A Coronal Mass Ejection Impacting Parker Solar Probe at 14 Solar Radii

The relationship between CME properties in the corona and their interplanetary counterparts is not well understood. Until recently, a wide spatial gap existed between the two regions, which prevented us from disentangling the spatial and temporal evolution of CMEs. NASA’s Parker Solar Probe (PSP) has imaged multiple CMEs since its launch in 2018, but these events either intercepted the spacecraft far from the corona or completely missed it. Here we describe one of the first CMEs observed simultaneously by remote sensing and in situ instruments, and compare the corresponding measured properties, such as orientation, cross section diameter, density, and speed. The CME encounter occurred on 2022 June 2, while PSP was around 14 solar radii from the Sun center. We reconstruct the CME with forward modeling and determine its morphology and kinematics. The reconstruction suggests that PSP misses the CME apex but encounters its flank. The encounter time matches the period when the PSP in situ measurements indicate the passage of a CME. We also reconstruct the flux rope diameter and orientation using the in situ magnetic field measurements. The results are consistent with the CME reconstruction from imaging data. The close agreement between remote sensing and in situ analyses suggests that discrepancies found in past studies are more likely associated with the CME temporal evolution. We also find that the magnetic field of the CME flank extrapolated to 1 au is well below the average solar wind background and likely indistinguishable from it. This point could explain past events where the CMEs' interplanetary counterparts were not identified.


Introduction
Coronal mass ejections (CMEs) are large-scale structures containing plasma and magnetic fields ejected from the Sun into the heliosphere.They are the main cause of multiple space weather disturbances, including the most extreme geomagnetic storms (Gosling 1993;Gonzalez et al. 1999).They are also associated with shocks, a key source of energetic particles (see, e.g., Webb & Howard 2012, and references therein).
A key point for understanding CME evolution in the inner heliosphere, as well as forecasting the space weather disturbances caused by CMEs, is associating the properties of CMEs and their interplanetary counterparts (frequently called ICMEs).A correct association is necessary to understand whether a CME reaches a given target, such as Earth's magnetosphere, and what is its magnetic field configuration at this point.
Over the past decades, this association has not always been straightforward.Multiple CMEs can be ejected in close proximity and may interact with each other.The large distances between the CME observations in the corona and their in situ counterparts in the inner heliosphere complicate the problem.CMEs are typically observed in the solar corona, particularly up to 30 R e .They can be observed farther from the Sun by imagers on board the Solar Terrestrial Relations Observatory (STEREO) mission (Kaiser et al. 2007) and the Solar Mass Ejection Imager (SMEI; Howard et al. 2013).For many CMEs observed in the past years, interplanetary observations are only available close to 1 au, with a few events also observed by planetary missions before 1 au (Luhmann et al. 2020, and references therein), or the Helios missions (Porsche 1981).
The CME−ICME connections have been a prominent issue in heliophysics for many years.Multiple efforts have been made to relate the CME properties with their in situ counterparts, including multiviewpoint observations and empirical models (Schwenn 1983;Sheeley et al. 1985;Lindsay et al. 1999;Gopalswamy et al. 2001;Schwenn et al. 2005;Nieves-Chinchilla et al. 2012, 2013;Möstl et al. 2014).Before the STEREO data became available, multiple studies tried to associate ICMEs measured around 1 au with their coronal counterparts.An extensive study was performed by Schwenn et al. (2005) using 181 halo CMEs observed by the LASCO coronagraph, and only 50.3% of the events were uniquely associated with ICMEs.It was the STEREO mission that clarified the close relationship between CMEs and ICMEs thanks to its unique imaging from the solar corona to 1 au.While the large connection between the two types of structures is not in doubt, the detailed properties, gleaned mostly by comparing image-based to in situ−based reconstruction, remain unresolved.Despite extensive studies, many cases have major discrepancies between the 1 au predictions and the actual observations remain unexplained (e.g., Manchester et al. 2017;Wood et al. 2017).They include the CME time of arrival, speed on arrival (see Vourlidas et al. 2019, and references therein), orientation, and size (Colaninno et al. 2013;Braga 2015;Wood et al. 2017;Braga et al. 2020).At this point the sources of errors in each case are not clear.Wood et al. (2017) associated ICMEs identified at L1 with their source CMEs observed by the twin STEREO spacecraft.Among 41 events at L1, only 31 were unambiguously connected to CMEs.The rest do not have clear coronal origins.Furthermore, they found that the CME orientation from imaging reconstructions differed from their in situ counterparts by more than 40°, on average.They also found discrepancies in the event duration.While the L1 measurements last on average 19 hr, CMEs extrapolated to L1 would last 55 hr, suggesting that the models overestimate the CME expansion rates and sizes.It is unclear whether these discrepancies are due to analysis errors, assumptions used on reconstruction methods, or temporal evolution.The cause of these discrepancies, however, could be traced back to multiple phenomena taking place in the corona and interplanetary medium, e.g., interaction with other events or CME rotation, deformation, deflection, and/or erosion.CMEs can rotate and deflect at a rapid rate during their early evolution (Lynch et al. 2009;Vourlidas et al. 2011;Isavnin et al. 2014).Their shape and propagation direction can also be changed if they interact with other CMEs or stream interaction regions (Liu et al. 2012;Temmer et al. 2012;Möstl et al. 2015;Mishra & Srivastava 2015;Isavnin 2016;Braga et al. 2022).Erosion is a process where part of the CME magnetic flux is lost due to reconnection while it propagates and interacts with the solar wind around it (Dasso et al. 2007;Ruffenach et al. 2012;Manchester et al. 2014).
The large distance between the remote sensing observations and the in situ counterpart was reduced thanks to the pioneering observations from the Parker Solar Probe (PSP; Fox et al. 2016) mission, launched in 2018 August.While before PSP we had only in situ observations above 0.3 au, the perihelia of this spacecraft range from 35 to 9 R e , with the shortest perihelion expected to be reached in 2024 December.This makes it possible to image the inner solar corona from within, hence allowing for simultaneous in situ and remote sensing observations of coronal events directed toward the spacecraft.
Since the PSP launch, multiple CMEs have been observed by its remote sensing instrument suite, called the Wide-Field Imager for Solar Probe Plus (WISPR; Vourlidas et al. 2016), a suite of two white-light heliospheric imagers that observe along the spacecraft ram direction.Their combined field of view (FOV) extends from 13°.5 to 108°from the Sun−spacecraft line.Despite WISPR observations, the CME in situ counterparts did not reach PSP (Howard et al. 2019;Hess et al. 2020;Liewer et al. 2020;Wood et al. 2020;Braga & Vourlidas 2021;Howard et al. 2022).Conversely, PSP obtained in situ measurements for some events, but mostly far from the solar corona (Lario et al. 2020;Nieves-Chinchilla et al. 2020;Kilpua et al. 2022;Pal et al. 2022;Ledvina et al. 2023).Some in situ signatures of a CME leg observed at 14 R e were reported by McComas et al. (2023), but they were not simultaneous with the remote sensing observations.The first CMEs with simultaneous remote and in situ observations were analyzed by Niembro et al. (2023) and Wood et al. (2023), but those encounters occurred in the inner heliosphere at 46 R e .
In this paper we describe and model one of the first coronal encounters of a CME.We have simultaneously remote sensing and in situ observations from PSP close to its perihelion.The event took place on 2022 June 2, when PSP's solar distance was 14 R e .We compare the CME properties derived from remote sensing with the in situ measurements.This event is of special interest owing to its proximity to the Sun and because the temporal evolution of the CME (and any possible change in configuration or complexity) is not a factor in such comparison.
This paper is organized as follows.In Section 2, we describe the remote sensing observations and the CME source region.We reconstruct the CME in 3D and derive its kinematics in Section 3. The magnetic field and plasma observations are described in Section 4. In Section 5, we use two models to determine the orientation and size of the CME using solely the magnetic field measurements and compare properties of these models with the CME reconstruction.In Section 6, we compare density measured in situ with remote sensing.Next, we extrapolate this CME to 1 au and compare it with ICME statistics (Section 7).Finally, we summarize our findings in Section 8.

