer adaptive optics [GLAO], multi-conjugate adaptive optics [MCAO], multi-object adaptive optics [MOAO], laser tomography adaptive optics [LTAO]) delivers uniform wavefront correction over large areas, achieved by sensing the atmospheric turbulence profile with multiple laser beams assisted with natural guide stars, and serving a very wide range of areas from galactic to

• extragalactic science. Examples include censuses of stellar populations in the Milky Way and Local Group, where AO significantly reduces the impact of crowding, and surveys of resolved morphologies and kinematics of distant galaxies whose apparent sizes are of order of the seeing disk and smaller. Wide-field AO has matured in the 2010s, with first systems now in science operations (e.g., Gemini Multi-Conjugate Adaptive Optics System [GeMS] on Gemini-South, Advanced Rayleigh Guided Ground Layer Adaptive Optics System [ARGOS] at the LBT).
• Extreme AO (ExAO) lies at the opposite end of the AO parameter space and aims to provide exceptionally high performance (Strehl ratios in excess of 90 percent) over a narrow field. Technical implementation of the ExAO concept is highly challenging. Coupled with high-contrast observations (>10⁵), the primary goal is the high-priority exoplanet science case. The challenges and scientific rewards have motivated massive efforts worldwide. Pathfinders are in operation (e.g., GPI on Gemini South, Spectro-Polarimetric High-Contrast Exoplanet Research Instrument [SPHERE] at the Very Large Telescope [VLT]), but significant development, notably in deformable mirror technology, speed of AO systems, and wavefront control, is needed in the 2020s to achieve the scientifically driven technical requirements (contrast of 10⁶ or better).
• Visible AO has high potential scientific return by opening up an entire wavelength regime to high angular resolution studies. The goal is to exploit the smaller diffraction limit (∝λ/D) of telescopes in the optical, yet both the coherence length and time decrease at shorter wavelengths (∝λ^6/5), requiring wavefront sensing at high spatial and temporal frequencies that are currently technologically challenging. This is an important developing area for the 2020s–2030s.

The panel received and reviewed RFI 2’s for Gemini North Adaptive Optics (GNAO) (at Gemini-North) and LIGER (at Keck). GNAO development was initiated as part of the NSF-funded program Gemini in the Era of Multi-Messenger Astronomy (GEMMA) for a queue-operated MCAO facility that will deliver diffraction-limited imaging over a ~1.5 arcmin field at Gemini North. The ultimate goal of transforming Gemini-North into a full AO telescope will require deployment of an Adaptive Secondary Mirror (ASM) and integration of multi-laser guide star wavefront sensors (WFSs) into the Acquisition and Guiding (A&G) unit. LIGER is an innovative next-generation AO-fed integral field spectrograph and imaging camera, fully funded through the Final Design Phase (until December 2020) and will take advantage of the NSF-funded Keck All-Precision AO (KAPA) system. With investments at the level of ~$15 million each for each project (equipment and labor), over timelines of 5–10 years to completion (2025 for LIGER, 2029 completion of ASM + A&G camera for GNAO), these programs are NSF MSRI-2 candidates.

The review of the programs (and science) white papers by the U.S. community, along with the RFI 2 documents, led the panel to conclude that the case for continued support of development over the next decade is strong:

1. AO and high-contrast imaging are key enabling technologies for high science priorities identified by the Exoplanets, Astrobiology, and the Solar System; Stars, the Sun, and Stellar Populations; and Galaxies science panels.
2. They play an essential role in boosting the scientific return and efficiency of existing facilities (e.g., Gemini, Keck, Magellan, DKIST) with modest-scale investment throughout the 2020s.
3. NSF mid-scale program opportunities have been identified (e.g., GNAO, Extreme Adaptive Optics on the Giant Magellan Telescope [GmagAO-X], LIGER) to nurture existing 6–10 m class telescopes.
4. Such investments in AO systems development are a key risk mitigation strategy for ELTs, whose full resolution and sensitivity potential can only be realized with AO; this is recognized as the most important technical risk for both GMT and TMT.

