New Observations Needed to Advance Our Understanding of Coronal Mass Ejections

Coronal mass ejections (CMEs) are large eruptions from the Sun that propagate through the heliosphere after launch. Observational studies of these transient phenomena are usually based on 2D images of the Sun, corona, and heliosphere (remote-sensing data), as well as magnetic field, plasma, and particle samples along a 1D spacecraft trajectory (in-situ data). Given the large scales involved and the 3D nature of CMEs, such measurements are generally insufficient to build a comprehensive picture, especially in terms of local variations and overall geometry of the whole structure. This White Paper aims to address this issue by identifying the data sets and observational priorities that are needed to effectively advance our current understanding of the structure and evolution of CMEs, in both the remote-sensing and in-situ regimes. It also provides an outlook of possible missions and instruments that may yield significant improvements into the subject.


Introduction
Coronal mass ejections (CMEs) are among the most spectacular eruptions in the solar system, consisting of copious amounts of plasma and magnetic field that are regularly expelled from the Sun. After erupting and as they travel through interplanetary space, CMEs tend to expand and interact with the ambient solar wind, resulting in large structures [measuring e.g. ∼0.3 AU in radial size by the time they reach 1 AU; Jian et al. 2018] that may have lost their twisted outer layers [Pal et al. 2021] and/or coherence [Owens et al. 2017] and are thus prone to deformations. The heliospheric evolution of CMEs may result in rotations, deflections, deformations, erosion, and re-configurations due to complex interactions of the ejected plasma with its surroundings [e.g., Manchester et al. 2017], including with other CMEs [e.g., Lugaz et al. 2017]. These aspects make CMEs extremely complex to fully characterize in 3D, in terms of both their morphology and magnetic configuration, especially in light of the limited observations that have been historically available.
For example, CMEs were first observed through remote-sensing data from a single viewpoint, i.e. Earth, with early observations of the Sun and its corona-including transient phenomena such as CMEs-made in the early 70s with e.g. OSO-7, Skylab, and groundbased observatories. The following Solwind and SolarMax (in the late 70s-early 80s) as well as SOHO (in the 90s) satellites brought significant improvements in the temporal and spatial resolution of solar data, but it was not until the launch of the STEREO mission in the 2000s that the first remote-sensing images away from the Sun-Earth line would be taken. STEREO consisted of twin spacecraft, STEREO-A and STEREO-B (leading and trailing Earth in its orbit, respectively), equipped with solar disk, corona, and heliospheric imagers in their remote-sensing suite. The availability of multi-point observations of the Sun and its environment has led to major advances in CME research, including improving our understanding of CME morphology [e.g., Thernisien et al. 2009;Wood & Howard 2009], better constraining CME propagation direction and speed through the solar corona as well as interplanetary space [e.g., Colaninno et al. 2013;Möstl et al. 2015], and enabling observations of CMEs that are "stealthy" from one viewpoint but evident from another [e.g., Nitta et al. 2021;Palmerio et al. 2021b]. Currently, remote-sensing imagers away from the Sun-Earth line can be found onboard STEREO-A (STEREO-B was lost in 2014), as well as the more recently launched Parker Solar Probe (heliospheric imagers only) and Solar Orbiter (disk, corona, and heliospheric imagers). The latter two, however, do not have their remote-sensing instruments operational at all times, leaving STEREO-A a unique and persistent viewpoint to support solar imagery from near Earth.
