2

Research, Observation, and Modeling Needs: The Sun and Heliosphere

A large share of the workshop’s panels was devoted to the research, observation, and modeling needs for various aspects of space weather, from the Sun and heliosphere to the magnetosphere and ionosphere–thermosphere–mesosphere (ITM) and on to the ground effects of space weather. These presentations specifically addressed three different parts of the statement of task:

• Examine trends in available and anticipated observations, including the use of constellations of small satellites, hosted payloads, ground-based systems, international collaborations and data buys, that are likely to drive future space weather architectures; review existing and developing technologies for both research and observations.

• Consider the adequacy and uses of existing relevant programs across the agencies, including NASA’s Living With a Star (LWS) program and its Space Weather Science Application initiative, the National Science Foundation’s (NSF’s) Geospace research programs, and the NASA-NOAA-NSF research-to-operations (R2O) and operations-to-research (O2R) programs for reaching the goals described above.
• Consider how to incorporate data from NASA missions that are “one-off” or otherwise nonoperational into operational environments, and assess the value and need for real-time data (for example, by providing “beacons” on NASA research missions) to improve forecasting models.

This chapter and the next two summarize discussions on those topics. Here in Chapter 2, the focus is on research, observation, and modeling needs for the Sun and heliosphere. Chapter 3 focuses on the research, observation, and modeling needs for the magnetosphere, ionosphere, thermosphere, and mesosphere, while the focus of Chapter 4 is the research, observation, and modeling needs to understand the ground effects of space weather.

In summarizing the presentations of the various panels, varying approaches are used in this and the other two chapters. In some cases the summary is done in a more traditional form, describing what each panelist said in turn. In other cases, in which there was significant overlap among the panelists’ presentations and the panel’s output became a group effort, the session summary is more of a synthesis and has only a few comments that are attributed to individual speakers.

Several sessions during the workshop were devoted to the issue of research, observation, and needs related to the Sun and heliosphere. Four keynote speakers reviewed various research needs, mainly relating to the Sun and heliosphere. A panel of presenters addressed model and observation needs related to the Sun, and a second panel took on the topic of model and observation needs related to the solar wind.

KEYNOTES

The four keynote speakers were Drew Turner of the Applied Physics Laboratory of Johns Hopkins University, Judy Karpen of NASA’s Goddard Space Flight Center, Noé Lugaz of the University of New Hampshire, and Kathryn Whitman of NASA’s Johnson Space Center. The speakers were given the following key questions for consideration:

• What do we need to understand to enable predictive capability of the following topics (shown below)?
• What are the research needs to make progress on that understanding?

Turner and Karpen were asked to discuss the state of space weather science and the adequacy of and gaps in existing space weather–related programs at the National Aeronautics and Space Administration (NASA), the National Science Foundation (NSF), and the National Oceanic and Atmospheric Administration (NOAA). Lugaz and Whitman were asked to address the following questions: “What do we need to understand to predict ICMEs (including stealth) and IMF Bz?” and “What do we need to understand to predict ‘all clear’ SEP periods?”

Turner’s talk focused on the key findings of a gap analysis conducted by the Johns Hopkins University Applied Physics Laboratory for NASA. The explicit task from NASA specified space-based observations can only be considered in the context of improved space weather predictive capabilities and the science of space weather (NASA 2021). Turner defined space weather as the intersection between the natural space environment as an object of fundamental scientific research and the threats posed to humans and technology systems that require risk mitigation by operational communities. Space weather activities span applied science and engineering as well as research-to-operations/operations-to-research (R2O2R) activities

The gap analysis Turner described breaks space weather effects into five categories: (1) direct drivers from the Sun itself, including solar flares and radio bursts; (2) radiation effects; (3) ionospheric effects, which affect radio signal propagation; (4) thermospheric expansion, which affects paths of satellites and other objects through the thermosphere; and (5) geomagnetically induced currents (GICs), which affect critical infrastructure on the ground. The NASA report characterized various space weather hazards according to their likelihood and their consequences, and it then combined that risk analysis with other factors, such as scientific merit, to prioritize current space weather observation gap categories. The top priorities in the resulting list were (1) solar and solar wind observations, including observations off the Sun–Earth line; (2) ionospheric key observables; (3) the solar wind in peri-geospace; (4) thermospheric key observables including ionospheric D- and E-region energetic particle precipitation and E- and F-region cusp and auroral precipitation; (5) ring current and radiation belt electrons; and (6) plasma sheet electrons and injections (or bursts) from cislunar into geosynchronous orbit (GEO) and medium Earth orbit (MEO) regions. Noting the importance of ground-based assets, Turner reminded the audience that the gap analysis was tasked to assess only space-based assets.

