National Academies Press: OpenBook

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

Chapter: 4 Optimizing the Science: Foundations

« Previous: 3 The Profession and Its Societal Impacts: Gateways to Science, Pathways to Diversity, Equity, and Sustainability
Suggested Citation:"4 Optimizing the Science: Foundations." 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 125
Suggested Citation:"4 Optimizing the Science: Foundations." 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 126
Suggested Citation:"4 Optimizing the Science: Foundations." 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 127
Suggested Citation:"4 Optimizing the Science: Foundations." 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 128
Suggested Citation:"4 Optimizing the Science: Foundations." 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 129
Suggested Citation:"4 Optimizing the Science: Foundations." 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 130
Suggested Citation:"4 Optimizing the Science: Foundations." 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 131
Suggested Citation:"4 Optimizing the Science: Foundations." 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 132
Suggested Citation:"4 Optimizing the Science: Foundations." 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 133
Suggested Citation:"4 Optimizing the Science: Foundations." 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 134
Suggested Citation:"4 Optimizing the Science: Foundations." 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 135
Suggested Citation:"4 Optimizing the Science: Foundations." 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 136
Suggested Citation:"4 Optimizing the Science: Foundations." 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 137
Suggested Citation:"4 Optimizing the Science: Foundations." 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 138
Suggested Citation:"4 Optimizing the Science: Foundations." 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 139
Suggested Citation:"4 Optimizing the Science: Foundations." 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 140
Suggested Citation:"4 Optimizing the Science: Foundations." 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 141
Suggested Citation:"4 Optimizing the Science: Foundations." 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 142
Suggested Citation:"4 Optimizing the Science: Foundations." 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 143
Suggested Citation:"4 Optimizing the Science: Foundations." 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 144
Suggested Citation:"4 Optimizing the Science: Foundations." 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 145
Suggested Citation:"4 Optimizing the Science: Foundations." 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 146
Suggested Citation:"4 Optimizing the Science: Foundations." 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 147
Suggested Citation:"4 Optimizing the Science: Foundations." 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 148
Suggested Citation:"4 Optimizing the Science: Foundations." 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 149
Suggested Citation:"4 Optimizing the Science: Foundations." 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 150
Suggested Citation:"4 Optimizing the Science: Foundations." 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 151
Suggested Citation:"4 Optimizing the Science: Foundations." 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 152
Suggested Citation:"4 Optimizing the Science: Foundations." 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 153

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.

4 Optimizing the Science: Foundations Building on Chapter 3, which describes the human investments and public impacts of the decadal survey’s program, and how to enhance and diversify them, this chapter focuses on the essential scientific foundations, specifically the resource infrastructure and the underpinnings of how scientists turn ideas and data into discovery. This chapter draws from the Enabling Foundations panel report and identifies challenges and opportunities for progress over the next decade. Astronomy and astrophysics have achieved breathtaking accomplishments over recent decades. Much of the credit for this sustained record of success can be attributed to diversified portfolios of investment by NASA, the National Science Foundation (NSF), and the Department of Energy (DOE), ranging from cutting-edge flagship observatories in space and on the ground to investments in mid-scale and smaller supporting facilities, and a foundation of support for the calibration, analysis, interpretation, and theoretical modeling of the rich data sets produced by these facilities. Although the decadal process primarily focuses on recommendations to the federal agencies that sponsor it, astronomy in the United States has also historically benefited from major investment by state universities, private universities, philanthropic foundations, and individual donors, in addition to the federal government. Private foundations and philanthropy play many roles in astronomy, from supporting observational facilities and large projects, to individual principal investigator (PI) support and postdoctoral fellowships (e.g. the Heising-Simons 51 Peg postdoctoral fellowships), and key seed funding for future projects. Over the last two decades private foundations are increasingly playing a key role supporting individual researchers, either directly through grants programs or through institutional support (e.g. the Kavli centers, Carnegie Observatories, or the Simons Center for Computational Astrophysics). For example, the Brinson, Guggenheim, Heising-Simons, Kavli, Moore, Packard, Research Corp., Simons, Sloan, and Templeton Foundations all provide significant funding to individual astrophysics researchers, including early career postdocs. This is critical support for U.S. astrophysics research in theory, computation, and support for archives in a time of increasing competition for federal funding. This foundation of smaller-scale investments often receives less attention in decadal planning exercises, but its importance cannot be overstated. The most complex and precise measurements would mean nothing without pipelines to calibrate and process the raw data, algorithms to analyze and interpret the data, theoretical calculations to provide a context for the results and help understand their implications, and support for people who do the science. Laboratory astrophysics measurements, data science and computational methods, and data archiving all play critical roles in turning photons, particles, or waves into scientific insights. These foundational programs have the potential to bring more people into the field through reducing barriers to participation by anyone through supporting their success, and through offering access to state-of-the-art tools, training, and facilities. Such programs provide the seed corn for the future innovators and leaders in the profession. 4.1 THE IMPORTANCE OF A BALANCED PROGRAM Federal astronomical research investments in the United States today can be subdivided into a few critical components. The largest federal investments are in major flagship observatories and facilities, PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-1

such as the original NASA Great Observatories (the Hubble Space Telescope, Chandra X-ray Observatory, Spitzer Space Telescope, Compton Gamma-Ray Observatory), and now the James Webb Space Telescope (JWST) and the Nancy Grace Roman Space Telescope, or, in ground-based facilities funded through the NSF Major Research Equipment and Facilities Construction (MREFC) line (e.g., Gemini, the Atacama Large Millimeter/submillimeter Array [ALMA], the Daniel K. Inouye Solar Telescope [DKIST], Vera Rubin Observatory, Laser Interferometer Gravitational-Wave Observatory [LIGO], and IceCube). A second critical component of the nation’s observational capabilities is the suite of smaller-scale facilities and dedicated survey instruments, with examples of the former including NASA Explorer-class missions and examples of the latter including the Sloan Digital Sky Survey (SDSS). Many of these mid-scale facilities are funded and operated by partnerships between public agencies, private institutions, and foundations. Both classes of facilities have delivered breakthrough discoveries including Nobel Prizes, along with thousands of smaller individual and team-led investigations that collectively have fueled the extraordinary success and growth of the field in recent decades. The other essential components of the national investment portfolio are the people who carry out and drive the science (addressed separately in Chapter 3), and the enabling foundation or infrastructure that supports research. The data produced by the aforementioned facilities would be lost without investment in this foundation: support for processing and archiving these data, analyzing and interpreting the data, theoretical modeling and simulations, and in many cases carrying out the laboratory and computational analysis of the atomic, molecular, and chemical signatures and diagnostics of the emitting processes. The aggregate investment in this foundational support comprises a small fraction of the overall agency portfolios, but it multiplies by several-fold the scientific yield from the facilities. The need for balance across this broad portfolio has long been recognized by the agencies themselves. A prime example is the wide range of mission size classes supported by NASA through its Astrophysics Science Mission Directorate, ranging from flagships to small and medium Explorers, and extending down to small satellites, CubeSats, and support for balloon and rocket payloads. This healthy mix enhances scientific balance, and cost effectiveness. It also provides pathways for new and early- career investigators to build scientific, technical, and leadership experience with progressively larger mission classes. Another commendable example is the provision in NASA mission lines for the costs of data processing, analysis, and archiving, and in the case of flagship missions, support for guest observer grants, which encourage the timely analysis, publication, and dissemination of the data and science resulting from the missions, including theoretical studies needed to interpret the data. The funding portfolio of the NSF Division of Astronomical Sciences (AST) is unlike that of most other NSF divisions, including those in the NSF Mathematical and Physical Sciences Directorate (MPS). Astronomy has a long history of capitalizing on shared, communal infrastructure that far exceeds the capability of any one individual institution, and that, for NSF facilities, is accessible by anyone. Under this model, a considerable fraction of the AST budget supports national observatories and facilities including the National Optical-Infrared Astronomy Research Laboratory (NOIRLab) and the National Radio Astronomy Observatory (NRAO), including U.S. participation in ALMA. These facilities support comprises about 75 percent of the total AST budget, with the remainder available to support individual investigator research, mid-scale infrastructure programs, and division-specific education activities. The fraction of research funding (<25 percent) is far lower than those of other MPS divisions (45-95 percent), but reflects in part the disproportionate number of shared national facilities in ground-based astronomy.1 The current model through which MREFC projects are funded for operation by divisions has resulted in an unbalanced program in NSF’s AST division that is not sustainable. The structural problem is addressed in Chapter 5 while this chapter focuses on balance in specific programs. Although the primary role for advising agencies on their funding portfolios on an ongoing basis rests with the various agency and National Academies-administered standing committees, it is appropriate that the decadal surveys assess in broad terms the impacts of the current balances, and where appropriate, 1 National Science Board, 2018, Study of Operations and Maintenance Costs for NSF Facilities, NSB-2018-17, Alexandria, VA, https://nsf.gov/pubs/2018/nsb201817/nsb201817.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-2

