G

Report of the Panel on Stars, the Sun, and Stellar Populations

INTRODUCTION

Stars are the main source of light in the universe. The scope of stellar astrophysics ranges from our host star, the Sun, to stars in galaxies so far away that we cannot distinguish them individually. Stars also span an immense range in mass and size. The least massive stars are barely distinguishable from giant planets and have lifetimes that exceed the age of the universe. The most massive stars have lifetimes shorter than the history of humans on Earth and die in spectacular explosions that seed the universe with heavy elements. Stellar astrophysics encompasses the study of single stars, stars in binary and multiple systems, and stars in bound clusters and associations that share common ages and compositions, a laboratory for investigating stellar evolution. Simple mass and energy conservation laws, nuclear fusion to generate energy, radiation and convection processes that transport that energy to the surface, and basic radiative transfer relations are sufficient to explain the most basic features of stars. Accurate modeling of these processes has allowed the study of stars to flourish and has facilitated a transition from the mere cataloguing of observable phenomena to detailed study of fundamental physics. However, stars are not just static, spherically symmetric objects that are solely described by one-dimensional models; they exhibit spatially and temporally complex phenomena, from rotation and magnetic field generation to pulsations and explosions. These phenomena determine how the stars populate and evolve across the Hertzsprung-Russell (HR) diagram. Thus, while stellar astrophysics is an old endeavor, new observations and advanced theory and simulations continue to reveal new phenomena that push our understanding of the fundamental physical processes that generate all of the visible light of the universe.

The scope of the Panel on Stars, the Sun, and Stellar Populations includes stellar structure and evolution, stellar activity and variability, brown dwarfs, ground-based solar astrophysics, resolved stellar populations including star clusters, nucleosynthesis, and chemical evolution. In the course of its work, the panel reviewed and incorporated the input of over 150 white papers submitted by the astronomy and astrophysics community addressing the preceding topics, as well as broader areas of astronomy in which stars are tools for studying fundamental physics, exploring the interstellar and intergalactic media, and probing distant galaxies. It is notable that most white papers had more than 10 authors and many were the result of organized

efforts by sections of the community. The panel had a rich source of ideas and analyses to draw on while developing its understanding of the needs and opportunities for advancing the science within its purview.

THE STATE OF THE FIELD

Before identifying the most critical problems in stellar astrophysics in the coming decade, the context is set by giving a brief overview of some of the most exciting discoveries in the past decade. This decade saw a resurgence of stellar astrophysics research fueled by time-domain observations by the National Aeronautics and Space Administration’s (NASA’s) Kepler space telescope and ground-based networks, large-scale spectroscopic surveys such as Apache Point Observatory Galactic Evolution Experiment (APOGEE) and Galactic Archaeology with Heliophysics Environmental and Radiation Measurement Experiment Suite (GALAH), sensitive and high-resolution submillimeter observations with Atacama Large Millimeter/submillimeter Array (ALMA), and the unparalleled astrometric precision of Gaia. It truly was the “decade of stars.” The study of the compositions, motions, ages, and multiplicity of stars has expanded from relatively small (≈10³–10⁵) to relatively large (≈10⁷–10⁹) samples exploring a large fraction of the Milky Way and its satellites, reaching to the brightest stars in galaxies throughout the Local Group and beyond. Our knowledge of the positions and motions of stars within the Milky Way has expanded ~1,000-fold in number and ~10,000-fold in volume. The 2010 decadal survey¹ noted that time-domain astronomy would be the new frontier for stellar astrophysics. Kepler has opened that frontier, invigorating the field of asteroseismology that provides a window into the hidden interiors and fundamental properties of stars across the HR diagram, and revealing new categories of variable phenomena related to stellar evolution, magnetospheres, multiples, and circumstellar environments.

We start with the Sun. The goal of modern solar physics research is to understand the entire Sun, from the core to the heliopause, in order to provide a holistic description of variations in its magnetic fields and the associated eruptive phenomena that can affect life on Earth. Observations during the past decade have revealed that magnetic fields are the source of energy for solar flares, the heating of the solar corona, and the acceleration of solar wind particles that shapes the entire heliosphere. Chromospheric data from ALMA have dramatically improved our understanding of this key interface layer between the photosphere and the corona (see Figure G.1). Helioseismic observations, data that allow us to “see” inside the Sun, provided the path for solving the solar neutrino problem in the 1990s. These data are now used to study changes inside the Sun and have revealed the beginning of a new solar cycle years before its first sunspots appear on the Sun’s surface. Advances in the next decade should provide more detailed understanding about how solar magnetic fields drive energetic phenomena at different spatial, temporal, and energy scales.

The past decade marked a revolution in stellar astrophysics through large-scale asteroseismology. Data from NASA’s Kepler satellite allowed, for the first time, systematic measurements of the radii and masses of thousands of individual stars, as well as age estimates. NASA’s Transiting Exoplanet Survey Satellite (TESS) mission is now continuing this process through asteroseismic measurements of bright stars across the entire sky. Asteroseismology’s greatest benefit is the access it provides to the internal structure and rotation of stars. The effect of internal rotation on the pulsation modes of evolved stars has allowed us to determine their core rotation rates precisely and has revealed that we do not fully understand angular momentum transport during stellar evolution. These data have shown that angular momentum transport from core to surface is more effective than current theories had predicted and have spurred the development of new interior evolution models. The ages provided by asteroseismic data have allowed us to determine stellar

___________________

¹ National Research Council, 2010, New Worlds, New Horizons in Astronomy and Astrophysics, The National Academies Press, Washington, DC.

ages quite precisely, which can be used to quantify the age-rotation rate relation (“gyrochronology”), particularly for older stars. Again, these data have indicated that modifications are needed for current stellar spin-down theories, to explain the observed reduction in the efficiency of magnetic braking as a star ages.