Remote Sensing Observations
The source of the CME is a filament eruption from an active region at S17E100.The region, later named NOAA Active Region 13029, was well observed by the EUVI images on from the Sun Earth Connection Coronal and Heliospheric Investigation (SECCHI; Howard et al. 2008) suite on board STEREO-A.The EUVI-A images in Figure 1(a) show a north-south oriented filament lifting off at around 05:50 UT.The filament ejection was followed by evacuation of the overlying corona and post-CME loop arcades.The CME front, marked by the arrow in Figure 1(b), indicates that the eruption was nonradial and the CME later emerges close to the solar equator.The Atmospheric Imaging Assembly (AIA; Lemen et al. 2011) on board the Solar Dynamics Observatory (SDO) observed only the off-limb signatures of the event.It showed the eruptive filament in projection and revealed that the eruption starts from the northern edge of the filament (Figure 1(c)).This is in agreement with the overall initial trajectory toward the solar equator suggested by the EUVI-A images.
The Large Angle Spectroscopic Coronagraph (LASCO; Brueckner et al. 1995) on board the Solar and Heliospheric Observatory (SOHO; Domingo et al. 1995) and the STEREO-A/SECCHI/COR2 coronagraphs observed the CME.Both spacecraft were located close to 1 au, with SOHO located around L1 and STEREO-A 28°eastward from the Sun-Earth line.The CME was first observed at 06:48 UT on LASCO C2 over the east limb.We also had east-limb observations from SECCHI COR1 from 06:55 to 12:00 UT.The heliospheric imagers from SECCHI/STEREO-A did not observe this CME, as expected for an east-limb event with the particular angular separation between STEREO-A and Earth.
The CME images in COR2 and C2 are suggestive of a threepart CME morphology (Illing & Hundhausen 1985;Vourlidas et al. 2013).A front and core are visible, but the intervening cavity is not well resolved, possibly due to overlapping structures.The inclined CME front (Figure 2) suggests a magnetic structure somewhat tilted with respect to solar north.Such a configuration would tend to obscure the three-part morphology.For visualization purposes, we indicate PSP's location by an asterisk over the COR2 animation on Figure 2.Even though the projected PSP location overlaps the CME in multiple images, this is not sufficient to confirm that the spacecraft is engulfed by the CME.PSP could be located anywhere along the line of sight.Thus, a CME reconstruction in 3D is needed to determine whether the encounter takes place or not.As follows in Sections 3 and 4, we confirm this point with both reconstruction and in situ data.
PSP also has remote sensing observations of this CME thanks to WISPR.The WISPR observations for this event are shown in Figure 3.The Sun is located beyond the left edge of the image.The event studied here is visible only across the bottom part of the WISPR FOV as indicated by the white arrows in Figure 3.The CME was not observed by WISPR from the ejection time until around 10:00 UT, when the CME was likely at smaller elongation angles than the WISPR FOV.During this period, WISPR observations were dominated by another CME in the FOV, which is not our focus here.The source of that CME and the CME itself are located westward of the Sun-Earth line as both SECCHI COR2 and LASCO C3 observe it on their west side.More details about this event are shown in Section 4.1.