On the other hand, in-situ measurements of the solar wind and interplanetary magnetic field-including transient phenomena such as the interplanetary counterparts of CMEs, also known as ICMEs-have been performed away from the Sun-Earth line already since the 60s and 70s, via mission programs such as Pioneer, Helios, and Voyager. However, given the large spatial scales involved, the likelihood of obtaining multi-point measurements of the same ICME is drastically reduced, so much so that the first analysis of the internal magnetic structure of CMEs was published only in the early 80s, using data from five different spacecraft between 1 and 2 AU [Burlaga et al. 1981]. Since then, multi-spacecraft studies of ICMEs have taken advantage of various heliophysics missions as well as planetary ones. These have enabled studies, among other topics, of the longitudinal variation [e.g., Farrugia et al. 2011;Kilpua et al. 2011] and radial evolution [e.g., Good et al. 2019;Salman et al. 2020] for selected ICME events. Nevertheless, such fortuitous spacecraft configurations are rather rare, and most analyses showcasing multi-point ICME measurements are characterized by arbitrary relative spacecraft geometries [e.g., Witasse et al. 2017;Palmerio et al. 2021a], which tend to complicate interpretation of an event, e.g. in discerning whether certain features are due to CME evolution in interplanetary space or to local distortions along the whole structure. Currently, in-situ measurements of the inner heliospheric environment are available from Earth, partially from Mars, as well as the STEREO-A, Parker Solar Probe, BepiColombo, and Solar Orbiter spacecraft. Given that each of these observers follows its own orbit around the Sun, studies that can address the internal structure and evolution of CMEs in interplanetary space via multi-point measurements are limited to those periods that are characterized by a propitious spacecraft configuration.
After over five decades of CME research (spanning over five solar cycles), we have obtained significant statistics on the characteristics and properties of CMEs from singlepoint (both remote-sensing and in-situ) observations [see, e.g., the CME and ICME catalogs of Gopalswamy et al. 2009a;Richardson & Cane 2010;Harrison et al. 2018;Jian et al. 2018;Nieves-Chinchilla et al. 2018]. Hence, it is clear that it is extremely challenging to reach a deeper insight on the intrinsic structure and evolution of CMEs based on a single viewpoint. To improve and expand our understanding in fundamental CME research, we need a set of dedicated observations that are aimed at treating CMEs as 3D structures that are in constant evolution and interaction with the ambient solar wind. In this White Paper, our goal is to identify such specific observations, in both the remote-sensing and insitu regimes, and address their benefits for the studies of CMEs and ICMEs in a holistic way. We will also elaborate on possible missions that would be able to meet the mentioned observational requirements and conclude by addressing our recommendations to the Heliophysics 2024 Decadal Survey Committee.

Remote-Sensing Observations
In terms of remote-sensing measurements, we will identify observational gaps that are crucial for a more complete understanding of CMEs and their heliospheric evolution in two main areas, namely direct imaging ( §2.1) and radio probing ( §2.2).

Direct Imaging
As mentioned in the Introduction, the launch of the STEREO mission in 2006 represented a major advancement for CME science from a remote-sensing perspective: For the first time, the same eruption could be observed on the solar disk from more than one viewpoint (for a total of three, i.e. Earth plus the twin STEREOs, depending on the source region location), and three coronagraph suites were concurrently operational, providing multi-point views of the solar corona. Additionally, both STEREO spacecraft were equipped with heliospheric cameras, constantly imaging the space between the Sun and Earth (and beyond). This has enabled, for example, the development of catalogs and stereoscopic studies of CMEs imaged by both STEREO probes in white light through the solar corona [Vourlidas et al. 2017] and interplanetary space [Barnes et al. 2020].
Despite the ground-breaking progress brought by the multi-viewpoint capabilities of the STEREO mission, all solar observations to date have been characterized by one common factor, i.e. they have all been performed from the vicinity of the ecliptic plane. This issue will be partially addressed by Solar Orbiter, which is planned to ultimately reach a heliographic latitude of 33 • during its extended mission, in July 2029 [Müller et al. 2020]. However, even Solar Orbiter's maximum elevation with respect to the Sun's equator is significantly closer to the ecliptic plane than to the solar poles, and thus the full potential of a polar imager will not be explored. Furthermore, the ever-changing longitudinal separation between the STEREOs (and Solar Orbiter) and Earth means that the advantages and availability of these additional viewpoints are inconsistent-especially after the loss of STEREO-B in October 2014 and with STEREO-A crossing the Sun-Earth line in August 2023. The benefits of sustained (fixed) observations from quadrature and polar views for tracking the solar wind and its transient phenomena (including CMEs) have been reviewed by Gibson et al. [2018, see also Figure 1]. Figure 1, with Parker Solar Probe (PSP; red) and Solar Orbiter (SO; green) orbits overlaid. The dashed orange line represents a sample "diamond" orbit with four spacecraft (gray cubes) 90 • apart, reaching as high as 75 • heliolatitude (Vourlidas, 2017). The colored patches on the 1AU sphere are the same as in Figure 1.