Understanding the connection between a star’s magnetic field and its stellar wind is of profound interest for understanding a star’s influence on its surroundings—that is, its astrosphere. Magnetic fields also affect a star’s sphere of influence. Consequently, over the past decade, the concept of the habitable zone (HZ), where water could be in liquid form on an orbiting planet’s surface, has now expanded to encompass the magnetic nature of stars and their influence on planetary evolution and the development of life. Understanding stellar activity and its high-energy emission, both as a risk to, and as a possible catalyst for life, is now a central concept in astrobiology. While we have yet to measure a stellar wind for a cool star (although upper limits have been obtained for Proxima Centauri), and we have as yet only a few plausible claims of coronal mass ejection (CME) observations from stars other than the Sun, stellar flares have been observed frequently by Kepler, TESS, and ground-based monitoring networks.

The past decade has seen improvement in techniques that allow us to directly measure the fundamental properties of stars—mass, radius, and luminosity—even at the extremities of the mass spectrum. Such measurements for stars with masses too low to ignite hydrogen (≈0.072 M_☉) have given us an empirical probe and new theoretical advancements on the equation of state of hydrogen-rich, electron-degenerate matter and the role of magnetic fields in stellar structure. Indeed, the mass-radius relationship for the lowest-mass stars cannot be explained without taking both effects into account. The discovery and census of Y-type brown dwarfs with temperatures as cool as 250 K has produced the first measurement of the field “stellar” mass function down to planetary masses (0.013 M_☉).

At the other end of the mass spectrum, there has been significant progress in our understanding of how high-mass stars evolve. Binary interaction has been shown to be a common and critical ingredient in the evolution of massive stars. Surveys show that interacting binaries are ubiquitous among these stars, accounting for at least 70 percent of the population, and that mass exchange can substantially modify their evolutionary trajectories. Furthermore, as massive stars experience substantial mass loss both on and after the main sequence, the uniformity and regularity of mass loss can significantly modify evolutionary

time scales. Extreme or episodic changes in loss rates can significantly hasten the evolutionary time scale to a star’s end phases. We have also learned that about 10 percent of massive stars possess strong surface magnetic fields (~0.3–20 kG) and high rotation rates (>200 km/s at the equator); both factors have a strong influence on mass loss and accretion. Stellar population synthesis models have improved by accounting for these factors and has led to a better accounting of the ionizing output from massive stars. Stellar evolution models are also better at quantifying the connection between the physical properties and multiplicity of massive star systems and the types of supernovae (SNe) they produce, as well as the types and masses of compact remnants they leave behind. The latter are particularly relevant for understanding the rate and nature of compact-object mergers that are now regularly being detected by the Laser Interferometer Gravitational-Wave Observatory (LIGO).

Time-domain surveys have opened up the study of the end states of stars through the detection of explosive, transient phenomena. All-sky surveys with nightly cadences have revealed several new classes of transients, including high-energy but short-duration (<1 day) relativistic explosions, long-duration but intermediate-luminosity red transients, and a myriad of faint thermonuclear SNe that range from calcium-rich transients to subluminous Type Ia and Iax SNe. These discoveries are not only enhancing our understanding of the variety of ways stars die by filling in the empirical picture of the end stages of stellar evolution, but they are also probing detailed stellar processes such as mass loss, binary evolution, and interactions with the interstellar medium (ISM).

The deployment of moderate-size spectroscopic surveys measuring stellar heavy-element abundances and velocities, combined with stellar astrometry from Gaia and ages derived from asteroseismic data, has extended our acuity of stellar populations beyond the solar neighborhood, resulting in major revisions in our understanding of the formation and chemical evolution of the Milky Way. Theories of galaxy formation suggest that galaxies such as the Milky Way grew by accreting gas and stars, and in the past decade we have discovered key facts about how matter accumulates and gas turns into stars to create the visible Milky Way. Data from Gaia have allowed the identification of a large, ancient accretion event, consisting of stars that dominate the inner stellar halo of the Milky Way. Ages of stars present in the Milky Way disk show that the disk was formed “inside-out,” with star formation starting earlier in the inner regions where gas first accumulated, and thus the Milky Way’s disk reached its current size gradually. Comparable photometric and spectroscopic studies of individual stars in Local Group dwarf galaxies are enhancing our knowledge of the origins and evolution of these satellite stellar populations.

The data on chemical abundances have shown that the fraction of stars whose chemical enrichment was dominated by elements formed in massive-star SNe, compared with those that show enrichment by both massive-star and white dwarf supernovae, varies enormously across the disk of the Milky Way. However, these groups cannot be separated clearly by age or overall heavy-element enrichment. There is no consensus yet on the cause of the abundance variations, but hypotheses include the radial movement of stars from the inner part of the Milky Way outward and a burst of massive star formation following the infall of fresh gas into the Milky Way.

Stellar astrophysics research over the past decade has also been revolutionized by large open-access data sets from numerous photometric, astrometric, and spectroscopic surveys. While many publicly funded heliophysics and space-based facilities have had open-data policies for many years, such policies are reasonably new for ground-based astronomy. The exploitation of the new extensive photometric and spectroscopic data sets has been greatly aided by new computational methods, such as machine learning and data-driven modeling, that are now regularly employed, as well as by publicly available software such as Modules for Experiments in Stellar Astrophysics (MESA) and Astropy. Additionally, improvements in molecular and atomic line lists, and continued work on 3D and nonequilibrium effects in model stellar atmospheres, have been important in the past decade. Nevertheless, inputs to stellar models are still quite

uncertain, particularly opacities of molecules and highly ionized atomic species. Opacities used in stellar models are often produced through computationally intensive calculations, which in some cases are difficult to calibrate and test. For molecules, the constraints are often numerical, with line lists exceeding 10¹⁰ transitions. For ionized atoms, the constraints are often experimental, requiring extreme testing conditions. For example, experiments at Sandia National Laboratory have revealed discrepancies between the measured and calculated opacities of iron, nickel, and chromium at conditions found at the base of the solar convection zone. The position of the solar convection-zone base is known precisely (0.713 ± 0.001 R_⊙) from helioseismic analyses, but in solar models this depends on opacities and models created with the latest estimates of solar heavy-element abundances that have convection-zone bases that are discrepant at the 11σ level. While discrepancies between theory and experiment were anticipated, the reasons for these discrepancies are not yet understood.