Reconstruction
The remote sensing observations alone are insufficient to determine the CME dimensions and position at a given time.Thus, we need to reconstruct the event to determine when it engulfs PSP.We do so using forward modeling with the graduated cylindrical shell (GCS) model (Thernisien et al. 2006;Thernisien 2011), an empirical model meant to reproduce the entire CME outer envelope, emulating the geometry of a flux rope structure.The model is formed by two conical legs connected to a curved tube reminiscent of a torus.This tube has a circular cross section that increases with height.Because of its shape, this is sometimes called the "croissant" CME model.Dimensions, position, direction, and orientation of the structure are manually adjusted to reproduce all observations available at a given time, including coronagraphs and heliospheric imagers such as WISPR.
The GCS model is described by six parameters.Three of them define the CME orientation: longitude f, latitude θ, and tilt angle γ.The other three parameters determine the CME dimensions, namely (i) h f , the front height; (ii) δ, the half-angle in the leg's conical region, which relates to the CME edge-on width; and (iii) α, the half-angle between legs' axes, which reflects the CME width face-on.
We adjust the CME parameters not only to constrain the model with remote sensing observations (as usual) but also to match the CME encounter duration suggested by the in situ measurements.We do so because we have them while the CME is in the coronagraph FOV, and our goal is to find a fit consistent with all observations.As coronagraphs observe brightness associated with electron density, we use the time profile of this parameter measured in situ.As explained in Section 4, we interpret density increases measured by PSP around 12:40 and 14:30 UT as the times when PSP enters and leaves the CME.We find this approach particularly helpful to constrain the GCS parameters at the time the CME leaves PSP, especially the longitude and angular width.While GCS is a well-known technique that has been used for more than a decade, we believe that this is the first GCS reconstruction performed to match not only the white-light observations but also the in situ crossing time.This innovative aspect gives us extra constraints to determine the parameters of the model.
We find the following parameters for the GCS model: f = − 72°, θ = 3°(both in heliographic coordinates), γ = 8°, α = 8°, and δ = 9°.The height we obtained is 17.4 R e at 12:00 UT and 20.2 R e at 14:00 UT, which results in an average CME apex speed of ∼350 km s −1 .This set of parameters reproduces the coronagraph observations before the encounter (from 08:00 and 12:00 UT) by changing only the CME height (Figures 4  and 5).Thus, our reconstruction is consistent with all observations, including the 1 au coronagraphs, WISPR, and the in situ density increase timing.
We compare the CME reconstruction with PSP position in Figure 6.The left panel represents the CME projection on the solar equatorial plane and reveals that PSP intercepts the CME flank only close to the conical leg end.The right panel is perpendicular to the solar equator, and it shows that PSP crosses the southern portion of the CME.
The CME encounter with PSP depends on the GCS parameters' errors.They arise because we adjust each parameter to reproduce the observations, and a range in their values results in virtually identical CME reconstructions projected in the observerʼs FOV.As the actual CME parameter errors are unknown, we estimate errors for this event from previous studies with observers in similar conditions.A recent study calculated the errors by comparing the parameters derived by multiple experts in CME reconstructions independently using synthetic images generated by a CME model ( Verbeke et al. 2023).While the parameters were known, they were not shared with the experts before the reconstruction.They found that events observed by two viewpoints separated by angles of ∼60°have the following errors: ∼2°for latitude, longitude, and tilt; 0.16 R e for height; ∼5°for α; and ∼2°for δ.Another study from Thernisien et al. (2009) estimated GCS errors for multiple CMEs with two viewpoints separated by angles between ∼40°and ∼70°.The mean errors they found are ∼4°for longitude, ∼1°.8 for latitude, ∼22°for tilt, and 0.48 R e for height.
Given PSP's orbit and the CME trajectory, the encounter would not happen if we had errors in latitude and/or δ larger than the CME angular extent in this dimension, which is 20°.This is not the case, as both errors add up to 4°, and PSP latitude is 6°above the southern CME portion.Conversely, errors in f and α are particularly relevant in this case, as PSP intercepts only the flank (Figure 6).Although these errors could bring some doubt on the encounter, the in situ signatures of ICMEs and their timing are consistent with an actual crossing (Section 4).