to identify areas where a more healthy balance could be achieved. Here “healthy” is regarded as a balance of program investments that optimize the overall scientific productivity and future sustainability of the enterprise. This follows the practice of recent decadal surveys, and some of the findings and recommendations here will echo those already noted in the 2000 and 2010 surveys. When addressing the question of balance, the Panel on an Enabling Foundation for Research and the steering committee identified a few critical areas where evolution in the funding balances within NSF AST and NASA Astrophysics has drifted into unhealthy territory, or, where the evolution of the research landscape itself has led to the need for enhanced investment in emerging disciplines. Chief among these are support for investigator grants for research and data analysis, and in the infrastructure support for data archiving, processing, and analysis as well as the related needs in computation, software, and data science. Previously identified under-investments in laboratory astrophysics and theory remain as critical needs. Each of these areas is addressed in the remainder of this chapter. 4.2 ENABLING SCIENCE THROUGH A HEALTHY INDIVIDUAL INVESTIGATOR GRANTS PROGRAM It is people who are the source of American scientific and technical prowess, and supporting those scientists is the way to realize the scientific visions that are put forward in Chapter 2, A New Cosmic Perspective. Access to world-leading facilities is not enough to produce science. Individual scientists need access to the financial resources that allow them to collect, analyze, and interpret data from those facilities. That funding unlocks the effort of scientists and trainees to explore new ideas or to execute the hard but important projects that drive the field forward. Without resources, however, scientists’ insights and talent lie unrealized and discoveries unmade. Chapter 3 emphasized the need to collect demographic information from researchers in external grant programs to assess indicators pertaining to outcomes of proposal competitions. A lack of data is apparent here as well; proposal success rates for only a few programs were available, and not always the most recent data. NSF noted that it is against their policy to release any information about proposals that have not resulted in awards; moreover, a recent policy prohibits the public release of proposal selection rates, so the number of submitted proposals and total request amounts were not made available to the survey. While NASA collects some data on proposers, the agency has only started to assess and evaluate it in a systematic way. Having these data would have better informed this report. Conclusion: The lack of publicly available data on proposal success rates by program and on aggregated metrics for who and what type of research is being supported (theory, facilities, laboratory investigations, investigator demographics, student vs. postdoc funding for example) hampers analysis, evaluation, and oversight. Recommendation: The National Science Foundation, NASA, and the Department of Energy should release data on proposal success rates on an annual basis, and should track metrics that allow them to analyze statistically what is being supported. 4.2.1 Bolstering the Individual Investigator Grants Program Funding for the majority of astronomy research flows through “individual investigator grants,” where the lead scientist proposes a specific project and asks for the needed resources (salary for trainees, summer salary for senior personnel in academic positions, computing, travel, etc.) to bring the project to fruition. These proposals take a variety of forms. Ground-based astronomy research and theory is funded through the NSF Division of Astronomical Sciences, with research for “individual investigators” provided through its Astronomy and Astrophysics Research Grants (AAG) program. NASA funds research that is PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-3

relevant to space-based astrophysics missions, primarily through the Astrophysics Research and Analysis Program (APRA), Astrophysics Theory Program (ATP), Astrophysics Data Analysis Program (ADAP), and the Exoplanet Research Program (XRP). NASA also provides support for data analysis for U.S. investigators who have successful Guest Observer (GO) proposals on some of its active missions, as well as funding preparatory work for some future missions. There are also funding opportunities for theory and archival work connected to specific missions (currently Hubble, Fermi, Swift, and Chandra). In this NASA GO funding model, a successful research proposal for observing time is considered sufficient to unlock funding when the observing proposal is approved and executed. For ground-based NSF funded research, however, the funding is unconnected to awards of observation time, even on NSF-funded facilities. The NSF AAG program is a cornerstone of the enabling foundation for research in astronomy and astrophysics in the United States. It supports research projects across nearly all subfields of the astrophysical sciences, and most of its funding supports individual investigators and their groups. Proposals are rigorously reviewed and the short funding durations for grants (typically 3 years) ensures that funding priorities reflect the most important scientific priorities in the field. The grants have led to discoveries that have transformed astronomy. For example, the work tracking stars in the immediate environment of the Milky Way Galaxy’s black hole (Figure 4.1) began receiving NSF funding in the early part of the 2000s, and was recognized with a Nobel Prize in Physics in 2020. Permission Pending FIGURE 4.1 The change in position of two stars at the galactic center around what is now confirmed to be a supermassive black hole at the center of the Milky Way galaxy. This data was taken with the Keck telescope over a time period of 1995-2014, combining two decades of speckle imaging and adaptive optics data. This work was supported by NSF AST individual investigator grants. These improved mass and distance estimates were crucial for cementing the black hole explanation for Sgr A*, the subject of the 2020 Nobel Prize in Physics. SOURCE: Boehle et al. (2016), http://www.astroexplorer.org/details/apjaa2b70f5. Preparing proposals for individual investigator grants is extremely time consuming, and—given the large impact a successful proposal has a large impact on a scientist’s output and career—the stakes are typically high. NSF AAG proposal success rates averaged 30-50 percent in the early 1990’s through the early 2000’s (Figure 4.2), during which time most scientists had an expectation that their work could be funded within a reasonable time frame, perhaps after one or two resubmissions. However, funding rates PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-4

began to decline to below 30 percent beginning in the mid-2000’s, and then even further to around 15 percent in the early part of the past decade. While funding rates have recovered somewhat, from 2010- 2018 the average AAG success rate has remained around 18 percent, far below the 30 percent success rate target recently identified by NSF as a goal for the foundation overall.2 FIGURE 4.2 Plots of NSF AST Astronomy and Astrophysics Grant Budget and proposal funding rate versus time, from 1990-2018 in real-year dollars. The increase in funding in 2009 originated from the American Recovery and Reinvestment Act. SOURCE: Based on data from R. Gaume, National Science Foundation, presentation to the steering committee on July 15, 2019. In 2015, a study group of the Astronomy and Astrophysics Advisory Committee (AAAC) investigated the impact of declining success rates at both NASA and NSF astrophysics programs on scientific productivity (Figure 4.3).3 They concluded that the decline in success rates was not related to changes in the average proposal quality or to the fraction of proposals judged to be highly deserving of funding. Reviewers are instructed to grade proposals on an absolute scale from E (excellent), V (very good), G (good), F (fair), P (poor), where an Excellent proposal is an “Outstanding proposal in all respects; deserves highest priority for support” and a Very Good proposal is a “high quality proposal in nearly all respects; should be supported if at all possible.” The fraction of proposals judged to be highly deserving of funding (VG, VG/E, E) has remained stable from year to year. However, the success rate of proposals ranked VG dropped from 45 percent in 2007-2008 to 25 percent in 2012. In other words, three out of four proposals that were judged as nearly certain to result in high quality science are rejected each year. The AAAC group’s quantitative analysis revealed that the major factor driving the increase in proposal oversubscription was that the budgets for these programs have not kept up with the increase in the number of (unique) proposers. This increase in the number of investigators tracks the overall addition 2 https://www.aip.org/fyi/2021/panchanathan-makes-case-nsf-expansion-appropriators 3 Cushman et al., 2015, “Impact of Declining Proposal Success Rates on Scientific Productivity,” AAAC Proposal Pressures Study Group, https://arxiv.org/abs/1510.01647. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-5

of researchers to the field, with no significant change to the mix of career stages over time. Individual budget items (typically dominated by salaries and tuition) have also increased in cost due to inflation, increasing the cost of funding a constant level of proposed effort. FIGURE 4.3 Historical NSF AST individual investigator grant statistics, from 2000-2015. There is a rise in the number of proposals submitted over this decade and a half, while the number of awards has not increased concomitantly. SOURCE: Cushman et al. (2015), Figure 1, https://arxiv.org/abs/1510.01647. Reproduced with permission. The decrease in AAG proposal success rate is attributed by a 2018 National Science Board (NSB) report in roughly equal measure to the increase in the number of submitted proposals and the decrease in available funds because of the increase in facilities operations costs with a nearly flat AST budget. 4 AST stands out in the MPS directorate for having both a low proposal success rate and spending the least amount of its budget on individual grants programs. Finding: There is a systematic tension between funding facilities’ operations and maintenance, and supporting the work of scientists able to turn data into discovery. The imbalance in AST has worsened over the last decade and will impact the ability to adequately support new facilities and new science going forward. This issue is developed in more detail in Chapter 5. 4 National Science Board, 2018, Study of Operations and Maintenance Costs for NSF Facilities, NSB-2018-17, Alexandria, VA, https://www.nsf.gov/pubs/2018/nsb201817/nsb201817.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-6