While the past decade has witnessed many significant advances in the study of the Sun, stars, and stellar systems, many outstanding questions remain. These include questions about the fundamental mechanisms driving magnetic field generation and its influence on internal structure, surface heterogeneity, wind generation, and environmental influence across the mass spectrum; the temporally and spatially dynamic properties of stars, and their internal to external manifestations; and the formation and co-evolution of stellar multiples, and their influence on the end states of massive stars. Moreover, with new technologies, computational tools, and theoretical developments, we have the opportunity to build a complex and comprehensive assessment of stellar populations throughout the Milky Way and beyond. The most promising scientific opportunities are described in the next sections.

THE SCIENCE PRIORITIES

The following sections describe what the panel believes are the science priorities of our field. The four science questions whose answers will bring transformative change to this field are (1) What are the most extreme stars and stellar populations? (2) How does multiplicity affect the way a star lives and dies? (3) What would stars look like if we could view them like we do the Sun? (4) How do the Sun and other stars create space weather? The discovery area that will give the highest science return is “industrial-scale” spectroscopy of 10⁹ stars in the Milky Way and beyond.

G-Q1. What Are the Most Extreme Stars and Stellar Populations?

Stars at the limits of mass, composition, pulsation properties, and rotation test the ability of theoretical models and also have profound implications for the evolution of galaxies. Stars that are not in equilibrium as they transition rapidly between more static states have rarely been seen, even though all stars go through such phases. Defining the extremities is also crucial for stellar systems because they can span a wide range in properties such as masses, rotation, magnetic fields, mass transfer rates, and so on. For star clusters and stellar populations, the range of the distribution of masses and frequency of multiple stars can be determined from observations, and these properties play a critical role in the evolution of the system. Over the next decade, HR diagrams of the Milky Way from Gaia and of the local universe with the James Webb Space Telescope (JWST), combined with theoretical explorations, will revolutionize our understanding of stars and stellar populations.

More than a century since its creation, the HR diagram—a plot of luminosity against surface temperature—remains one of the most powerful tools for understanding stars and stellar systems. A star’s location on the HR diagram encodes essential physics, including the star’s current energy source, internal structure, elemental abundances, and circumstellar environment. Patterns in the HR diagram reveal deep truths about how

stars evolve, as well as the ages and chemical abundances of the environments that host the stars. Mapping and interpreting the deep connections between the HR diagram and stellar physics have impacts throughout astrophysics, from reionization in the early universe, to the history of galaxies, to the fundamental physics of degenerate matter. In order to answer the question above, we need to determine the fundamental properties and demographics of stars across the HR diagram and what these properties imply for the evolution of stars.

The past decade has transformed our characterization of the HR diagram within the Milky Way. The astrometric Gaia mission measured distances to millions of individual stars, placing them securely on the HR diagram and revealing stars that are unquestionably in unexpected regions of the HR diagram. Cross-referencing the Gaia HR diagram to extensive photometric and spectrographic catalogs (many the products of the past decade’s investments) has made it possible to identify theoretically predicted but rare phases of stellar evolution, such as the aftermath of mass transfer between binary stars. It is in this mapping between the observable features of the HR diagram and the fundamental parameters of stars that the next decade will shine. We now know empirically how stars of different types populate the HR diagram, but the connection to physical parameters and the underlying stellar physics is often limited in precision, completeness, accuracy, dynamic range, and theoretical understanding. The lack of precise stellar masses and radii remains a limiting factor for exoplanet characterization, for studies of the equation of state for degenerate matter, and for understanding the stellar variables that anchor much of the extragalactic distance ladder. For example, to study the influence of convection on stars, and in particular how to implement it in models, requires radius estimates to better than 5 percent. Understanding the inflated radii of M dwarfs requires even better precisions, of the order of a few percent. At the low-mass end, mass estimates to better than 10 percent are needed to understand the transition between brown dwarfs and stars. Lack of accurate masses, binarity, and rotation limits our knowledge of stars that explode as supernovae or erupt as other violent astrophysical transients. Incomplete stellar samples limit our knowledge of the initial mass function and may miss rapid evolutionary stages that dominate a stellar population’s overall luminosity, supernova rate, and dust production. Precise physical parameters are also needed to calibrate atmospheric and evolutionary models that are used to make age determinations and to trace chemical enrichment of stellar populations. All of these issues are least understood at the extremes of stellar mass and chemical composition, preventing a comprehensive theoretical understanding of stellar interiors, atmospheres, nucleosynthesis, and evolution.

The path forward lies in harnessing both old and new approaches on a large scale. Fundamental measurements of mass and radius can be made directly, but are typically limited to relatively bright stars for which asteroseismic or interferometric data can be obtained or to rare varieties of stars, such as eclipsing binaries. Expanding this to statistically powerful samples across the HR diagram can be done by expanding the eclipsing and non-eclipsing binary sample and extending direct interferometric radius measurements to single stars at low masses and beyond the solar neighborhood. Continuing a vigorous program of asteroseismology, which has revolutionized understanding of solar-type and more massive stars, will also extend precision calibrations to lower masses and additional stellar types. These results can be used to calibrate indirect methods to infer mass and radius (i.e., theoretical models). Microlensing mass measurements with the Nancy Grace Roman Space Telescope (formerly the Wide-Field Infrared Survey Telescope [WFIRST]) will extend calibrations across the full stellar mass spectrum, provided that lensing stars can be detected and photometrically anchored to the HR diagram. The Laser Interferometer Space Antenna (LISA) could help in determining the radii of the extreme low-mass end of stars.