Magnetic Field and Plasma Parameters
In this section, we describe the in situ magnetic fields and basic plasma parameters measured by PSP.These time profiles are shown in Figure 7.We use the magnetic field measurements from FIELDS (Bale et al. 2016) with temporal resolution of ∼0.22 s.The proton moments, such as speed and temperature, are obtained using the SPAN-I distributions (Livi et al. 2022) of the Solar Wind Electron Alpha and Protons (SWEAP) instrument suite (Kasper et al. 2015).To have an accurate estimation, we use electron densities obtained via the quasithermal noise (QTN) technique (Moncuquet et al. 2020) using FIELDS data.The resolution of electron densities is ∼3.5 s.
One indication of the CME encounter is the increase in the magnetic field intensity between 11:50 and 14:00 UT even after we normalize it by the square of the solar distance (Figures 7(a)).Moreover, the magnetic field components show rotation, which is a signature of ICMEs (see, e.g., Zurbuchen & Richardson 2006, and references therein).
Despite the magnetic field increase and possible rotation, the period between 11:50 and 14:00 UT contains no other clear ICME signatures, rendering the definition of the boundaries somewhat ambiguous.One possible interpretation could be that the CME encounter begins after the discontinuity at around 12:50 UT.The period before it could then be some kind of sheath-like region, but it lacks the increase in density that is expected for a pileup region.Moreover, most of the period after 12:50 UT does not have multiple ICME signatures.An exception is between 13:30 and 13:50 UT.The magnetic field fluctuation diminishes, suggesting a reduction in the turbulence levels.The plasma beta parameter decreases below 0.1 inside the CME, indicating that the plasma is magnetically dominated, satisfying most ICME identification criteria (see Zurbuchen & Richardson 2006, and references therein).On the other hand, this 20-minute period lacks clear magnetic field increase and rotation, so we cannot apply the usual methods to reconstruct the flux rope orientation (Section 5).
The proton bulk speed of ∼300 km s −1 is consistent with the CME flank speed we get from the GCS reconstruction (Section 3).We derive an average CME apex speed of ∼350 km s −1 , and ∼300 km s −1 for the flank, which is the region encountered by PSP.
This event is not preceded by a shock or other discontinuity.This is unsurprising since the CME speed of ∼300 km s −1 is close to the ambient solar wind speed measured ahead of this event.Shocks are generally observed only when a transient structure velocity with respect to the background solar wind exceeds the Alfvén speed.This relative speed is close to zero here, as the proton bulk speed is similar inside the CME and in the solar wind ahead of it.We estimate that the Alfvén speed in the region ahead of the ICME ranges from ∼200 to ∼400 km s −1 depending on the density and magnetic field intensity of each plasma portion.We observe two discontinuities inside the ICME, one at around 12:44 UT and the other at 14:07 UT.Both have a sharp magnetic field depletion combined with sudden density and speed increases.Although the discontinuities resemble slow-mode shocks (see, e.g., Richter et al. 1985;Farrugia et al. 2001), rotation or tangential discontinuities might have similar signatures.Thus, a more detailed study is needed to confirm or reject this point, which is outside the scope of this work.
The two discontinuities inside this ICME suggest that this is a complex event.Complexity normally refers to the degree of deviation from the "standard" configuration that we expect for the CME counterparts, such as a magnetic cloud structure or a flux rope (Winslow et al. 2022).As CMEs may erode, deform, deflect, or kink/rotate as they propagate in the interplanetary medium, there is an inference that CMEs would become more complex as they propagate away from the Sun.This process was suggested as a possible mechanism for discrepancies between CME and their corresponding ICMEs.While this increase is reported in some cases (Scolini et al. 2023), there is no definitive answer about the relation between complexity and solar distance (see Winslow et al. 2022, and references therein).
For instance, an ICME with a high level of complexity was observed before by PSP at 53 R e (Nieves-Chinchilla et al. 2020).As the measurements from the current event are taken at around 14 R e , complexity may not arise exclusively from interplanetary propagation.
The discontinuities and other deviations of this CME from the "standard" configuration might also be a result of the flank encounter.Previous studies show that some CMEs do not have clear in situ counterparts when the observing spacecraft passes through the flanks, missing the central flux rope (Kim et al. 2013;Luhmann et al. 2020, and references therein).In fact, only about one-third of ICMEs observed at 1 au have clear in situ signatures, such as rotation of their magnetic field and low plasma beta (Jian et al. 2006;Nieves-Chinchilla et al. 2019).Some ICMEs may not include clear signatures owing to the interaction with another structure, such as a high-speed stream.Winslow et al. (2021) studied two events observed in longitudinal conjunction; one kept its structure quite unchanged, but the other had major changes owing to interaction with structures such as high-speed streams.This led to an ICME observed at Mercury as a very organized structure becoming complex at 1 au.Palmerio et al. (2022) investigated a CME observed at Earth that missed Mars owing to the interaction with a high-speed stream, causing deflection and rotation.If any interaction happened for the current event, it would affect the CME for a much shorter period than in previous studies mentioned.Thus, interaction with a highspeed stream is unlikely to produce major changes in the ICME in situ signatures observed by PSP.
The discontinuity observed around 14:00 is followed by an approximately 30-minute period of decreased magnetic field intensity and increased density.This behavior resembles heliospheric plasma sheet regions identified in previous PSP studies (Lavraud et al. 2020;Palmerio et al. 2024).We also identified a magnetic field reversal at 17:30 UT, which we interpret as a heliospheric current sheet crossing.The proximity of these two structures to the CME may suggest that some interaction between them takes place, but further work is needed to verify this point, and this is outside the scope of this article.