Some of the highest-priority observational gaps could be addressed with a near-future system-of-systems approach, through combining a network of state-of-the-art observatories, supporting infrastructure, and the space weather centers that develop the forecasts (see Figure 2-1 for persistent observations needed to fill the gaps). For example, to complement L1 monitors Turner suggested dedicated L4 and L5 monitors that could provide more observations of the solar disk, solar corona, and solar wind. The trailing L5 monitors are more useful for measuring potentially geoeffective active regions and solar wind stream structures,

while the leading L4 monitors are more useful to monitor the solar energetic particles. The backside of the Sun L3 observations offer a number of distinct advantages for longer-term forecasting but are challenged by complicated communication requirements. Turner also suggested additional solar wind monitors between L1 and the magnetospheric boundary, which could provide better magnetospheric specification. Finally, persistent observations of the radiation environment and of the lunar and cis-lunar environments are needed for astronaut and infrastructure safety. In addition to the above, he said, comprehensive ionosphere, thermosphere, and radiation belt monitoring and science are needed in order to improve space weather predictions. Turner also noted that within the next decade, it is expected that the thin shell in the thermosphere will host around 10,000 spacecraft. While they will need accurate space weather information, they could also be used to provide more knowledge of this region.

Turner said that filling these observational gaps will require a systems approach with coordinated, concurrent measurements. Creating such a system will require a long-term strategy coordinated over agencies as well as dedicated and supported implementation plans. Observing system simulation experiments (OSSEs) can be invaluable in providing cost–benefit analyses with respect to model performance (e.g., the L3/L4/L5 observations of solar magnetic field versus investments into transport models and helioseismology from the far-side). Tools provided by data assimilation techniques and machine learning can be used to fill in the data gaps.

In the next talk, Karpen said that to improve space weather predictions it will be necessary to better understand the relevant physical phenomena (i.e., the solar wind structure and turbulence; the nonlinear response of the magnetosphere; the ionospheric conductances, winds, and outflows; and the coupling between the neutral atmosphere and ionosphere). Furthermore, researchers will need to figure out how to couple both regimes (e.g., solar wind to magnetosphere) and physical transitions (e.g., magnetic field to plasma-driven dynamics, sub- to super-Alfvénic, collisional to collisionless plasmas) across spatial and temporal scales. She also suggested that space weather would benefit from collaboration with plasma physics researchers, who are already taking advantage of numerical and computational advances, and with Earth scientists, whose atmospheric models are important for understanding how space weather is driven from below.

Karpen argued that idealized models will need to be advanced through the use of data assimilation and model coupling as well as through better data-driven models and those incorporating continuous coverage of the solar magnetic field. The models should use the advances made in faster and larger massively parallel computational systems. The Community Coordinated Modeling Center (CCMC) and the International Space Weather Action Team (ISWAT) are making progress on standardizing the validation and verification of research models prior to transition. However, there is a serious gap in training and employing new computational physicists in heliophysics. If space weather research is to make advances that help improve space weather forecasting, cooperation between NASA and NOAA will be crucial. Undertaking the complex task to couple Earth system models with the space weather models will also require additional funding.

Lugaz started from the notion that knowledge of the full plasma and magnetic field measurements upstream of Earth is needed in order to forecast geomagnetic storms and the radiation belt enhancements. Measuring these upstream of Earth at L1 gives about 1-hour lead time. Knowing these parameters as the plasma leaves the Sun would give the longest lead times, but that is not currently feasible. Coronal mass ejections (CMEs) are the source of the strongest interplanetary magnetic field (IMF) and thus the most critical for space weather forecasts. Other magnetic field–driven impacts are typically created by corotating interaction regions (CIRs) or more complex and very difficult to predict events, such as CME–CME or CME–CIR interactions. Ordered according to decreasing lead time and increasing accuracy, the primary tools for forecasting and prediction are modeling and simulations, remote sensing observations, and in situ measurements (see Figure 2-2).