K.4.3 Extreme-Precision Radial Velocities

Mass is the most fundamental property of a planet, and knowledge of a planet’s mass is essential for two reasons. First, a measurement of a planet’s mass is necessary to constrain its bulk density and, in turn, infer something of its composition and ultimately its formation. Second, a planet’s mass is a key input for interpreting spectroscopic features in its atmosphere, since the atmospheric scale height depends on the planetary surface gravity, in addition to the mean molecular weight and temperature. There is keen interest in studying terrestrial planets, including those orbiting at Earth-like insolations around Sun-like stars, motivating a push for mass measurements to the sensitivity required for such worlds.

The National Academies Exoplanet Science Strategy (ESS) had two key findings related to precise radial velocities, both echoed by the Panel on Exoplanets, Astrobiology, and the Solar System:⁹

ESS Finding: The radial velocity method will continue to provide essential mass, orbit, and census information to support both transiting and directly imaged exoplanet science for the foreseeable future.

ESS Finding: Radial velocity measurements are currently limited by variations in the stellar photosphere, instrumental stability and calibration, and spectral contamination from telluric lines. Progress will require new instruments installed on large telescopes, substantial allocations of observing time, advanced statistical methods for data analysis informed by theoretical modeling, and collaboration between observers, instrument builders, stellar astrophysicists, heliophysicists, and statisticians.

The panel advocates that together NASA and NSF address the grand challenge of achieving the precision required to measure the masses of terrestrial planets orbiting Sun-like stars, which implies a single measurement precision of 10 cm/s and control of systematics at the level of 1 cm/s. While such measurements will be done from the ground, they are inextricably linked to the scientific success of numerous current and proposed missions—namely, the legacy Kepler/K2 data set, the ongoing TESS mission, and a future direct imaging mission. For a direct imaging mission, such precise radial velocities will identify terrestrial exoplanets orbiting nearby stars, determine when they are situated at quadrature, and remove degeneracies from the interpretation of atmospheric spectra features. NASA has tackled formidable technology challenges in the past in pursuit of its scientific goals, and the panel advocates here that this same coordinated effort be brought to the EPRV challenge. The panel concurs with the recommendation from the Exoplanet Science Strategy¹⁰—namely that “NASA and NSF should establish a strategic initiative in extremely precise radial velocities (EPRV) to develop methods and facilities for measuring the masses of temperate terrestrial planets orbiting Sun-like stars.” Following this recommendation, NASA and NSF jointly commissioned a community-based Extreme Precision Radial Velocity Working Group, which recently presented the blueprint for a strategic EPRV initiative.¹¹

K.4.4 Massively Multiplexed OIR Spectrographs

There is very strong support for massively multiplexed spectroscopy across many sectors of the science community. The survey received five white papers related to the topic (SDSS-V, MegaMapper, SpecTel, DESI, and FOBOS), and this capability was highlighted in multiple science questions by the Compact Objects,

___________________

⁹ National Academies of Sciences, Engineering, and Medicine, 2018, Exoplanet Science Strategy, The National Academies Press, Washington, DC.

¹⁰ Ibid.

¹¹ NASA Exoplanet Program, “NASA-NSF Extreme Precision Radial Velocity (EPRV) Initiative,” https://exoplanets.nasa.gov/exep/NNExplore/EPRV.

Cosmology, Galaxies, Interstellar Medium (ISM), and Stars panels. The survey also received an RFI 1 from Maunakea Spectroscopic Explorer (MSE) and RFI 2’s from MegaMapper and FOBOS. Furthermore, numerous past reports have heavily emphasized this capability. The support stems from the strong feeling that large-format spectroscopy is required to maximize return on investment for the large-area imagers coming online in the 2020s.

Many science panels emphasized this capability. The Panel on Cosmology emphasized large-volume, large-scale structure studies and the need for photometric redshift calibration. The Panel on Stars, the Sun, and Stellar Populations discussed “industrial-scale spectroscopy” in the context of galactic archeology studies of dwarf galaxies, stellar streams, and the galactic halo, as well as spectroscopically exploring stars with unprecedented photometric data (e.g., from Gaia). The Panel on Galaxies considered the importance of galactic archeology as well as statistical galaxy evolution studies and IGM tomography, and the Panel on Compact Objects and Energetic Phenomena discussed transient follow-up and dynamical searches for compact objects.