Is the CME−ICME Association Unambiguous?
To determine whether the PSP in situ measurements are associated exclusively with the CME under study here, we check whether other events could reach PSP in the same period.We inspected SECCHI, LASCO, and WISPR images up to 1 day before the current CMEs.We found two candidates, as follows.
The first event was observed by LASCO starting on 2022 June 1 as a faint CME at around 17:00 UT in C2 and at 20:00 UT in C3.The position angle of this event suggests that it is located entirely to the north of PSP.Thus, we do not expect in situ signatures associated with this CME.
The second event is the CME mentioned in Section 2 that occupies most of the WISPR FOV on 2022 June 2.It was also observed by LASCO C2, C3, and COR2 as a west-limb event, and later it enters the FOV of heliospheric imagers on STEREO-A.From these observations we can conclude that the CME is directed westward from the Sun-Earth line.As PSP was over the eastern limb on 2022 June 2, this event cannot be intercepted by PSP on this day.Thus, the CME we reconstructed in Section 3 is the only one in the right direction to encounter PSP.
We inspected the in situ measurements on the hours preceding and following the ICME, and we did not identify any other event.Therefore, as we found only one CME directed toward PSP on 2022 June 2, the CME−ICME association is unambiguous.

Deriving the Flux Rope Orientation from the Magnetic Field
In this section, we derive the flux rope orientation using solely in situ magnetic field measurements.Our goal is to compare it with the orientation derived by GCS reconstruction, which relies on imaging instruments.
We use two well-established flux rope models to derive the flux rope axis orientation and cross section at PSP, namely, the circular-cylindrical (CC) model and the elliptical-cylindrical (EC) model (Hidalgo et al. 2000;Nieves-Chinchilla et al. 2016, 2018).Both models assume an axially symmetric magnetic field with twisted lines to derive the flux rope axis orientation, impact distance, and cross section size.CC assumes a circular cross section and EC an elliptical one.The EC model is a generalization of the CC model and has two additional parameters: the ratio between minor and major cross section axes, and the ψ cross section major-axis angle with the spacecraft trajectory.
The magnetic field components derived using EC and CC are compared with measurements in Figure 8, and their parameters are listed in Table 1.EC reproduces the measurements reasonably well and better than CC, which underestimates the magnetic field intensity in the first half of the event.The better performance of EC for this event is also reflected in the correlation coefficients calculated between each model and observations (see Table 1, last row).As expected for this type of modeling, the short-scale variations cannot be reproduced, including the sudden changes in the magnetic field, such as the discontinuities mentioned in Section 4. We compare the EC and CC flux rope orientations with the GCS in Figure 9.The black (purple) line centered on PSP represents the EC (CC) flux rope axis.As the orientation is derived from the magnetic field observed in situ by PSP, we understand it as representative of the local portion of the CME axis, which corresponds to approximately the top of the legs.The length of the axis is arbitrarily chosen for visualization purposes, as neither model has parameters related to this dimension.Overall, the flux rope orientation from EC and CC agrees with the CME reconstruction.EC and CC models suggest that the flux rope has a small elevation angle with respect to the solar equator, with its main axis approximately perpendicular to the Sun-Earth line (Table 1).We emphasize that EC and CC angles are defined differently than we do in GCS.While the angles on the latter refer to the CME apex direction, EC and CC angles refer to the axis perpendicular to the cross section in the portion of the structure observed in situ.The longitudes of EC and CC depend on the particular portion of the CME crossed, and they are approximately perpendicular to the GCS longitude close to the CME apex only, which is not the case here.Despite the differences in longitude, the tilt angle from EC and CC models is comparable to the GCS tilt angle.In this event, we found a low tilt in all cases.The GCS tilt (8°) is close to the EC tilt (7°; see Table 1).
We also compare the GCS cross section with EC and CC.As the GCS radius is variable, we take the point intercepting PSP rather than the apex.According to Figure 9, this corresponds to the widest portion of the conical legs.At this point, the cross section radius is given by h tan δ.We calculate this using the CME parameters and their error estimates from Section 3. The results are shown in Table 2. Overall, the GCS cross section is consistent with CC and EC when we consider the error estimates.This result seems to be better than those from Nieves-Chinchilla et al. ( 2012) and Wood et al. (2017), which have major discrepancies in the diameter.While the remote sensing and the in situ observation are taken at the same time here in this study, the in situ observations from the previous papers were taken at 0.5 and 1 au, typically hours after the last remote sensing observations available.This result suggests that the temporal evolution between the corona and 1 au might explain the discrepancies observed.

