Source and Propagation of a Streamer Blowout Coronal Mass Ejection Observed by the Parker Solar Probe

STA observed a CME in COR2 on 2018 November 10. In COR1, the streamer belt brightened starting on 2018 November 9 at 10:20 UT and extended through the FOV for several frames. In Figure 2, panel (a) shows the bright front of this structure in COR1 as it nears the outer edge of the FOV at 12:05 UT. The online movie version for COR1 shows the expansion of the streamer and the evolution of the structure moving through it. Since the FOV of COR2 overlaps with that of COR1, we are able to track the structure into the FOV of COR2. Panel (b) of Figure 2 shows it entering COR2 at 12:08 UT. In Figure 3, we show a few EUVI frames from the 24 hr prior to the COR2 observation in panel (b). In the general region of interest, there is lots of activity in the form of low-density structures lifting from the Sun (indicated by the arrows). However, in EUVI, there is no obvious signature on the disk near the potential source region that we could identify as a potential source of flare activity associated with this event. The development of these structures across the EUVI FOV is best appreciated in the animation available online. Due to the nonradial nature of structures in this region of the corona, many of the structures observed at ±20°–30° from the equatorial region appear to follow a path that leads to the structures observed in COR1 and COR2, which is consistent with past observations during solar minimum configurations (Kilpua et al. 2009). Based on this and the fact that the streamer brightens in COR1 before the structure enters the FOV of COR2, we theorize that the CME formed at or below the inner FOV of COR1 (i.e., below 1.4 R_s). Proving or disproving this requires observations of the disk behind the Sun during this event, which do not exist.

A height–time map was generated for EUVI, COR1, and COR2 observations and is shown in Figure 4. The map was generated for a radial slice at an angle of 85° (counterclockwise from north). In it, a bright structure that lifted off in EUVI at the end of 2018 November 9 and reaches the edge of the FOV on 2018 November 10 is shown. This structure then appears to cross into the COR1 FOV and into the COR2 FOV at the end of the day. The structure appears to be fainter in COR1, at least near the inner edge, which may be a result of the difference in structure appearance between EUV and coronagraph observations. By the same token, structures have a different appearance in observations from inner coronagraphs when compared to those from outer coronagraphs, in this case, COR2. That being said, the method we used to process the data sets (N. Alzate et al. 2020, in preparation) suppressed high-frequency noise and low-frequency variations to allow us to track different structures as they cross all three FOVs.

Zoom In Zoom Out Reset image size

Tracking the front of the CME in the STEREO images enables the determination of the speed and duration of the CME, which are important characteristics to study the CME dynamics. We used the method described by Lara et al. (2004) in which the coronal brightness is computed at several heliospheric distances and in 11 position angle (PA) directions separated by 05 centered around the central position angle (CPA) for each available image. With this, it is possible to obtain the brightness as a function of time and distance.

With a Gaussian fit to the brightness as function of time, we compute a duration of the CME of Δt = 2.5 hr, and, by following the maxima, we plot the height in terms of time at the brightness maximum and the outer edge of the CME (nose). These acceleration profiles are shown in Figure 5. The nose of the CME was found to be moving at ${V}_{{\mathrm{CME}}_{0}}=249\pm 5$ km s⁻¹ and the brightest part of the CME at ${V}_{{\mathrm{STR}}_{0}}=243\pm 5$ km s⁻¹ at R_inj = 6 R_⊙. These velocity measurements are consistent with the observed mean speeds of SBO-CMEs during solar minimum (Vourlidas & Webb 2018). The gradual swelling of streamers caused by steadily injected magnetic energy will lead to low ejection speeds of the CME itself.