Massively multiplexed spectroscopy is required to fully realize the primary science goals of the Rubin Observatory, the Roman Space Telescope, Gaia, and other surveys. Happily, investment in existing 2–10 m capabilities would achieve a large fraction of the main science goals through the continuation of projects like SDSS V, DESI, U.S. access to the Prime Focus Spectrograph on Subaru beyond their nominal missions, or investment in a next-generation instrument for an existing telescope, like FOBOS on Keck.

A dedicated facility would, of course, provide advantages over relying solely on existing infrastructure. Most glaring is the lack of high spectral resolution (R ~20,000) multi-object spectrographs. Both the Panel on Galaxies and the Panel on Stars, the Sun, and Spectral Populations emphasized the power of this modality for galactic archeology work. MSE and SpecTel presented plans to the panel for such a mode. MegaMapper offers the widest field of view, as required for any cosmological application. Both SpecTel and MSE have larger apertures, which would be powerful for galaxy evolution and cosmological applications. Last, a Southern telescope (SpecTel or MegaMapper) will have obvious benefits for follow-up in the era of the Rubin Observatory.

In all cases, the United States could envision playing a significant role in these projects through a MSRI-2-level investment, which could provide up to about 20 percent of the cost of a project like MSE, SpecTel, or up to about 50 percent of MegaMapper, perhaps split with DOE. However, the time scales are such that the panel strongly encourages investment in some combination of next-generation SDSS V, DESI, and Subaru Prime Focus Spectrograph (PFS) through the 2020s, which could then be followed by investment in a larger dedicated facility. The panel does not believe an MREFC-level of funding is warranted in this decade.

K.4.5 NOIRLab and the OIR System

The U.S. ground-based OIR research community supports and benefits from a remarkable array of general purpose and specialized facilities of all aperture sizes in the 2 m to 10 m range. Financial support comes from a heterogeneous collection of federal, non-federal, private, and international sources. The scientific passion and technical ingenuity that led to these capabilities enabled U.S. global leadership for more than 100 years. Since the late 1950s, this so-called OIR System has benefited greatly from NSF financial support for major observatories, telescopes, instruments, and data systems.

The formation of NOIRLab is a welcome, long-awaited outcome recommended in whole or part by several national reports, including the 2000 and 2010 decadal survey reports. The newly constituted center provides (1) critical end-to-end (photons-to-science) infrastructure for wide-field optical imaging (Blanco/DECam, Rubin Observatory) and optical spectroscopy (Mayall/DESI); (2) time-domain research (public

brokers, Southern Astrophysical Research Telescope (SOAR), Gemini North, Gemini South, data pipelines, etc.); (3) access, visualization, and analysis of massive object catalogues with millions to billions of objects; and (4) wide-field, high-spatial resolution imaging at the 8 m scale (Gemini North and South). On behalf of NSF, NOIRLab operates and maintains mountain-based research parks in Arizona and Chile for facilities built, deployed, and operated by university-based science collaborations. It also facilitates the long-term development and retention of human capital. Indeed, many of the key managerial and technical leaders for the DKIST, Rubin Observatory, GMT, and TMT projects were initially trained at one of the NOIRLab constituent parts. Last, NOIRLab is the focus for strategic community discussion and planning at NSF’s behest (see Table K.3).

Beyond NOIRLab, there are several facilities that are de facto national assets, with the W.M. Keck Observatory being an example. Sustaining a university-based instrumentation and technology development program in support of the OIR System is critically important to retaining and supporting a technical workforce, as well as training students for careers in all branches of science, technology, and engineering. Equally important is continued support to the community for development and maintenance of data analysis and management infrastructure, from Astropy to cloud-based systems needed to work with petascale multi-dimensional object catalogs to science-case driven software systems needed for time-domain, multi-messenger research.

K.4.6 OIR Interferometers

The renaissance in stellar astrophysics in the 2010s was driven by new data and capabilities that enabled enormous advances in precision measurements of fundamental physical properties of stars: masses, radii, and luminosities. The first science priority of the Panel on Stars, the Sun, and Stellar Populations is measuring fundamental stellar properties across the Hertzsprung–Russell (H-R) diagram, with an emphasis on precision stellar masses and radii. Ground-based OIR interferometry is the key enabling technology, with increased polarimetric sensitivity providing significant opportunity for new discovery. The 2010s saw significant advances in OIR interferometry, with CHARA and NPOI becoming fully operational and scientifically productive. During this same time, there was steady progress in construction of the ambitious Magdalena Ridge Observatory Interferometer (MROI) configurable imaging interferometer. The OIR

TABLE K.3 Critical Leadership Roles That NOIRLab Will Play in the Decade Ahead

interferometry community is growing, and has presented a unified vision of how to advance science with these three complementary facilities in the coming decade.

