Multiwavelength Observations of the Formation and Eruption of a Complex Filament

Figure 1 provides an overview of the event. Panel (a) presents the GOES 1–8 Å SXR flux, which shows a flux increase beginning at about 16:57 UT, peaking at 17:08 UT at a level of C2.8. This pre-eruption flux increase is associated with the formation of an active region filament (ARF) in AR 11429. The SXR then declines until 17:15 UT at which point it increases rapidly to a level of M5.4 at 17:27 UT. It is during this time period that the ARF and MFR erupt, driving a fast CME. The SXR emission reaches its maximum of M8.4 at 17:45 UT, followed by an extended decay phase.

Zoom In Zoom Out Reset image size

SDO/AIA EUV integrated light curves are shown in Figures 1(b), (c), where the emission has been summed over the AR (Figure 2) and normalized to its pre-flare levels. Figure 1(b) shows light curves of cool plasma in the 304 Å bandpass and relatively cool coronal plasma in the 171 and 211 Å bandpasses. As we will show in detail in Sections 3.2–3.4, the four peaks (as indicated by red arrows) in these light curves are associated with the following processes: the interaction between two discrete filaments F1 and F2; the confined eruption of a filament that subsequently forms as a result of the interaction of F1 and F2; the eruption of the ARF; and the evolution of post-flare loops. In contrast, Figure 1(c) shows the light curves of emission from hotter plasma in the 94, 131, and 335 Å bandpasses. Of these, only the 131 Å emission shows a clear precursor signature. None of the hot bandpasses shows an impulsive signature associated with the filament eruption.

Zoom In Zoom Out Reset image size

Figure 1(d) shows HXR light curves as measured by RHESSI in the 3–12 (right y-axis), 12–25 (right y-axis), and 25–60 keV (left y-axis) photon energy ranges. A precursor phase can be identified in the X-ray energy band of 3–12 and 12–25 keV. The later phase of filament and MFR eruption and the early phase of the post-eruption were not observed as RHESSI went into nighttime. RHESSI resumed observing at about the time of the SXR maximum at 17:45 UT. Figure 1(e) shows the RSTN radio flux variation at 4.995 GHz, which may be regarded as a qualitative proxy for HXR emission. While minor activity at 5 GHz is apparent during the precursor phase, clear impulsive peaks occur during the filament eruption and the subsequent rise to the maximum.

Figure 1(f) shows the evolution of the EUVI emission as viewed from STA. Red diamonds show the 171 Å count variation and asterisks indicate the position of the filament heights by tracking the relevant features in STA. As seen by STA, the two filaments (F1 and F2) are hard to distinguish; the gray cross indicates the height of F1 or F2. Red and blue asterisks later indicate the height of the upper filament and the ARF, respectively. The velocity deduced by the height–time plot is shown in Figure 1(f) as a plot inset from 16:45–17:30 UT. F1 and F2 show a slight increase in the very beginning of the filament formation phase, followed by a relatively obvious increase in height as a result of the confined eruption of the upper filament (see Section 3.1). A rapid increase in height during the main eruption is consistent with the flux increases seen by SDO and GOES. As was the case for the AIA emission in cool plasma (panel b), the EUVI-A flux also increases during the ARF formation and eruption phase. Subsequently, however, it decreases during the AIA main flare phase, which may result from the coronal dimming during the filament eruption from the EUVI-A point of view. The erupting filament moves out of the EUVI-A field of view (FOV) at around 17:30 UT.

Figure 2 shows AR 11429 at about 16:43 UT, prior to precursor activity and the eruption. Although still flare productive, AR 11429 was already in its decay phase by 2012 March 10. The AR is composed of two well-separated sunspot groups: the western part and the eastern part. The line-of-sight magnetic field is overlaid on each panel, where yellow (red) contours represent the negative (positive) magnetic polarities (contour levels: ± 20 and ± 800 Gauss). Figure 2(a) shows that the western sunspot group is composed of compact sunspots with opposite polarities, while the eastern group is comprised of several small positive sunspots enclosed by a penumbra followed by a compact negative sunspot. The PIL, indicated by the almost coincident yellow and red lines, extends through the entire AR from the northeast to the southwest with a length of 1.4 × 10⁴ Mm.

The long PIL is associated with different features in EUV images. In the 171 Å image, the PIL in the eastern sunspot group is outlined by the aligned fibrils. Even through there is no obvious filament seen in this bandpass, the fibrils aligned along the PIL can also be identified in the 304 Å image (left bottom in Figure 2(b)), showing the likely location of the FC (Gaizauskas 1998). The area outlined by the green box shown in the center of Figures 2(a), (b) shows the presence of a dark arch-shaped feature. By contrast, there are no aligned fibrils or filament along the PIL in the western sunspot group, where the ARF later forms and erupts. Figure 2(c) shows AR 11429 in the 131 Å bandpass, indicating relatively hot plasma, where bright loops are seen aligned with the PIL in the central and western parts of the AR.

Figure 2(d) shows the enlarged area outlined by a green box in Figures 2(a), (b) at 193 Å recorded at 16:47:43 UT. It shows that the arch-shaped dark feature is composed of two small crossing filaments. One is narrow and dark, indicated by white arrows and is labeled F1. The other is more diffuse, indicated by red arrows and is labeled F2. A black arrow indicates where F1 and F2 cross. The blue arrow indicates an EUV brightening. The western branch of F1 is composed of several dark threads along the PIL. Red curves outline the outer boundary of F1 and F2.