Interpreting the HR diagram also requires accounting for the many factors beyond mass that influence stellar luminosities and atmospheric temperature. These factors may include age, chemical composition, non-equilibrium atmospheres, rotation, binarity, and so on. Direct age measurements can come from nucleo-cosmochronology for old metal-poor stars, lithium depletion for young fully convective stars and brown dwarfs, asteroseismology, and age-sensitive abundance indicators. These age standards can cali-

brate gyrochronology, and HR-based age-dating methods can anchor theoretical stellar evolution models. Accurate interpretation of the HR diagram requires stellar samples spanning a wide range of ages and elemental abundances, spectral data outside the optical band, better atomic and molecular opacities, improved non-equilibrium atmosphere modeling and retrieval methods, and an accounting for evolutionary effects (e.g., self-enrichment).

Rotation and multiplicity remain significant sources of uncertainty in interpreting the HR diagram, and there is a pressing need to determine stellar rotation rates with synoptic time-domain observations, asteroseismology, and spectroscopy. Such observations will help improve models of how stellar rotation evolves, as well as models of binary co-evolution. The other unknown is the magnetic field. While proxies of chromospheric magnetic activity have been monitored for decades, measurements of magnetic fields and their configurations remain rare. These measurements are becoming more accessible through spectropolarimetry and Zeeman Doppler Imaging (ZDI), but the resources remain scarce.

Expanding the depth of our knowledge of the physics that underpins the HR diagram not only deepens our knowledge of stellar processes, it also unlocks the HR diagram’s power for all of astrophysics. The HR diagram remains the most accessible constraint on stars’ properties and evolutionary state, given the greater observational demands on spectroscopy or asteroseismology. Completing the mapping between the HR diagram and internal stellar physics leverages information that is only observationally available in the Milky Way or its satellites. This mapping will be critical as we resolve individual stars at mega-parsec distances with JWST and 30 m class telescopes, and study more distant stellar populations through their integrated light.

G-Q2. How Does Multiplicity Affect the Way a Star Lives and Dies?

Most stars orbit other stars. Those in very widely separated systems are crucial probes of star formation processes and coeval laboratories for stellar structure and evolution studies, and constrain the properties of the Milky Way’s dark matter. Those in very closely separated systems have fates that are intertwined, altering each other’s evolution, mass loss, and final outcomes. Close binary interaction can manifest itself as transients associated with mass transfers and mergers. Binary co-evolution fundamentally influences nucleosynthesis, governs supernovae rates, affects gravitational wave production, and modifies stellar lifetimes. Binary co-evolution also influences the structure of planetary nebulae, the demographics of compact objects as well as our interpretation of the HR diagram. While the past decade has driven home the ubiquity of these phenomena, the demographics and mechanisms of binary co-evolution remain weakly constrained. Better observational samples and theoretical modeling of these phenomena are the route to addressing some of the more challenging questions, such as the fate of close binaries that do interact.

Observations to date show that over 70 percent of stars above 8 M_⊙ are in multiple systems. Being bright, these can outshine other stars; such systems are believed to dominate the UV light from external galaxies. High-mass stars in binaries can interact before becoming supernovae, as the η Car system demonstrates. Many multiple systems have at least some mass transfer or are in close contact with one another, while others co-evolve through a poorly understood common envelope configuration or even merge. There can be substantial changes to the emergent radiation of a star if its outer layers are stripped off by a companion. Indeed, binary interactions in very metal-poor stars may provide enough high-energy photons to contribute substantially to the reionization of the universe. Stripped envelopes and other mass-loss debris surround a dying star and can interact with supernova ejecta. Most of the diversity in core-collapse supernova light-curves may be owing to interactions with circumstellar material rather than the underlying engine; to understand the physics of the latter, we must understand the phenomenology of the former. For these reasons, it is imperative that we correctly determine the fraction of high-mass stars in binaries,

particularly close binaries for which co-evolution effects are most prominent, and measure the outcomes of binary co-evolution, such as mass-transfer rates.

Stars with masses less than 8 M_⊙ can also have close companions, creating the progenitors of cataclysmic variables and Type Ia SNe. The progenitors of these transients co-evolve in at least one common envelope phase. Depending on the initial masses, separations, and system mass-loss rates, they evolve into either a single white dwarf star with mass accretion from an evolved main sequence or red giant companion, or into a pair of white dwarf stars. As with massive stars, the details of what occurs in the common-envelope phase are not well understood, but they are critical for determining whether the outcome is capable of creating a Type Ia SN, and whether we observe it earlier as an AM CVn star, a super soft X-ray source, or a double white dwarf system. Stars between 0.5 and 1.5 M_⊙ are prime candidates for habitable worlds because of their long lifetimes and relatively benign astrospheres. Roughly two-thirds of Sun-like stars are in binaries; thus, understanding how planet formation and planetary-system evolution is affected by the gravitational and radiation influences of other stars in a system is crucial. We know that planet formation in binary systems is possible, but it may be disfavored. Depending on the proximity of a stellar companion and the stellar mass ratio, the companion can influence the dynamical evolution of a planetary system through processes such as Kozai-Lidov oscillations. Understanding the role of multibody interactions in planetary systems requires identifying the frequency of such systems and the correlation of stellar binarity with the architecture and orbits (inclinations and eccentricities) of planets.