The panel received and analyzed an RFI 1 from CHARA and RFI 2’s from each of Navy Precision Optical Interferometer (NPOI) and MROI. These provided further information on the technical and performance goals for the projects, and presented detailed budgets and schedules to achieve those goals during the 2020s. These formed an important component of the panel’s discussions to assess the scale of the required investments.

The panel concludes that continued U.S. scientific and technical leadership in ground-based optical interferometry requires strategic investment during the 2020s. Based on the RFI responses, modest ($1 million to $2 million per year) investments would promote continued growth of the interferometry user community and fund new technology development efforts. Mid-scale investments in CHARA and NPOI at the high end of MSRI-2 would enable implementation of larger telescopes, longer baselines, and advanced beam combination technologies needed to deliver the greater angular resolution and photometric sensitivity required to achieve the science goals of the Panel on Stars, the Sun, and Stellar Populations. A final phase of the envisioned CHARA upgrade (replacement of all existing 1 m inner-array telescopes with new AO-equipped 2 m telescopes) would require an additional mid-scale investment during the 2030s. Achieving the potential of MROI would require an MREFC-level investment to bring the full 10-telescope array into science operation by the end of the 2020s.

The full triad of U.S. OIR interferometers is deeply complementary: no one system can accomplish the science of all three by themselves. However, NSF is unlikely to have sufficient funds to support all three. Thus, the U.S. interferometry community and funding agencies would be best served by formulating a plan to realize this goal collectively instead of through internecine competition. This will strengthen U.S. leadership and make it an important component of a balanced U.S. ground-based OIR portfolio.

K.4.7 Solar Physics

The Sun is the only star observable with high spatial resolution and the only star for which global synoptic observations are made continuously. First light of the 4 m DKIST telescope in early 2020 promises significant progress in the coming decade for understanding detailed physical processes in the photosphere and the low corona related to the causes of flux emergence, the dynamo that drives stellar activity cycles, the mechanisms of coronal heating and solar wind acceleration, the fundamental process of magnetic reconnection, including the triggers for sudden release of stored magnetic energy in the star’s atmosphere, and the effects of stellar activity on the habitability of exoplanets around stars more or less like the Sun. DKIST first-light instruments have breakthrough capabilities. Plans for the next generation can begin after initial observations indicate the most important capabilities for future instrumentation.

To fully address the survey’s science goals for solar and stellar astrophysics, very-high-resolution observations in the restricted field of view of DKIST would need to be supplemented by global synoptic measurements. Two critical measurement gaps exist that require investment in the coming decade: measurements of the magnetic field in the global corona and greatly improved synoptic observations of the solar photosphere. Such observations of the Sun are also critical to the national priority to better understand and predict space weather and its effects at Earth and elsewhere.

Present coronal observations are mostly limited to intensity/density. Magnetic measurements are required to develop better physics of the corona and for understanding energy storage and release related to explosive events. The Coronal Solar Magnetism Observatory (COSMO), a proposed upper mid-scale project, will measure the global corona using a 1.5 m refractive coronagraph. The project passed a preliminary design review in 2018 and can provide synoptic observations of the coronal magnetic field both above the limbs and for certain structures on the solar disk. The panel received and reviewed an RFI 1 from COSMO.

The current ground-based solar synoptic network has two aging components: the Global Oscillation Network Group (GONG) and Synoptic Optical Long-Term Investigations of the Sun (SOLIS). The six-site GONG network has limited capability owing to modest spatial and spectral resolution, and the more capable SOLIS instrument suite has only one site that is not currently operating. Space-based observations currently fill this gap, but have finite life expectancy and measure quantitatively the magnetic field at only one height in the atmosphere. Continuous measurements of velocity, magnetic field, and intensity at multiple heights in the photosphere and above are required. Currently a conceptual design, ngGONG would be a global network (six) of highly capable solar observatories, including coronagraphs at three sites.

