National Academies Press: OpenBook

Pathways to Discovery in Astronomy and Astrophysics for the 2020s (2021)

Chapter: Appendix H: Report of the Panel on an Enabling Foundation for Research

« Previous: Appendix G: Report of the Panel on Stars, the Sun, and Stellar Populations
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 342
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 343
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 344
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 345
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 346
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 347
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 348
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 349
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 350
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 351
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 352
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 353
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 354
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 355
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 356
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 357
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 358
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 359
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 360
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 361
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 362
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 363
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 364
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 365
Suggested Citation:"Appendix H: Report of the Panel on an Enabling Foundation for Research." National Academies of Sciences, Engineering, and Medicine. 2021. Pathways to Discovery in Astronomy and Astrophysics for the 2020s. Washington, DC: The National Academies Press. doi: 10.17226/26141.
×
Page 366

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

H Report of the Panel on an Enabling Foundation for Research H.1 INTRODUCTION The Astro2020 steering committee charged this panel with the following tasks: (1) summarize the current state of resources and support, (2) identify major challenges, and (3) make suggestions to the Astro2020 committee on the topics of computation, simulation, data collection, and data handling; funding models and programs; laboratory astrophysics; and general technology development programs.1 To address its charge, the panel relied on the many valuable white papers submitted by the scientific community, presentations at its three panel meetings, previous National Academies of Sciences, Engineering, and Medicine studies, interactions with other science and program panels, and member expertise. The sections on Explorers and mid-scale projects were based on work of interpanel task forces that drew from the other relevant prioritization panels. H.2 INCREASING THE INVESTMENT IN THE ENABLING FOUNDATION People are the enabling foundation of scientific advancement. The key outcome from many of the programs that are the subject of this report, ranging from the theory programs to the suborbital program, is the development, training, and support of scientists. By investing in programs that enable people with the broadest possible range of backgrounds to contribute to scientific advancements, reducing barriers to entry, and providing access to state-of-the-art tools, training, and facilities, the profession’s scientific productivity will be maximized. It is people who are the source of U.S. scientific and technical prowess. In the long history of the decadal surveys, this is the first panel to be explicitly charged with focusing on the “enabling foundation.” However, previous decadal surveys have discussed many of the issues raised in this report. Astro2010 and the subsequent midterm assessment emphasized the important opportunities enabled by the Explorer program and mid-scale programs at the National Science Foundation (NSF). Astro2010 stressed the importance of investments in theory, technology development, and laboratory astrophysics. However, many of the augmentations in funding for these programs recommended in Astro2010 were not realized. This lack of investment in the enabling foundation limits the profession’s ability to reap the benefits of its investments in telescopes and delays the development of key technologies. Over the past decade, there has been significant growth in investment in instrumentation and mission capabilities. While the NASA astrophysics budget has grown by 40 percent from fiscal year (FY) 2010 to FY 2020, the investment in people through the NASA research portfolio comprising the Astrophysics Research and Analysis Program (APRA), Astrophysics Theory Program (ATP), Astrophysics Data Analysis Program (ADAP), and Exoplanet Research Program (XRP) has grown only 17 percent, slower than the consumer price index and much slower than the growth of costs at universities and research institutes. With the number of proposals nearly doubling over the decade, the success rate and the inflation-adjusted funding levels of grants (i.e., the investment and support of people) have 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. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-1

declined. The success rate for NASA proposals during the period FY 2003 to FY 2010 averaged 28 percent. This rate declined to 22 percent from FY 2011 to FY 2019. Meanwhile, the absolute funding per grant grew modestly from $496,000 to $570,000 over this period, somewhat slower than the consumer price index and significantly slower than the cost of graduate school tuitions. The combination of current underfunding and significant opportunities in the coming decade motivates the panel’s suggestions to refocus investments on enabling foundations. The panel suggests that a 20 percent increase in the funding (above inflation) for these programs would restore the success rate to historical levels in support of researchers, students, and postdoctorates and match the overall growth in the astrophysics program. At NSF, there has also been a significant investment in instrumentation through both the Division of Astronomical Sciences and the Major Research Equipment and Facilities Construction (MREFC) program. The Vera C. Rubin Observatory, Atacama Large Millimeter/Submillimeter Array (ALMA), and Daniel K. Inouye Solar Telescope (DKIST) will produce a flood of data that requires increased investment in the enabling foundation. Between 2010 and 2017, the average Astronomy and Astrophysics Research Grants (AAG) success rate was 18.3 percent, significantly lower than the physics program and lower than AAG success rates in previous decades. NSF did not provide information on the success rates for the past several years for the field as a whole nor for different demographic groups or subfields. More recent information on success rates and funding levels would have better informed this report. The low success rate for proposals has resulted in members of the community operating under extreme stress. This has resulted in significant long-term consequences that inhibit the community’s ability to accomplish its goals and to retain talent. The obvious fact is that with proposal success rates so low, outstanding people and teams proposing to do outstanding science are not funded. The National Science Foundation’s Astronomy and Astrophysics Advisory Committee (AAAC) identified the threat of the falling success rates in its 2015 report and warned of the risk of a runaway effect with researchers constantly resubmitting as the success rates for proposals dropped.2 Another AAAC study3 suggests an optimal success rate of 30–35 percent, far above the current rate for the AAG program. Researchers at all stages of their careers are unable to take full advantage of the massive investments in large infrastructure projects, which puts U.S. researchers at a disadvantage compared with their peers in other countries. The barrier to entry is especially damaging to young researchers whose highly ranked proposals might get turned down multiple times during their critical first years. People from underrepresented groups are also disproportionately affected by these low success rates.4 Improvements can be made in the way grants are awarded. For example, in the NSF AAG program, a typical 3-year grant will fund a graduate student and, perhaps, 1 month of summer salary. With a 3-year cadence and an 18 percent success rate, it is essentially impossible to support a student through the completion of a Ph.D. The grants do not support any costs that are critical to lowering the barrier to entry for women and people from underrepresented groups, such as child care and moving expenses. The panel suggests that the funding agencies could encourage parental leave by requiring that fringe benefits have some minimum benefits, including leave. These stresses and exclusionary practices are limiting the current and future diversity, vibrancy, and productivity of the field. These new costs to institutions could be absorbed by adjustments in the overhead rate agreements. Other fields are, of course, concerned about their grants programs as well. A recent Report of the 2019 Committee of Visitors—Division of Physics5 for NSF detailed the state of the grants programs in a number of areas. It paints a picture of a healthy and vibrant community. Focusing on the grants program 2 AAAC, 2016, Report of the Astronomy and Astrophysics Advisory Committee, NSF, Arlington, VA, https://www.nsf.gov/mps/ast/aaac/reports/annual/AAAC_2015-16_Report.pdf. 3 P. Cushman, J.T. Hoeksema, C. Kouveliotou, J. Lowenthal, B. Peterson, K.G. Stassun, and T. von Hippel, 2015, “Impact of declining proposal success rates on scientific productivity,” https://arxiv.org/abs/1510.01647. 4 See Appendix N, “Report of the Panel on State of the Profession and Societal Impacts,” for detailed discussion of barriers to entry for underrepresented groups. 5 NSF, 2019, Report of the 2019 Committee of Visitors, NSF Committee of Visitors, Arlington, VA, https://www.nsf.gov/mps/advisory/covdocs/PHY_2019_COV_final_report.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-2

in the Division of Physics (PHY) offers an unsettlingly stark contrast to astronomy. A few examples include  Over the course of 4 years, 24 out of 38 proposals by new young investigators were funded in the Gravitational Physics program.  In Quantum Information Science (QIS) the Atomic, Molecular, and Optical (AMO)- Experimental/AMO-Theoretical/QIS programs have success rates of 47 percent, 35 percent, and 30 percent, respectively. The report went on to explain that proposal pressure was to blame for the QIS program being “lower than the others.”  Elementary Particle Physics “has success rates sometimes as low as 25 percent … and turns away excellent, fund-if-possible proposals.”  In Nuclear Physics, “The funding success rates are approximately 45 percent for Theory and 38 percent for Experiment.” While the report concedes that these numbers “may sound relatively high,” it goes on to argue the researchers are underfunded and strong proposals were declined. As astronomy moves to an era of large instruments, the comparison to physics is appropriate. Both have major investments in common infrastructure and both have large and active theoretical and experimental communities. While the average and median grant sizes in the Division of Astronomical Sciences (AST) and PHY are almost identical (~$600,000 and $350,000), there are almost twice as many awarded in PHY. Given the disparity in the success rates, it is relatively easy to argue that the astronomy grants program is underfunded in an absolute sense. NSF physics invests 68 percent of its budget in research and 35 percent of its budget in facilities. NSF astronomy invests 21 percent of its budget in research and 78 percent of its budget in facilities. These numbers do not include the large MREFC investments in the Rubin Observatory and DKIST. In summary, the astronomy grants program is underfunded both in an absolute sense, and relative to physics. While the NSF budgets for AST and PHY are almost identical, PHY partners with the Department of Energy (DOE), which provides significant support for some major facilities, allowing for more support of the research program. NSF AST supports most to all of the major facilities used by its researchers, resulting in significantly less resources available for the research program. Recognizing the role that AST plays in supporting facilities, the panel suggests that NSF consider a significant augmentation of the AST program in support of research. AST’s large and growing investment in facilities and declining funding for research grants makes it an outlier among NSF directorates. A recent NSB study6 showed that the fraction of funding for facilities remained flat over the past 20 years for the Division of Materials Research at 10 percent, Division of Physics at 20–30 percent, and the Directorate of Geosciences at 30 percent. AST devoted 60 percent of its budget to facilities in 2002–2016, and this fraction is projected to grow to 80 percent by 2023. During the past decade, the international community has constructed (or begun construction on) a series of major observatories that are poised to deliver potentially transformative observations: the JWST, ALMA, the Rubin Observatory, Euclid Telescope, DKIST, and the Nancy Grace Roman Space Telescope. These telescopes will begin producing exabytes of data. However, the data alone are not enough. A significant investment is needed in the people who will develop the theoretical framework, build the archives, develop the software, train the community on the new products and tools, create and implement the computational methods, make the vital laboratory measurements, and analyze the data to produce these transformative results. While support for operations and science analysis are part of the budget for large projects at NASA and DOE, these costs are not part of project budgets at NSF. 6 NSF, 2018, Study of Operations and Maintenance Costs for NSF Facilities, NSB-2018-17, NSF National Science Board, Arlington, VA, https://www.nsf.gov/pubs/2018/nsb201817/nsb201817.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-3