Despite the prevalence and importance of multiple-star systems, we have not yet mapped their demographics sufficiently. Among other parameters, we need to know the orbital properties and mass ratios of multiple-star systems across age, mass, mass-loss rates, and composition. While we know that multiplicity is more common in high-mass than in low-mass stars, we have incomplete information on the statistics of systems with extreme mass ratios, long orbit periods, and very low masses (e.g., brown dwarfs). The first investigations of how multiplicity depends on composition have indicated that low-mass metal-poor stars have higher rates of multiplicity. Further investigation of this population is necessary because multiplicity at very low metallicities is relevant to the ionizing flux, nucleosynthesis, carbon abundances, and pollution signatures of Pop III stars. Measurement of multiplicity in systems losing large amounts of mass is needed to forecast mass-loss rates for broader stellar populations. Information on systems with more than two stars is also lacking, despite their potential importance in Type Ia production and star-planet dynamics. With current and upcoming survey capabilities, we are poised to map multiplicity in exciting new ways, using astrometry from Gaia, gravitational waves from LIGO and LISA, and synoptic photometry from the Zwicky Transient Facility (ZTF), Legacy Survey of Space and Time (LSST), TESS, and others. The addition of spectroscopic capabilities is crucial to the prospects for discovery and characterization.

Efforts in theory and computation need to go hand in hand with observations. The vast majority of stellar structure and evolution calculations assume single stars. There are very few modeling efforts that take binarity into account, and fewer still that take both rotation and binarity into account. Given that such an effort will need multidimensional models of multiple, interacting stars, there needs to be support for new code-development in this field.

G-Q3. What Would Stars Look Like if We Could View Them Like We Do the Sun?

With the exception of Betelgeuse, a supergiant that has a radius almost 1,000 times that of the Sun, we cannot easily observe features on a star’s surface. As a result, stars are typically treated observationally as featureless points of light and theoretically as spherically symmetric objects. However, observations of our own Sun demonstrate that stellar surfaces are complex and dynamic, and sometimes not even spherical. Spots, flares, tidal distortions, mass-loss, rotation, and internal convection all break the spherical symme-

try of stars. Observations in the past decade have revealed the pervasiveness of these effects, which have confounded our ability to understand the fundamental properties of stars and their surroundings. Advances in computing and computational methods are now beginning to make sophisticated 3D models of stellar interiors and atmospheres a possibility; this will enable greater clarity on a myriad of stellar asymmetries.

Physical processes that break interior symmetry, such as rotation and meridional flows, cause mixing across chemically inhomogeneous interior layers, and this in turn alters how these stars evolve. These effects can only be approximated in currently used 1D models, often leading to conflicts with precise observations. Convective heat transport is another source of error in models. Helioseismic and asteroseismic data have already shown that 1D approximations of convection do not model the surface layers of stars correctly. More worryingly, free parameters in 1D approximations of convection and magnetic field generation can lead to incorrect predictions of stellar radii. Helioseismic data also reveal extra mixing below the solar convection zone, a feature not present in standard 1D models. Asymmetric processes affect the later evolutionary stages as well. For instance, the internal mass distribution of white dwarfs, which informs models of supernovae and chemical enrichment, depends on the sizes of the cores in their progenitors. However, core size is affected by convective overshoot, rotationally induced mixing, and related instabilities that are inherently 3D in nature, and cannot be understood with a 1D approach. 3D models are also needed to properly understand the interplay between stellar convection, rotation, and magnetic field generation.

At the stellar surface, where photons can escape and reach us unimpeded, stellar asymmetry manifests itself in the form of spots: large areas of enhanced magnetic field and generally reduced brightness, with enhanced brightness in the surrounding “plage.” These dark spots were the features Galileo identified as blemishes on the Sun. In the past decade, we have witnessed how varied the nature of star spots can be. Even stars with masses similar to our Sun can have dramatic differences in spot sizes and surface distributions. For lower-mass stars, spots often occupy a much larger fraction of the stellar surface and are not limited to low latitudes. Over the next decade, it will be critical to develop a comprehensive understanding of how star-spot sizes and distributions are driven by the underlying magnetic structure, and physical properties, of stars of all types. In terms of theory, complete 3D magnetohydrodynamic models of stellar dynamos are essential to understand the details of stellar magnetic configurations and their time variation, particularly in fully convective stars.

The Sun remains our best model in understanding surface inhomogeneities, and new data from the Daniel K. Inouye Solar Telescope (DKIST) will revolutionize the field by allowing us to infer details of magnetic field strengths and orientations, as well as magnetic processes such as reconnection, in the solar atmosphere. These discoveries need to be expanded to encompass other stars through high-sensitivity stellar spectropolarimetry, which measures a star’s magnetic field structure and plasma properties. Because spot asymmetries can masquerade as exoplanet signals, velocity-resolved stellar surface spectra (Doppler imaging) will provide a critical reference for disentangling exoplanet and stellar signatures, essential for the robust detection of habitable Earth-like planets.

In the cool atmospheres of brown dwarfs, where magnetic spot formation can be inhibited, surface asymmetries arise instead from condensate cloud structures (see Figure G.2). The complexity of the observed light curves of cool brown dwarfs points to global dynamic processes similar to those observed in the solar system giant planets, which may include thermochemical instabilities that can trigger (or mimic) cloud structure. Better understanding of brown dwarf cloud structure and composition through spectroscopic monitoring and 3D circulation and condensation modeling will also contribute to our understanding of exoplanet atmospheres, which have also shown evidence of cloud-induced variability.

Away from the stellar surface, asymmetries are conspicuous as non-isotropic, episodic, and clumpy mass outflows that emerge from stars of all masses. Low-mass stars generally lose large amounts of mass only during their late evolutionary phases on the asymptotic giant branch (AGB), whereas high-mass stars

lose mass throughout their entire lives. Among the challenges we face in interpreting stellar outflows are those of understanding the rate of radiation-driven mass loss in high-mass stars, understanding the wind-launching and dust acceleration mechanisms in AGB stars, and deriving robust empirical estimates for mass loss in the late stages of massive star evolution (luminous blue variables, red supergiants). These observational effects need to be matched to reliable theoretical predictions. Because mass loss has a deterministic influence on the evolution of evolved stars, it is increasingly important to understand the nature of clumpy, episodic, and sometimes eruptive states of mass loss. These are the far more dominant modes of mass loss than the weaker winds typically associated with mass loss in solitary luminous stars. Current models have difficulty in predicting mass loss. There is a need for the development of quantitative theoretical predictions of mass loss rates in red supergiants and the underlying physics driving them. This is one of the main issues in understanding the evolution of massive red supergiants like Betelgeuse.