The CME Density Derived by Remote Sensing and
In Situ In this section, we derive the CME electron density from coronagraph and heliospheric imager observations.Our goal is to compare it with the density measured in situ by PSP.
Once we have the CME reconstruction completed and its parameters defined, we get its dimensions and volume using the equations from Thernisien (2011) and some algebra, which we detail in the Appendix.The only CME parameters needed are the height at each time (h F = 17.4 R e at 12:40 UT and h F = 20.2R e at 14:30 UT) and the angular width edge-on and face-on, which are represented by κ = 0.16 and α = 8°.Doing so, we get a volume of R 226 7.6 10 m 3 28 3 ( ) ´at 12:40 UT.Once the volume is determined, we need the CME mass to calculate the density.This parameter is routinely calculated for all CMEs observed by the LASCO coronagraphs, and it is available in the CDAW CME catalog (Yashiro 2004).A detailed description of the method is provided in Vourlidas et al. (2010).For this event, the LASCO CME mass is 1.092 × 10 15 g.As the mass calculation assumes 90% of hydrogen and 10% of helium (Hildner et al. 1975;Vourlidas et al. 2010), we assume the same composition here to derive the number density.Thus, we get 7280 cm −3 at 12:40 UT and 4653 cm −3 at 14:30 UT.This decrease is explained by the increase in the CME volume over time combined with a Now we compare this result with the electron number density measured in situ (Figure 7(f)).Within the CME period (from 12:40 to 14:30 UT), the density varies significantly, and we get an average of 8028 cm −3 .
We attribute the differences to particularities of each observation type.The density calculated using coronagraph observations considers the entire CME, extending to latitudes beyond the region probed by PSP.Conversely, the in situ measurements reflect only the CME portion probed, which corresponds to the flanks and may include part of the sheath region.Within the CME encounter period, the density is higher in two periods (between 12:45 and 13:30 UT and between 14:00 and 14:28 UT) and lower between 13:31 and 13:47 UT, which corresponds to the interval with clearer ICME properties, such as low plasma beta and low magnetic field fluctuations.In the latter, the density is ∼2670 cm −3 .We also speculate that the CME density in the flanks might differ from other CME regions.While the CME propagates and expands, different portions may reach background solar wind plasmas with distinct properties.As the ambient plasma is expected to play a role in the CME expansion and force balance, deformations may occur (Odstrcil et al. 2004;Owens 2006;Manchester et al. 2014;Braga et al. 2022).We speculate that this could result in local variations, which may explain differences between in situ and imaging densities.

How This Event Compares to Others
To our knowledge, no other spacecraft at farther solar distances measures this event in situ.We do not expect in situ observations from STEREO-A or L1, as the CME does not reach the red and green lines in Figure 6, which represent these two spacecraft.Solar Orbiter is located approximately 180°w estward from the Sun-Earth line on this date, far from the CME path.As only PSP in situ measurements are available for the current event, we extrapolate the CME in time until it reaches 1 au and compare its properties with 1 au averages in Table 3.
One major difference between this event and many ICMEs measured at 1 au is the shorter spacecraft encounter time.While the current event is measured in situ for approximately 2 hr, most 1 au events last more than 10 hr (Richardson & Cane 2010).This difference can be easily explained.
First, the CME expands as the solar distance increases and the ambient pressure decreases.If the CME expands selfsimilarly following the GCS model up to 1 au, its diameter would be 60 R e (0.26 au) at its apex.According to an extensive study from Richardson & Cane (2010), the average 1 au diameter is 0.33 ± 0.01 au, suggesting that the current event would be below average but not an outlier.We also estimated the 1 au diameter using the empirical expression R 0.78±0.10, where R is the solar distance (Bothmer & Schwenn 1998).By applying this expression, we derive a cross section diameter of 0.12-0.21au at 1 au, which is smaller than the self-similar expansion.This result also suggests that the current event's size is below average.A second explanation for the short duration might be the spacecraft path within the CME.Longitudes of L1 observatories do not change considerably while they probe a structure.As CMEs typically do not deflect in longitude, a 1 au probe usually crosses ICMEs radially.On the other hand, PSP's path inside the CME has a significant component perpendicular to the radial direction.PSP moves from −63°to −57°Stonyhurst longitude while inside the CME.
The current event seems to be below average not only in size but also in magnetic field strength.If we take measurements between 13:30 and 13:50 UT (approximately 500 nT) and we consider that it decreases with the square of solar distance, we get 2 nT at 1 au.As the mean magnetic field of ICMEs at 1 au is 10.1 nT, and only a few events display magnitude below 4 nT (Richardson & Cane 2010), this event lies in the lower range of the magnetic field intensity.Moreover, it is unlikely to be discernible at 1 au, where the average solar wind magnetic field is 6.35 ± 0.01 nT (Richardson & Cane 2010).This event illustrates a case where the connection between CME and in situ counterparts would be problematic to make, as the CME  is clearly observed from multiple viewpoints, while its 1 au would be most likely imperceptible, possibly leading to the incorrect conclusion that the event missed a 1 au target spacecraft.
While the magnetic field measurements of the current event are below average, they may not reflect the peak intensity in the CME magnetic field.This happens because PSP does not probe regions close to the CME central axis, where flux rope models expect the highest magnetic field (Marubashi 1997).Thus, the in situ measurements may not reflect the peak magnetic field of this event.
We also compare the current event's density with 1 au averages.According to Bothmer & Schwenn (1998), it is ∝R −2.4±0.3 .By using this expression with the density observed in the period with clearer ICME properties (2670 cm −3 ), we find a 1 au average density of 9 cm −3 , ranging from 5 to 19 cm −3 when considering errors.As the average ICME density at 1 au is 6.9 ± 0.2 cm −3 (Richardson & Cane 2010), the current event density is likely above average.