The panel suggests that any large investment in new facilities include resources for operations, a data pipeline, supporting theoretical work, and analysis from the inception of the project. This support will grow the research community using these facilities and enhance their scientific return. DKIST, ALMA, and the Rubin Observatory are very large investments. As noted in the National Academies mid-decadal review:7 NSF and the National Science Board should consider actions that would preserve the ability of the astronomical community to fully exploit the Foundation’s capital investments in ALMA, DKIST, LSST, and other facilities. Without such action, the community will be unable to do so because at current budget levels the anticipated facilities operations costs are not consistent with the program balance that ensures scientific productivity. This passage was quoted in a recent National Science Board (NSB) study.8 The panel suggests as guidance that the annual budget of the grants programs be augmented by at least 1 percent of the cost of the construction cost of a space-based project, roughly what is currently done for flagships such as JWST and the Roman Telescope, and 2 percent of the construction costs of a ground-based project. The larger percentage reflects the lower costs of construction on the ground. The panel also suggests that operations (including software support) be part of the MREFC budget for these projects. H.2.1 Investment in Archives and Joint Analyses Archiving centers will continue to provide a critical enabling infrastructure for collecting, curating, documenting, providing community training, and making the data sets accessible. These tasks will be even more essential in the coming decade. Archives that reliably catalogue events and objects in the sky have always formed a foundation for astronomical research. In the modern era, these archives are digitally curated in databases, and the past, currently ongoing, and future missions from space and the ground across all wavelengths will generate on the order of 500 petabytes by the end of 2030, several orders of magnitude more astronomical data than has been collected in human history. The simulations needed to interpret these data will generate comparable sized data sets. Although the data volume is dominated by a small number of major missions and surveys, the small- and mid-scale data sets are challenging enough in terms of scale and complexity to also require special attention to their archiving needs. In the past decade, the majority of scientific papers based on data from large missions and surveys are archival analyses. In the coming decade, these archival analyses will become even more important and their technical implementation more challenging. With the missions and surveys planned for the 2020s and beyond, enormous opportunity exists to greatly multiply the scientific return beyond their core goals, using them to address a broader set of both foreseen and unforeseen questions. Illustrative examples of these opportunities include the following:  Combining images of the sky at the level of the pixels of the maps produced for the major ground and space imaging programs, including but not limited to the Rubin Observatory, the Euclid satellite, the Roman Telescope, the eROSITA mission, and future cosmic microwave background (CMB) experiments, which would allow powerful new constraints based on gravitational lensing at both cosmological and galactic scales, photometric redshifts, galaxy evolution, and motions in the solar system and the Milky Way. This analysis will require 7 National Academies of Sciences, Engineering, and Medicine, 2016, New Worlds, New Horizons: A Midterm Assessment, The National Academies Press, Washington, DC, p. 8. 8 NSF, 2018, Study of Operations and Maintenance Costs for NSF Facilities, NSB-2018-17, NSF NSB, Arlington, VA, https://www.nsf.gov/pubs/2018/nsb201817/nsb201817.pdf, p. 22. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-4

investments beyond those of individual projects to enable this enhanced science return. High- quality cosmological simulations will also be needed to enable these joint analyses.  Combination of time-resolved and target-of-opportunity observations across facilities, large and small, and wavelengths. If there exists a close enough integration of access methods, formats, and meta-data to efficiently and properly analyze the joint data set, this would allow new windows on transient and variable phenomena.  Timely analyses of multi-messenger phenomena (across electromagnetic, gravitational wave, and particle detectors), which would allow tests of fundamental physics, the nature of black holes, and cosmology, and triggering of follow-up observations, if they can be performed rapidly and robustly. The panel outlines a framework for structuring the U.S. astronomical archiving system so that it addresses astronomy’s most compelling science goals. The existing federally funded archiving centers are essential. Operating from the Space Telescope Science Institute (STScI), NOIRLab, the Infrared Processing and Analysis Center (IPAC), the National Radio Astronomy Observatory (NRAO), the Smithsonian Astrophysical Observatory (SAO), the Gravitational Wave Open Science Center (GWOSC), and other locations, they focus on different types of data and have developed unique expertise. The combination of their data, meta-data, documentation holdings, the access tools they have developed, and most importantly, their scientific and technical staff is critical to maximize the scientific return from the profession’s investments in new telescopes. Nevertheless, the panel’s use cases illustrate that separating archives by wavelength, mission, and funding agency limits the science. Because they are funded by different agencies, including NASA, NSF, and DOE, with different policies and science goals, the coordination necessary to enable new joint capabilities is hampered. Technological developments deriving from the commercial hardware and software industry are also emerging that can enhance the profession’s approach to archives. Today, most astronomical data centers operate their own hardware; however, cloud services (both computing and storage) are an increasingly affordable and flexible way of handling astronomical data. A decade ago, the cutting-edge archives allowed complex server-side queries (e.g., in Structured Query Language [SQL]), which typically would be followed by a data download and local analysis by the end user. Today, science platforms exist or are being developed at most centers (most often relying on Jupyter notebook deployment) to provide more complete server-side analysis tools so that almost all analyses can be performed on the server, with limited if any data download to the end user. A key technological tool behind these science platforms is the “container,” an encapsulation of the full operating system of a system that can be duplicated and operated on any hardware. The proliferation of open source software and software development resources has expanded the suite of tools available to astronomers for building, accessing, and analyzing data from archives. These developments over the past decade have swept into astronomy from the wider world of software and data, and archiving centers are individually preparing to take advantage of these and future, unforeseen developments as they arise in the next decade. Nevertheless, the archive centers have limited ability to adopt these new technological developments in a coordinated and synergistic manner. Insufficient resources are focused toward very few funded efforts in building shared infrastructure among them, again partly owing to their separation across NASA, NSF, and DOE. A major challenge facing the U.S. astronomical archive system is the need to coordinate its federated archives to address the science needs of the 2020s. The development of standard protocols and tools to implement them over the past 15 years has been a necessary but not sufficient effort to enable this coordination. The panel suggests a more proactive and robust approach is needed to maximize the scientific return from the profession’s investments in instruments, telescopes, and satellites. To face this challenge, the panel envisions an Astronomical Data Archiving System (ADAS) to serve as an umbrella organization to coordinate the activities of the existing archive centers, with the goals of increasing their effectiveness in achieving their missions and opening up new opportunities that PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-5

would otherwise be impossible. Models in other fields such as earth science, in the international astronomical community, and in previous and ongoing U.S. virtual observatory efforts can provide lessons in addressing the growing needs and remaining challenges in astronomical archiving. Critical considerations in designing this new coordinated system would include the following:  Preserving the roles, responsibilities, expertise, and funding streams that define current centers and missions.  Enabling career paths for archive scientists that allow them to move within the ADAS “family.”  Providing the resources and mission to initiate and lead community-wide efforts focused on broadening participation through education, training, citizen science, and curriculum development in computational skills, software development, and data science.  Providing the resources and mission to pursue opportunities to develop common resources and shared expertise across centers.  Enabling science-driven efforts to perform analyses across boundaries of missions and centers.  Providing support for users to access the simulations needed to interpret the data;  Having the capability to bring new centers under the ADAS’s umbrella and to coordinate/collaborate with international partners.  Incorporating data from smaller projects that lack the support of a major center and provide support for data archiving/preservation for these projects.  Providing mechanisms for the U.S. astronomical community to contribute input to the planning and prioritization of ADAS activities.  Continuing to support journals and access to journals. The NASA Astrophysics Data System, which provides free bibliographic access to 13 million publications records, plays an important role in astronomical research, and will continue to be essential in the coming decade.  Developing common policies for data storage and transfer specifications. With funding outside that of the existing individual centers, ADAS could provide the resources to fulfill its mission to make all centers more interoperable and to enable cross-center scientific analysis and to coordinate complementary training programs. To enable discovery in the 2020s, the panel suggests that the ADAS, the archive centers, the funding agencies, the major NASA-funded missions, and NSF-funded and DOE-funded projects address numerous other challenges through funding of a number of data-related efforts:  Science-ready data products and APIs could be built into the funding of all new missions and projects, utilizing the available common infrastructure and protocols of the ADAS and/or the archive centers.  Support for software infrastructure efforts specific to astrophysics, both community-led and those led by archive centers.  Initiation of new programs and/or supporting community-led efforts in training and education, enabling contributions from a broader range of scientists.  Preservation of data from smaller or nonfederal projects.  Standard, machine-readable methods for preservation of meta-data and documentation.  Direct support for ensuring that important analyses preserve their replicability, through providing needed support for the software engineering effort necessary to do so. The archiving of astrophysical simulations of theoretical predictions deserves special mention here as an important challenge to which the archives can contribute. “Simulations” includes both PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-6