Understanding the evolution of various astrophysical phenomena, from exoplanet atmospheres to solar-like magnetospheres to the interaction of core-collapse SNe with circumstellar medium, requires an appreciation of the inherent asymmetries present from the interiors to the surfaces to the winds of stars. These require study in a comprehensive 3D manner.

G-Q4. How Do the Sun and Other Stars Create Space Weather?

The Sun and other stars affect their environments in numerous ways, from the interaction of stellar winds, flares, coronal mass ejections (CMEs) and other forms of mass loss with surrounding disks, planetary bodies, stellar companions, and the interstellar medium, to the creation of planetary nebulae and supernova remnants and the end stages of stars. We broadly interpret these phenomena as space weather, expanding on the traditional definition of this term to incorporate a star’s influence on its environment throughout its life cycle. Understanding how a star generates these phenomena helps to explain how stellar environments evolve over the star’s lifetime and beyond. Identifying the physical processes involved in these interac-

tions informs a broad range of current astrophysical problems, from stellar feedback in galaxy evolution, to the formation and retention of atmospheres on planets. Now is the time to address this question as new capabilities allow us to observe embedded hot stars in star-forming clouds, CMEs beyond the Sun, and the diversity of exoplanets orbiting other stars.

The physical scale of a star’s influence on its environment is a function of its mass, with both quiescent and transient effects playing significant roles. Stellar radiation across the electromagnetic spectrum can have profound effects on the evolution of structure—for example, the photoionization of star-forming regions by the UV emission of a single massive star. Stellar mass loss through steady, episodic, or transient processes also influences the star’s environment. The radiation-driven winds of massive stars interact with the circumstellar environment, creating nebulae filled with gas and dust. A massive star’s evolution depends on its mass loss rate, which in turn depends on metallicity as the wind acts on the highly ionized metals produced by the star, and on magnetic field strength, now measured in 10 percent of hot stars. The winds of low-mass stars are more elusive, but the solar system provides an essential laboratory for understanding these processes. We know that the magnetically driven solar wind deflects the tails of Sun-grazing comets, drives atmospheric mass loss, and interacts with planetary magnetospheres, generating aurorae. The multiphase solar wind must have an acceleration mechanism beyond thermal expansion. The steady, fast solar wind originates from open magnetic field lines near the polar regions; the origins of the slower, more variable solar wind are still unclear but seem to be related to the opening of field lines associated with active regions and coronal loops. Synoptic surveys are essential for modeling these processes. Solar flares probe fundamental particle acceleration, although the specific nature of magnetic acceleration requires further advances in both modeling and radio polarimetry. CMEs, containing lower-temperature, higher-density material, and generating the most damaging aspects of space weather for Earth, often accompany flares, although our understanding of the relationship is incomplete. Critically, all of these energetic phenomena originate from magnetic processes on the Sun, and we will soon be able to measure details of the magnetic phenomena powering these events with DKIST.

For low-mass stars, similar phenomena seem to occur, although often with very different properties. Stellar flares on active stars extending to the lowest stellar masses, observed from the radio to the X-ray, can have high-energy luminosities up to five orders of magnitude larger than flares on the Sun. However, current observations tend to be biased toward the nearest and/or most active stars. In the next decade, we need to characterize transient energetic phenomena systematically, as a function of stellar type, mass, age, metallicity, and rotation rate, and to explore these phenomena across the electromagnetic spectrum. Direct observation of stellar winds and CMEs will help constrain these effects. Possible paths forward include CME detection through photoelectric absorption of a star’s X-ray spectrum and high-spatial/high-throughput X-ray imaging of the asteropause shock to measure wind kinematics.

Understanding the energetic phenomena driven by stellar magnetism requires exploration of the diversity of stellar magnetic surface structures. Magnetic field maps constructed using ZDI show that nondipole fields can dominate the surface field structures of some stars, but measurements have been made for only a few systems. The poles of the young main sequence star AB Dor appear to have the same polarity, based on contemporaneous Doppler imaging and X-ray spectroscopy and light curves. X-ray studies with high spectral resolution and throughput, combined with ZDI, are needed for a larger sample of low-mass stars at various evolutionary stages.

Intertwined with stellar magnetism is the long-standing question of what heats the stellar corona. There are two equally plausible mechanisms within steady-state magnetic loops: magnetic wave heating and flare-like impulsive heating scaled down to nanoflares. Distinguishing between these mechanisms requires systematic characterization of stellar coronal properties across a range of energies, as well as theoretical modeling that extends down to low-mass stars. Indeed, evidence from the few stars that could

be observed spectroscopically with current X-ray technology suggests a diversity of coronal heating mechanisms, and as such the situation may be far more complex than currently appreciated. Very active stars show high electron densities in relatively quiescent coronae, densities observed only in flares on the Sun. Furthermore, the decline in optical/X-ray magnetic emission and its decoupling from persistent (and variable) radio emission at the lowest stellar and substellar masses suggests a profound change in magnetospheric structure and heating mechanisms. Coherent pulsar-like radio pulses and highly polarized emission suggest a connection with Jovian-like auroral mechanisms, but many mysteries remain. The detailed mechanisms of magnetic emission around low-mass stars have clear implications on the habitability of close-in orbiting exoplanets.

Ultimately, processes interior to the star generate the magnetic dynamos observed through their energetic surface phenomena. The growing interest in exoplanet atmospheres and potential habitability mandates a better theoretical understanding of stellar magnetism and its effects throughout a star’s system. Detailed magnetohydrodynamic dynamo models of the Sun and other stars spanning a range of physical properties are needed to explore these effects. These include investigation of potentially novel dynamo processes in fully convective stars. However, dynamo models are not enough. These need to be coupled to models of radiative transport through a star’s exosphere to predict emergent phenomena.