FIGURE 8 | Interesting viewpoints shown as in
designs. This means that the instruments operate for limited amounts of time per orbit (10 days for PSP, 30 days for the SO remote sensing payload). Other complications include restricted data volumes due to the limited number of downlinks and onboard storage, data latency on the order of months, and angular configurations between the two spacecraft and Earth that are continuously changing.
These operational constraints reduce the length of polar magnetic field observations by SO, for example, to only a few days per orbit (up to 10 days per orbit by the end of the extended mission). The shortness of this time period will limit helioseismology studies (Löptien et al., 2015). The data volumes and short observing periods also do not allow for synoptic observations or consistent 360 • coverage of the solar atmosphere from the SO imagers. PSP imaging is restricted to a wide field heliospheric imager providing context for the in-situ payload but has no disk imaging (for thermal reasons).
It is undeniable that both missions will provide unique data and views of the coronal and heliospheric environment, and enhance the sophistication of multi-viewpoint analysis, far beyond STEREO. However, they are necessarily limited to the science achievable by the instruments they have on board, and in their ability to obtain the sustained measurements needed for longer-time-frame studies and space-weather monitoring (see Table 4). We now consider how next-generation extra-SEL mission concepts might address the remaining gaps between the observational capabilities of our existing and planned missions, and the outstanding science questions raised in this paper.

Missions for the Future
Several white papers describing extra-SEL mission concepts were submitted during the last Solar and Space Physics Decadal Survey activies, and were also summarized in the 2014 Heliophysics Roadmap. Variations of those concepts were submitted for the Next Generation Solar Physics Mission call for ideas. The majority of these white papers are not in the published literature, but a subset, with emphasis on helioseismology science, is discussed by Sekii et al. (2015). Here, we present some representative concepts for mission architectures that emphasize sustained measurements, and so fill most, if not all, of the gaps indicated by Table 5 (depending on instrument payload).

Quadrature Mission
The idea of a mission to the Lagrange L5 point has been explored in several concepts in recent years (Webb et al., 2010;Gopalswamy et al., 2011;Vourlidas, 2015;Lavraud et al., 2016). It has obvious advantages for space-weather research and forecasting, such as more accurate speed measurements for Earth-directed CMEs and increased coverage of solar irradiance and the photospheric magnetic boundary for operational models. It also addresses many of the open science questions that particularly benefit from sustained measurements, such as probing the solar interior more deeply, and observing the evolution of structures over time, as well as 3D modeling of CMEs and their source regions. Such analyses depend upon the existence of complementary SEL observations.
From the mission design perspective, injection toward the L5 (and L4) points is relatively straightforward but requires significant V and hence a large rocket. There is considerable trade space: orbit around L5 vs. stationed at L5, drift vs. direct injection, and travel time vs. mission length, to name a few. Indeed, it is also possible to put something at 90-120 • off the SEL utilizing a launch into a geo-transfer orbit and electric propulsion.
An L5 concept was studied for the Solar and Space Physics Decadal Survey, and estimated to cost above 600M(FY14) for a standard spacecraft with multiple instruments. An operational space-weather mission to L5 could be cheaper if it had a reduced payload-at minimum, a coronagraph and a magnetograph-as was the case studied by Trichas et al. (2015). As of the writing of this paper, the European Space Agency (ESA) is in Phase-A development for an operational L5 mission.