instrument simulation given the astronomical input (e.g., the image of the sky) and the astrophysics simulation of physical laws that might produce a specific image of the sky under some theoretical framework. Here, the focus is on simulations that are essential for statistical analyses of the observations. The panel suggests that archive centers continue to develop the software and tools to allow the instrument simulations of instrument behavior. As both simulations and observations grow in scale, the ability to co- locate observational data with astrophysical simulations will prove necessary, and the panel suggests that archives develop partnerships with the high-performance computing centers and simulation groups necessary to provide this service. The panel suggests that archives of simulations ensure that software for all curated simulations is versioned, traceable, and can be used to replicate the simulation if necessary. The new data sets from astronomical facilities at all scales in the 2020s will lead to numerous new discoveries and breakthroughs in understanding of astrophysics and physics. The breadth and depth of this new science, how broad and inclusive the community is that contributes to it, and the profession’s ability to take advantage of new and unexpected opportunities depends on a well-funded and well- coordinated archiving system designed for the coming decade and beyond. This suggestion is in line with the recommendations made in the NASA Science Mission Directorate’s Strategy for Data Management and Computing for Groundbreaking Science 2019–20249 document. Indeed, the panel envisions that this new system will address the Open Data/Open Software, High-End Computing, and Advanced Capabilities areas in addition to the items included in the Archives Modernization area. NASA’s strategy could be enhanced by a stronger emphasis on community education. A detailed study incorporating input from the astronomy community and from the existing centers will be needed to appropriately scope and define this system. The panel envisions that the system would require very roughly 50 full-time employees (FTEs), which would include astronomers, software engineers, and other staff in proportions that require a focused future study. This system would add ~$10 million a year in operating costs above the costs of the existing archives. The panel suggests that this be a supplement to existing programs. This preliminary estimate is based on considering the number of FTEs working on data management, processing, and distribution at the major current and planned facilities, which add up to hundreds of FTEs. To coordinate these systems in an effective manner requires a sufficiently large investment of central effort, which motivates the considerable investment envisioned here (although fractionally this investment is smaller than Earth Science Data and Information System (ESDIS), a similar system in Earth Sciences). It is essential that, irrespective of the goals laid out above, the ultimate scope of activity and the expectations of any ADAS-like system be tuned appropriately to its available resources. While some of the ADAS FTEs would be located at the existing larger centers, some may be best co-located with smaller projects. As discussed below, the panel suggests considering an approach where NASA serves as the lead agency for an interagency supported archiving program, and NSF and DOE are the lead agencies for providing access to high-performance computing resources to the broader national community. H.2.2 Software Software development is an essential part of almost all aspects of astronomy, and software developers, perhaps better called “software instrument builders,” are an essential part of the astronomy community. However, neither are sufficiently funded or supported by existing structures. The profession has entered an era in which the ultimate success and impact of major programs will be equally dependent on software and hardware development. As such, software development needs 9 NASA, 2019, Science Mission Directorate’s Strategy for Data Management and Computing for Groundbreaking Science 2019–2024, Washington, DC, https://science.nasa.gov/science-red/s3fs- public/atoms/files/SDMWG_Full%20Document_v3.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-7

to be included in budgeting, planning, and career development. In this panel report, software includes data reduction pipelines such as Astropy and analysis packages developed by large projects, large software projects such as Modules for Experiments in Stellar Astrophysics (MESA)10 or Enzo,11 and codes used by individuals or small groups to produce results in published papers. Software development has evolved significantly over the past several decades. Large teams of people often work together to write and develop software. For high-performance computing, these large teams are essential in part because of the increasing complexity of the problems being solved, and in part because of the increasing complexity of computer hardware. Today’s astronomical software must take advantage of heterogeneous computing systems currently available, from graphical processing units (GPUs), to multi-core processors running in collections of thousands of nodes, to standard general- purpose CPUs running on laptops. For both theoretical models and data analyses, the complexities of the systems modeled demands large codes developed by teams with a wide range of expertise. The practices for developing software as a team are quite different from developing software as an individual. The panel envisions that future astronomical training will include best practices for developing code as part of a larger team and for a diverse range of computing hardware. The Image Reduction and Analysis Facility (IRAF),12 developed and supported through federal funding for three decades through multiple avenues, demonstrated how a workhorse software system, freely available, could be immensely empowering to the community to make use of all types of data from many types of instruments and telescopes. However, the lack of federal support for the modernization of this stalwart resulted in a gaping hole in the software infrastructure available for data reduction and analysis. This hole has been partially filled by the mostly volunteer-run Astropy Project and the Astropy package. However, the Astropy Project may need to abandon its support and development of its community software, in the same way the IRAF project was eventually forced to do, because of the lack of sufficient and reliable funding. These examples are illustrative of how the normal grant funding structure works poorly in the context of building software infrastructure intended to undergird most modern astronomical software. Any effort to create infrastructure like that depends on continuous fundraising efforts with short time horizons, without any path to earning the longer term commitments that are necessary for longer term planning and stability. These infrastructure development efforts require reliable federal funding along with the software and the people maintaining it. By investing in software training for the entire astronomy community and in astronomical software developers, the profession can build the tools for transformational astronomy in the coming decade. This will require support for making code open source and the maintenance of large codes being developed by both individuals and by broad community efforts. These investments will improve reproducibility and reduce unnecessary duplication of codes. Astronomical software development is training a generation of people who are finding exciting opportunities outside astronomy. Without more funding opportunities and career tracks within astronomy, it is challenging to retain these software builders. The 2018 National Academies report Open Source Software Policy Options for NASA Earth and Space Sciences13 makes important observations: SMD [Science Mission Directorate] needs to foster a new culture of openness and encourage a social norm of sharing and collaboration, in part by incentivizing the development of OSS [Open Source Software] in the academic community through the use of targeted grants, fellowships, and prizes. The move toward openness is also facilitated by the establishment and use of open source 10 MESA, “MESA home,” http://mesa.sourceforge.net, accessed July 26, 2021. 11 The Enzo Project, “Home,” https://enzo-project.org, accessed July 26, 2021. 12 IRAF Community Distribution, “Home,” https://iraf-community.github.io/, accessed July 26, 2021. 13 National Academies of Sciences, Engineering, and Medicine, 2018, Open Source Software Policy Options for NASA Earth and Space Sciences, Washington, DC, The National Academies Press, p. 3. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-8

libraries (code and tools used by programmers when writing software) to collect and disseminate community software. The 2020 NASA Research Opportunities in Space and Earth Science (ROSES) call14 is responsive to this part of the 2018 National Academies report through its funding of open source software and open source tools, frameworks, and libraries. However, with funding at a modest level, this is only a small step toward supporting these programs. The National Academies report suggests the possibility of requiring software management plans for all proposals. The report further suggests that proposals that involve releasing software include a budget description for software development, documentation, distribution, support, publications, and maintenance. The addition of software maintenance plans in proposals encourages open code and longer term code maintenance by the code authors themselves and allows the peer review process to set the pace of culture change. While there are costs associated with maintenance, there are also significant savings in having the continued usability of code developed at high cost. Hardware and software environments are constantly evolving, and in the face of this changing environment the astronomical software community needs to embrace sustainable software solutions. “Containerization,” for example, is currently an effective solution for codes to remain operable. Containerization is an encapsulation of the full operating system so that the code can be operated on any hardware. The panel has identified the implementation of containerization as one of the goals of the ADAS described above. Replicability and reproducibility is an essential part of the scientific process. For software, this requires supporting, incentivizing, and educating the community about best practices that preserve the software and input files needed to reproduce and replicate analyses and results. This will likely need to involve coordinated efforts among investigators, collaborations, science libraries, publishers, archives, repositories, and the federal agencies. The American Astronomical Society (AAS) journals publication of software papers and partnership with the Journal of Open Source Software, the availability of Github repositories, and the European Zenodo repository are all encouraging developments. Improving standards for citing software15 encourage proper crediting of work, document the ingredients used in an analysis, and are essential to enabling replicability and reproducibility of results. NASA’s Science Mission Directorate’s Strategy for Data Management and Computing for Groundbreaking Science 2019–2024 presents a forward-looking vision in this area. The NSF Division of Astronomical Sciences (AST) could collaborate with the Computer and Information Science and Engineering division (CISE) to develop its strategic plans for supporting astrophysics for software development and data management. Similarly, DOE Cosmic Frontiers could work with the Advanced Scientific Computing Research (ASCR) program to develop a strategy for its astrophysics projects. These strategies could include training in software engineering, computer science, and programming practices for the astronomical software development community. H.2.3 Theoretical Astrophysics Theory often drives fundamental new discoveries, as well as informing the design and operation of new observations. Support for both individual investigators and theory networks through the agency 14 NASA, 2020, “Research Opportunities in Space and Earth Sciences—2020 (ROSES-2020),” NASA HQ, Washington, DC, https://nspires.nasaprs.com/external/viewrepositorydocument/cmdocumentid=735966/solicitationId=%7BBCEE336 B-D550-CCBA-1C8C-7A866DB06F45%7D/viewSolicitationDocument=1/FULL%20ROSES-2020_Amend45.pdf. 15 D. Bouquin, G. Muench, K. Cruz, D. Chivvis, and E. Henneken, 2019, “Citing Astronomy Software: Inline Text Examples,” AstroBetter, https://www.astrobetter.com/blog/2019/07/01/citing-astronomy-software-inline-text- examples/. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-9