G-DA. DISCOVERY AREA: “INDUSTRIAL-SCALE” SPECTROSCOPY

Building on the scientific progress from large-scale, time-domain photometric surveys over the past decade, the panel sees that considerable advancement can be made by greatly expanding spectroscopic surveys in breadth, sensitivity, precision, and cadence across the full electromagnetic spectrum—in other words, extremely large-scale or “industrial-scale” spectroscopy. This capacity will be accomplished through advancements and investment in instrumentation and facilities, improvements and standardization of spectroscopic reduction and analysis techniques, archival storage and broad community access to data products, development of novel approaches to explore the highly multidimensional data sets that will emerge from these efforts, and support for laboratory astrophysics, including theory and experiment.

Astronomy became astrophysics with the first spectrum. Spectroscopy determines compositions, magnetic field strength, space motion, rotation, multiplicity, planetary companions, surface structure, and other important physical traits. Industrial-scale spectroscopy expands current capabilities in spatial, spectral, temporal, and sample-size dimensions, with a higher sensitivity that enables deeper, farther, and faster observations. Spectroscopy is too important to continue to be a “follow-up” of photometric surveys. In the next decade, spectroscopy will be the dominant discovery tool for astronomy. The need for photometry will not go away; we will still need Kepler-like stable time-domain photometry that is combined with the sky coverage of TESS; we also need Spectro-Photometer for the History of the Universe, Epoch of Reionization, and Ices Explorer (SPHEREx)-like broadband spectral fluxes.

In the X-ray regime, new advances will come from expanding the rate and wavelength coverage of spectroscopic observations. Increases in throughput by two or more orders of magnitude, and increases in spectral resolution by factors of 2 to 5, will open up discovery space through measurements of thermal broadening and new line-ratio diagnostics. Athena does not have all the characteristics needed for work on stars. While Athena’s microcalorimeter will provide high-resolution spectra for the most energetic phenomena such as stellar flares, studies of quiescent and/or lower energy phenomena, such as coronal heating and absorption of CMEs, would benefit from high spectral resolution at lower energies; these can be obtained with grating spectrometers. Furthermore, Athena does not have sufficient spatial resolution to disentangle individual stars in star-forming regions or other crowded fields. The current sample of stars with X-ray grating observations is less than 200, so an increase in sample size by two orders of magnitudes

or more will dramatically change the field. For low-mass stars, high spectral resolution for dozens of M dwarfs will probe the magnetic transition to fully convective stars. For high-mass stars, time and spectrally resolved X-ray line profiles will probe the clumpiness of shocks in the wind and constrain mass loss rates.

Past UV spectroscopic samples are small, both in number and in wavelength coverage. Extending our reach to the extreme UV opens up discovery space in stellar exopheres, magnetism, and CME detection on stars other than the Sun by allowing the study of lines such as FeIX, FeX, FeXI, and so on. Providing multiplexing capability in the near UV is critical for observing those few stars in large samples that trace nucleosynthetic enrichment from the first generation of stars, as well as from merging neutron stars.

Moderate increases in the number of optical and near-IR stellar spectra provided by Milky Way surveys on 2–3 m class telescopes, and by Local Group surveys on 10 m class telescopes, have already changed the fields of stellar physics, Milky Way archaeology, and the origin of the elements. Further breakthroughs in these areas will be fueled by (1) very large samples—10⁸–10⁹ stars—on current telescopes and (2) wide-area surveys on even larger telescopes. These observations will yield results on extremely faint stars, as well as on extremely rare (e.g., hyper-metal poor) stars. In combination with Gaia astrometry, we will more fully understand how chemical evolution works on galaxy-wide scales in the Milky Way. Non-U.S. projects such as the WHT Enhanced Area Velocity Explorer (WEAVE) and the 4-metre Multi-Object Spectrograph Telescope (4MOST) will be a step in this direction, but neither will reach the scale that can transform the field completely; an order of magnitude increase in distance sensitivity would permit robust statistical studies of populations that are currently under-sampled and will also probe flare energetics to the point where they merge with quiescent-scale emission and solar-like flares.

Much progress can be made using broadband radio measurements. Radio wavelengths can be used as a probe of particle acceleration. A continuous frequency coverage from 10–400 GHz on sufficiently large baselines can measure mass-loss rates in evolved stars. For the Sun, increasing observational cadence without sacrificing signal to noise will enable us to measure spectral changes quickly enough to study solar flares, CMEs, and shocks.

We know the expected outcomes of industrial-scale spectroscopy. However, just as extensive time-domain photometry revealed unknown categories of transients, transiting debris, unusual variables, and megaflares on solar-type and low-mass stars, we anticipate that changing the scale of spectroscopy will reveal new, unanticipated phenomena. Even when we know the general physics that can be probed with these observations, orders-of-magnitude improvement in sample size will provide new, unanticipated insights about the universe, as the gains made from asteroseismology have shown. Information across many bandpasses—for example, with SphereX in the mid-IR—will further enrich this discovery area. The full impact of a decade of spectroscopy-driven discoveries will require a multiwavelength approach to observations and modeling. Because there are few nationally supported facilities with any of the capabilities highlighted above, this discovery area will require investment to reach its potential.

REQUIRED CAPABILITIES

Given the wide range of stellar temperatures, brightness, and environments, the potential advances identified in this appendix require a multifaceted approach encompassing observing, computing, and laboratory capabilities and resources (see Table G.2).