Conclusions
We analyze a CME observed simultaneously by remote sensing and in situ instruments in the corona (∼14 R e ) on 2022 June 2.We summarize our results as follows: 1. WISPR observes the CME as a diffuse structure when engulfing PSP.It does not observe the CME in the hours before the arrival because of the orbit and FOV of the instrument.2. PSP intercepts a CME portion close to the top of one leg, ∼10°of longitude away from its apex.3. We detect signatures of ICMEs such as increased magnetic field (after normalizing by PSP solar distance), magnetic field rotation (particularly along the tangential direction), and lower fluctuation levels.4. When we compare the 1 au expended values of this CME, we find that the current event has below-average magnetic field and cross section radius but above-average density. 5.The CME orientation derived by the reconstruction agrees with the flux rope orientation we derived solely from magnetic field measurements using the EC and CC models.We found that the axis from both methods is   almost parallel to the solar equator and to the Sun-Earth line.This agrees with the orientation of the CME portion engulfing PSP we derive with GCS. 6.The CME speed derived by the reconstruction for the flank (∼300 km s −1 ) is consistent with the in situ measurements.7. The CME density from remote sensing observations is 7280 cm −3 at 12:40 UT and 4653 cm −3 at 14:30 UT owing to its expansion.The in situ electron density varies significantly inside the CME with an average value of 8028 cm −3 .As PSP probed only the CME flank, we speculate that the difference could be explained by local density changes.8.The studied CME would have been indiscernible in 1 au magnetic field measurements if it were encountered as a flank because its expected magnitude would be 2 nT, well below average solar wind values.This observation illustrates a possible reason for inconsistencies in the association between CMEs and ICMEs reported in the past.9.The CME is rather complex since it has two discontinuities.As the relation between complexity and solar distance is not totally clear at this point, one may argue that CMEs are more complex close to the Sun.On the other hand, the flank hit may also explain this point.We expect that future events observed by PSP may help us clarify this issue.
We do not find any major discrepancy between the in situ measurements and those derived from the CME reconstruction.The level of agreement between the two types of measurements in our work is better than that found in previous studies, e.g., in Braga (2015), Wood et al. (2017), andColaninno et al. (2013).The reason behind this improvement seems to be the removal of the time lag between the in situ measurements and the time of the coronagraph images used for the 3D reconstruction.This suggests that the CME temporal evolution from the upper corona to 1 au is likely a reason for the discrepancies reported in previous studies.
Additional events with simultaneous observations from WISPR and in situ instruments from PSP are expected in the near future as the solar cycle activity increases.This will help us disentangle the spatial and temporal evolution of CMEs, better understand their interaction with the solar wind, and eventually improve the prediction of their arrival time.France).STEREO/SECCHI data are available for download at https://secchi.nrl.navy.mil/.The FIELDS experiment on the Parker Solar Probe spacecraft was designed and developed under NASA contract NNN06AA01C.We acknowledge the SWEAP team led by J. Kasper for use of SWEAP data.The authors acknowledge CNES (Centre National d Etudes Spatiales), CNRS (Centre National de la Recherche Scientifique), the Observatoire de PARIS, NASA, and the FIELDS/RFS team for their support to the PSP/SQTN data production, and the CDPP (Centre de Données de la Physique des Plasmas) for their archiving and provision.The SOHO LASCO CME catalog is generated and maintained at the CDAW Data Center by NASA and The Catholic University of America in cooperation with the Naval Research Laboratory.SOHO is a project of international cooperation between ESA and NASA.LASCO was constructed by a consortium of institutions: the Naval Research Laboratory (Washington, DC, USA), the Max-Planck-Institut für Aeronomie (Katlenburg-Lindau, Germany), the Laboratoire d'Astronomie Spatiale (Marseille, France), and the University of Birmingham (Birmingham, UK).