grant programs is essential for the health of theory programs; however, funding levels have remained flat and proposal success rates have been dropping throughout the past decade. Theoretical astrophysics has grown to encompass analysis and interpretation of data, computational methods that enable investigation of complex physical systems, as well as the more traditional pencil-and-paper calculations, a growth that at times blurs the boundary between theory and observation. Theory has long played an outsized role in discovery in astrophysics. In some cases, entirely new observational programs have been developed to test theoretical predictions, often with spectacular success. For example, measurements of fluctuations in the cosmic microwave background radiation have provided remarkable new insights into fundamental properties of the universe. Theory is crucial for the interpretation of most observations—for example, the theory of stellar structure for asteroseismology data, or orbital mechanics for exoplanetary systems. Last, theory is used to develop essential new tools and frameworks for analysis of large and complex data sets, especially those resulting from surveys. In recognition of the important role of theory, the Astro2010 decadal survey recommended a new approach for support involving augmentation of existing grants programs and the creation of new ones. A modest ($8 million per year) augmentation was recommended for the NSF Astronomy and Astrophysics Research Grants (AAG) program, which supports investigators in all areas of astronomy including theory. An additional important recommendation for theory in the Astro2010 survey was the creation of a new interagency funding opportunity for Theory and Computation Networks (TCAN). This program was in recognition of the increasingly complex nature of modern research problems, and the increasing reliance on ever more complex software and numerical methods for their solution. The program was initially proposed to support 5-year grants jointly funded by all three agencies, but the DOE declined to participate, owing to their already strong support of computing through, for example, the INCITE program. Both NSF and NASA did initiate a TCAN program,16 but only for 3-year projects, and NSF discontinued the program after only one solicitation. For the future, several issues remain important. The NSF AAG program is a crucial vehicle for funding new independent and novel investigations in astronomy and astrophysics, but especially in theory. The panel suggests that this program continue to be supported in the face of continuing budget pressures. Second, the original intent of the TCAN program was to promote cross-agency and cross- disciplinary teams to tackle challenging problems. The implementation of the program, while providing welcome support for theory, did not achieve this vision. Some of the factors that contributed to this lack of success included cross-agency support for the program lasting only 1 year, with NASA left to solely support the program in all subsequent years, and funding for the program being shifted from other programs, rather than being allocated as new funding. If there is an opportunity for new initiatives in the future, a revived TCAN program that truly promotes interaction across agencies (involving, e.g., both ground- and space-based facilities), and across disciplines (including astrophysics, applied mathematics, computer and data science, physics, etc.), could have significant impact. Last, there remain concerns regarding barriers to funding cross-disciplinary projects such as large code development projects (involving, for example, computer science and astronomy), and multi- messenger astronomy (involving physics and astronomy). A significant (25 percent) increase in the funding for the NASA Astrophysics Theory Program (ATP) was recommended by the Astro2010 survey, but unfortunately this also was not realized. Instead, not only did funding remain flat, but also the program moved to a 2-year cycle of proposal. This cadence negatively impacts career development across the community. The panel suggests that the agency implement at least this augmentation of 25 percent and that the program resume its annual cadence. 16 NSF and NASA, 2013, “Theoretical and Computation Astrophysics Networks (TCAN),” NSF 13-512, https://www.nsf.gov/pubs/2013/nsf13512/nsf13512.htm. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-10

H.2.4 High-Performance and High-Throughput Computing Computation, both in theory and in data science, has emerged as foundational for essentially every topic in astronomy and astrophysics. Increasing capability in computing infrastructure is crucial for scientific progress. DOE and NSF are significantly increasing their capacity in high-performance computing over the coming decade, while NASA is expanding at a slower rate. High-performance computing (HPC) involves undertaking detailed calculations at high speeds using large supercomputers. HPC tasks require a large amount of compute power for solving complex problems (e.g., simulations of physical processes). High-throughput computing (HTC) enables relatively simple computational tasks to be undertaken in a highly efficient way (e.g., processing and analysis of very large data sets). In the past decade, computation, both in theory and data science, has unarguably emerged as foundational for essentially every topic in astronomy and astrophysics. Numerical simulations and big data analysis have become increasingly sophisticated, and their role in astrophysics has correspondingly experienced enormous growth. Despite approaching the limit of Moore’s law, computational power has also been growing steadily, with exascale supercomputers expected to become publicly available as early as 2022, and the potential for significant expansion in these capabilities by 2030. With the increasing sophistication in software, analysis, and computational capability, there is enormous potential and opportunity for scientific discovery in the coming decade. HPC enables discovery through simulations of processes such as the formation and evolution of stars, planets, galaxies, the universe, and gravitational wave events. HTC enables discovery using large data sets including investigations using archived observational data, joint pixel processing of complementary observations, and analysis of large, simulated data sets. Continued and expanded support for increasing capability in computing infrastructure and in people with sufficient expertise in HPC and HTC is crucial for scientific progress. The size of observational and synthetic data sets has consistently increased over time, from terabytes to petabytes, and soon to exabytes. In this era of Big Data, there is an emerging opportunity to use publicly available cloud computing for cost effective solutions, rather than hosting huge hardware resources and numerous proprietary facilities. The utility of cloud computing for HTC is currently clearer than for HPC, although this may be changing. The panel suggests that the funding agencies continue to explore the potential of cloud computing for a range of efforts and to provide support for projects to utilize cloud computing where appropriate. In the coming decade, DOE and NSF have plans to significantly increase their capacity in HPC, while NASA plans to expand at a slower rate. To ensure sufficient HPC resources for missions and to ensure that the community has sufficient access to HPC facilities over the coming decade will require either coordination among the funding agencies, or an increased expansion of HPC facilities at NASA. However, funding that supports work across agencies is sparse. As noted in Section H.2.3 above, the Astro2010 decadal survey recommended the development of a TCAN program, which was intended to be a collaboration between NASA and DOE for space-based astronomy and NSF and DOE for ground-based astronomy. A TCAN program was initiated between NASA and NSF, but has since become a NASA-only program. The panel suggests that reinvigorating a focused collaboration between all three funding agencies will enable the most efficient use of resources and will facilitate rapid development in key advanced computing areas that are currently experiencing only moderate progress owing to lack of support. In the past decade, HPC simulations have become integral to theoretical modeling, forecasting and survey formulation, and in the eventual analysis and interpretation of observational data. Developing and exploiting the software to undertake these tasks not only requires specialized facilities with large computing and storage resources, but also people with extensive expertise in both computer science and astrophysics. The potential barriers to access for individual investigators is of great concern to the community, and several white papers were submitted to highlight these concerns, particularly for those at institutes without preexisting relationships with the large HPC and HTC facilities. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-11

The panel suggests that the agencies increase their investment in training in the use of HPC and HTC facilities. In order to ensure appropriate training to undertake HPC and HTC programs and more equitable access to HPC and HTC facilities, it is important that extensive training be available at the undergraduate and graduate level. In addition, the panel suggests that more support be provided for training directly from the facilities—including workshops and internship programs—for all career levels. Coordination among the funding agencies would benefit the broader national program. The panel suggests an approach that involves NASA taking the lead in supporting archiving, and DOE and NSF taking the lead in providing HPC supercomputing facilities to scientists across astrophysics. H.2.5 Data Science and Machine Learning The interaction between astronomy and data science is a fruitful two-way exchange. Data science advances enable new insights in astrophysics. Rich astronomical data sets with underlying physical symmetries can push technology development in data science. Support for ongoing training in new data science techniques will enrich the return from large data sets and advance both data science and astronomy. Over the past decade, data science has advanced dramatically. Machine learning techniques are playing an increasingly important role in astrophysics, and this trend is likely to continue into the future. Over the past few years, there have been multiple joint data science/astrophysics faculty appointments. Many universities are adding new courses in this field. Both undergraduates and graduate students are pursuing joint degrees in programs that did not exist in 2010. Astronomical data offers many opportunities for data science research and is already proving to be a valuable data set. For example, Stalzer and Mentzel17 ranked the Sloan Digital Sky Survey (SDSS) as the sixth most influential paper in Big Data, just behind Shannon’s classic information theory. Astronomical data is valuable for data science for many reasons:  The data is open access, has no commercial value, and is free of the many ethical issues associated with other kinds of image data. In contrast, images of faces scraped from the Internet often do not have permissions, are used for photo surveillance, and are often racially biased samples.  Astronomical data is rich and ranges from images, tables, graphs, and uneven time series to multi-dimensional grids.  Data in the physical sciences is structured with individual particles, planets, and stars interacting in particular ways and with well-understood symmetries—data structures that differ from the widely studied images and sequences in other areas. This rich structure has already inspired early work in graph neural networks and geometric deep learning.  Astrophysicists have high-fidelity simulators that capture mechanistic causal models that describe both the astronomical phenomenon (e.g., the evolution of large-scale structure) and the astronomical processes (e.g., observations of gravitationally lensed galaxies by the Rubin telescope). In recent years, astrophysicists and data scientists have developed numerous new techniques for likelihood-free inference, advances in density estimation, implicit generative models, and probabilistic programming. These techniques are now being used across a wide range of fields (e.g., particle physics, chemistry, and neuroscience) and are part of an emerging new area spanning machine learning and the physical sciences.  Because it is possible to simulate data, it is possible to query whether a model is overfitting the data. Since the underlying physics is known for many astrophysical data sets, it is possible to learn whether artificial intelligence (AI) is learning the true underlying rules. This is a 17 M. Stalzer and C. Mentzel, 2016, A preliminary review of influential works in data-driven discovery, SpringerPlus, 5:1266, https://doi.org/10.1186/s40064-016-2888-8. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-12