An innovative approach based on a fractionated spacecraft concept was also proposed by Liewer et al. (2009). Instead of a monolithic spacecraft carrying a multitude of instruments, the authors proposed the launch of a set of cubesats or minisats each carrying a single telescope along with a minisat carrying a standard antenna to relay communications from the constellation to Earth. The advantages include (1) lower launch costs through extensive use of hosted payload opportunities, (2) measurement persistence, since failed cubesats or telescopes could be replaced with another launch, and (3) redundancy, since a single spacecraft failure would not take down the whole constellation. The disadvantages include (1) the necessarily reduced size of instrument payloads, leading to restrictions on aperture size and other performance metrics, (2) data acquisition limitations-since cubesats have less powerful radios and smaller antennas, and (3)  These considerations are not only valid for solar disk and coronal observations, but also for heliospheric imaging. First of all, an imager from Earth's perspective has been lacking since 2011, when the Solar Mass Ejection Imager onboard the Coriolis spacecraft was deactivated [this issue will be addressed by the PUNCH mission to be launched in April 2025, Deforest et al. 2022]. As shown by Amerstorfer et al. [2018], a heliospheric imager can provide advantageous measurements to track CMEs that are even directed toward the observer itself (e.g., an imager at L1 would be useful to study Earth-directed CMEs). Additional heliospheric cameras from quadrature and/or polar vantage points would enable stereoscopic analysis of the evolution of CMEs long after they have left the outer corona. Moreover, the PUNCH mission will be equipped with photometric capabilities, allowing for studies of the 3D structure of the solar wind and its transients. Given the difficulties in tracking CMEs through the helio-sphere even with an optimally-placed imager or a pair of imagers [Lugaz 2010], single-view polarization measurements can assist in overcoming the problems of stereoscopy by resolving the CME leading edge and identifying substructures within CMEs. The benefits of polarized heliospheric imaging for CME research have been discussed by DeForest et al. [2016].
Finally, a crucial capability that was not realized on the STEREO suite of remote-sensing instruments is represented by magnetographs. In addition to being able to link CMEs and their internal structure to the magnetic configuration of their source regions [e.g., Palmerio et al. 2017], one of the largest sources of uncertainty in heliophysics magnetohydrodynamic (MHD) modeling is the structure and evolution of the coronal magnetic fields [Wiegelmann et al. 2017]. In fact, all (realistic) MHD models of the corona and inner heliosphere require the full-Sun surface-field as a boundary condition. The development of flux transport models [e.g., Hoeksema et al. 2020] has been largely driven by the lack of magnetograph coverage away from the Sun-Earth line. Being able to observe the full emergence, evolution, and decay of active regions on the far side of the Sun would be extremely beneficial in terms of quantification of the magnetic flux, energy, and helicity budgets associated with CME magnetic source regions [e.g., Vourlidas et al. 2020b]. Additionally, an increased/full coverage of solar surface magnetic field measurements would also improve model performance on CME and CME-driven shocks [Jin et al. 2022].

Radio Probing
An additional aspect of CME research is represented by remote-sensing observations that do not consist of "images" in the strictest sense. These measurements are usually realized at radio wavelengths and enable probing of different characteristics of the interplanetary medium "from a distance" via groundbased facilities or space-based missions.
The most widely used applications in this sense comprise observations of CME-related radio emissions in the form of noise storms and bursts [see the review by Vourlidas et al. 2020a]. In particular, so-called Type II radio bursts are used to track CME-driven shocks through the solar corona and inner heliosphere [e.g., Lara et al. 2003;Gopalswamy et al. 2009b]. When radio spectra from two separate locations are available, Type II sources can be triangulated to reconstruct shock fronts in 3D [e.g., Magdalenić et al. 2014]. This technique, however, can have large uncertainties that are dependent on the signal-to-noise ratio (S/N), instrument crosscalibration, and the broad angular width of the radio emission-generally restricting its use to near-Sun frequencies (MHz). Improvements in detector S/N as well as the frequency coverage and number of observing spacecraft may increase the utility of the technique for tracking ICME-driven shock propagation, especially if used in conjunction with simultaneous white-light coronal/heliospheric imagery.