Appendix The CME Volume
Here we describe the expressions we used to calculate the CME volume.Once we reconstructed the CME, we can determine its dimensions using three CGS parameters (α, δ, and h f ).
As defined in Thernisien (2011), the CME is formed by two conical legs and a toroidal portion connecting them.Thus, the CME volume V CME is given by where V l and V t are the volumes of each leg and the toroidal portion, respectively.Each leg is a conical region with height h and half-angle δ.Its base radius is h tan d.Hence, the leg volume is given by The leg height h is related to the CME front height h f we derived in Section 3 by .This expression can be easily derived using Equations (4), ( 6) and (30) from Thernisien (2011).
The toroidal region has a circular cross section with variable radius R, which is maximum at the apex and reduces as we get closer to the legs.From Thernisien (2011)  where β is the angle with origin at the center of circle that the toroid formed between a given portion in the toroid and the direction perpendicular to the front axis and parallel to the plane formed by the two legs.This angle spans from −α to π + α.We determine its volume V t integrating numerically the area of the cross section πR 2 along the toroidal region: The volume is totally determined using the CME height, its half-angle α, and the aspect ration κ.A complete geometrical description of the model is available in Thernisien (2011).

Figure 2 .
Figure 2. Running-difference observations from SECCHI COR2-A.We represent the PSP position by the asterisk.An animation is available in the online version of the journal.The animation corresponds to the period between 06:01 and 15:56 UT on 2022 June 2.The real-time play time of the animation is 8 s. (An animation of this figure is available.)

Figure 3 .
Figure 3. Observations from WISPR-inner on board PSP.The CME under study is marked by the white arrows as another CME dominates the FOV.An animation is available in the online version of the journal.The animation corresponds to the period between 03:00 and 23:45 UT on 2022 June 2.The real-time play time of the animation is 14 s.(An animation of this figure is available.)

Figure 4 .
Figure 4.The CME observations on 2022 June 2 09:00 UT (middle row), its GCS modeling projected over each observation (top row), and its synthetic images (bottom row).

Figure 5 .
Figure 5.The CME observations on 2022 June 2 12:00 UT (middle row), its GCS modeling projected over each observation (top row), and its synthetic images (bottom row).

Figure 6 .
Figure 6.The CME 3D reconstruction (represented in gray) as a function of time compared to PSP position (blue circle).The left panel corresponds to the solar equatorial plane, and the right panel shows a meridian plane.The red and green lines indicate the directions toward STEREO-A and Earth, respectively.The region inside the green line indicates the CME portion located in PSP orbit plane.The blue line indicates future PSP positions.An animation is available in the online version of the journal.The animation corresponds to the period between 08:00 and 15:00 UT on 2022 June 2.The real-time play time of the animation is 1.5 s. (An animation of this figure is available.)

Figure 7 .
Figure 7. Magnetic field and plasma parameters observed by PSP from 2022 June 2 08:00 to 17:00 UT.From top to bottom: (a) magnetic field intensity; (b-d) its radial, tangential, and normal components (one panel for each); (e) solar wind proton bulk speed; (f) electron density; (g) plasma beta parameter; and (h) solar distance.

Figure 8 .
Figure8.The comparison of the CC (blue line) and EC (magenta line) models with the magnetic field measurements (black circles).The panels, from top to bottom, represent the magnetic field intensity and its radial, tangential, and normal components.

Figure 9 .
Figure 9.Comparison of the ICME axis orientation from the CC (black line) and EC (magenta line) models with the CME reconstruction using the GCS method (gray and green).The red dashed line in the left panel represents the GCS model axis.

Table 1
Flux Rope Magnetic Configurations Derived Using in Situ Measurements and

Table 2
Comparison of the CME Cross Section Radius Derived by GCS, CC, and EC Note.The numbers in parentheses indicate the minimum and maximum radii we get considering the errors in GCS parameters.

Table 3
Comparison of 1 au Properties Expected for the Current Event with Averages b Bothmer & Schwenn (1998).
C.R.B. acknowledges the support from the NASA STEREO/SECCHI (NNG17PP27I) program and NASA HGI (80NSSC23K0412) grant.A.V. and G.S. are supported by WISPR Phase-E funds and NASA grant 80NSSC22K0970.V. K.J acknowledges support from the Parker Solar Probe mission as part of NASA's Living with a Star (LWS) program under contract NNN06AA01C.Parker Solar Probe was designed, was built, and is now operated by the Johns Hopkins Applied Physics Laboratory as part of NASAʼs Living with a Star (LWS) program (contract NNN06AA01C).Support from the LWS management and technical team has played a critical role in the success of the Parker Solar Probe mission.The Wide-Field Imager for Parker Solar Probe (WISPR) instrument was designed, was built, and is now operated by the US Naval Research Laboratory in collaboration with Johns Hopkins University/Applied Physics Laboratory, California Institute of Technology/Jet Propulsion Laboratory, University of Gottingen, Germany, Centre Spatiale de Liege, Belgium and University of Toulouse/Research Institute in Astrophysics and Planetology.The Sun Earth Connection Coronal and Heliospheric Investigation (SECCHI) was produced by an international consortium of the Naval Research Laboratory (USA), Lockheed Martin Solar and Astrophysics Lab (USA), NASA Goddard Space Flight Center (USA), Rutherford Appleton Laboratory (UK), University of Birmingham (UK), Max Planck Institute for Solar System Research (Germany), Centre Spatiale de Liége (Belgium), Institut d'Optique Theorique et Appliquée (France), and Institut d'Astrophysique Spatiale (