FIGURE 4.4 Money spent on grants to investigators as a function of the total NSF Division budget, for divisions within the Geosciences (GEO) and Math and Physical Sciences (MPS) directorates, for FY2018. The color coding indicates into which directorate the division falls, while the size of the symbol is proportional to the proposal success rate (in FY2020). Dashed lines indicate the fraction of the division budget devoted to research. The AST division stands out both for the low fraction of the division’s budget used for research, as well as the low proposal success rate (24%). NOTE: AST=Astronomical Sciences; PHY=Physics; CHE=Chemistry; DMR=Materials Research; DMS=Division of Mathematical Sciences; OCE=Ocean Sciences, AGS=Atmospheric and Geospace Sciences; EAR=Division of Earth Sciences. SOURCE: R. Osten, NSF. based on data from https://www.nsf.gov/about/budget/fy2020/pdf/27_fy2020.pdf, https://www.nsf.gov/about/budget/fy2020/pdf/26_fy2020.pdf, https://www.nsf.gov/funding/funding- rates.jsp?org=MPS, https://www.nsf.gov/funding/funding-rates.jsp?org=GEO. The AAAC report noted that the historical proposal success rates of 30-35 percent achieved for NSF AAG funding prior to and including FY 2003 was a healthy competitive environment, where the average proposer faced a manageable level of risk (~30 percent) of no funding after three attempts. Over the entire foundation, roughly 30 percent of proposals are ranked highly meritorious, and recent initiatives by the NSF Director are focused on achieving a grant proposal success rate of 30 percent.5 Other NSF divisions that share common features with astronomy, such as physics and oceanography both being heavy users of the MREFC line, have higher proposal success rates than the astronomy division and devote a larger fraction of their budget to supporting individual investigator grants (Figure 4.4). Differences in culture between different scientific fields may also contribute to this disparity, as some fields deprecate multiple proposal submission by a particular research group responding to a proposal call. Success rate alone is not the only factor to consider, as grants need to be of sufficient size to carry out the proposed science project. 5 https://www.aip.org/fyi/2021/panchanathan-makes-case-nsf-expansion-appropriators PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-7

As pointed out in the Enabling Foundations for Research report, many programs within the NSF Physics Division (PHY) have success rates higher than 30 percent. Such levels also offer opportunities for early-career researchers proposing for the first time to have a realistic chance of success. The same AAAC study showed that when success rates dropped to lower levels first-time proposers fared even less well, with success rates as low as 7 percent. Chronic underfunding carries the additional risk of stagnation of the field. The analysis above has focused on NSF grants, which are the sole way of publicly funding peer- reviewed science with ground-based telescopes. As noted in the Enabling Foundation report, while oversubscription rates for some NASA programs are healthier (e.g., APRA), support for other programs (e.g., ATP for theory) also appears to suffer from similar high proposal pressure and underfunding. The Enabling Foundation report suggested a 20 percent increase in funding above inflation for all individual investigator grants programs to restore success rates to a healthy competitive environment. This underfunding has also impacted equity within the field. Ensuring adequate funding enables the whole community to reap the benefits from federally funded facilities. In the absence of public funding, scientists at wealthy institutions may still be able to tap into other sources of funds to support their research (while benefiting from institutional support for graduate students or endowed postdocs), or to carry out preparatory work to ensure successful proposals. However, these paths are largely out of reach for those at less affluent institutions, shutting them out of using the facilities the nation has invested in. This wastes decades of investment in, training of, and effort by flourishing scientists—by not providing the conditions in which they can execute their plans for science. Conclusion: Robust individual investigator grant funding is crucial to meet the science challenges described in the Cosmic Ecosystems; New Messengers, New Physics; and Worlds and Suns in Context science themes. The historical proposal success rate for individual investigator grants within the NSF AST division of around 30 percent, realized at the start of the millennium for astrophysics programs, strikes an appropriate balance between a healthy competitive environment and a good chance of eventual success with resubmission. By any of several metrics which can be used to judge a healthy level of competition for individual research grants—dollar amounts spent on research, percentage success rates, fraction of high-quality proposals being rejected—the current state of individual investigator grants within the NSF AST division is not at a healthy level. Increasing grant funding is also required to ensure more equitable access to resources. Recommendation: The National Science Foundation should increase funding for the individual investigator Astronomy and Astrophysics Research Grants by 30 percent in real dollars (i.e., above the rate of inflation) over 5 years from 2023-2028 starting with the fiscal year 2019 budget inflated appropriately. This will have the effect of restoring success rates to a healthy competitive level. This funding augmentation is needed to reach the 30 percent proposal success rate goal, justified both from analysis of other programs and areas at NSF as well as being consistent with NSF’s stated goal. 4.2.2 The Importance of a Healthy Theory Foundation Theory is crucial in astrophysics, as both a mechanism for driving new discoveries and a framework for interpreting essentially all signals received from space. The focus of modern theoretical research has increasingly expanded from traditional pencil-and-paper calculations to complex computer simulations and sophisticated statistical analyses. Understanding prize-winning discoveries such as the Cosmic Microwave Background anisotropy and gravitational waves from merging black holes would not have been possible without the conceptual framework provided by theory. Indeed, breakthroughs in PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-8

theoretical predictions of the characteristics of the gravitational wave signal produced by black hole mergers were critical to the Nobel Prize-winning discovery based on LIGO detections (Figure 4.5). The science themes presented in Chapter 2 demonstrate that theory and observation are intertwined, necessitating a multi-pronged approach to addressing these important topics: a theoretical understanding of how gas of different temperatures and densities can co-exist in galactic outflows of differing velocities, for example, is essential to examining the processes that link matter inside of galaxies with its outside environs. As another example that relates to the priority science area of Pathways to Habitable Planets, theoretical calculations of planetary atmosphere chemistry and evolution will be needed to interpret biosignature gases detected in exoplanet spectra. This theoretical research lays the groundwork for designing new observational programs and planning for new facilities. FIGURE 4.5 LIGO gravitational wave data from its two observatories at Livingston, Louisiana, and Hanford, Washington, from the first gravitational wave detection of merging black holes. The lower panels shows the LIGO gravitational wave signal from its two observatories at Livingston, Louisiana (in blue) and Hanford, Washington (in orange, shifted by 7 milliseconds), from the first gravitational wave detection of merging black holes. In the top panel, the signal from the Livingston Observatory is shown with a numerical theoretical model for two merging black holes, each about 30 times the mass of the Sun, lying 1.3 billion light-years away. SOURCE: Caltech/MIT/LIGO Laboratory, https://www.ligo.org/detections/GW150914.php. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-9

Low funding rates at both NASA and NSF have affected the ability to carry out theoretical investigations. For NASA’s ATP program, which funds theory relevant to NASA’s missions, proposal funding rates dropped from 17 percent in 2010 to 14 percent by 2013.6 Since 2015, the program has moved to a 2-year proposal cadence, but proposal success rates still remain low (22 percent in FY 2019).7 When the longer cadence is coupled with low success rates, scientists have little realistic expectation that their research will be funded while it is most relevant, if it is ever to be funded. The NSF AAG program discussed in the section above is a crucial vehicle for funding new independent and novel investigations in all fields of astronomy and astrophysics, but especially in theory. Lack of data on success rates of different proposal types prevents an assessment of how well this program supports theoretical investigations. Recommendations for augmenting support for theoretical investigations have appeared in multiple previous astronomy decadal surveys, signaling a perennial under-commitment to supporting this mode of scientific inquiry. Most recently, the Astro2010 recommended that funding for NASA’s Astrophysics Theory Program be increased by 25 percent, but instead the budget remained flat, and the calls for proposals slowed to a 2-year cadence. When coupled with current extremely low proposal success rates, these changes have particularly hurt the career development of pre-tenure theorists. The Panel on an Enabling Foundation for Research took up the earlier suggestion of a 25 percent increase in funding, and additionally suggested that the ATP be returned to an annual cadence. New Worlds, New Horizons also recommended the creation of a new inter-agency funding opportunity called Theory and Computation Astrophysics Networks (TCAN). This program was intended to respond to the facts that (1) theoretical and computational problems have reached a scale and complexity that requires more sustained funding of larger teams than a standard NSF AAG or NASA ATP grant; and (2) that many of the most important theoretical problems transcend the artificial boundaries of the three agencies (NSF, NASA, and DOE). The resulting TCAN concept would have supported 5-year programs that would be jointly supported by all three agencies. However, after one funding cycle only NASA has continued to participate in the program. Finding: A strong foundation of theoretical research remains critical for interpreting astrophysical observations and planning new facilities, but past decadal survey recommendations for supporting theory have not been implemented. Conclusion: Theoretical investigations are necessary to extract the full scientific intent of new and existing facilities, and funding for such studies must increases to recover from limited funding in the past. Recommendation: Given the foundational importance of theory to the astronomical enterprise, NASA’s Astrophysics Theory Program should resume an annual cadence, and receive a 30 percent funding augmentation in real(inflation-adjusted) dollars over 5 years from 2023-2028 starting with the fiscal year 2019 budget inflated appropriately. 4.2.3 Maximizing the Science through Large Programs Utilizing Ground-Based Facilities. The method of funding data analysis for NSF-supported projects is significantly different, and at times more problematic, than the funding for NASA projects. Observing time on a NASA facility is accompanied by funding to support data analysis and the creation of high-level data products. This ensures that observers have the resources they need to accomplish the proposed science. For example, the 6 E. Scannapieco, NASA, presentation to EF panel, and NASA written communication to the steering committee on February 3, 2020. 7 Ibid. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-10