Another way of probing CME properties at radio wavelengths is realized via Faraday rotation measurements, i.e. of the rotation of the polarization plane when linearly polarized radiation propagates through a magnetized plasma, e.g. from a CME [see the review by Kooi et al. 2022]. In particular, Faraday rotation can be used to remotely probe CME magnetic fields [e.g., Wood et al. 2020, see also Figure 2]. This technique, however, has not been explored to its fullest potential yet, since it requires the presence of a background transmitter of (at least partially) linearly polarized light, which may be either a natural radio source (e.g., a pulsar, a nebula, and/or a galaxy) or a spacecraft. This aspect poses limitations to the availability of line-of-sight observations for a given CME. Given that it is impossible to constrain the positions of galactic and extra-galactic sources, utilizing spacecraft as background radio sources may be a promising opportunity. Past observations have suffered from the lack of suitable probes in optimal positions and/or from the limited availability of ground-based telescope operations; nevertheless, some studies have successfully detected CMEs using radio signals from Pioneer 9 [Levy et al. 1969], Helios [Bird et al. 1980], and

Figure 4
Corona and CMEs on 2 August 2012 as observed with the L points are the LOS to the radio sources (a) just before occultation; and 0900 by CME-1, CME-2, and CME-3, respectively; (c) during and CME-2; and (d) during occultation of 0843 by CME-2 only. 08 solid curves (LE-1 and LE-3) and dashed curves (LE-2) represent the the Earth side and far side of the Sun, respectively. These figures ar LOS and CME geometries onto the two-dimensional LASCO-C3 ima (LASCO-C3 vantage point) in Figure 5. The photosphere appears as th occulting disk, and the horizontal axis is the heliographic Equator wi the LASCO public archive: sohowww.nascom.nasa.gov.
images. Because CME-1 and CME-2 overlap, we have ou bright outer loops (LE-1 and LE-2, respectively) in this fi LASCO-C3, CME-1 is in the foreground, and the leading e line in Figure 4. CME-2 is in the background, and the lead dashed line.
The quasar 0842 had the largest range in impact param LOS was located near the heliographic Equator; the heliog point decreased from 11.8 • to 11.2 • and the Carrington lon 247.4 • . As may be seen in Figure 4, 0842 was occulted by C and continued to be occulted by this CME for the duration MESSENGER [Jensen et al. 2018]. These limitations and challenges could be addressed by a multi-spacecraft mission to be used as a constellation of multi-frequency background transmitters, allowing to remotely probe CME magnetic fields along multiple lines of sight.
In fact, these considerations are valid also for the interplanetary scintillation technique [e.g., Manoharan 2010; Jackson et al. 2020], which is used to study fluctuations in the intensity of a radio source due to solar wind irregularities, allowing e.g. to remotely probe CME density [e.g., Lynch et al. 2002].

In-situ Measurements
In terms of in-situ measurements, we will identify observational gaps that are crucial for a more complete understanding of CMEs and their heliospheric evolution in two main research regimes, namely those dedicated to studies of the large-scale ( §3.1) and smallerscale ( §3.2) structure of ICMEs. Note that here we consider "small scales" to correspond to angular separations of less than ∼6 • at ∼1 AU, i.e. absolute distances of less than ∼0.1 AU.
interplanetary shock type (see e.g., Berdichevsky et al., 2000, Figures 1 and B, parts (a) and (b)). We then search near the shock ramp for the regions upstream/downstream of the shock that are most consistent with their belonging to the same magnetic plasma domain upstream/downstream of the shock compression. Having identified the upstream/downstream regions, we evaluate the shock normal, its error, and the velocity of propagation in the interplanetary medium as shown for ''The p-ave Technique'' (Berdichevsky et al., 2000, Appendix A1) in such a way that the shock solution is sturdy and gives the best solution consistent with the Rankine-Hugoniot equations (see Berdichevsky et al., 2001). The results obtained are listed in Table 3.
At STB: Several hours ahead of the MC there is in the observer frame (RTN coordinates) a fast forward interplanetary shock moving at a speed of 390 km s À 1 and oriented 121 westward. The direction of this shock is consistent with its being driven by the high speed stream observed later, on November 20. There are two reasons for dismissing the shock as one being driven by the MC: (i) the shock speed 5 450 km s À1 , the MC's leading speed, and (ii) the orientation of the MC axis appears to be oblique to the shock normal, in contrast to well-known cases of shocks driven by MCs (see e.g., Lepping et al., 2001).