much more significant test of a model than cross-validation and is important for making models safe and actually improving the understanding of science.  Building theories for the physical world is a potentially less ethically fraught implementation of AI on classifying text or images. While much of the work in data science has been focused on images, because they require no domain knowledge and are useful for online advertising and customer analysis, physical systems may be a better path toward Artificial General Intelligence (AGI). Data science offers powerful new tools for studying astronomical data and astrophysical systems. Machine learning has already shown significant success as a tool for identifying anomalies in data. These techniques could lead to transformative discoveries from the new data sets expected to become available in the 2020s. Machine learning has the potential of increasing the amount of information obtained from astronomical data sets by enabling modeling of complex nonlinear phenomena and instrumental effects. If machine learning can be successfully used to model multi-scale phenomena, it could open up the ability to more accurately simulate a wide range of astronomical processes from planet formation to galaxy formation. As this is a rapidly evolving field, the panel suggests that funding agencies use the grant programs and existing data centers to initiate and support ongoing training for astronomers as well as broad opportunities to enable a diverse group of scientists to apply and teach these techniques. H.2.6 Laboratory Astrophysics Laboratory astrophysics is essential for enabling science across astrophysics, and new laboratory measurements are essential to realize the full potential of recent and imminent major observatories (ALMA, JWST, GMT/TMT, etc. ) targeting stars, planets, star and planet formation, and high-energy phenomena. If the aim is to understand the structure and evolution of stars, galaxies, and the universe as a whole through the observations from future facilities, laboratory astrophysics will be required. Since astronomical systems span an enormous range of densities and temperatures, developments in laboratory astrophysics have the potential to stimulate developments in chemistry and physics. Currently, there are relatively few groups in the United States that are making the needed laboratory astrophysics measurements, and the prospects for establishing new groups is limited by an overall small funding envelope. This needs to be addressed to maximize the scientific return of the major astronomical investments in the 2020s. To expand the field of laboratory astrophysics and to ensure that existing expertise is transferred to a new generation, the overall funding envelope needs to be increased, and barriers to entry need to be removed. Despite limited resources, laboratory astrophysics has been instrumental in advancing astrophysical discoveries in the past decade. In the search for our interstellar chemical origins, the 2010s delivered first identifications of aromatic organics, and chiral molecules, and the first inventories of organic molecules at the onset of planet formation. These results were obtained because of new spectroscopic line lists. Complementary laboratory work revealed that many of these organics can form in icy grain mantles at close to 0 K, and that complex, prebiotically interesting organic molecules are thought to be ubiquitous during star and planet formation. New laboratory data has also been key to characterize the atmospheres of exoplanets; experimentally determined molecular line opacities at high temperatures have enabled retrievals of water abundances and constraints on atmospheric carbon/oxygen ratios, while haze formation experiments have been key to elucidate what kind of hazes and clouds may form on different kinds of exoplanets. Laboratory astrophysics has also been instrumental in advancing the fundamental understanding of the underlying physics governing stars. For example, in the early 2010s, a discrepancy between the theoretical convection zone boundary within the Sun and the value implied by asteroseismic data was identified. This became known as the solar convection zone boundary PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-13

problem and represents a lack of understanding in the physics that govern the closest star, the Sun. The Sun is the benchmark by which all other stars across the Milky Way and beyond are understood. Recent laboratory measurements of the opacity of iron at the temperature and density of the solar convection zone boundary revealed large discrepancies between the observed and theoretical opacities, implying that theoretical stellar opacity calculations are far from correct. It is because of these laboratory measurements that the solar convection zone boundary problem has been reduced significantly. Laboratory astrophysics was identified in the Astro2010 decadal report as “vital for optimizing the science return from current and planned facilities,”18 especially in the ALMA and JWST era. Yet, they found that “support and infrastructure for laboratory astrophysics are eroding both in the National Laboratories and in universities,” and they recommended that “the funding through APRA that is aimed at mission-enabling laboratory astrophysics should be augmented at a level recommended by this scientific assessment … a notional budget increment of $20 million over the decade may be required.”19 This augmentation was not implemented. As a result, the community is now in the age of ALMA and (soon) JWST, without many of the required laboratory measurements, which will severely limit the science return from these observatories if not addressed in the 2020s. ALMA and JWST have incredible capabilities to explore the interstellar medium and star and planet formation, and in the case of JWST, to characterize planets. However, interpreting this data requires laboratory and computationally generated databases of dust and molecule opacities, complex-refractive indices of condensed matter particles (aerosol analogs), spectroscopic lines, collisional cross sections, and gas and solid-state reaction rates. At present, these are all woefully incomplete, and there is a small number of active laboratories that contribute to them. The 2020s will also see a number of large observational surveys focusing on stellar astrophysics, which together address the fundamental astrophysical problem of stellar properties, such as the detailed chemical compositions, masses, and ages. In the era of upcoming photometric (Rubin Observatory, Skymapper, etc.) and large high- to low-resolution spectroscopic surveys (SDSS-IV, SDSS-V, 4MOST, WEAVE, GALAH, Gaia-ESO, etc.), astronomers will not be limited by data in the pursuit of stellar astrophysics, but rather by a lack of laboratory measurements needed to interpret the data. While these fundamental parameters are crucial for stellar astrophysics, they are also important in a wide range of astrophysics ranging from exoplanet science to galaxy formation. The availability of relevant laboratory atomic, molecular, and optical (AMO) data, such as highly accurate wavelengths, transition probabilities, photoionization cross sections, line broadening parameters, and collisional cross sections, will be critical for maximizing the scientific return of these surveys, observatories, and missions, which together represent a significant investment of U.S. astronomy resources. At higher energies, the scientific return from proposed high-resolution X-ray spectroscopic missions, like Athena, will not be able to capitalize on their high resolution without new atomic data including collisional and photoionization cross sections. Potential diagnostics of density, temperature, ionization, abundances, and so on will not be realized without improved laboratory data on transition energies, electron impact ionization collision strengths, photoexcitation, and ionization. Laboratory astrophysics is also a required foundation to enable science on a range of scales—from as small as dust grain growth to the solar convection boundary problem, to understanding the shock physics of supernovae. Laboratory astrophysics is mainly funded via grants by NASA Astrophysics Research and Analysis (APRA) and Astrophysics Data Analysis Program (ADAP), and the NSF AST. However, the number of awards approved across all of these programs is small, and they have been declining since the early 2000s. This put the laboratory astrophysics field under severe pressure during a time when there are growing needs from the ALMA and JWST, and stellar and exoplanet astrophysics communities. The 18 National Research Council, 2010, New Worlds, New Horizons in Astronomy and Astrophysics, Washington, DC, The National Academies Press, p. 32. 19 National Research Council, 2010, New Worlds, New Horizons in Astronomy and Astrophysics, Washington, DC, The National Academies Press, pp. 220–221. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-14

panel suggests that going forward it is critical that agencies continue to fund the currently active groups, train the next generation of laboratory astrophysicists, and lower barriers of entry into the field. The panel suggests increased investment in laboratory astrophysics. This increase could come in many forms, such as (1) increased investment in the laboratory portion of the NASA APRA program; (2) inclusion of a special funding line of laboratory astrophysics in missions and facilities, possibly through NASA Phase E funding and something analogous for NSF-supported facilities; (3) an increase in the funding for the NSF Major Research Instrumentation (MRI) grants and removal of the institutional cost- sharing requirement; and (4) an increase in the funding for interdisciplinary but laboratory astrophysics- centered workshops, internship, and professional development programs at the National Laboratories and other laboratory astrophysics centers around the United States that will interface with universities. While these remedies are not the only possibilities, they represent four clear actions that can be taken to strengthen laboratory astrophysics and enable the United States to make the most of its astronomical investments and achieve the scientific aims of the next decade. H.2.7 Technology Development Investment in astrophysical technology development is broadly enabling for astrophysics and enhances U.S. technological competitiveness. Astrophysical discovery and technological advancement march hand-in-hand, with new technology opening new and unexpected windows on the universe. Consequently, the astrophysical community has a core strategic need to mature new technologies to the point where they can be flown on NASA missions or deployed on ground-based instruments. Conversely, there is clear synergy between the astrophysical community’s needs and expertise, and those of broader society. To seek life on other worlds, astronomers require essentially noiseless, nearly quantum limited detectors in the UV, visible, and IR. Many of these same properties are needed for quantum computing and information science. Robotics, automation, and advanced manufacturing enable building the next generation of telescopes on the ground and in space—and are likewise strategically important to U.S. technological competitiveness. Astronomers routinely contribute to—and draw on— advances in materials science for their electro-optical sensors, optical components, and system engineering to build ultra-precise instruments. H.2.7.1 NASA NASA tracks risk using Technology Readiness Levels (TRLs). A TRL is an integer in the interval [1,9], with TRL 1 roughly corresponding to an idea with supporting data up to TRL 9 denoting proven, flight-heritage hardware. All key technologies are generally required to be TRL 6 early in the life of a project. For strategic missions (JWST, Roman, etc.), TRL 6 is required to pass the Preliminary Design Review (PDR). For Explorers, SmallSats, and CubeSats, all technologies are generally required to be TRL 6 or higher upon selection for implementation. In addition to directed funding, NASA uses two grant programs to support technology development. These are the Astrophysics Research and Analysis (APRA) Program and the Strategic Astrophysics Technology (SAT) Program. APRA is open to all TRLs, whereas SAT focuses on the “mid- TRL” range from 3 to 6. APRA is broad-based, and not specifically focused on strategic missions. It encompasses Suborbital/Suborbital-Class Investigations, Detector Development, Supporting Technology, and Laboratory Astrophysics. During the period 2014–2018, about 50 percent of new APRA funding went to the Suborbital Program. The remainder was split with about 20 percent each going to Detector Development and Technology Development, and about 10 percent to Laboratory Astrophysics. Although the typical $200,000 to $400,000 per year award amounts for Technology Development are appropriately PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-15