Hubble Space Telescope mission allocates nearly a third of its annual operations budget to associated research funding; this in turn supports a robust rate of published papers, which peaked in 2019 at 1014 refereed papers that year. The lack of equivalent funding for time awarded on NSF-supported ground- based facilities means that researchers must apply for funding through the AAG proposal process, typically after the observing time is approved. This multi-step process regularly produces delays of two or more years between when observations are approved and when funded scientific personnel can formally join the analysis. The process is also highly inefficient, requiring at minimum two separate rounds of proposals and reviews, once for a telescope time allocation committee and once for a grants panel. Even with higher AAG proposal funding rates, the time delay and gauntlet of multiple proposal reviews add significant inefficiencies that hamper the scientific output of the most powerful facilities. Finding: Associating research funding for data analysis and production of high level data products with awarding of observing time ensures that observers have the resources they need to accomplish the proposed science. The NASA model does not translate easily to the ground, however, where weather and observing conditions add significant uncertainty to program completion rates, particularly for smaller observing programs. It would, however, be appropriate for and significantly increase the scientific impact of the subcategory of large projects that exist for current and future MREFC-class astronomical facilities. These MREFC-funded facilities are a significant investment of federal dollars and are motivated by a few major scientific objectives. Large or key projects are the programs that have been established by peer review to be the most important science priorities for a given telescope. They require a large investment in observing time and there is often an expectation of dissemination of results through paper publishing, and catalogs or data releases. These programs are typically given the observing resources to reach a high degree of completion. There are only a small number of these large programs for MREFC-class astronomical facilities, with typically three to five large programs approved per year on each, with not all programs having significant U.S. participation. Large programs are sufficiently competitive that funding panels are naturally reluctant to award funding to an ambitious proposal that has not yet actually been approved. Thus, projects are not likely to have their challenging processing and analysis funded in a timely manner, which delays the science deemed to be especially important. The net result is that enormous barriers exist for producing the most compelling, legacy science, particularly in the early stages of large projects, when data collection and reduction is most intensive, but funding is 2 or more years out of reach. Allowing these large programs to immediately submit supporting budgets to the NSF AAG would give U.S. investigators a way to quickly ramp up reduction and analysis, and to focus their energy on producing science rather than yet another proposal to the AAG program. This approach would also help to attract a larger, more diverse user base to NSF facilities, given that lack of funding is a larger barrier for scientists from under-resourced institutions, who are not well-positioned to tackle ambitious projects that would need years of up-front effort before NSF funding can be secured. Recommendation: The National Science Foundation Division of Astronomical Sciences should establish a mechanism of associated research funding for data analysis and production of high level data products for large principal investigator-led programs on MREFC-scale astronomical facilities in order to accelerate the scientific output and maximize the timeliness and community impact of these key large projects. Given the small number of large MREFC-scale programs, this recommendation could be accommodated within the AAG increase. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-11

4.3 BREAKING DOWN CROSS-AGENCY BARRIERS Astrophysical questions increasingly transcend traditional wavelength, division, and agency boundaries. This richness in scientific perspective unfortunately is accompanied by logistical complications in funding and project management. Historical boundaries between organizations can raise a particularly high barrier to fundamentally interdisciplinary efforts, such as solar physics. Within the last decade or so, the emergence of new subject areas has magnified these inconsistencies: the study of exoplanet atmospheres requires knowledge not only of temperatures and pressures appropriate there, but also knowledge gleaned from studies of planets in the Solar System, and indeed detailed understanding of processes at work on Earth; within the framework of NASA’s Science Mission Directorate, this endeavor potentially spans three distinct directorates. The rise of multi-messenger astronomy unites information gathered across the electromagnetic spectrum, from ground and space, as well as new carriers like gravitational waves (Figure 4.6) and neutrinos, requiring a breakdown of traditional funding silos to fully realize the science return. These new ways of approaching science can potentially be funded by multiple entities, but simultaneously risk not being funded by any if no group feels they “own” the science. These barriers can be transcended through dedicated programs like the Windows on the Universe (WoU) Initiative (which jointly reviews proposals between NSF’s physics and astronomy divisions), or like the TCAN theory program recommended in the 2010 decadal survey (Section 4.2.2). These programs need thoughtful guidelines and execution to be successful in practice. NASA has started taking steps to identify disciplines needing interdivisional research and/or interagency partnerships and coordinating technology development across multiple disciplines. Foundational cross-agency issues affect fields with new and exciting results like neutrino astrophysics, gravitational wave astronomy, and particle astrophysics. These are relevant to astronomy research but are funded through other divisions. Neutrino astrophysics and gravitational wave astronomy are primarily funded out of NSF PHY and/or Office of Polar Programs. As described in more detail in Chapter 7, the survey committee is endorsing (but not ranking) the IceCube-Gen2 neutrino large facility and technology development for next generation gravitational wave observatories largely because of the benefit to the field of astronomy. Conclusion: Effective mechanisms to fund cross-cutting research at NSF, NASA, and the DOE would accelerate scientific results. 4.4 SOLAR PHYSICS Solar physics is directly relevant to astronomy. As the nearest star, the Sun is both a key calibrator for our understanding of stellar astrophysics, and a unique laboratory for understanding magnetism and its coupling to mass, which is relevant through the universe. The Sun is also an important input to understanding Earth’s climate and space weather. In the next decade, solar observations and theory will be key ingredients in understanding the Earth-Sun connection and its implications for the co- evolution of stars and planets throughout the Milky Way Galaxy, particularly given the impact that eruptive events and high energy emission of light and particles can have on planetary atmospheres. While the new high-resolution capabilities of DKIST will surely transform our understanding of the Sun, there remains a pressing need for complementary global measurements of the entire Sun and its magnetic activity. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-12

FIGURE 4.6 The multi-messenger nature of the detection of the kilonova 170817, first detected in gravitational waves and gamma ray bursts, and shortly thereafter in many other wavelengths. In the left panel, green contours indicate location determination from gravitational wave detectors (LIGO in light green, LIGO-Virgo in dark green); light blue contours delineate likely regions using triangulation from time delays between gamma-ray satellites Fermi and the INTErnational Gamma-Ray Astrophysics Laboratory (INTEGRAL); and dark blue contours trace the Fermi Gamma-ray Burst Monitor localization. The insets on the right show optical images of the host galaxy NGC 4993 10.9 hours after the merger taken with the Swope telescope (top right) and a DLT40 pre-discovery image 20.5 days prior to the merger (bottom right). SOURCE: B. P. Abbott et al. 2017, “Multi-messenger Observations of a Binary Neutron Star Merger,” The Astrophysical Journal Letters, 848 L12. doi:10.3847/2041-8213/aa91c9. Observations of the Sun depend on facilities spanning multiple federal agencies, even different directorates within the same agency, and these groups take advice from different decadal surveys. Solar ground-based observations are done with a mix of solar-dedicated facilities (such as the Mauna Loa Solar Observatory [MLSO] and the Expanded Owens Valley Solar Array [EOVSA]), as well as general purpose astrophysics facilities like the JVLA and ALMA that have solar-capable instruments. The field of ground- based solar physics is funded by two different NSF divisions: Astronomical Sciences (within the Mathematical and Physical Sciences Directorate) as well as Atmospheric and Geospace Sciences (residing in the Geosciences Directorate). Space-based heliophysics research at NASA is the domain of the Heliophysics Division, distinct from the Astrophysics Division. The National Oceanic and Atmospheric Administration (NOAA) also oversees space weather prediction capabilities and is another federal agency relevant to the subject. The direction for investments in space-based assets are prioritized by the solar and space physics decadal survey process, while this astronomy and astrophysics decadal survey committee advises only the division of Astronomical Sciences at NSF about ground-based solar physics. This mix of different solar observation regimes, each controlled by a separate decadal process, was a topic of attention in Astro2010, which had recommended that NSF work with several communities to determine the best route to a balanced and effective ground-based solar program maintaining PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-13