At wind: Ahead of the MC there is a fast forward interplanetary shock moving at a speed of 450 km s À 1 radially away from the sun. This shock appears to be driven by the MC because (i) its velocity is the same as that of the MC's leading edge, and (ii) the shock orientation is orthogonal to orientation of the MC axis, supporting the idea that the upstream solar wind is pushed as the MC displaces itself as a whole away from the sun, consistent with the interpretation in other studies (see e.g., Lepping et al., 2001).
The arrival of the high speed stream is accompanied by the presence of a fast reverse interplanetary shock. This relatively

Large-Scale Studies
CMEs are extremely large heliospheric structures, expanding to several times the size of the Sun already when traveling through the solar corona and reaching typical radial widths of ∼3 AU at heliocentric distances of ∼15 AU [Richardson et al. 2006]. Accordingly, studies that aim to address the global morphology, magnetic configuration, and plasma distribution of ICMEs tend to rely on multispacecraft measurements over large heliospheric distances. However, only a handful of dedicated heliophysics missions have been launched away from Earth during the past ∼5 decades (Helios, Ulysses, STEREO, Parker Solar Probe, and Solar Orbiter). Hence, the research community has been taking advantage of data from planetary missions (both during cruise phase and after orbit insertion), such as MESSENGER to Mercury, Venus Express to Venus, MAVEN to Mars, Juno to Jupiter, and Cassini to Saturn. This has led to a wealth of multi-spacecraft studies utilizing measurements "not originally meant for CME science" [e.g., Mulligan et al. 1999;Nieves-Chinchilla et al. 2012;Witasse et al. 2017;Good et al. 2019;Kilpua et al. 2019;Davies et al. 2020Davies et al. , 2021Davies et al. , 2022Palmerio et al. 2021c] to accom-pany those based more strictly on heliophysics missions [e.g., Skoug et al. 2000;Farrugia et al. 2011;Winslow et al. 2021;Lugaz et al. 2022, see also Figure 3], as well as studies focused on ICMEs at other planets than Earth [e.g., Winslow et al. 2015;Lee et al. 2017].
Nevertheless, a handful of spacecraft (whether intended for heliophysics or planetary studies) each following its own orbit around the Sun across the vast heliosphere signifies that most multi-point CME encounters are fortuitous detections that are realized over arbitrary relative separations of the various observers, making it particularly challenging to systematically characterize CME structure and evolution. For example, if two in-situ locations are separated by ∼1 AU in heliocentric distance and ∼30 • in heliographic longitude, it becomes nearly impossible to attribute structural and compositional differences to radial evolution, to longitudinal variations, or to both. Several works have analyzed the radial evolution of CMEs based on a few encounters characterized by nearradial alignment of the involved spacecraft [Good et al. 2019;Vršnak et al. 2019;Salman et al. 2020], but most multi-point events cannot rely on such rare configurations [e.g., Möstl et al. 2022]. Thus, measurements that are aimed to directly address multi-point observations of CMEs-e.g., via a series of probes with well-defined separations in longitude/latitude and heliocentric distance-are likely to bring outstanding advancements in the field of heliophysics.
Another important issue in CME in-situ research is represented by the absence of significant progress in many in-situ instrumentation capabilities in the past thirty years. Measurements and cadence provided by ACE and Wind for more than 25 years are often considered to be "good enough" for many studies, whereas higher-cadence compositional and energetic particle data, as well as advancements in radio measurements as highlighted above are necessary to make significant progress towards answering withstanding science questions.

Smaller-Scale Studies
Studies of the structure of CMEs over smaller distances (≲0.1 AU, corresponding to ≲6 • at 1 AU) have been even less frequent. Since the STEREO mission was launched during solar minimum, only a couple of events were observed in situ before the relative separation between the twin spacecraft (and between each of them with Earth) became too large for analysis of the smaller-scale structure of CMEs [e.g., Liu et al. 2008;Kilpua et al. 2009;Mulligan et al. 2013]. Some works have taken advantage of the presence of both the ACE and Wind spacecraft near Earth. For example, Möstl et al. [2008] performed an optimized flux rope reconstruction for a CME detected in November 2003 using data from both satellites, with ACE at L1 and Wind in the dawn direction closer to Earth's magnetotail. During 2000-2002, Wind performed prograde orbits, yielding separations with ACE (at L1) of ∼0.01 AU, which allowed Lugaz et al. [2018] to study variations in the ejecta magnetic field for 21 ICMEs, and Ala-Lahti et al.