In terms of modeling and simulations, the key questions revolve around the ways that active regions emerge and evolve and around which active regions and coronal structures result in (fast) CMEs and how and when. These questions can be addressed by different approaches, depending on whether the forecast is made before or after the eruption has occurred, but in all cases computational limitations require using a combination of full physics description and approximations.

Long (over 4 days) lead time predictions are based on physics-based simulations driven by solar observations; better accuracy with lead times of 1 to 3 days can be obtained by combining remote sensing observations either with physics-based models using solar observational inputs or with empirical models. However, complex events, superstorms, and possibly solar maximum periods in general are all likely to require full simulations that incorporate multiple aspects that are currently non-standard; for example, variable solar wind, successive CMEs and CME interactions, or other unusual conditions.

Lugaz stated that remote sensing observations are key inputs to all modeling and simulations. The key science questions include the formation and variability of CIRs, CME evolution, and more complex events such as CME–CME interactions and stealth CMEs. In order to make progress in this area, two or three distinct observational views are needed, with a view from L5 (trailing the Sun–Earth line) being particularly valuable. Observations of the polar magnetic fields (which are best done out of the ecliptic plane) are key to understanding high-speed solar wind and the related non-CME geoeffective events, as well as understanding the CME sheath regions and CME deflection.

Regarding in situ measurements, Lugaz suggested merging different approaches, including data assimilation (of both remote sensing and in situ observations), ensemble simulations, parameterization of simulations,

and machine learning techniques. However, the most accurate way to forecast the IMF polarity affecting Earth is to measure it in situ in the solar wind upstream of Earth. Such observations have been primarily done at L1, but multiple measurements are needed to accurately characterize CMEs and other solar wind structures.

The key science questions that can be advanced using in situ observations involve mesoscales, for example, evolution of the CMEs and the solar wind on time scales of less than 1 day or over the 1-hour travel time from L1 to Earth’s bow shock, and determining the relevant length-scales of CMEs and other solar wind structures. Prior studies have shown that at times, there can be significant differences in the solar wind measured at L1 and at Earth’s bow shock. Addressing these issues requires multi-spacecraft measurements inside L1; rideshares and small satellites would be perfectly suited for these investigations. Unfortunately, as several speakers pointed out, the lack of coordination and systemic approach between missions results in only random conjunctions.

In the panel’s final presentation, Whitman focused on forecasting solar energetic particle (SEP) events as well as “all-clear” conditions (those when SEP fluxes are sufficiently low to allow extravehicular or other activities impacted by space radiation). She noted that operations on the International Space Station (ISS) take place at low Earth orbit inside Earth’s protective magnetosphere, which reduces the SEP threat to individuals. However, with planned missions farther out in space, prediction of both SEP events and all-clear SEP periods will become increasingly important. Furthermore, she said, the definition of “all-clear” is not consistent across the community. NOAA requirements are based on particle fluxes >10 MeV, while the Space Radiation Analysis Group (SRAG) augments that with an additional requirement for >100 MeV particles; research groups apply a variety of threshold levels to multiple energy ranges. She noted that a set of standard SEP “all-clear” definitions would ensure better usability for operations as well as facilitate cross-model validation and comparisons.

A major difficulty in predicting SEP events is that the largest events are associated with M- and X-class flares and fast CMEs, which occur only rarely; thus, the data sets are too sparse for statistical modeling. The pre-eruption forecasting is done by combining the likelihoods that an M- or X-class flare will occur, that a fast and wide CME will be produced, and that an SEP event will be produced and propagate to the location of interest. Thus, improvements in forecasting any of these phenomena should increase the reliability of SEP forecasts (Figure 2-3).

Post-eruption forecasts use information about flares, CMEs, radio waves, electrons, and protons in various combinations to estimate SEP characteristics before the particle intensities increase. Unfortunately, most models are currently unable to make predictions prior to the onset of a well-connected SEP event, and post-eruption forecasts suffer from data latency issues, manual determination of the CME parameters, and occasionally analysis run time. Machine learning holds a lot of promise in improving forecasts, particularly for pre-eruption models, Whitman said.

Whitman’s final comments reflected a common theme in the session—the importance of understanding the entire system of systems and the need for high-cadence coronagraphs and magnetographs, extreme ultraviolet imaging, and in situ magnetic field, and energetic particle measurements at L1, L4, and L5 (Figure 2-4). Whitman closed by saying, “It is critical that operationally supported, high-cadence, reliable and accurate space weather data streams for all phenomena relevant to SEP production are publicly available for operations and the deployment and development of models that require real-time observations.”