sized for individual investigators supporting perhaps a postdoctorate or a few students, in practice APRA awards are often too small to allow investigators to partner with industry. Yet, having industrial partners is essential for developing key enabling technologies including advanced optics and detectors. The SAT program was created in response to Astro2010 as part of addressing a mid-TRL gap from TRL 3–TRL 6 that was perceived to exist for strategic missions. SAT solicitations are highly responsive to “technology gaps” that strategic missions identify with substantial input from the community in Program Analysis Groups (PAGs). Although the PAGs solicit input from individual investigators, in practice the needs of large, strategic missions hold more sway. Although the SAT program addresses the needs of large, strategic missions, a notable mid-TRL gap still exists for Explorers, SmallSats, and CubeSats. Typical APRA award amounts are insufficient to support Detector Development or Supporting Technology development by a PI-led team working with the major technology vendors. The NASA Space Technology Mission Directorate (STMD) has historically provided support for developing long-range technologies relevant to astrophysics. In the past several years, this support has been eliminated. This has led to a decrease in support for challenging long-term technology. The APRA Program offers the best opportunities for developing highly innovative but risky new technologies. Unfortunately, many excellent proposals cannot be selected at current funding levels. Increased funding to the APRA program would increase the likelihood of developing transformative new technologies for NASA. H.2.7.2 NSF NSF funds technology primarily through the Division of Astronomical Sciences Advanced Technology and Instrumentation (ATI) program and the Mid-Scale Innovations Program (MSIP). The Major Research Instrumentation Program (MRI) occasionally funds technology development as well. The MSIP program was initiated in response to a recommendation from the Astro2010 report. Since 2012, the MSIP program has awarded $146 million to 19 projects. While originally intended to support programs up to $40 million, the largest MSIP award to date has been for $17 million (NANOGrav Physics Frontier Center), with all but three funded programs less than $10 million. Not all of this funding is directed toward technology development, because MSIP also funds construction and operations costs for mid-scale projects. Since 2014, the combined ATI and MRI programs have funded an average of 13 proposals, totaling $7.5 million per year. The award amounts range from $30,000 to $4 million (a large MRI). The wide range of funded award amounts reflects the program’s effort to address a range of development activities from simple studies to incremental advancement to deployed instruments on telescopes. However, the average award of $575,000 over 3 years cannot be expected to yield significant technology progress. The panel suggests expanding the NSF ATI and MSIP programs to enable a significant increase in the number and size of awards to enable substantial technology progress. The panel suggests that an increase of 10 percent per year over the next decade will ensure a healthy program. H.2.8 NASA Suborbital Program The suborbital program consistently returns fast-turnaround, cutting-edge science; provides important technology development for future programs; and trains the next-generation of researchers, technologists, and program managers. The suborbital program is critical to maintaining the health of university laboratories capable of carrying out space missions. The suborbital program fills a critical niche by delivering science that is impossible to do from the ground, and does it much more cost effectively than orbital missions. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-16

The suborbital program remains a key part of NASA’s portfolio, addressing a wide variety of high-profile scientific problems, developing and testing technology important for future missions and training the next generation of instrumentalists and project leaders. The suborbital portfolio is broken into two components: high-altitude ballooning and sounding rockets. The Enabling Foundation panel received a white paper and met with representatives of both communities who presented the program status. H.2.8.1 Balloon Program The introduction of new super-pressure balloons has enabled the exploration of new, more ambitious science missions with significant science returns. The panel suggests that increasing the number of payloads and flights would take better advantage of this capability. While NASA’s high-altitude balloon program has been active for more than 50 years, it has been anything but static. The program has seen a significant evolution in capabilities in response to the user community needs. The Columbia Scientific Balloon Facility (CSBF) is directed by the Balloon Program Office. The program offers a wide array of capabilities ranging from single-day “conventional” flights to Long Duration Balloon (LDB) flights lasting up to 60 days, to the new Super Pressure Balloon (SPB) flights with predicted flight times up to 100 days with extreme altitude stability. They support launch operations from Texas, New Mexico, Sweden, New Zealand, Australia, and Antarctica along with the occasional “remote” launch site. The access to a near-space environment (35–40 km) with such a wide variety of options for duration and sky coverage provides a critical resource to the community. This is demonstrated by the breadth of payload science goals and the hundreds of publications and well over 100 Ph.D.s awarded over the past decade. There is also a small but vibrant program to engage future scientists through “piggyback” instruments on existing payloads. Maintaining a robust balloon program over the next decade and beyond will require optimization of resources and capabilities within the program:  There remains a significant technology barrier to new PIs entering the ballooning community. An effort to provide guidance, information, and common hardware and software could greatly facilitate the entry of new researchers into the program. This could include pairing new (or prospective) PIs with more experienced groups to facilitate the transfer of experience and skills. These groups could also be engaged to supply “common” technology such as star trackers and power systems.  A more formalized outreach program perhaps engaging existing PIs could greatly expand the successful piggyback program to smaller undergraduate institutions throughout the country.  The SPB capability has advanced significantly over the past decade to the point where 1400 kg payloads are possible. However, the altitudes are limited to 33.5 km, which is too low for many science programs. Developing balloons to achieve higher altitudes and more mass would significantly broaden their utility.  The number of flights has been dropping over the past several decades. While this has been offset by longer flight durations (total days in the air), a robust program over the next decade will need to expand its user base. This would translate into more funded payloads coupled with more launch opportunities. By encouraging lower cost “conventional” flights with an emphasis on technology development, NASA can increase the size and diversity of the ballooning community.  Balloon launches place a heavy burden on the CSBF personnel responsible for coordinating all aspects of the program besides the payload itself. Many are required to spend up to 8 months per year supporting launches. This results in high turnover of personnel with highly specialized skill sets. Increasing the number of launches while reducing the burden on the CSBF personnel will be a challenge. It may involve a combination of more trained personnel PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-17

with more flights per campaign, such as expanding the number of launches per season in Antarctica and Wanaka with increased infrastructure investment. H.2.8.2 Sounding Rocket Program Like the balloon program, the sounding rocket program has been consistently returning science and technology results while training next-generation instrumentation builders over many decades. The rocket program provides crucial access to the space environment with apogees of 250–350 km with 300 seconds above 160 km. This is essentially the only path available to develop and test technologies where even the small amount of residual atmosphere at balloon altitudes is unacceptable. The payloads concentrate in the infrared, ultraviolet, and X-ray. Access to these altitudes comes with relatively short flights and limited payload masses (<220 kg). There are approximately 10 active astrophysics sounding rocket payloads with 4 to 5 flights per year. Compared to the balloon program, the science is not as varied and the direct scientific impact is not as deep. However, unlike balloons, sounding rockets reach space. Therefore, the technology development enabled by the sounding rocket program is more directly traceable to NASA missions, which collectively have significant impact. The sounding rocket program differs from the balloon program in that the power, telemetry, and pointing systems are provided to the experimental team. This is possible because the sounding rocket capabilities are very well defined. The result is that the barrier for entry by new PIs is, in theory, lower than for the balloon program. However, there has been very little in the way of increased capability for the sounding rocket vehicles (longer flights, larger payloads). While such increased capability could lead to expanded science return, there appears to be limited pressure from the community to do so. H.2.9 NASA Explorers Programs NASA’s planned cadence provides good balance for the space portfolio.20 Pioneers and SmallSats will provide new rapid response opportunities for broad multi-messenger and multi-wavelength observations and technology development.21 The Pioneers program will likely allow teams to complete and fly missions with schedules tailored to the needs of the projects (not limited to the APRA 5-year maximum funding cycle). The panel suggests that a mid-decadal review of the status of the Explorers programs and the impact of the new SmallSat and Pioneers initiatives would be appropriate. The NASA Explorers Programs (Table H.1), Medium-Class Explorers (MIDEX), Small Explorers (SMEX), and Missions of Opportunity (MOs), provide consistently excellent scientific returns for a relatively moderate investment and the ability of rapid response to new scientific and technical breakthroughs over a broad range of wavelengths. Explorers enable discoveries with multi-wavelength observations complementing flagship and perhaps future probe missions. The Astrophysics SmallSat program (started in 2018) and Astrophysics Pioneer program (anticipated in 2021) will provide new rapid response opportunities for broad multi-messenger observations and technology development. All of these programs provide opportunities for strategic workforce, scientific, and technical development that ensure the long-term success of NASA’s scientific aims. Explorers provide an ideal training ground for the next generation of space experimentalists. There are often university-based, and as such are key to sustaining the university groups that have launched many space science careers. 20 This section was written with substantial input from an interpanel committee of Angela Olinto (Chair), Megan Donahue, Charles Hailey, Bruce Macintosh, Amy Mainzer, Bernie Rauscher, Mark Saunders, and Evgenya Shkolnik. 21 CubeSats in development: CUTE, SPARCS, BurstCube, SPRITE, and BlackCAT. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-18

TABLE H.1 Status of Astrophysics Explorers, Current and Under Development, and Announcement of Opportunity (AO) Dates Mission AO (Launch) Class Mission AO (Launch Class Year Plan) Year NuSTAR 2003 (2012) SMEX SPHEREx 2016 (2023) MIDEX NICER 2011 (2017) MO ARIEL (CASE) 2016 (2028) MO TESS 2011 (2018) MIDEX ESCAPE or COSI 2019 (2025) SMEX IXPE 2014 (2021 plan) SMEX Dorado or LEAP 2019 (TBD) MO GUSTO 2014 (2021 plan) MO (To be selected) 2021 (2028) MIDEX XRISMa 2008 (2022 plan) MO (To be selected) 2021 (TBD) MO a XRISM is the successor to JAXA/NASA Hitomi (launched in 2016), which failed after a month in orbit. NASA’s contribution to Hitomi was selected as an Explorer MO in 2008. NASA Astrophysics plans an Explorers program cadence of two MIDEX, two SMEX, and four MOs per decade. In addition, NASA plans a cadence of 5 to 10 Pioneers and about 10 SmallSats (i.e., CubeSats <6 U)22 per decade. The SmallSat program deployed one CubeSat in 2018 (HaloSat) and has five in development for deployment from 2020 to 2024. The Pioneers program will include larger SmallSats (CubeSats >6 U), major balloon payloads, and modest International Space Station attached payloads (with a $20 million FY 2020 cost cap, not including launch). Pioneers are designed to fill in the gap between the Astrophysics Research and Analysis (APRA) program (typically <$10 million) and Explorers MOs (<$35 million for SmallSats). Explorers’ most recent cost caps were $250 million (FY 2017) for MIDEX, $145 million for SMEX (FY 2020), $75 million (FY 2020) for associated MO for Small Complete Missions and Partner Missions of Opportunity, and $35 million (FY 2020) for SmallSat MO, all without the launch vehicles. Explorers, Pioneers, and SmallSat programs are openly competed through peer-review processes. The CubeSat program is the newest addition to NASA’s program. While not strictly sub-orbital, there are parallels to both the balloon and sounding rocket programs. CubeSats are volume-limited in their current implementation up to 12 U. The CubeSat includes a science instrument with a bus. The bus can be commercially provided and includes power, communications and pointing. Like the sounding rocket program, this commercially provided infrastructure lowers the barrier to CubeSat entry to new PIs, although the cost is borne by the PI. The management of a CubeSat instrument program could be a 22 A Unit (U) is a standard volume unit for a CubeSat, a cube 10 cm on a side. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-19