[2020] to analyze differences in the structure of CME-driven sheaths for 29 events. These works resulted in evaluation of the typical scale lengths associated with various solar wind structures (see Figure 4). For example, Lugaz et al. [2018] concluded that the scale length of longitudinal magnetic coherence inside CME ejecta lies around 0.25-0.35 AU (14-20 • at 1 AU) for the magnetic field magnitude, but around 0.06-0.12 AU (3-7 • at 1 AU) for the magnetic field components.
These results highlight that there exists a region of the parameter space left largely unexplored, i.e. that corresponding to radial separations of 0.005-0.050 AU and angular separations of 1-12 • . This 'mesoscale' region could be researched further during mid-to-late 2023, when STEREO-A will be positioned close to the Sun-Earth line, or via orbital maneuvers of the existing assets at L1, or via a dedicated multi-spacecraft mission.

Improvements to Bridge Current Gaps
The critical "gaps" identified in both the remote-sensing (Section 2) and in-situ (Sec-lower coherence in Bx. However, we found weak or no correlation between magnetic field measurements and different shock parameters. As the low-pass filtered magnetic field data also hint that there is a coherent embedded global magnetic field in the ICME sheath (Figure 4), we conclude that extensive physical mechanisms, such as the field line draping around the ICME ejecta, are plausible explanations for the observed differences in the scale lengths between the magnetic field components. Analysis of our results suggests that field alignments in the ICME sheaths are oblique to the radial direction, and we noted that the maximized total correlation has a displacement from the time lag giving the shock alignment ( Figure 1b). Possible variations in defining the shock transition could cause this. Another possibility is that alignments formed in the draping of the magnetic field are aligned to the surface of the ICME leading edge and not the shock plane (Kataoka et al., 2005). Fixed sheets of magnetic field direction are then measured by the spacecraft with a lag that differs from the lag of aligning the shock boundaries, which further implies the plausible importance of the draping in explaining the presented observations. Our observation of the double-peaked distribution in Figure 1b coincides with this discussion.
In this study, we have discovered that magnetic fields in the ICME sheath are more coherent than what they are in the solar wind. To illustrate this, we sketch in Figure 6 an ICME complex in Earth-centered Figure 6. Sketch of an ICME complex in Earth-centered interplanetary space in the ecliptic plane. The ICME sheath is preceded by an interplanetary shock (dark blue curve) and driven by ICME ejecta, bounded by orange curves, within which there is a flux rope illustrated with an exaggerated twist. The ICME complex is modeled as arcs of a circle by taking the average angular width of the ICME ejecta given by Zhao et al. (2017) and the average radial width reported by Kilpua et al. (2017) for the sheath. Blue lines show interplanetary magnetic field (IMF) that has 45°Parker spiral angle at the Earth's distance from the Sun. The sheath is occupied by magnetic fluctuations and the field lines drape around the ICME ejecta. Also, turbulent progress of the fluctuations is exemplified by the eddies within the sheath. Scale lengths of the solar wind (Richardson & Paularena, 2001), ICME sheath (Table 1), and ICME ejecta (Lugaz et al., 2018) are illustrated in the y-direction. The near-Earth space is shown in the zoomed box where red, and black curves indicate the bow shock and magnetopause boundaries that are estimated by using the models given by and Merka et al. (2005) and tion 3) observational regimes can be addressed with a conscious, focused commitment on the part of NASA, NSF, NOAA, and other federal funding agencies to facilitate multi-point and multi-spacecraft heliophysics and planetary science missions, support for multi-probe data analysis and modeling research programs, and an innovative crossdisciplinary approach to maximize the scientific return from missions beyond their primary scientific objectives. First of all, it is evident from the discussion in the previous sections that CME science (but more generally solar and heliospheric physics) has been largely lacking support for dedicated multi-spacecraft capabilities. The first such mission in the inner heliosphere is HelioSwarm, which was selected for flight as recently as 2022 to study turbulence in the solar wind (and is expected to launch no earlier than 2028), while spacecraft constellations have been employed for magnetospheric studies for over two decades (e.g., with Cluster, THEMIS, and MMS). Missions and assets that can enable systematic, multi-point studies of CMEs and related phenomena in both the remote-sensing and in-situ regimes will be a prerequisite for breakthrough science and understanding to be achieved over the next decade. This much-needed progress can be realized via a multitude of observing strategies, such as coordinated spacecraft at the Lagrange L4 and L5 points (or even including the L3 point), constellations in heliocentric orbits and/or in solar polar orbits, and "swarms" of smaller probes (e.g., cubesats) clustered upsteam of Earth and/or other planets. A number of mission concepts relevant to these goals have been proposed over the past few years [e.g., Vourlidas et al. 2018Vourlidas et al. , 2020bBemporad 2021;Allen et al. 2022;Telloni 2022].