Zoom In Zoom Out Reset image size

In situ plasma data were collected by the Solar Wind Electrons Alphas and Protons (SWEAP) instrument suite (Kasper et al. 2016) on board PSP, which is currently orbiting the Sun with an initial perihelion of 35 R_Sun and an eventual perihelion of 9.86 R_Sun. The SWEAP instrument suite is made up of the Solar Probe Cup (SPC), SPAN-A (ions and electron measurements), and SPAN-B (electron measurements). The SPC (Case et al. 2020) is a Faraday cup that measures the temperature, density, and proton velocity at 0.825 s during encounter and every 28 s outside of encounter. The data plotted here are at a 28 s cadence. The SPAN electrostatic analyzers (Whittlesey et al. 2020) provided electron measurements used to create pitch angle distributions (PADs). The magnetic field data were collected by the FIELDS instrument suite (Bale et al. 2016). It provides the magnetic field vector components and magnetic magnitude. We plot 1 minute averaged magnitude data generated by the FIELDS team. This is an official data product and can be found on CDAWeb. It uses triangular averaging. The native cadence of the magnetic field data varies from  2.3 to 293 samples s^–1 and can also be found on CDAWeb. For derived parameters such as plasma beta, we interpolated the FIELDS data to the SPC data. For large-scale structures such as this CME, the 1 minute data are more appropriate, as the full-resolution data would have multiple peaks and valleys associated with wave activity at short timescales that are not representative of longer intervals and general plasma characteristics discussed in this manuscript.

Figure 6 plots the in situ data from 2018 November 11–12. The CME is measured initially by PSP FIELDS and SWEAP at 23:50:45 UT on 2018 November 11 at a distance of 54.7 R_Sun or 0.25 au. This onset is marked in the plot by the leftmost black vertical dashed lines. The Radial, Tangential, Normal (RTN) coordinate system components of the proton speed data are plotted in panels (e) and (f). The average speed of the CME is v = 390 km s⁻¹. The proton density is plotted in panel (g). The average density of the CME was ρ= 177 cm⁻³. The proton temperature is plotted in panel (h). The average temperature of the protons in the CME was found to be 7.3 × 10⁴ K. The plasma beta calculated using the data is plotted in panel (i). The β during the CME is always much less than 1. There is no clear shock structure associated with the CME. The start time of the CME was found by examining the increase in magnetic field that coincided with a rotation in field direction. The plasma data were also used as a marker of the start time with a lower temperature and the presence of bidirectional electrons.

The electron PADs are in panels (j) and (k) of Figure 6. The upper PAD panel has the raw flux represented, and the lower panel normalizes the flux at a particular angle to the mean flux over all pitch angles. Bidirectional electrons are known to be signatures of a CME and closed magnetic field configuration that either ends in a loop or is tied back to the Sun (Gosling et al. 1987). Bidirectional electrons are present when there are electrons at 0° as well as 180°, indicating that electrons were traveling both away from and toward the Sun. During the passage of the flux rope, there are bidirectional electrons present (for a more in-depth discussion of the flux rope structure, see Nieves-Chinchilla 2020, in this issue).

The combination of the in situ and remote sensing data gives information on the timing and characteristics of the CME. Utilizing this information, we can project the CME back to its source and learn about its coronal origin. In this work, the WSA model (Arge & Pizzo 2000; Arge et al. 2003, 2004) is used to derive the coronal field using a coupled set of potential field–type models. The first is a traditional magnetostatic potential field source surface (PFSS) model, which determines the coronal field out to the source surface height. The PFSS solution serves as input into the Schatten current sheet model to provide a more realistic magnetic field topology of the upper corona (e.g., from 2.5 to 21.5 R_Sun). An empirical velocity relationship (Arge et al. 2003, 2004) is then used to derive the solar wind speed at the outer coronal boundary. The model then propagates solar wind parcels outward from the end points of each magnetic field line, determining their time of arrival at PSP and thus providing the magnetic connectivity between 1 R_Sun and PSP magnetic field observations. The WSA model uses a simple 1D modified kinematic model, which accounts for stream interactions by preventing fast streams from bypassing slow ones (Arge et al. 2004).