Program Balance: The COSMO, ngGONG, and DKIST second-generation projects are expected to compete against other Division of Astronomical Sciences (AST)–wide projects in a more robust NSF midscale project line. DKIST was an MREFC project that addresses important science goals that could not be achieved in any other way. Funding DKIST operations is a burden for the facilities budget of NSF/AST, but funding for operations of major new facilities needs to be addressed more broadly at NSF. The existing suite of U.S. solar ground-based-OIR observatories includes the 1.6 m Goode Solar Telescope (GST, a DKIST precursor), the Mauna Loa Solar Observatory (MLSO, a COSMO precursor) and a few smaller facilities based at universities (see the National Academies 2020 Solar and Space Physics Midterm Assessment¹²). Continued investment in these facilities provides critical measurements enabling new science, continuity of unique long-term synoptic observations, and capabilities for instrument development. Improving programmatic balance, including adequate/increased support for science analysis was a major theme of the 2013 Solar and Space Physics decadal survey. Coordination with in situ and remote-sensing space-based instrumentation, including the Parker Solar Probe, Solar Orbiter, and Solar Dynamics Observatory, is essential to address survey goals.

K.4.8 Technology Development: Astrophotonics

Well-established technologies developed in the past decades for fiber-optics telecommunications and other commercial applications are now emerging as potentially revolutionary strategies for a new generation of astronomical instruments. With micro-electro-mechanical-systems (MEMS) already adopted, for example, by the James Webb Space Telescope (JWST), the new frontier is represented by astrophotonics, which uses fibers and optical waveguides built into solid-state devices to manipulate light collected by a telescope.

The current portfolio spans a wide range of devices, including multi-core fibers, photonics lanterns, photonics spectrographs, complex Bragg gratings, on-chip beam combiners, pupil remappers, and laser frequency combs. The GRAVITY instrument at the Very Large Telescope Interferometer (VLTI) and the Michigan Infrared Beam Combiner (MIRC) instrument at the Center for High Resolution Astronomy (CHARA) have successfully adopted on-chip beam combiners to combine the light collected by their 4 × 8 m and 6 × 1 m telescopes, respectively. The possibility of obtaining extremely high-precision radial velocities, of the order of a 10 cm/s or better, as well as direct imaging of exoplanets—two of the main science cases for the U.S. ELT system—may largely rely on the maturity of single-mode fibers and on-chip nulling interferometers.

Strengthening the coordination between the most active astrophotonics research groups in the United States would optimize resources and facilitate the passage from laboratory research to industrial partnership. This could be done through the creation of a distributed, multi-disciplinary Institute of Astrophotonics to coordinate the teams working in this field. The more coordinated approach adopted by Europe (Germany

___________________

¹² National Academies of Sciences, Engineering, and Medicine, 2020, Progress Toward Implementation of the 2013 Decadal Survey for Solar and Space Physics: A Midterm Assessment, The National Academies Press, Washington, DC.

in particular) and Australia has led to success and leadership in this field. A few tens-of-millions of dollars of funding over the next decade would be needed to significantly advance this technology and reestablish U.S. leadership in astrophotonics. Cost-savings for translational programs in astronomical applications can derive from the availability of past fundamental research, industrial prototypes, and available technological infrastructure for hardware production.

The versatility, compactness, and light weight of photonics devices make them ideal for space science applications. NASA, therefore, may want to invest and support these developments. Increasing the coordination between NSF and NASA, as well as other stakeholders like DOE, the National Institute of Standards and Technology (NIST), and so on would create a coherent and synergistic program that optimizes resources and avoids duplications, taking into account the different strategic goals of the various agencies.

K.4.9 Time-Domain Astronomy and Multi-Messenger Astrophysics

In the past decade, we witnessed the birth of multi-messenger astrophysics (MMA), and the important role that ground-based OIR telescopes played in identifying and characterizing the electromagnetic counterpart to gravitational waves from a binary neutron star merger. The healthy ecosystem of optical telescopes, including National Optical Astronomy Observatory (NOAO)–access facilities (2 m Las Cumbres Observatory Global Telescope (LCOGT)/Faulkes, 4 m Blanco, 4.1 m SOAR, and 8 m Gemini-S) enabled a prompt localization and detailed characterization of the optical/infrared kilonova light curve and its spectrum, showing the telltale signatures of the products of r-process nucleosynthesis and their subsequent radioactive decay.