challenge, and it appears to the panel that moving to the sounding rocket model of a NASA-provided BUS would dramatically increase participation. H.2.9.1 Diversifying the Explorer Program Explorer missions tend to be led by teams composed mostly of white m from a limited set of universities and NASA centers. Using gender as a marker of diversity, a study23 of Explorer-class proposals from 2008–2016 finds that the participation by women in the leadership and science teams “is well below the representation of women in astronomy and astrophysics as a whole.” PIs, leadership teams, and instrument builders of selected Explorer missions are less diverse than even the astronomy and astrophysics community. To remove barriers to access, the panel suggests that efforts to diversify Explorers and SmallSat mission leaders and teams be enhanced. The panel suggests that NASA require proposing teams to demonstrate diversity in their teams. A team’s diversity includes the diversity of institutions, geographical locations, stages in career, and underrepresented populations. Given the extraordinarily important role that the Explorers program plays in scientific discoveries, technology demonstration, and in the training of space scientists at all levels, it is vital that Explorers reflect the diversity of the workforce to which the scientific community aspires. There are substantive barriers to entry to the Explorers program that are unique to the costly and complex development of satellite experiments, and those barriers affect the diversity of teams and scientific ideas. One barrier is the need for considerable engineering resources and expertise that are currently (and deliberately) underfunded and that must be provided by the proposing institutions. The panel suggests that NASA implement a new program of Concept Maturation Studies (CMS) for future Explorer missions. The new program would select 5 to 10 CMS, for future SMEXs, MIDEXs, and MOs, placing primary emphasis in the evaluation on the proposed science, instrument concept definition and maturation, and team diversity, with less emphasis on management, detailed engineering, and cost. It is important that CMS awards are not a precondition to proposing or being selected for an Explorer call. The panel suggests that the CMS be solicited and funded separately from the normal Explorer AO process, but with the same cadence as Explorer AOs per decade. The level of effort required for responding to the CMS AO is comparable to an APRA sub-orbital proposal, but with a more extended discussion of science and diversity of the proposing team, the primary criterion in this pre-step 1 program. The reduction in emphasis on technical, management, and cost maturity would properly put the focus on science, diversity, and technical innovation, allowing institutions with fewer resources to compete effectively in the formal AO process. The panel suggests 5–10 fully funded studies, each funded at the $1 million to $2 million level. At the discretion of the PI, a CMS budget could involve resources at NASA centers (such as the Integrated Design Center at Goddard Space Flight Center and TeamX at the Jet Propulsion Laboratory) or other institutions with comparable expertise. This would enable proposals with well-defined science goals to access appropriate engineering support, even if they come from universities that do not have appropriate expertise in house. The panel suggests that NASA strengthen its support for sustained workshop/workforce development experiences where people with exciting scientific ideas are selected to come to NASA centers and work with engineers to help develop their ideas and their background knowledge in context with current and near-term technologies. The PI Launchpad24 is an example of such a program. Its goal was to train people interested in developing their first flight mission proposal but have no idea where to start. Continued, ongoing support to develop and run workshops and more intensive internship programs 23 J. Centrella, M. New, and M. Thompson, 2019, “Leadership and Participation in NASA’s Explorer-Class Missions,” white paper submitted to the Astro2020 decadal survey, https://arxiv.org/abs/1909.10314. 24 NASA, “PI Launchpad Workshop Content,” last update July 29, 2021, https://science.nasa.gov/researchers/pi-launchpad. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-20

of this nature is essential to lowering the barriers to entry into flight projects and to increasing equity, diversity, and inclusion in the community. NASA centers wield a strong influence on who is selected to propose, so opening their doors to new and expanded sources of ideas will expand the breadth and innovation of NASA’s science. H.2.10 NSF Mid-Scale Programs The mid-scale program has enabled many exciting projects that have had high impact, have led to important scientific discoveries, and have contributed to training students.25 In response to Astro2010, the 2012 NSF Portfolio Review,26 and as one of NSF’s Ten Big Ideas, the NSF MPS Directorate and the Division of Astronomical Sciences (AST) have taken several steps to support mid-scale programs. Mid-scale programs are generally understood to be those with total costs between the funding scale of the Major Research Instrumentation (MRI) program (currently $4 million) and the Major Research Equipment and Facilities Construction (MREFC) program (currently $70 million). An important component of the support for mid-scale programs within both AST and PHY is the Mid-Scale Innovations Program (MSIP), which has been solicited biyearly since 2014. A key review criterion for MSIP awards is the value and benefit to the U.S. astronomical community. MSIP generally has provided $2 million to $10 million total awards to each project, and can support a variety of activities (science project operations, facility, development, or open access capability). Through its first three cycles, it has competitively awarded $114 million to 18 distinct projects that span a diverse range of projects covering almost the entire spectrum available to ground-based astronomy, including 11 projects in radio and 7 in optical astronomy. In addition, NSF awarded $9 million of MSIP funding to the Dark Energy Survey without an open call for proposals, and with only one proposal considered. A more recent program that supports mid-scale projects is the NSF-wide Mid-Scale Research Infrastructure (MSRI) program, which has been solicited once starting in 2019. MSRI is restricted to fund only design and construction, not operations. The first track of the program (MSRI-1) funded programs up to $20 million and in AST has issued a total of $16.7 million to two astronomy projects, the Event Horizon Telescope (EHT) and development for CMB-S4. The second track of this NSF-wide program, commonly referred to as MSRI-2, calls for implementation proposals with total project costs in the range from $20 million to $70 million. According to NSF, the long-term intent is for MSRI-2 to cover project costs over a range extending up to $100 million. In addition, NSF has awarded a number of other projects at mid-scale without open solicitations, through special project funding, often to take advantage of internal NSF opportunities for the benefit of AST and to form partnerships with other agencies to leverage their resources. While NSF did not provide requested data on the overall level of NSF investment in such programs, an example of one such enabling partnership is the NN-EXPLORE project to support extreme precision radial velocity (EPRV) work at the 3.5 m WIYN telescope. To date, $8.8 million has been funded from NSF and roughly $15 million to $20 million from NASA. This funding follows a recommendation of Astro2010 to support EPRV, as well as fulfilling the need to support NASA’s TESS mission. In order to assess the health and impact of the mid-scale program, the Enabling Foundations panel of the Astro2020 decadal survey formed a working group that solicited input from other panels, as well gathering info from NSF. Taken as a whole, the investments made by NSF in mid-scale projects through the MSIP, MSRI- 1, and MSRI-2 programs have indeed provided substantial benefits to the U.S. astronomical community 25 This section is based on a report from an interpanel committee: James Stone (Chair), Michael Blanton, Jenny Greene, David Kieda, Andrea Lommen, Dan Marrone, and David Silva. 26 NSF, 2012, Advancing Astronomy in the Coming Decade: Opportunities and Challenges, National Science Foundation Division of Astronomical Sciences Portfolio Review Committee, https://www.nsf.gov/mps/ast/portfolioreview/reports/ast_portfolio_review_report.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-21

and fulfilled important and numerous research objectives that cannot or would not be addressed through the AAG, ATI, and MRI grants programs or as part of MREFC projects. Based on an analysis of listed refereed publications available from the websites of the 18 MSIP awardees, they have at least 500 publications published through March 2020. Several MSIP programs provide publicly available astronomical data, and it is likely that these have generated even more publications that are not tracked through the websites. There are several important results enabled by MSIP worth listing:  Public release of data from the Dark Energy Survey, Zwicky Transient Factory, and the HyperSuprimeCam survey, providing important benefits to the U.S. community not only in their own right, but also as mid-scale forerunners of the Vera Rubin Observatory.  The Event Horizon Telescope’s image of the supermassive black hole at the center of M87 has provided important insights into space-time in the strong gravity regime, the extreme conditions in accreting plasma, and has captured the imagination of the public across the world.  Direct measurements of the sizes of 300 nearby stars from the Center for High Angular Resolution Astronomy has vastly improved astronomers’ ability to test stellar structure and evolution models, as well as image surface features such as flares and starspots.  With its higher angular resolution, measurements of the cosmic microwave background by the Atacama Cosmology Telescope Polarization (ACTPOL) survey have provided an independent check of the Planck best-fit cosmology. Moreover, POLARBEAR’s measurement of the B mode power spectrum is an independent determination of the gravitational lensing signal. The proposal pressure for the MSIP program has been strong, with a total of 87 proposals submitted in the three cycles, for a funding rate of ~20 percent. The funded projects include new instruments for optical and radio telescopes that come with public access/data, enhancements to national facilities, numerous dedicated cosmological experiments, and open survey programs. Virtually none of these projects fit within the funding envelope of the ATI, MRI, or AAG programs, so the presence of this funding mechanism has been an essential component of astronomical progress in the 2010s. The panel notes clear signs of strain in this program. Although NSF did not provide any information about the unfunded programs, it is notable that the largest funded projects have budgets that are less than 30 percent of maximum cost specified in the program solicitation ($40 million in the first cycle, $30 million thereafter), a situation that is unimaginable in other AST mechanisms. The MSIP has been unable to fund at least one proposal near its cost cap—CCAT—despite its identification as the top medium-cost ground-based program in Astro2010. There are other examples of programs known to have failed to fit within the MSIP despite strong external reviews or strong subsequent performance, including the Frequency-Agile Solar Radio telescope, which has been ranked highly by multiple decadal surveys in solar and space physics and astronomy and astrophysics, and SDSS-IV, which has yielded a valuable community resource and hundreds of publications without MSIP funding. The sense of the committee is that the opportunity to productively support mid-scale opportunities is not nearly saturated. Expansion of the MSIP program would be rewarded with proportionally increased scientific productivity of the field. NSF is the de facto federal steward for the general health and welfare of U.S. ground-based astronomy. Judging from the white papers submitted by the U.S. community to Astro2020, there is great scientific opportunity to be enabled by a strong mid-scale investment program. Moreover, many of these opportunities can or will leverage partnerships with other NSF divisions, federal agencies and private philanthropy. It is also clear that NSF support needed to realize these opportunities is often greater than $20 million per project. All three of these observations (opportunity, partnership potential, NSF investment needed per project) were validated repeatedly over the past decade. The creation of MSRI-2 in 2019 for projects in the $20 million to $70 million dollar range was a PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-22