Synchronic photospheric field maps used as input to the WSA model were generated using the Air Force Data Assimilative Photospheric Flux Transport (ADAPT) model (Arge et al. 2010, 2011, 2013) driven by Global Oscillations Network Group (GONG) magnetograms. ADAPT utilizes flux transport modeling (Worden & Harvey 2000) to account for solar time-dependent phenomena (e.g., differential rotation, meridional and supergranulation flows) when observational data are not available, making it particularly useful for studying this far-side event. Since ADAPT is an ensemble model, it provides 12 possible states (i.e., realizations) of the solar surface magnetic field, ideally representing the best estimate of the range of possible global photospheric flux distribution solutions at any given moment in time. When coupled with ADAPT, WSA derives an ensemble of 12 realizations representing the global state of the coronal field and spacecraft connectivity to 1 R_Sun for a given moment in time, providing an estimate of the uncertainty in the ADAPT–WSA solution. The best realization is then determined by comparing the model-derived and observed initial mass function (see Figure 3 of Szabo et al. 2020) and solar wind speed. While a magnetostatic potential field–based solution cannot capture time-dependent solar phenomena such as CMEs, ADAPT–WSA is used in this work to derive the coronal field pre-eruption and, in combination with in situ and remote sensing observations, deduce the solar source region of the SBO-CME. For a complete description of ADAPT–WSA and the parameters specific to the following model output, see Szabo et al. (2020).

Zoom In Zoom Out Reset image size

Figure 7 shows the global corona and PSP connectivity to 1 R_Sun during corotation with the Sun, timestamped to best represent the coronal field as the underlying closed field of the SBO-CME source region begins to expand upward through the FOV of STA/COR1. Prior to this eruption, PSP corotates with a mid-latitude coronal hole of negative polarity shown in Figure 7 at approximately 335–340 Carrington longitude (CL) during Carrington rotation (CR) 2210. This coronal hole is comparable in size to STA/EUVI observations on 2018 November 14–18 when it rotates onto disk center after the eruption. According to ADAPT–WSA simulations, this coronal hole begins to close down and shrink after 2018 November 8 (see the coronal hole at perihelion in Figure 2 of Szabo et al. 2020) to its state in Figure 7. The PSP comes out of solar corotation, crosses the HCS, and magnetically connects to another mid-latitude coronal hole of positive polarity, shown in Figure 7 between 270 and 300 CL (CR 2210). The model-derived HCS crossing is ∼1 day earlier when compared to observations (Figure 3 of Szabo et al. 2020), likely due to the CME observed at PSP that is not captured by the model solution as discussed previously. The location where PSP traverses this helmet streamer (i.e., crosses the HCS) is highlighted with a blue oval in Figure 7.

Zoom In Zoom Out Reset image size

Based on the location of PSP near the time the event is visible in COR1 (i.e., approximately −1° HEEQ latitude, 335° CL in Figure 7) and the observations of the event in COR1 (Figure 2(a)), there are only two plausible source regions of this eruption. One is the mid-latitude helmet streamer (HS) that PSP is connected to prior to the eruption (Figure 7, region in blue oval), and the other is the HS formed between the northern polar coronal hole and the negative mid-latitude coronal hole that PSP was connected to during corotation (Figure 7, region in green oval). To deduce the SBO source region, we first determined when this event becomes visible in STA/COR1. The first evidence of brightening and field expansion in COR1 is at approximately 02:00:00 UTC on 2018 November 10. We then aligned the WSA 3D coronal field solution with the STA COR1/COR2 FOV for this time period, shown in Figure 8, to see if either helmet streamer aligned with STA observations of this event on the east limb.

Zoom In Zoom Out Reset image size

Figure 8 shows a large helmet streamer on the east limb highlighted in green that corresponds to the region in the green oval in Figure 7. Assuming nearly radial propagation, the structure moving through the COR1 and COR2 FOV (Figure 2(a) and (b)) traces back to this model-derived HS cavity. This structure spans approximately 800–900 Mm at its 1 R_Sun base between the two coronal hole boundaries. The photospheric field beneath the helmet streamer cavity and in the surrounding region is rooted in the quiet Sun with an average photospheric field strength of $| {B}_{\mathrm{LOS}}| \lt 10$ G according to ADAPT–GONG photospheric field observations. The identified source region is consistent with the description of SBO-CMEs in Vourlidas & Webb (2018), where these eruptions occur without the presence of an active region and result from the slow buildup of shear over time along the polarity inversion line, likely due to differential rotation.