multidisciplinary ties. In the current decadal process (which can only recommend ground-based components of the solar observation program), only three white papers describing solar facilities were submitted to this decadal survey.8 These mid-scale ground-based solar projects emerged with favorable reviews from the OIR and RMS panel reports, however the solar mid-scale projects are not considered amongst the strategic initiatives called out in Chapter 7. The survey committee was not able to give proper perspective as to how these facilities will support and enhance the broad range of multi-agency activities currently underway in solar physics, as these are the domain of the solar and space physics decadal survey. Advancing these myriad scientific goals is most efficiently done utilizing a comprehensive approach. Conclusion: The most appropriate role for future astronomy and astrophysics decadal surveys is to comment on the value of ground-based solar physics projects for astronomy and astrophysics scientific priorities. For consideration of these projects in the context of the full range of multi- agency activities in solar physics, the solar and space physics decadal survey is the more appropriate body to prioritize and rank them. 4.5 THE DATA FOUNDATION Through much of the past 150 years, the majority of astronomical observations were held by individuals or institutions, archived on photographic plates or data tapes. By 2020 this landscape has been completely transformed. All modern data are digital, and a significant portion are archived in publicly- accessible on-line data libraries. The scientific importance and impact of these archives is fundamental. For the past 15 years, for example, publications from archival use of data from the Hubble Space Telescope have outnumbered those by the original proposing teams (Figure 4.7), with comparable numbers of citations,9 and other major facilities are seeing similar trends. The empirical evidence it that curating scientific data in well-organized archives enables multiple repurposings and extends the useful lifetime of the data (Figure 4.8). Astronomy is entering a second wave of this data revolution, with increasing numbers of survey facilities largely dedicated to producing archival data sets from the outset, which are subsequently shared by thousands of users for myriad individual scientific projects. For two decades, the Sloan Digital Sky Survey has been a ground-breaking precursor of this new mode of survey astronomy. In space, NASA’s Infrared Astronomical Satellite and Wide-Field Infrared Survey Explorer’s all-sky surveys created data sets that have lasting value to this day. More recently, the European Space Agency (ESA) Gaia observatory, which is measuring precise positions and proper motions for a billion stars, has revolutionized Milky Way and stellar astrophysics (Chapter 2). Although it was fully built and supported by Europe, its data archives are openly accessible worldwide, and have supported hundreds of investigations by U.S. astronomers in the 5 years since the first data release. In the coming decade the Vera Rubin and Nancy Grace Roman Observatories, the highest-priority ground and space projects in the 2010 decadal survey, respectively, will provide comparably rich data sets, which promise to revolutionize time domain astronomy and promise breakthrough discoveries across a wide range of astrophysical disciplines. They will also bring unprecedented volumes of data—of order 500 Petabytes (500 million Gigabytes) by the end of 2030 collectively across all observatories and missions, several orders of magnitude more astronomical data than has been collected in human history. When combined with the increasing availability of data from other, mid-scale facilities, the very nature of the observational research enterprise is evolving. In short, progress in astronomy requires fully preparing for the next phase 8 ngGONG (Hill et al. 2019BAAS…51g..74H), COSMO (McIntosh et al. 2019BAAS...51g.165M), FASR (Bastian et al. 2019astro2020U..56B). 9 Space Telescope Science Institute, https://archive.stsci.edu/hst/bibliography/pubstat.html, accessed May 18, 2020. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-14

of the on-going transition away from targeted observations to large public data sets, in order to maximize the science returns from current and upcoming facilities. FIGURE 4.7 History of publications arising from observations taken with the Hubble Space Telescope. The green curve gives the trend of refereed papers originating from the original proposing Guest Observer (GO) team; papers in purple have no overlap between the original proposing team and paper authors and indicate purely archival (AR) research uses; the aqua curve indicates a mix of GO and archival researchers. The fourth category indicates papers for which the assignment into the other bins cannot be made. The rate of archival paper production has outpaced that of GO paper production from the early days of the observatory, a product of the open archives and pipeline data processing. SOURCE: R. Osten, based on data available from STScI, https://archive.stsci.edu/hst/bibliography/pubstat.html. Along with the increasing importance of surveys and large data volumes, a related revolution in computational astrophysics is underway. Numerical simulations are playing an ever-growing role in modeling the physics of planets, stars, interstellar clouds and plasmas, galaxies, and the universe itself. Numerical simulation has become an essential skill set for many theoretical astrophysicists. The outputs from these simulations represent a valuable resource, but currently are rarely made publicly available, and will comprise a very significant data volume. Although many theorists and modelers use publicly available codes, far fewer people write or maintain them. The data revolution has also transformed the manner in which many astronomers conduct the majority of their research. Many observational astronomers rarely observe in person at a telescope, instead spending the bulk of their time developing methods to carry out sophisticated analyses of large online data sets. The algorithms used to process the data and create the results then become as important as the underlying observations. Mechanisms to share software, such as code-sharing and revisioning through Github and providing tutorials with worked examples through Jupyter notebooks, enable reproducibility and lead to further levelling of the field for access to and improvements on the motivating PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-15

science. Directly linking papers and the data contained therein in archives also strengthens the connection between the resultant science and the input observations. These revolutions are of course not unique to astronomy and astrophysics, and span many fields. “Harnessing the Data Revolution” is one of NSF’s “Ten Big Ideas,”10 initiatives in which NSF plans to make significant investments. NASA has recently convened a “Big Data Task Force”11 and released the Science Mission Directorate’s Strategy for Data Management and Computing for Groundbreaking Science 2019-2024.12 The recommendations for building the Data Foundation for Astronomy and Astrophysics presented in this section align well with these efforts. 4.5.1 Data Archiving, Curation, and Pipelines The importance of archiving, curating, and facilitating the use and analysis of these rich data sets has long been recognized by NASA and NSF, and numerous programs exist to support these areas. For NASA these include mission-specific support (Figure 4.8), support for archival data centers, and individual investigator programs such as the Astrophysics Data Analysis Program (ADAP) and Exoplanets Research Program (XRP). NSF supports data curation at its national observatories, and mandates a plan for managing data and sharing the products of funded research in individual investigator programs through its general AAG program. The vast network of private ground-based facilities have much more variable levels of archiving and accessibility. The question is whether the current suite of programs is fit for the 2020’s and beyond. While recognizing the many successes of the current programs, the decadal survey will focus on areas where modest investments can produce major scientific payoffs. Virtually every celestial photon collected by a telescope is a precious resource, capable of contributing to future discoveries. This legacy value can be maximized through investing in infrastructure that enables facilities to collect and reduce these data in a uniform manner, and that archives data to be easily retrievable, with the eventual goal of making the data publicly available. The need for high-level data processing is also being driven by the increasing complexity of instrumentation (e.g., integral-field spectrometers and multi-object spectrographs) in space and on the ground. The importance of joint analysis of observations from different facilities and wavelengths, and of sophisticated archiving with associated science platform tools, will grow dramatically over the next decade. A prime example is the measurement of cosmological constraints on dark energy and other parameters in the coming decade, which will rely heavily on the joint processing and analysis of data from the Euclid (ESA), Roman, and Rubin observatories. As detailed in previous chapters, the tremendous interest in multi-wavelength, multi- messenger, and time-domain analysis will pose new challenges over the next decade, as will carrying out any science project with unprecedented volumes of data. The current state of these data archives varies considerably, but the general trend is for an increasing role of archival data in scientific pursuits. The decade just completed saw an expansion of archive capabilities both on the ground and in space (Figure 4.7, 4.8, 4.9, 4.10, 4.11), and this is only expected to grow in the coming decade. The remote nature of space facilities mandated effective data storage from the outset, and perhaps not surprisingly, well-managed archives are available for nearly all major NASA missions. These data have a long duration impact: data taken from the early days of the Hubble Space Telescope still find productive uses in refereed papers nearly 30 years after initial acquisition. Seventy percent of data archived from early in the life of the Chandra X-ray Observatory appear in four or more publications (Figure 4.8). The situation for ground-based facilities is much more mixed. Large facilities such as those built and operated by the International ALMA Observatory and the European Southern Observatory (ESO), as well as surveys such as SDSS and the Panoramic Survey Telescope and Rapid Response System (Pan- 10 https://www.nsf.gov/news/special_reports/big_ideas/index.jsp. 11 https://science.nasa.gov/science-committee/subcommittees/big-data-task-force. 12 https://science.nasa.gov/science-red/s3fs-public/atoms/files/SDMWG%20Strategy_Final.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-16

STARRS) have established archives of quality rivaling those of the space observatories. These archives have been major contributors to the scientific productivity of those endeavors.13 Figure 4.9 and 4.10 detail the steady increase in archival usage of ESO Paranal telescopes and the ALMA Observatory, respectively; in both cases roughly a third of papers are now produced using at least some archival data. The effort put in by the ALMA Observatory to create data reduction and calibration pipelines has the result that currently 95 percent of data is calibrated and imaged;14 in addition to enhancing the archival utility of the data, such steps reduce barriers to entry for new users and widen ALMA’s user base. NOIRLab hosts an Astro Data Lab as a centralized hub for archiving and disseminating observations from U.S. nighttime OIR observatories, with emphasis on large surveys and data discovery tools, and the NASA-funded Keck Observatory Archive curates Keck data at the NASA Exoplanet Science Institute (NexScI). Such examples, however, have been the exception rather than the rule. Several factors account for this situation. Few privately-supported U.S. ground-based observatories are financially positioned or structurally incentivized to provide fully-reduced data products, and for older public facilities like the JVLA or VLBA with complicated data processing, such a goal may simply not be possible for all data in spite of best efforts (see Figure 4.11 for the increasing trend in archival usage from the JVLA). And while some facilities place their data into public archives, these resources are often difficult to tap. The net result is an opportunity lost, for the scientists who could be exploring data immediately rather than spending months reducing it or making new observations, for the observatories that invested in instruments whose data are underused, and for the science that could be done if that data could be easily accessed. FIGURE 4.8 The percentage of data published as a function of time for data taken from the Chandra X-ray Observatory archive, demonstrating the impact of a well-organized archive. Data here is quantified as exposure time. Here, 70 percent of the oldest data sets have four or more publications using the data. SOURCE: Courtesy of the Chandra Data Archive operations team. 13 M. Romaniello, M. Arnaboldi, C. Da Rocha, C. De Breuck, N. Delmotte, A. Dobrzycki, N. Fourniol, W. Freudling et al., 2016, The growth of the user community of the La Silla Paranal Observatory science archive, The Messenger, 163(5), http://www.eso.org/sci/publications/messenger/archive/no.163-mar16/messenger-no163-5-9.pdf. 14 ALMA, “Processing,” accessed on May 18, 2021 https://almascience.nrao.edu/processing/science-pipeline. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-17