very welcome development. While it is too early to evaluate the MSRI-2 return-on-impact for astronomical research, MSRI-2 has a healthy oversubscription rate in terms of proposal submission and dollars requested suggesting increased funding for this program would have high scientific impact across several fields, not just astronomy and astrophysics. Initial outcomes have awarded successful MSRI-2 proposals at or near the upper funding bound. Time will tell, but a range of project sizes, enabling projects by more groups, would be a desirable outcome with high scientific impact judging from ideas presented to the Astro2020 panels. In the MSRI-2 and MREFC programs, projects with widely disparate budgets spanning different divisions and directorates are reviewed together. This could potentially limit mid-scale projects in AST, especially given an apparent internal NSF culture to homogenize total award funding across all divisions. Continuation of a healthy AST-only MSIP program would be very wise if the astronomical sciences community envisions many exciting projects with budgets in the mid-scale range in the coming decade. How big is “mid-scale”? Following the path of MSRI-1 ($4 million to $19 million) and MSRI-2 ($20 million to $69 million), a new MSRI-3 program with a range of $70 million to $150 million may be warranted. The latter could be split out from the MREFC funding line and indeed could be called MREFC-1 with larger projects going to MREFC-2. Given the larger aspirations of the entire NSF- supported research community, it may be time to separate major projects (under $500 million investment by NSF) from even larger projects ($500 million and above). Last, at the lower end, there is still much scientific opportunity to be realized in astronomy and astrophysics at the $5 million to $20 million range enabled by MSIP. The panel recognizes the importance of having as few siloed funding opportunities as possible; a healthy MSIP program seems well-warranted. In order to keep costs down, funding for project management has sometime been reduced at the proposal level to such a degree that the success of the effort is endangered. Moreover, while the MSIP program provides funding for operations, the MSRI-1 program does not, highlighting yet one more reason why the MSRI-1 program is not a replacement for MSIP. The importance of recruiting and supporting a strong management team with sufficient resources for operations is often crucial for success; the panel suggests that this aspect of the program be emphasized more in evaluations. The NSF funding at mid-scale has been a mix of competitive and noncompetitive programs. Competitive programs follow a best practice including openly advertising calls for proposals, multiple considered proposals, and peer review in proposal evaluation. Both the competitive and noncompetitive awards have led to important scientific advances. Circumstances at times will justify future noncompetitive awards, but the panel suggests that the accepted best practice employed by competitive programs be the norm. Last, the success of the few interagency projects funded at the mid-scale over the past decade suggests that further, and closer, interagency cooperation and funding opportunities could greatly benefit the science that can be supported in the coming decade. There are more worthwhile projects proposed to the AST MSIP program than can be funded at the current budget levels. The panel suggests enhancing the support of a mid-scale program funded entirely within Astronomical Sciences. H.2.11 Programmatics of Ground-Based Resources The disconnect between ground-based observing time and funding for those projects is a significant inefficiency in the current system, a barrier to scientific progress, and a hindrance to the science return on infrastructure investments. One of the benefits to winning observing time on a NASA instrument is access to a grants program that supports the analysis of that data once acquired and the dissemination of results. But funding is not a benefit of winning time on ground-based facilities run by NSF. The panel identifies this disconnect as a significant inefficiency in the current system. Investigators are often caught in a situation where a grant review panel could be skeptical about a proposed project being awarded the needed PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-23

telescope time to complete the project and where a telescope allocation committee could be skeptical about a team having the necessary resources to make the most of the proposed observations. This situation is a net drain on the entire system because projects effectively need to be reviewed twice (by a telescope allocation committee and a grant review committee) and because some teams have data/observing time but insufficient resources to properly collect, analyze, and disseminate any results. This situation is particularly concerning because it has a disproportionate impact on teams with access to fewer institutional and personal resources. In order to maximize the scientific return of their investment in ground-based observing facilities, the panel suggests that NSF explore funding travel costs and publication costs for U.S.-based teams that have been awarded observing time on public facilities. These costs could be viewed as part of telescope operating costs. Observations from a telescope must be published; otherwise, they have no scientific value. The panel suggests that NSF consider directly funding journals that publish observations from public telescopes. If this were done through directly funding the journals rather than small grants to universities, this would reduce overhead. If this program included a grant component, it could extend the student support programs already implemented for ALMA and NOAO observations.27,28 In order to maximize the scientific return of their investment in ground-based observing facilities, the panel suggests that NSF explore funding travel costs and publication costs for U.S.-based teams that have been awarded observing time on public facilities. These costs could be viewed as part of telescope operating costs. Observations from a telescope must be published; otherwise, they have no scientific value. The panel suggests that NSF consider directly funding journals that public observations from public telescopes. If this were done directly through funding the journals rather than small grants to universities, this would reduce overhead. If this program included a grant component, it could extend the student support programs already implemented for ALMA and NOAO observations.29,30 H.3 CONCLUSIONS Over the past decade, the United States has made major investments in astronomical instrumentation, and these instruments are poised to produce data that could transform our understanding of the universe. However, this transformation will require that astronomers, physicists, and computer scientists produce the tools needed to analyze the data, make the laboratory measurements needed to interpret the data, and develop the theory and simulations that enable new paradigms. Investing in the enabling foundation is an investment in the people that will do this transformative science. It is also an investment in the people who will develop the technologies that will enable the observatories, satellites, and instruments of the future. These investments remove barriers and ensure retention to entry to create a more diverse astrophysics workforce. As advocated in the National Science Board’s Vision 2030:31 At the post-secondary level, the U.S. must embrace a pathways model to workforce development. Because entry into the STEM workforce is not always via a linear high school–university– workforce path, the U.S. must offer individuals, from skilled technical workers to Ph.D.s, on- 27 National Radio Astronomy Observatory, “Student Observing Support (SOS) Program,” last update June 20, 2021, https://science.nrao.edu/opportunities/student-programs/sos. 28 National Radio Astronomy Observatory, “Financial Support,” last update December 16, 2016, http://ast.noao.edu/observing/financial-support. 29 National Radio Astronomy Observatory, “Student Observing Support (SOS) Program,” last update June 20, 2021, https://science.nrao.edu/opportunities/student-programs/sos. 30 National Radio Astronomy Observatory, “Financial Support,” last update December 16, 2016, http://ast.noao.edu/observing/financial-support. 31 National Science Foundation, 2020, Vision 2030, NSF National Science Board, NSB-2020-15, Alexandria, VA, https://www.nsf.gov/nsb/publications/2020/nsb202015.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-24

ramps into STEM-capable jobs. The U.S. must also deepen partnerships between educational institutions and the business sector to prepare Americans for the industries of the future and support reskilling and upskilling of incumbent workers so that they can better navigate rapid changes in the world of work. In order to lead in 2030, the U.S. also must be aggressive about cultivating the fullness of the nation’s domestic talent. Although the proportion of Black and Hispanic representation in S&E [science and engineering] jobs rose slightly from 1995 to 2017, these groups remain underrepresented compared to their proportion in the general population. Over the past two decades, the number of women in S&E occupations has doubled. Yet despite comprising over half of the college-educated workforce, as of 2017, women account for just 29 percent of the S&E workforce. The demographics of astrophysics mirror the rest of science and engineering. By creating more paths to entry into the field, by encouraging retention, and by enhancing programs that attract a diverse range of students, astrophysics will attract a richer range of talents. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION H-25

Next: Appendix I: Report of the Panel on Electromagnetic Observations from Space 1 »
Pathways to Discovery in Astronomy and Astrophysics for the 2020s Get This Book
×
Buy Prepub | $104.00 Buy Paperback | $95.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

We live in a time of extraordinary discovery and progress in astronomy and astrophysics. The next decade will transform our understanding of the universe and humanity's place in it. Every decade the U.S. agencies that provide primary federal funding for astronomy and astrophysics request a survey to assess the status of, and opportunities for the Nation's efforts to forward our understanding of the cosmos. Pathways to Discovery in Astronomy and Astrophysics for the 2020s identifies the most compelling science goals and presents an ambitious program of ground- and space-based activities for future investment in the next decade and beyond. The decadal survey identifies three important science themes for the next decade aimed at investigating Earth-like extrasolar planets, the most energetic processes in the universe, and the evolution of galaxies. The Astro2020 report also recommends critical near-term actions to support the foundations of the profession as well as the technologies and tools needed to carry out the science.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!