Having identified the coronal origin of the SBO-CME, we can now model the propagation from that source out to PSP and beyond. We examine the overall propagation through a simple solar minimum environment, as well as looking at the propagation of a flux rope model that allows for insight into the possible formation of energetic particles. We studied its evolution using a numerical simulation that followed the flux rope structure from its launch through its arrival at PSP (R_PSP = 0.2543 au) in order to compare the simulated parameter time profiles with PSP in situ data.

The numerical simulation was done using the 2D hydrodynamic, adiabatic, and adaptive grid code YGUAZÚ-A, developed by Raga et al. (2000) and used previously to simulate the propagation of CMEs in the solar wind by Niembro et al. (2019). The code integrates the 2D Cartesian coordinate Euler equations for a full ionized plasma with a mean molecular weight μ = 0.62, in which the gas is assumed to be composed in mass fractions of the Sun by 0.7 H, 0.28 He, and 0.02 the rest of the elements.

In this model, the computational domain is filled by an isotropic flow with a mass-loss rate ${\dot{M}}_{\mathrm{SW}}$, an initial speed ${V}_{{\mathrm{SW}}_{0}}$, an initial temperature of 10⁵ K, and with ${C}_{V}=1/(\gamma \,-1)=1.5$ (with $\gamma =5/3$) as the specific heat at constant volume. Then, the SBO-CME is imposed as a sudden change of the flow (to ${V}_{{\mathrm{CME}}_{0}}$ and ${\dot{M}}_{\mathrm{CME}}$) at the injection radius ${R}_{\mathrm{inj}}=6$ R_⊙ during Δt and within a solid angle that corresponds to ϕ. Afterward, the solar wind resumes to V₀ and ${\dot{M}}_{\mathrm{SW}}$.

This model was used because we have a solar minimum simplified configuration of the inner heliosphere. The CME moves radially, which means that the effects of curvature and the magnetic field are neglected. To study the propagation, hydrodynamic codes give reliable insights into the dynamics of CMEs in the solar wind. ENLIL models were compared to this model in Niembro et al. (2019), and the results with ENLIL are similar due to the radial propagation of the CMEs. The results are consistent between both models, as the magnetic field contribution can be neglected and its contribution is within the uncertainties.

The initial SBO-CME speed ${V}_{{\mathrm{CME}}_{0}}=249$ km s⁻¹, mass M_CME = 2.1 × 10¹⁵ g, duration Δt = 2.5 hr, and angular width ϕ ∼ 20° were computed following Lara et al. (2004) and from the height–time plot shown in Figure 5 assuming the injection time on 2018 November 10 23:54 UT at R_inj = 6 R_⊙, in which COR1 STA was tracking the SBO-CME. For the CME mass, we estimate the mass at each pixel as m = (B_obs/B_e) × 1.97 × 10⁻²⁴ g, with B_obs the observed brightness and B_e the brightness of a single electron (Colaninno & Vourlidas 2009; Vourlidas et al. 2010). Here B_e was computed using the model described by Billings (1966). We calculate the total mass by summing the values in the region, limited as per the model of Thernisien et al. (2009), in the image that contains the CME. With this value, we calculate the CME mass-loss rate to be ${\dot{M}}_{\mathrm{CME}}=7.6$ × 10⁻¹⁴ M_⊙ yr⁻¹. The mass-loss rate is a minimum value due to the assumption that the mass is in the POS and not in a 3D structure. Due to the narrow nature of the CME, the error in mass estimate could be as high as 50%, as reported by Colaninno & Vourlidas (2009).

For this particular event, we considered that the ambient solar wind conditions are slightly different before and after the SBO-CME. The initial conditions of the ambient solar wind prior the SBO-CME (${V}_{{\mathrm{SW}}_{01}}$ and ${\dot{M}}_{{\mathrm{SW}}_{1}}$) expected at R_inj were derived using the estimated PSP in situ measurements of V_PSP = 380 km s⁻¹ and n_PSP = 200 cm⁻³ computed during the interval of time from 2018 November 11 23:45 UT to 2019 November 11 23:50 UT. By solving the equations described in Cantó et al. (2005), we estimate that at R_inj, the speed of injection is ${V}_{{\mathrm{SW}}_{01}}=371.54$ 372 km s⁻¹. Then, we compute a mass-loss rate of ${\dot{M}}_{\mathrm{SW}}=\mu \times$ 4π ${R}_{\mathrm{PSP}}^{2}$ N_PSP V_PSP = 2.4 × 10⁻¹⁴ M_⊙ yr¹. For the ambient solar wind behind the CME, we used the average PSP measurements of V_PSP = 400 km s⁻¹ and n_PSP = 200 cm⁻³ computed during the interval of time from 2018 November 12 06:37 UT to 2018 November 12 07:35 UT and found that this flow was injected at R_inj with ${\dot{M}}_{\mathrm{SW}}$ and ${V}_{{\mathrm{SW}}_{02}}=391.97$ –392 km s⁻¹.