Other practical ways to maximize science return include the incorporation of simultaneous remote-sensing and in-situ capabilities on all heliophysics flagship missions that are set to fly through the solar wind, wherever possible. This would allow for an increased coverage of the solar atmosphere and/or extended corona in terms of imaging, and for the presence of an additional in-situ monitor scattered throughout the heliosphere. The temporal cadence of each instrument should reflect the timescales of the processes they are supposed to shed light upon: For example, coronagraph imagery from SOHO/LASCO (12-min cadence) and STEREO/COR2 (15min cadence) is often of limited use especially in the case of fast CMEs, which can evolve dramatically over much shorter timeframes, particularly in the range ∼1-10 R ⊙ . The selection of new missions should aim to explore presently unknown regions (e.g., the solar poles) and enable novel science methodologies (e.g., using a spacecraft network as radio transmitters) but should, at the same time, guarantee continuity of the existing assets and data products.
Furthermore, synergies between heliophysics and planetary science should be facilitated and even encouraged. In addition to the multi-point CME studies enabled by planetary missions mentioned in Section 3.1, there are e.g. several recent examples of novel planetary science results made with Parker Solar Probe data during its flybys of Venus [e.g., Collinson et al. 2021;Pulupa et al. 2021;Wood et al. 2022]. Another example is the community-driven, heliophysics working group formed to analyze the in-situ magnetic field and particle data being returned during BepiColombo's cruise phase [Mangano et al. 2021]. Such cross-disciplinary studies may be even more successful (and less fortuitous) if NASA (and other agencies) prioritize developing multi-spacecraft and multi-viewpoint capabilities at the highest levels of mission selection, funding, planning, and implementation. Two relatively straightforward action items that have the potential to maximize future scientific returns are (1) the inclusion of magnetometer, solar wind plasma, and energetic particle instruments on upcoming and future planetary science missions, and (2) to provide the funding and support for making science-quality cruise phase data readily available to the science community.

Summary and Recommendations
In this White Paper, we have outlined the current status of CME observational research and identified the current gaps that need to be filled in order to reach a more complete understanding of their internal structure, properties, and evolution. Our recommendations for the Heliophysics 2024-2033 Decadal Survey Committee are: • Prioritize multi-point inner heliospheric observations via coordinated spacecraft and/or constellations of probes • Improve simultaneous remote-sensing and in-situ coverage over the solar atmosphere/corona and the inner heliosphere • Observe the Sun and its environment from novel viewpoints, e.g. away from the ecliptic • Encourage CME studies that target heliospheric evolution over a range of radial and longitudinal separations • Address observational gaps in the CME 'mesoscale' region • Enable higher-cadence remote and in-situ measurements relevant to CME science • Select new missions with the aim to explore new regions, but without losing continuity of the existing assets • Include heliophysics-relevant instrumentation on planetary missions, to maximize cross-disciplinary studies and science return