THE SUN

The Solar Panel, moderated by committee member Pete Riley, consisted of panelists Todd Hoeksema of Stanford University, Sarah Gibson of the University Corporation for Atmospheric Research, Cooper Downs of Predictive Science Inc., Craig DeForest of the Southwest Research Institute, Valentin Pillet of the National Solar Observatory, and Phil Chamberlin of the University of Colorado.

A common theme from the panel was that the single most important measurement for predicting and understanding the state of the corona and its temporal evolution is global (i.e., 4π steradian), time-dependent monitoring of the photospheric magnetic field, which provides the magnetic boundary condition of the solar atmosphere. This is how all models of the solar corona and solar wind are driven, both space weather/predictive models and first-principles/magnetohydrodynamics models. The time-dependent state of the corona affects the formation of the solar wind, stream interaction regions, and CMEs. This inner boundary condition is also needed for physical understanding of fundamental processes, such as magnetic energy storage and release. Current observational views of the polar magnetic field are limited, and observations from the far side of the Sun are missing. Current full-surface boundaries are created either by a series of static snapshots over 2 weeks to view the full solar disk or else by invoking models that artificially evolve the observational data to the part of the Sun that is not visible. Thus the boundary is not well represented, the time dependence of the solar corona is not well constrained, and global models are not based on any true representation of the solar surface at any given time (Figure 2-5).

Polar magnetic field measurements of the Sun are critical to understanding the solar wind and the magnetic topology into which a CME propagates. Unlike the backside of the Sun, which eventually rotates around to the front, the polar magnetic fields have never been measured without a severely limiting observing angle. Best estimates of the polar magnetic fields are provided by the Hinode spectropolarimeter, which, however, are only sporadically available. Instrumental limitations of other full-disk data (e.g., from the Solar Dynamics Observatory’s Helioseismic and Magnetic Imager [SDO/HMI] and the Global Oscillation Network Group [GONG]) do not provide consistent estimates of polar magnetic fields, and the resulting range of credible estimates is large, which then produces vastly different predicted heliospheres. Even measurements from the Solar Orbiter will not fully meet the observational needs, as measurements would be needed from both poles simultaneously.

Another common theme of the panel was the need to track the evolution of CMEs in three dimensions as they travel toward Earth. As solar wind and CME structures evolve between the Sun and Earth, current L1 coronagraph measurements are incapable of distinguishing between narrow, fast CMEs and wide, slow CMEs (Figure 2-6). Panelists suggested several possible observing scenarios, including multiple viewpoints (e.g., stereoscopy between an L1 measurement and off-Sun–Earth-line measurements) of a CME during its propagation. Several speakers mentioned the value of a polar coronagraph and heliospheric imagers, as the polar view provides a top view for all CMEs traveling near the ecliptic plane where the planets reside and allows detection of the longitudinal structure of CMEs, corotating interaction regions, and shocks. A near-term solution will be offered by the Polarimeter to Unify the Corona and Heliosphere (PUNCH) mission, which will use polarized white-light coronagraphs and heliospheric imagers to provide three-dimensional information and the chirality of the CME magnetic field, which will provide information of the IMF polarity upon arrival at Earth. The ratio of polarized light to total brightness provides information on the distance of the plasma feature from the observer; time series of this ratio will then provide three-dimensional information on the plasma structure.

The panel pointed out that most current CME propagation models do not include the magnetic field. Magnetohydrodynamic (MHD) models include the magnetic field but are computationally expensive given the large system size and resolution needed to resolve the structure in scales relevant to Earth. Modeling magnetized CMEs requires starting with an appropriate estimation of the magnetic field at the solar boundary, which has its limitations discussed above. The ideal solution may be a combined modeling and data assimilation approach that has been successfully used in Earth sciences applications, such as hurricane tracking.

In addition to the global photospheric magnetic field, global imaging of the middle corona was identified as an important missing piece to understand the Sun. For example, understanding the global restructuring of the corona after a CME requires global imaging of the entire corona (i.e., measurements from different viewpoints). In addition to the vantage point issues, the middle corona (1.5-4 solar radii) represents an observational gap in the range of imagers from extreme ultraviolet (EUV) monitoring the solar surface and white light coronagraphs recording the outflowing plasma at higher altitudes. As this

gap coincides with the critical region where CMEs are accelerated, it creates uncertainties in physical models of CME initiation. Future solar imagers should eliminate such sampling gaps either by extending the fields of view of EUV imagers higher or by reducing the noise in white light measurements close to the solar surface.