FIGURE 4.9 Graph of the growth of refereed papers using archival data from European Southern Observatory’s (ESO) La Silla Paranal Observatory, as a percentage of the total number of papers published that year. Blue indicates papers for which there is overlap between the paper authors and the proposing PI and Co-Is. Black indicates no overlap (i.e. purely archival usage), and green is intermediate, where a combination of purely archival and purely PI data is used. In the last complete year for which statistics are available (2020), more than a third of all papers used archival data in some format. SOURCE: Retrieved from ESO, http://telbib.eso.org/index.php?boolany=or&boolaut=or&boolti=or&yearfrom=1996&yearto=2021&boolins=or&bo oltel=or&site=Paranal&search=Search. Courtesy of the ESO Telescope Bibliography (telbib), maintained by the ESO Library. FIGURE 4.10 Archive data usage for refereed papers reporting science results from the Atacama Large Millimeter/submillimeter Array (ALMA) over the 2010-2021 time period. Roughly one third of all papers produced have utilized archival data, either alone or in combination with PI data. SOURCE: ESO/ALMA, http://telbib.eso.org/statistics/archive.php?boolany=or&boolaut=or&boolti=or&yearfrom=2010&yearto=2021&bool ins=or&telescope[]=%22ALMA%22&booltel=or&site=Chajnantor&fl=telescope,datastatus&stats=arc&query_stats =year%3A%5B2010+TO+2021%5D+and+site%3AChajnantor+and+%28telescope%3A%22ALMA%22%29. Courtesy of the ESO Telescope Bibliography (telbib), maintained by the ESO Library. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-18

FIGURE 4.11 Paper production from NRAO’s Very Large Array/Jansky Very Large Array detailing the evolution of PI usage and archival usage of VLA/JVLA data in producing refereed papers. This does not include papers from surveys such as the NRAO VLA Sky Survey (NVSS) or Very Large Array Sky Survey (VLASS). SOURCES: (background image), https://public.nrao.edu/news/the-very-large-array-astronomical-shapeshifter/. (main image), R. Osten, based on data from L. Utley, NRAO. This untapped collection of observations not yet being archived can be seen as a tremendous opportunity to extract more science from the U.S. ground-based system. With appropriate strategic planning and modest financial investments, creating and archiving science-ready data products should offer a multi-fold return on the science from ground-based facilities. Scientists’ limited time and grant support could be focused entirely on analysis and discovery, and the impact of a telescope’s observations could span decades, as photons are reused for science that was unimagined at the time they were collected. Increasing access and the quality of archival observations can also serve as a powerful agent towards broadening participation in the profession, because they bring cutting edge data to any individual with internet access (even the public via Citizen Science initiatives), with minimal barriers to entering the active research community. This democratization of science through archive access will continue in the next decade.15 Finding: As demonstrated by space missions, and supported by archiving efforts at ESO and ALMA, readily accessible data in both raw and reduced form from ground-based telescopes can greatly multiply their scientific impact, even more so if pipelines are available to produce processed data. 15 J. E. G. Peek, V. Desai, R. L. White, R. D’Abrusco, J. M. Mazzarella, C. Grant, J. L. Novacescu, et al., 2019, Robust archives maximize scientific accessibility, white paper submitted to the Astro2020 decadal survey, https://arxiv.org/abs/1907.06234. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-19

With the emerging roles of mega-survey facilities such as the Vera Rubin Observatory, multi- wavelength and multi-messenger astrophysics, and time domain astronomy in this decade, there is an even greater need for data discovery and analysis across multiple archives; this motivates the need for coordination of those archives. The Enabling Foundations panel and this committee considered how to best address this need. The panel proposed establishing a cross-agency umbrella organization called the Astronomical Data Archiving System (ADAS) to coordinate the activities of the existing astronomical data centers and set priorities for new investments. The National Virtual Observatory (NVO) effort was undertaken from 2007-2014 with similar goals of enhancing archive interoperability, but with a significantly different structure and implementation from the proposed ADAS. In partnership with a similar International Virtual Observatory organization, the NVO achieved a number of important successes including the creation of a common set of data formatting and metadata standards, a first generation of data retrieval and exploration tools, and enhancing communication between the many individual data centers in the United States and around the world. NSF support ended in 2014, however, and any new organization of this type would need to build on the lessons learned from the NVO experience, and the experiences of the many archiving centers that have been operating independently over the last two decades. It will be important to preserve the expertise and resources of the existing data centers, in part by providing flexible and stable career paths for archive scientists and software developers. It is also critical to provide a centralized channel for input from the U.S. astronomy community on prioritization of data archiving efforts, so that the work of the archive centers remains in touch with the science needs of the community. The system as envisioned by the Enabling Foundations panel could also address cross-agency strategic planning in the related areas of software development, high-performance/high-throughput computing, archiving and curating data from theoretical simulations, and community training in related areas. An important component of creating effective archives is coordinating with cross-agency and international archiving services to develop best practices and interoperability. While the International Virtual Observatory Alliance continues, the lack of a U.S. national coordinating effort hampers efficient communication between the various national and international funding agencies and institutes which produce and archive astronomical data. Progress will come from an end-to-end approach that considers the entire flow of data from the instrument, to the archive, to analysis and publication. Increasing the prevalence of both science-ready data products and effective archives is best achieved if done hand-in- hand with each other. Making codes publicly available will help to minimize redundancy, encourage the adoption of common standards, and promote applications using multiple data sets. The survey committee endorses the importance of the goals of the proposed ADAS articulated by the Enabling Foundations Panel, but concluded that the appropriate scope of this effort and the details of the form it would take require further study, led by NASA and NSF, with possible participation from DOE. Recommendation: NASA and the National Science Foundation should explore mechanisms to improve coordination among U.S. archive centers and to create a centralized nexus for interacting with the international archive communities. The goals of this effort should be informed by the broad scientific needs of the astronomical community. The U.S. ground-based OIR system is distributed amongst public and private facilities and therefore needs special consideration. For three decades many of the needs in this area were served by the Image Reduction and Analysis Facility (IRAF), a freely-available workhorse software system funded by multiple federal streams. 16 Originally developed and maintained by the National Optical and Infrared Observatories (NOAO) in the 1980’s, lack of funding for modernizing the software resulted in its evolution to a community-supported platform on GitHub. This ground-up community-based approach has since become the dominant mode for providing common data pipelines and analysis tools across ground- 16 Image Reduction and Analysis Facility, “Community Distribution,” accessed on May 18, 2021 https://iraf- community.github.io/. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-20

based OIR astronomy. Prime examples include The Python Spectroscopic Reduction Pipeline (Pypelt) which has been adopted for multiple telescopes and instruments for at least 11 observatories,17 and the broad suite of application Python-based software collected by the Astropy project.18 Although these community based efforts have done much to fill the gaping needs for up-to-date pipelines and software, some may need to abandon their support, in the same way the IRAF project was eventually forced to do, because of the lack of sufficient and reliable funding and a stable workforce of contributors. 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 this type of infrastructure relies 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. NSF could help to provide foundational support for these efforts, for example by incentivizing the developments of pipelines and archiving by requiring a reduction and distribution plan as part of its funding for instrumentation, and/or by being open to joint proposals by multiple investigators or observatories to fund adapting and running existing pipelines. Funding for tailoring and operating these pipelines could be granted to observatories in exchange for their willingness to distribute their raw data and science-ready data products to a public archive; the cost of this investment is a small fraction of all the money invested in these facilities already, offering a highly-leveraged scientific opportunity. Recommendation: The National Science Foundation and stakeholders should develop a plan to address how to design, build, deploy, and sustain pipelines for producing science- ready data across all general-purpose ground-based observatories (both federally and privately funded), providing funding in exchange for ensuring that all pipelined observations are archived in a standard format for eventual public use. 4.5.2 Software Development Astronomy has entered an era in which well-designed and well-constructed software can be as important for the success of a project as hardware. Examples of highly complex software include pipelines that reduce data for telescopes (e.g., Astropy), data analysis packages, and codes that simulate physical processes, such as stellar evolution (e.g., Modules for Experiments in Stellar Astrophysics [MESA]), N-body (Galaxies with Dark Matter and Gas Interact [GADGET]), or hydrodynamics codes (e.g., Enzo). In addition, advanced statistical techniques and Machine Learning are playing a growing role in reducing large data sets in physics and astronomy, and can also require complex codes. Increasingly, many software packages are developed by large teams, and must make use of heterogeneous types of hardware platforms, from general purpose CPU’s running on laptops to large multi-core computing clusters that make use of massive parallelization and graphical processing units (GPUs). Despite the increasing importance of software development and developers for the advancement of the field, neither are sufficiently funded or supported by existing structures. Moreover, people who have strong software development skills are critical for the field, yet are likely to have many career opportunities outside of astronomy. Indeed, this is true throughout the physical sciences. Professional tracks available for scientists who choose to specialize in developing scientific software infrastructure, which might not be readily supported through traditional tenure-track faculty positions, could aid in the retention and development of these individuals. These positions can be supported through national labs, science centers, observatories, or in research positions at universities, with the understanding from 17 J. X. Prochaska, J. F. Hennawi, K. B. Westfall, R. J. Cooke, F. Wang, T. Hsyu, F. B. Davies et al., last update May 19, 2020, “PyPelt: The Python Spectroscopic Data Reduction Pipeline,” arXiv.2005.06505v2. 18 The Astropy Project, “Homepage,” accessed on May 19, 2021, https://www.astropy.org. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-21