Heading into the next decade, we expect a rich landscape of discoveries in time-domain astronomy (TDA). These range from wide-field surveys such as the NSF/DOE Rubin Observatory in the optical and new wide-field capabilities in the radio, to gravitational-wave observatory networks like Advanced Laser Interferometer Gravitational-Wave Observatory (LIGO) and Kamioka Gravitational Wave Detector (KAGRA), and ground-based all-sky very high-energy neutrino and gamma-ray detectors like IceCube and High-Altitude Water Cherenkov Observatory (HAWC), as well as space-based gamma-ray burst detectors. Opening up these new windows of discovery will place heavy demands on the ground-based OIR ecosystem, for localization, classification, and the characterization of transients.

In many cases, a tiered approach to follow-up OIR observations of MMA and TDA discoveries will be required, engaging the full range of the OIR telescope ecosystem: from wide-field and/or rapid slewing 1–4 m aperture telescopes for candidate discovery and filtering, to 6.5–10 m aperture telescopes for spectroscopic classification, to an extension of Rubin Observatory LSST operations for deep, wide-field target-of-opportunity gravitational-wave counterpart searches, and ultimately to ELTs for detailed characterization from spectroscopy and late-time light curve evolution. Similarly, the new landscape of TDA surveys, most notably Rubin Observatory LSST, will yield a population of persistent, variable sources that will demand flexible telescope scheduling for time-sensitive observations.

One of the biggest challenges in this exciting era of MMA and TDA will be the efficient, selective, and prompt allocation of follow-up observing resources. Specific needs will include (1) upgrades to enable robotic telescope operations (e.g., LCOGT); (2) coordinated, dynamic telescope scheduling software (e.g., Astronomical Event Observatory Network [AEON]); (3) real-time communication and data analysis infrastructure (e.g., TOMs); (4) automated data reduction and calibration (e.g., using Astropy); (5) machine-learning enabled classification (e.g., Antares); and (6) systems for rapid incorporation and reprioritization from initial follow-up observations taken (e.g., TreasureMap). With investment in these cyberinfrastructure tools and robotic telescope operation and scheduling capabilities, we can fully exploit the existing ecosystem of OIR telescopes to handle this new treasure trove of transients and multi-messenger discoveries

in the next decade. Furthermore, by enabling the participation of private telescopes in an NSF OIR-lab follow-up network, we can further expand U.S. OIR access and coordination for effective follow-up in the MMA and TDA era.

Last, just as spectrum management at radio frequencies has been a key NSF mission for many years, the time may have arrived for NSF to take a more active role in managing the use of the optical window. For decades, all major OIR observatories have faced challenges from background light generated by ground-based lights of various kinds. In the past 12 months, OIR astronomy is suddenly facing an almost existential challenge from mega-constellations of satellites in low Earth orbit (LEO). Federal action and support are needed immediately, or an entire scientific field may be crippled in an unrecoverable way.

K.5 CONCLUSIONS

The exciting cases presented by the science panels show that even in this era of panchromatic/multimessenger astrophysics, observations with ground-based OIR telescopes will continue to be essential for addressing many of the most important scientific questions we will face in the decades to come. These telescopes are not only invaluable in their own right, they are also essential in fully unlocking the potential for discovery and understanding provided by the other windows on the universe.

The U.S. community will need to successfully deal with some extraordinary challenges if it is to maintain a position of leadership in ground-based OIR astronomy. In fact, ground-based OIR astronomy is so intimately and intricately woven into the fabric of the discipline that researchers face challenges in maintaining leadership in astronomy as a whole. These challenges are twofold: to create a U.S. ELT program that is fully competitive and leverages the existing bi-hemispheric investment in astronomical facilities, while at the same time providing the resources needed to exploit the powerful suite of existing facilities in the current decade and beyond.

Meeting these challenges requires a sea-change in the way the U.S. ground-based OIR community and its federal/state/private funding sources work together. Without a vigorous partnership between the various components of this system, the United States will not be able to remain a leader. We need to learn from the past and face the future with boldness and vision.