A longer-term goal is to develop the capability to routinely produce maps of the vector magnetic field in the photosphere, chromosphere, and corona. This would be a significant advancement over the current situation where the fields above the photosphere are inferred from models. In particular, observations of filament channels and magnetic neutral lines would allow observing the energy storage and release processes. Furthermore, measuring the chirality in the chromosphere at magnetic neutral lines would help track magnetic helicity build-up and provide knowledge of the characteristics of solar eruptions upon liftoff.

Understanding the global solar dynamo processes will not only allow better understanding of the overall 11-year solar cycle, it will also allow assessment of where and when “active nests” on the solar surface will occur, how much flux will appear in an emerging region, how the emerging flux will interact with existing flux patterns, or how complex the structure will be. Progress on these topics requires monitoring the solar interior, which can only be done through indirect helioseismology measurements; this would also benefit from multiple viewpoints.

The solar cycle and solar dynamo are important for the long-term space weather predictability (how active will the next solar cycle be?) and the coupling of space weather to atmospheric weather and climate. Measurements of the meridional circulation (measured via magnetic fields) can offer insights into the solar cycle because meridional circulation may act like the solar cycle “clock.” However, a challenge for the observations is that these phenomena should be measured consistently over long (solar cycle) time scales.

Solar irradiance is a key input for ITM driving and thus needs to be measured continuously. Solar EUV irradiance and soft X-rays are the primary energy input into the upper atmospheres of Earth and other planets, including Mars. As Chamberlin said, “It’s the photons! Not just the plasma!” There are no specific warnings either for short-term irradiance variations caused by solar eruptive events (flares) or for longer-term solar activity variations. Yet, irradiance variations have an immediate impact on the atmosphere, especially at low latitudes. Solar eruptive events are not understood to a great enough degree to predict their impact on irradiance, proxy models do not capture spectral variations, and physics-based irradiance models are computationally expensive. Irradiance measurements from L5, complementing those from the Sun–Earth line, would provide information on evolution of active regions and help estimate the CME mass via coronal dimming measurements.

THE SOLAR WIND

The Solar Wind Panel, moderated by committee member Nicholeen Viall, followed the Solar Panel. Its panelists were Joe Borovsky of the Space Science Institute; Vic Pizzo of NOAA’s Space Weather Prediction Center; Stuart Bale of the University of California, Berkeley; Nick Arge of NASA’s Goddard Space Flight Center; and Erika Palmerio of Predictive Science, Inc.

Continuing the common theme, the Solar Wind Panel noted the need for continuous, global monitoring of the Sun and the solar corona. Basic understanding of the solar wind and improving the solar wind models will require solar observations, as discussed above, as well as remote sensing and in situ measurements of the solar wind itself. Global coverage will require a fleet of monitors similar to those needed for solar observations (e.g., L4, L5, far-side, and over the solar poles). The panel members generally believed that global coverage is more important than increasing spatial resolution (Table 2-1), and it was noted

TABLE 2-1 Critical Data Products Needed to Improve, Validate, and Constrain Coronal and Solar Wind Models

NOTE: 3D = three-dimensional; A = ahead; ACE = Advanced Composition Explorer; B = behind spacecraft; B.C. = boundary conditions; CODEX = Coronal Diagnostic Experiment; DSCVR = Deep Space Climate Observatory; EUV = extreme ultraviolet; Ne = electron density; PHI = Polarimetric and Helioseismic Imager; PSP = Parker Solar Probe; PUNCH = Polarimeter to Unify the Corona and Heliosphere; SDO/HMI = Solar Dynamics Observatory’s Helioseismic and Magnetic Imager; SOHO = Solar and Heliospheric Observatory; SolO = Solar Orbiter; ST = STEREO = Solar Terrestrial Relations Observatory; SW = Solar wind; V&C = validate & constrain; WL = white light.
SOURCES: Nick Arge, NASA, presentation to the workshop; data from C.N. Arge, S. Jones, C.J. Henney, S. Schonfeld, A. Vourlidas, K. Muglach, J.G. Luhmann, and S. Wallace, 2021, “Multi-Vantage-Point Solar and Heliospheric Observations to Advance Physical Understanding of the Corona and Solar Wind,” Heliophysics 2050 white papers, https://www.hou.usra.edu/meetings/helio2050/pdf/4056.pdf, and A. Posner, C.N. Arge, J. Staub, O.C. StCyr, D. Folta, S.K. Solanki, R.D.T. Strauss, F. Effenberger, A. Gandorfer, B. Heber, C.J. Henney, J. Hirzberger, S. Jones-Mecholsky, P. Kuehl, and O. Malandraki, 2021, “A Multi-Purpose Heliophysics L4 Mission,” Space Weather 19:e2021SW002777, https://doi.org/10.1029/2021SW002777.