funding agencies that proposals for funding software infrastructure may look more like “instrumentation” proposals than standard PI grants. As discussed in the Open Source Software Policy Options for NASA Earth and Space Sciences report, funding for software maintenance and for open-source software projects, which have been transformative for astronomical science over the past decade, could pay major dividends in the future. 19 Finding: Software development has become an essential part of every sub-field of astronomy. However, software developers and large software development efforts are not adequately funded or supported by existing structures. 4.5.3 High Performance and High Throughput Computing Computation has assumed an increasingly pervasive role throughout astronomy and astrophysics, from theoretical simulations of physical processes to sophisticated data analysis. Access to and expertise in the use of specialized computing facilities has therefore become ever more integral to the scientific process, and thus requires on-going investments and training over the coming decade. High-performance and high-throughput computing resources (HPC and HTC, respectively) are playing an increasingly important role in astrophysical research (Figure 4.12), with the former being critical for simulations and the latter for analysis of large data sets. HPC is a major part of a computational astronomer’s climate footprint, hence motivating the use of efficient options. Industry-provided options for HTC currently exist through cloud computing, and are often cost effective solutions to astronomical needs. However, as the size of data analysis problems expand, the cost trades could potentially become unfavorable, and a publicly-funded alternative may be more cost effective than relying on private industry to provide cloud computing. Funding programs may need to adapt to this rapidly changing trade space, while also ensuring that mechanisms exist for proposals to fund cloud computing access, rather than more traditional purchases of computing hardware. DOE and NSF have announced plans to significantly expand their HPC/HTC capabilities over the coming decade, while NASA plans a more modest expansion. NSF computing resources are also available without NSF support, but this is not true for NASA computing resources. Just as Section 4.3 emphasized the need for inter-agency funding opportunities to support science that transcends agency boundaries, it is essential for agencies to provide opportunities for access to HPC/HTC computing resources for cross-cutting projects. Developing the specialized codes that are competitive for allocations on large national computing facilities requires expertise in both computer science and astrophysics, as well as pre-existing access to facilities that can be used for code development and testing. These requirements can pose a significant barrier to entry for scientists at institutions that do not have access to this expertise or these facilities. Support for training (from NSF, NASA, or national laboratories) can be an effective means of helping to level this playing field. This support could take many forms, for example through small grants to individuals, support for training workshops or schools, or training opportunities through NSF and/or NASA centers. 19 National Academies of Sciences, Engineering, and Medicine, 2018, Open Source Software Policy Options for NASA Earth and Space Sciences, The National Academies Press, Washington, D.C., https://doi.org/10.17226/25217. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-22

FIGURE 4.12 Comparison of density through the central halo of a galaxy in a standard resolution (left) and high resolution (right) simulation at a redshift where the majority of star formation is occurring, demonstrates the utility of high performance computing simulations for advancing understanding of complex processes like the factors affecting galaxy formation and evolution. SOURCE: Adapted from Molly S. Peeples et al 2019, “Figuring Out Gas & Galaxies in Enzo (FOGGIE). I. Resolving Simulated Circumgalactic Absorption at 2 ≤ z ≤ 2.5,” The Astrophysical Journal, 873 129. © AAS. Reproduced with permission. doi:10.3847/1538-4357/ab0654. 4.5.4 Data Science and Machine Learning 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, universities have created multiple joint data science/astrophysics faculty appointments, and are adding new courses. Both undergraduates and graduate students are pursuing joint degrees in programs that did not exist in 2010, and NSF is increasingly investing in “big data” across all subfields. Astronomical data offer many opportunities for data science research. For example, a key paper in the Data-Driven Discovery Initiative by the Moore Foundation ranked the SDSS as the 6th most influential work in data-driven discovery, just behind Shannon’s classic information theory.20 Astronomical data are valuable for data science for many reasons: the data sets are rich and openly available, well-structured and well-modelled. These have led to 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. Data science offers powerful new tools for studying astronomical data and astrophysical systems. Machine learning has already shown significant success at providing tools for identifying anomalies in data, and can speed up parameter estimation in large data sets by significant factors (Figure 4.13). These techniques could lead to transformative discoveries from the new data sets available in the 2020s. Machine learning has the potential to increase the amount of information obtained from astronomical data sets by enabling modeling of complex non-linear phenomena and instrumental effects. If it can be 20 M. Stalzer, 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 4-23

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. Finding: Data science, including applications of machine learning, will play an increasing role in astronomical research over the coming decade. Incorporating training in this area at the graduate level and beyond will better prepare researchers regardless of whether they pursue careers in astrophysics or in other STEM fields. FIGURE 4.13 Hubble Frontier Fields image of the galaxy cluster Abell 370, illustrating numerous arcs resulting from strong gravitational lensing of background galaxies as their light passes through the massive cluster and is subsequently distorted. Machine learning has demonstrated an ability to identify strong lensing arcs orders of magnitude faster than the current state of the art (references in Ntampaka et al. 2019)21 and is an example of the impact of using deep learning techniques for model parameter estimation in large data sets. SOURCE: Space Telescope Science Institute, https://frontierfields.org/. NASA, ESA, and J. Lotz and the HFF Team (STScI). 4.5.5 Laboratory Astrophysics At its core, the science of “astrophysics” is built around the assumption that the observational data which astronomers and astrophysicists collect are all produced by understandable, physical processes that are the same throughout the universe. As observations are pushed to rarer or fainter spectral features, or to previously unexplored systems or physical conditions, the understanding is being increasingly limited not by the quality of the data themselves, but by the limited information about the underlying physical parameters. Thankfully, many of the needed parameters can be measured here on Earth. Laboratory measurements are needed to determine the oscillator strengths of atomic, ionic, and molecular transitions; 21 M. Ntampaka, C. Avestruz1, S. Boada, J. Caldeira, J. Cisewski-Kehe, R. Di Stefano, C. Dvorkin, et al., 2019, The role of machine learning in the next decade of cosmology, white paper submitted to the Astro2020 decadal survey, https://arxiv.org/pdf/1902.10159.pdf. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-24

reaction rates for the interactions that control the abundances of astrophysically relevant gases, ices, solids, and high-energy tracers; and the complex surface chemistry and optical emission and scattering processes that are increasingly relevant for understanding solid material in the ISM, protoplanetary disks, planetary atmospheres and planetary and stellar interiors. These experiments can be challenging, since conditions found in astronomical settings span a wide range of conditions that can be difficult to match in Earth-based laboratories. Spectral surveys in the far-infrared and submillimeter with Herschel and ALMA have detected more than 100,000 spectral lines from molecular clouds, star forming regions, and the center of the Milky Way (see Box 4.1). Many of these lines are from a few abundant molecules, such as methanol. However, even these spectral “weeds” contain useful information on conditions in the interstellar medium, such as temperatures, densities, pressures and radiation fields. Molecules of potential prebiotic interest, such as amino acids, have complex spectra requiring detections of multiple lines with a common excitation. Infrared spectra from JWST will have spectral features from aromatic hydrocarbons, which contain information on radiation fields and annealing processes. Studies of exoplanet atmospheres will need chemical reaction rates for a broad array of chemical species and conditions. Understanding the chemistry of protoplanetary disks is the first step toward grasping the composition and evolution of planetary systems. More generally, interpreting these spectral lines requires advances in laboratory astrophysics. Despite limited resources, examples abound of areas in which laboratory astrophysics has been key to advancing astrophysical discoveries over the past decade (Box. 4.1). In the search for humanity’s 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 absolute zero temperature, 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 significantly revising constraints on convective mixing in the Sun and other stars. If astronomy aims to understand the structure and evolution of stars, galaxies, and the universe as a whole through observations from future facilities, laboratory astrophysics will be required. BOX 4.1 Applications of Laboratory Astrophysics The history of buckminsterfullerene, or “buckyballs,” illustrates the interdependence of theory, laboratory work, and astronomical observations. One of the theoretical motivations that led to the discovery of these soccer ball-shaped carbon molecules was a desire to understand the diffuse interstellar bands. The origins of these broad absorption features in astronomical spectra remained elusive for nearly a century, although large carbon molecules in interstellar gas clouds were considered likely candidates. Laboratory experiments in the 1980s led to the identification of C60 in the emission spectra of soot. The solid phase of C60 in astronomical spectra was first identified in the mid-infrared spectrum of a star in 2012. Since then, absorption spectra of heavily reddened stars have revealed multiple transitions of singly ionized C60, confirming its presence as one of the carriers of the diffuse interstellar bands (Figure 4.1.1). The spectral signature of fullerenes could not have been identified without theoretical and laboratory studies of soot. In the era of ALMA, JWST, and proposed future facilities, with which astronomers will have the capability to study the chemical origins of exoplanetary systems and to detect molecules in exoplanetary atmospheres, the ability to spectroscopically identify complex PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-25