In Figure 9, we show the results of the model (green solid line with colored star symbols) overplotted on the SWEAP/SPC in situ thermal plasma data (black solid lines). YGUAZÚ-A allowed us to tag the different plasma components, which we have marked with colored symbols: the solar wind in front of the SBO-CME in blue, the SBO-CME in red, and the solar wind behind the SBO-CME in green. The colored symbols allow us to hydrodynamically identify these different plasma components and determine the arrival time of the flux rope.

Zoom In Zoom Out Reset image size

With black dotted lines, we show the initial and final times of the SBO-CME observed at PSP (from 2018 November 11 23:50 UT to 2018 November 12 06:36:19 UT) and the arrival of the flux rope (2018 November 12 04:15 UT). With the numerical model, we predict the arrival of the end of the flux rope in the slow SBO-CME. The model requires a density transition from the CME (red symbols in Figure 9) and the ambient solar wind (green symbols) to indicate that the perturbation (SBO-CME) is present within the ambient solar wind. In this case, due to the fact that it is a slow CME, the flux rope interacts with the solar wind behind it. Using the model results, we find that the flux rope arrives 30 hr after launch (shown in the figure with a dashed red line) with a speed of 383 km s⁻¹ with a difference of 0.33 hr and 2 km s⁻¹ from the SWEAP/SPC measurements.

It is important to notice that due to the fact that we are not modeling the CME magnetic field structure, our density prediction during the CME is not comparable to the observations. However, before and after the CME, our results accurately predict the solar wind density value.

With the simulation, we also predict that at 14 R_⊙, the CME reaches a speed of 374 km s⁻¹, which can be compared with the observations plotted in Figure 5. According to the observations, at 14 R_⊙, the CME reaches a speed of 347.61 ± 37.14 km s⁻¹ (∼375.58 ± 23 km s⁻¹ with the second-degree polynomial), which is, within the uncertainties, comparable with the one we obtain with the simulation at this particular distance.

The CME discussed here is a very weak event, and no shock-sheath system can be identified in the images during the eruption of the CME or in the in situ plasma data at PSP. Energetic particles were seen at PSP around the arrival time of the SBO-CME (details in Giacalone et al. 2020, this issue). Here we have modeled this CME from the source to PSP using a forward modeling technique that is detailed in Rouillard et al. (2020) to explore the possibility of the solar energetic particle (SEP) being associated with this SBO-CME. The flux rope structure assumed in the present work is a bent toroid with an exponential increase of its cross-sectional area from footpoint to apex. Narrow CMEs have been shown in past STEREO studies of previous events to correspond to a flux rope in the shape of a bent toroid such that its major axis is located in a plane containing the line of sight of the observing spacecraft (e.g., Thernisien et al. 2009). Additionally, the flux rope model has an internal magnetic field structure described analytically inside the envelope of the CME.

The forward modeling technique proceeds via manual iterations until a good fit of the flux rope is obtained to the available observations. The fitting procedure is essentially very similar to the one proposed by Thernisien et al. (2009) and Wood et al. (2010). Unfortunately for this event, we do not have imaging from multiple vantage points because the event was very weak and erupted on the far side of the Sun; therefore, it could not be seen by near Earth–orbiting spacecraft. Hence, compared to the COR-2A images, the model assumes that the CME erupts close to the POS. The imaging observations were not sufficient to confirm the longitudinal extent of the legs of the flux rope, and this is an uncertainty of the present fitting. Figure 10 shows a set of coronagraphic observations from COR-2A in which we overplot the 3D modeled flux rope.