Long-term global and synoptic monitoring of the Sun, in optical and radio wavelengths, coupled with detailed DKIST and space-based observations, is necessary for dynamic helioseismology and magneto-graphic monitoring of the full solar disk. The construction of the 4-m DKIST is almost complete, and this telescope will start collecting science data soon, providing us with detailed studies of magnetic interactions. However, observing with DKIST will be like watching the Sun through a microscope. The telescope will

provide observations of magnetic fields in exquisite detail, but these observations need to be put into a global context, and this will require new synoptic facilities that measure the global magnetic fields of the Sun at high cadence, as well as helioseismic observations that will allow us to determine associated changes in solar structure and dynamics below the photosphere. Radio facilities are needed to probe changes in the chromospheric magnetic field. Coronagraphs capable of measuring coronal magnetic fields are required to obtain a complete picture of the entire magnetic Sun; current coronagraphs measure only intensity.

As far as asteroseismology is concerned, the large pixel sizes, and short time series, of TESS means that the progress that was being made in characterizing the global properties and internal structure of stars has slowed down. One way to make progress is to have another space mission that can make high-cadence observations of stars down to V > 16 magnitude for more than 4 years, with pixel sizes small enough to ensure contamination-free observations of dense stellar systems like cores of star clusters.

Highly multiplexed, panchromatic, spectroscopic surveys are needed to obtain the precise temperatures, luminosities, elemental abundances, and velocities—and their variance over time—for the number and variety of stars to be investigated in the coming decade. Multiplexed instrumentation matches the scale of current photometric (≈10¹¹ sources) and astrometric (≈10⁹ sources) samples and enables the discovery of rare stellar classes and short evolutionary phases that push the limits of astrophysical theory. High-sensitivity (large-aperture) spectroscopy extends stellar measurement far beyond the Milky Way and enables study of intrinsically rare stars that are unlikely to be close. Except for the case of very faint stars, where extremely large telescopes (ELTs) are required, the key to obtaining such large numbers of spectra is new instruments that can be put on 4–10 m telescopes. Panchromatic spectroscopy, facilitated by advanced UV and X-ray facilities, probes the full range of stellar phenomena, including coronal heating and mass ejection processes in the X-ray, mass loss from massive stars and atomic abundances of metal-poor stars in the UV, and the molecular chemistry of very cool and highly embedded stars in the infrared. Multi-epoch observations map stellar binaries and reveal invisible companions, probe stellar interiors through precision asteroseismic measurements, and unveil the dynamic atmospheres of cloudy brown dwarfs and massive evolving stars. Concurrent advances in spectropolarimetry are needed to map stellar magnetic field structures, particularly for the lowest- and highest-mass stars whose interior structures differ considerably from the Sun’s and whose interior dynamos remain poorly understood.

High-resolution imaging and interferometry are needed to resolve individual stars and stellar systems at relevant scales: 10–100 milliarcsecond optical and infrared imaging will isolate stars below the main-sequence turnoff in the Local Group, resolve and map the orbits of tight binaries, and enable proper motion selection in distant clusters; 10–100 microarcsecond near-IR imaging will provide parallaxes of cool and deeply embedded stars at kpc scales. Sub-microarcsecond interferometry will directly measure stellar radii down to substellar masses, monitor structural distortion in evolved stars, and resolve massive multiples in distant clusters. High-resolution imaging and imaging spectroscopy, and spectropolarimetry in the optical and infrared, will create direct probes of thermal and magnetic properties of the Sun.

Sensitive global telescope networks that enable frequent or continuous monitoring at optical and infrared wavelengths are needed to resolve stellar behavior over a wide range of time scales. These include stellar flares over seconds to minutes; stellar rotation, and magnetic spots or cloud structures over hours to days; and binary orbits and supernova progenitor and post-explosion evolution over months to years. These networks are also critical for prompt study of transient events, including mergers, tidal disruptions, SNe/gamma-ray bursts (GRBs), and gravitational wave events.

These observational capabilities have to be matched with investment in laboratory capabilities to provide the necessary atomic data to characterize highly ionized metal atoms, molecular opacity data at low temperatures, and pressures spanning the ISM to cool stellar atmospheres. As the solar metallicity problem shows, interpretation of stellar spectra is subject to systematic errors, and a thorough study of atomic transi-

tions under different densities and temperature is needed to resolve these issues. Also needed are studies to model the equations of state of stellar interiors, particularly degenerate stellar interiors. Investment is also needed to advance high-performance computing for 3D modeling of stellar interiors, atmospheres, and binary-star evolution, and to process and perform real-time analysis of the petabytes-per-day data flows anticipated in future surveys.

Maximizing future science capabilities and outcomes goes beyond investment in facilities; theoretical and numerical studies are also needed to advance our understanding. The next decade of astronomical research will rely on advanced software tools to analyze and interpret massive observational and theoretical data sets. The Image Reduction and Analysis Facility (IRAF) and Astronomical Image Processing System (AIPS) have respectively served the optical/infrared (OIR) and radio communities well over the past 40 years, but some tools, like IRAF, are not supported any more, and the community-based, mainly unfunded, efforts to make equivalent python packages (PyRAF, Astropy) still lack critical functionality, largely owing to lack of dedicated support from the funding agencies. Similar funding support is needed for development of codes for numerical simulations.

There need to be improvements in open-access computing facilities to enable computationally intensive analysis by the broader research community. Current models of access to national supercomputers are limited to extremely computationally intensive theoretical work. Such facilities are not available for intensive-data analysis.

But perhaps the most important developments needed are long-term archives of astronomical data. The current model for funding archives is haphazard, and very often the future of data of discontinued missions is unknown. There needs to be a system in place that can archive the data to expand their utility over time and scale, including maintaining the archive even after a program is discontinued. Also important are open data policies. Astrophysics missions such as Kepler and TESS, and ground-based observatories such as Global Oscillations Network Group (GONG), have shown how open access policies significantly increase the scientific output of research investment. For maximum scientific return, it is necessary for other publicly supported facilities, whether ground- or space-based, to make their data public in a reasonable amount of time.

TABLE G.1 Science Questions and Discovery Area

TABLE G.2 Summary of the Capabilities and Future Needs to Address the Science Questions and Discovery Area