chemical species in space, including prebiotic molecules, is critically important. A robust program of laboratory astrophysics to support these investigations is essential. FIGURE 4.1.1 (upper left) Image of buckminsterfullerene, C60, indicating the linked carbon atom structure. (upper right) Spitzer spectrum of possible C60 features in the infrared spectrum toward the binary XX Oph. This is the first detection of C60 in the solid phase. Horizontal lines indicate the location of features in C60 smoke and gaseous C60. (lower panels) Absorption spectra from Cordiner et al. (2019) with the Hubble Space Telescope, comparing those of reddened and unreddened stars. Red curves denote broadened laboratory spectra of transitions of C60+. SOURCE: Upper left: Buckyball graphic in the public domain, from Benjah-bmm27, https://commons.wikimedia.org/wiki/File:Buckminsterfullerene-perspective-3D-balls.png. Upper right: A. Evans, J.Th. van Loon, C.E. Woodward, R.D. Gehrz, G.C. Clayton, L.A. Helton, M.T. Rushton, S.P.S. Eyres, J. Krautter, S. Starrfield, and R.M. Wagner, 2012, Solid-phase C60 in the peculiar binary XX Oph?, Monthly Notices of the Royal Astronomical Society: Letters 421(1): L92-L96, doi: 10.1111/j.1745-3933.2012.01213.x, by permission of the Royal Astronomical Society. Lower panels: Adapted from M.A. Cordiner et al., 2019, Confirming interstellar C60+ using the Hubble Space Telescope, Astrophysical Journal Letters 875: L28, doi:10.3847/2041-8213/ab14e5, © AAS, reproduced with permission. The 2020s will also see an even greater focus on stellar astrophysics, with “industrial scale spectroscopy” combining with data from Gaia aimed at obtaining complete inventories of stellar properties, such as detailed chemical compositions, masses, and ages. In the era of upcoming photometric (Vera Rubin Observatory, Skymapper,22 etc.), and large high- to low-resolution spectroscopic surveys (SDSS-IV, SDSS-V, the 4-metre Multi-Object Spectrograph Telescope [4MOST], the William Herschel Telescope Enhanced Area Velocity Explorer [WEAVE], Galactic Archaeology with 22 Keller, S. C. et al., 2007, Publications of the Astronomical Society of Australia, 24, 1. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-26

HERMES [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 high-resolution X-ray spectroscopic missions, such as XRISM and Athena, will not be able to capitalize on their high-resolution capabilities without new atomic data including collisional and photoionization cross sections and dielectronic recombination rates. Potential diagnostics of density, temperature, ionization, abundances etc. 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. A prime topic for the next decade, constraining the heavy elements produced in the electromagnetic counterparts to neutron star mergers (kilonovae) requires understanding the spectra of rapid neutron capture heavy elements such as neodymium and other rare-Earth metals (Figure 4.14). There are insufficient laboratory measurements of line strengths and wavelengths for these elements so current models that predict and interpret observations rely on theoretical atomic structure calculations. Additional laboratory measurements would be very valuable and would also inform abundance measurements of neutron rich elements in stellar spectra. FIGURE 4.14 Comparison of the infrared spectrum (from Gemini-South) of the electromagnetic counterpart to the binary neutron star merger GW170817 (black) to theoretical models of radioactively powered kilonova emission (red; see Ch 2.2). In the models, based on theoretical atomic structure calculations and radiation transfer, the broad absorption features are produced by a collection of highly Doppler-shifted transitions of neutron-rich heavy elements including neodymium and cerium. More detailed laboratory data on the atomic transitions of these and other neutron-rich elements are needed for a complete understanding of the spectra of neutron star merger counterparts and the heavy elements they produce. The region of strong absorption by Earth’s atmosphere is indicated by the gray box. SOURCE: From R. Chornock et al 2017, “The Electromagnetic Counterpart of the Binary Neutron Star Merger LIGO/Virgo GW170817. IV. Detection of Near-infrared Signatures of r-process Nucleosynthesis with Gemini-South,” The Astrophysical Journal Letters, 848 L19. © AAS. Reproduced with permission. doi:10.3847/2041-8213/aa905c. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-27

Laboratory astrophysics was identified in the Astro2010 decadal survey report as “vital for optimizing the science return from current and planned facilities,” especially in the ALMA and JWST era.23 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.”24 Currently laboratory astrophysics is supported through grants from the NASA APRA and ADAP programs and NSF AST, as well as some support from DOE and national labs for laboratory astrophysics. However the number of awards approved across all of these programs is small and they have been declining since the early 2000s.25 A search of NSF AST awards over 2015-2019 revealed 15 grants funded in laboratory astrophysics for a total of $6.2 million, and $12.4 million in funding from APRA over the same period. Despite the Astro2010 recommendation above for a $2 million per year increase in APRA funding for laboratory astrophysics, grant funding has remained essentially constant over the decade.26 Given the growing need for laboratory data and the relatively small investment required relative to the costs of the facilities supported, enacting the Astro2010 recommendation is more important than ever. It is important to add however that simply allocating more grant funding by itself will not be sufficient to address the entire problem. This research is most effective when the laboratory researchers have close ties to the astrophysical users of the experiments and data, but high start-up costs (typically $2 million or more) and the cross-disciplinary nature of the subject often leave university astronomy departments reluctant to hire new faculty in this area. Agency support for early-career faculty, similar for example to the NSF Faculty Development in Space Sciences (FDSS) program, could incentivize departments to invest in this field. Coordination and high-level prioritization of prime areas for future funding could also be effective. NASA for example already facilitates such an exercise through its Laboratory Astrophysics Workshop, but most of the resulting priorities are set by the researchers in the field. Broadening a similar exercise to include the user communities for the laboratory and computational data would be an important step towards ensuring that the precious funds are optimized to address the most pressing needs for interpreting current and future observations. Finally, for large flagship missions and MREFC-scale NSF facilities which rely heavily on laboratory data, including provision for these essential activities into the project budgets could be very cost effective and would naturally focus the laboratory work on the most urgent scientific needs for those facilities. Conclusion: Laboratory astrophysics is essential to the interpretation of astrophysical data from facilities such as JWST, ALMA, and future facilities like the ELTs. Research in this area needs to be regarded as a high priority. The existing approaches are not sufficiently advancing the field. Recommendation: NASA and the National Science Foundation should (1) convene a broad panel of experts to identify the needs for supporting laboratory data to interpret the results from the new generation of astronomical observatories, (2) identify the national resources that can be brought to bear to satisfy those needs, and (3) consider new approaches or programs for building the requisite databases. This panel should include experts in laboratory astrophysics as well as representative users of the data, who can best identify the highest-priority applications. 23 NWNH, p. 32. 24 NWNH, p. 220-221. 25 See Section 4 of Nave et al. 2019; Atomic data for astrophysics: Needs and challenges. Bulletin of the AAS 51(7). 26 2018 NASA Laboratory Astrophysics Workshop: Scientific Organizing Committee Report; https://baas.aas.org/pub/2020i0202/release/1?readingCollection=1bea0260. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-28

4.6 SUMMARY “Research is formalized curiosity. It is poking and prying with a purpose. It is a seeking that he who wishes may know the cosmic secrets of the world and they that dwell therein. ” Zora Neale Hurston, Dust Tracks on a Road (1942) This report lays out a roadmap for reaching the destinations for discovery in Chapter 2, A New Cosmic Perspective. Just as humans are fundamental to the success of research endeavors (as argued in Chapter 3), so too is the infrastructure for supporting that “poking and prodding with a purpose.” These are foundational components to the astronomical research endeavor, without which no steady footing can be assured. Important components of this foundation are threatened, however, due to perennial underinvestment. Writing the present chapter brought to light multiple examples of previous decadal survey recommendations that remain unfulfilled. While the weakness of vital parts of the foundation has not prevented the extraordinary scientific advances of the past decade, as with any foundation, continued neglect and erosion of the foundation will continue to undermine the entire enterprise over the longer term. Realizing the opportunities that can be achieved with appropriate funding and focus is the best way to ensure that the science destinations are reached, so that new courses can be charted efficiently. Ultimately understanding the connected cosmos through that “formalized curiosity,” and reaching the ambitious decadal goals—unveiling the drivers of galaxy growth, new windows on the dynamic universe, and pathways to habitable worlds—requires more than big new machines. It requires people to translate observations into discoveries, theoretical studies to connect the observational clues, experiments in the laboratory and with the computer to interpret the data and the theory, and digital libraries of these precious data which meet the needs for the twenty-first century. Finally, support for big machines and big projects needs to be balanced with support for the individual researchers who are the wellsprings of scientific creativity and discovery. A few well-targeted, modest investments in the enabling research foundation will restore a healthy balance to the overall portfolio and maximize the scientific return. PREPUBLICATION COPY – SUBJECT TO FURTHER EDITORIAL CORRECTION 4-29

Next: 5 Evaluating and Balancing the Operational Portfolio »
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!