Zoom In Zoom Out Reset image size

The flux rope model is performed to the outer extent of COR-2A, where we find that the CME acceleration phase has nearly ended and the CME speed is almost constant (380 km s⁻¹). The results of the CME kinematics are presented in Figure 2 of the extended data in McComas et al. (2019). By propagating the flux rope at a constant speed of 380 km s⁻¹ from the time the CME exits the COR-2A FOV to PSP, we "predicted" an impact of the CME at PSP on November 12. The predicted arrival time and the magnetoplasma properties of the CME from this flux rope model are in very good agreement with those measured in situ by the FIELDS and SWEAP instruments, as seen in Figure 6 and the model speed of v ≈ 383 km s⁻¹. The CME modeling presented here provides a good description of the CME evolution from the upper corona to PSP.

To investigate the possible formation of a shock associated with this SBO-CME, the above flux rope model is utilized to model the behavior from the COR2 FOV to PSP. To derive the 3D properties of the modeled coronal pressure/shock wave that could have formed around the expanding flux rope, we use methods presented in previous papers (e.g., Rouillard et al. 2016; Plotnikov et al. 2017; Kouloumvakos et al. 2019) but adapted in this study to use a flux rope rather than an ellipsoidal structure typically used for more energetic events. We start with the sequence of the regularly time-spaced flux ropes and consider a set of grid points distributed over the surface. From the sequence of the time-spaced flux rope grid points, we calculate the 3D distribution of velocity vectors at each point on the surface of the flux rope. The 3D speed is plotted in Figure 11 (panel (a)).

Zoom In Zoom Out Reset image size

As already stated, for this weak CME event, we do not observe the clear sheath and shock structure typically observed in more energetic CMEs (see, for example, Kwon & Vourlidas 2017; Kouloumvakos et al. 2019). So we assume that if a weak shock formed, it should be located near the surface of the flux rope due to the CME interaction with the background solar wind. That could result in a very weak and narrow sheath region. Then we use a model for the background solar wind to compute the wave properties and determine if there are regions where a shock forms. We exploited the plasma and magnetic field properties from the magnetohydrodynamic coronal and solar wind model produced by Predictive Sciences Inc. (courtesy of Jon Linker and Pete Riley). The Magnetohydrodynamic Around a Sphere Thermodynamic model is an MHD model developed by Predictive Sciences Inc. that makes use of the photospheric magnetograms from SDO/HMI as the inner boundary condition of the magnetic field and includes detailed thermodynamics with realistic energy equations accounting for thermal conduction parallel to the magnetic field, radiative losses, and parameterized coronal heating (Lionello et al. 2009; Riley et al. 2011). Details of the model and synoptic figures of the cubes used in this study can be found on the PSI site.¹⁷ We have also uniformly scaled the values of density, magnetic field, and temperature to match as well as possible the solar wind properties measured by the PSP SWEAP and FIELDS instruments just before the CME impacted the spacecraft.

Combining the flux rope fitting technique with the background solar wind model with the methods presented by Kouloumvakos et al. (2019), we can compute at each point on the surface of the flux rope the fast magnetosonic Mach number (MFM; panel (a)), the density compression ratio (panel (c)) by solving the Rankine–Hugoniot relations, and the magnetic field obliquity with respect to the flux rope shock normal (θ_Bn; panel (d)). We also traced in blue in Figure 11 the trajectory of the nominal Parker spiral. Figure 11 shows that throughout its transit to PSP, the CME did not produce a shock in the region magnetically connected with the probe; the region where the Parker spiral (blue line) intersects with the flux rope remains back in panel (c), and the Mach number is always below 1. However, the calculations show that weak shock solutions with Mach numbers greater than 1 but below critical values do appear over a region at the western flank of the flux rope that are not connected with PSP. The subcritical shock in these regions is driven by the lateral expansion of the CME moving at less than 50 km s⁻¹ in the orthoradial direction in a background solar wind flow that is moving predominantly in the radial direction. The ability of slow CMEs to compress the background solar wind and produce weak shocks in response to the expansion of the flux rope has been discussed and studied in past papers analyzing STEREO data (Rouillard et al. 2011; Lugaz et al. 2017).