(see above discussion on solar observations and models) that the accuracy of the initial and boundary conditions at the Sun are critical for the accuracy of the solar wind models.

The optimal number of measurement points was discussed, with the panelists generally believing that there is a limited return on adding more than a couple of spacecraft in the ecliptic plane. While plans for wider longitudinal spread of solar and solar wind observations exist, the panel pointed out the need for high-latitude observations for both the photospheric magnetic fields and off-ecliptic views of CMEs. Thus, some panel members advocated prioritizing a polar viewpoint as opposed to additional in-ecliptic monitors. Major advances in observing polar fields and CME tracking could come from polar view missions, such as Ulysses; four spacecraft could provide continuous coverage of both poles, which would also allow tracking CME deflections, rotations, deformations, and interactions throughout the inner heliosphere. Furthermore, polar imagers would illuminate the extent and causes of the deviations of the IMF direction from the nominal Parker spiral, thereby improving understanding and predictions at Earth orbit. However, polar operational missions are expensive; Pizzo suggested that the missions will need to be designed to have very focused goals and expectations to manage the cost. Panelists also mentioned additional methods to track the solar wind and CMEs, specifically in situ measurements between the Sun and L1 and Faraday rotation measurements from the ground.

Two corollaries to the importance of multiple viewpoints and global coverage are that the intercalibration of instruments on board those spacecraft is crucial and that robust methods are needed for

incorporating the data into the models (see also Chapter 5). Magnetograph data, for example, vary in spatial resolution and polarization sensitivity, in the spectral lines observed, and in sampling patterns, making inter-calibrating magnetograph data a notoriously difficult task.

Global modeling of the coupled system is crucial, as understanding CMEs and SEPs requires a holistic Sun-to-heliosphere model. As CMEs and SEPs propagate within—and are transported through—the solar wind, it is necessary to have a model that simultaneously captures the states of the solar corona, the solar wind, and the CMEs. CME–CME interactions are known to influence their evolution, but the models are currently not well constrained. Furthermore, the lack of quantitative measurements of the magnetic field within CMEs makes it difficult to develop a full MHD model of CME propagation for operations (without data to serve as model input, constraints, and validation points). The panel noted that the modeling and predicting of the CME magnetic field are progressing and need to start to transition to operations (including validation). Likewise, some SEP models have real-time capabilities, but they need to be tested and their performance evaluated for real-time forecasting.

Finally, accurate measurements of the solar wind that actually hits Earth’s magnetosphere are needed to understand—and ultimately predict—the driving of magnetospheric dynamics. Measurements of the solar wind from L1 can deviate from those measured at the magnetopause due to triple aberration and inherent structures in the solar wind. Triple aberration (Figure 2-7) is caused by the combination of the motion

of Earth around the Sun, the non-radial component of the solar wind flow, and the interplay of magnetic structures, which propagate outward along the Parker spiral faster than the bulk solar wind plasma. As the solar wind flow velocity vector varies by about ±5 degrees over the 220 R_E distance from L1 to Earth, it causes a deflection that is about ± 30 R_E, or roughly the width of the magnetosphere. This would indicate that a plasma parcel going through L1 might completely miss hitting Earth’s magnetosphere due to triple aberration only. The challenges can be solved by a solar wind monitor between L1 and Earth. To address the inherent structures in the solar wind, the best solution would be to develop a better physical understanding of their causes. For example, it is not currently known how much of the structure is created at the Sun as the solar wind is formed, versus how much is generated en route through turbulence. To best disentangle the origin of the solar wind structures, improved measurements of elemental composition should be leveraged to differentiate between the structures created at the Sun versus structures created during propagation through the heliosphere (see also the section on new architectures in Chapter 6).

REFERENCES