STA observed this event on the limb with a temporal resolution of about 75 s in 171 Å. Seen as a limb event, and in the absence of sky background, STA allowed the eruption and its morphology to be seen more clearly than in the higher-time-resolution SDO/AIA data seen against the background AR. The EUVI-A solar limb is outlined in Figures 2(a), (b) as black and white curves, respectively. It shows that the solar limb as seen by STA cuts across F1 and F2 near their crossing point. We now discuss the detailed evolution and eruption of the ARF in three separate phases: (1) the ARF formation phase, from 16:46–17:15 UT; (2) the ARF eruption phase, from 17:15–17:27 UT; and (3) the post-eruption phase for times later than 17:27 UT. Our focus here is primarily on the first two phases. The high spatial and temporal resolution data from SDO and STEREO provide the opportunity to give a rather complete picture of the formation and eruption of an ARF.

Figures 3 and 4 show the details of the formation of the ARF using joint AIA and EUVI-A images with the same pixel size. This process also can be seen more clearly in animation of Figures 3 and 4. In order to highlight the morphology at different times and different wavelengths, both homomorphic filtering (Oppenheim et al. 1968) and histogram equalization (Pizer et al. 1987) were used in the image processing. Each row shows images at approximately the same time. The first two columns show STA and SDO 171 Å images. The next two columns show SDO 211 and 131 Å images, respectively. In the first row of Figure 3, F1 and F2 are indicated by green and red arrows in panels (k) and (f). F1 is a compact dark thread, while F2 is a more diffuse feature, but with an obvious twisted fine structure to the west (as indicated by a white arrow in panel (f)). The location where F1 and F2 cross is indicated by a black arrow in Figure 3(k). Taking the F1 and F2 crossing point as a reference point, the left and right parts of F1 are labeled LF1 and RF1, and the left and right parts of F2 are labeled as LF2 and RF2.

Early transient activity is first evident in the EUV as brightenings at ∼16:46:30 UT and somewhat later in the UV at 1600 Å. The first row in Figure 3 shows the morphology of the brightenings at ∼16:48:30 UT in which their emission in 1600 Å is overlaid on a 171 Å image as red contours (Figure 3(f)). The contours demarcate areas where the intensity is >0.2 times the maximum 1600 Å intensity. As indicated by the blue arrow in the same panel, the initial EUV brightening is near the location where F1 and F2 cross. The 1600 Å contours is co-spatial with the EUV brightenings. The 1600 Å contours with additional EUV brightenings to the east and west of the main brightening, as indicated by purple arrows, appear as an arch-shaped structure, which coincide with the location of F1 and F2. Together, the early UV/EUV brightenings resemble a possible arch-shaped structure that we label L1. As viewed by EUVI-A, the initial EUV activity is characterized by a slow rise of the filament that begins around 16:48 UT with an average upward speed of ∼17 km s⁻¹ (Figure 1(f)). A dark filament is indicated by red arrows as a small arch as shown in Figures 3(a), (b). Compared with AIA observations, F1 and F2 are hard to identify clearly in EUVI-A observations during the filament formation phase. We suggest that the dark feature identified in Figures 3(a)–(c) is the counterpart of both F1 and F2. Likewise, the counterpart to the EUV brightening in the AIA image sets cannot be identified in EUVI-A images, likely because of the occulting foreground limb in the latter. The second row of Figure 3 shows that the loop-like feature L1 continues to evolve with the UV/EUV brightenings extending east and west. During this period, additional EUV counterparts to L1 in 171, 211, and 131 Å appear (indicated by blue arrows) near the crossing point of F1 and F2, and in between RF1 and RF2. EUV brightenings are first apparent as viewed from STA at approximately 16:51 UT (Figure 3(c); red arrows), on both sides of the dark filament, while the AIA observations show continuing extension of the UV/EUV brightenings along L1.

Around 16:53:30 UT, as shown in the fourth row of Figure 3, the configuration of F1 and F2 changes, indicating that a possible reconnection process occurred. Specifically, the images show that two new filaments LF1-RF2 and LF2-RF1 appear with L1 located in between them. The LF1-RF2 filaments appear to form a long filament overlying a shorter filament that forms from LF2-RF1. We refer to, and label, these newly formed filaments the upper filament (UF) and lower filament (LF). The high temperature bandpass of 131 Å shows that the UF appears as a reversed-ω-shaped structure at 16:54:35 UT (Figure 3(s)), with the sunken structure spatially consistent with the EUV brightening and the crossing of F1 and F2. The newly formed UF can be seen clearly a few minutes later in Figure 3(t). However, its right part is obscured in the 171 and 211 Å images (Figures 3(j), (o)) by bright EUV emission. The LF, which is invisible in 131 Å can be seen in 211 and 171 Å images by careful examination as indicated by black arrows in Figures 3(n), (o).

As indicated by the gray arrows in Figures 3(j), (o), and (t), a portion of the UF to the west experiences a partial confined eruption (labeled CE), beginning at roughly 16:57 UT. The eastern part of the UF remains stable (white arrows, Figure 3(t)). The evolution of the CE is shown in Figure 4. Note the change in the FOV between the top two rows and the bottom three rows. At about 17:01 UT, the initial EUV brightening under the CE as recorded by AIA appears in the EUVI-A 171 Å image as a compact bright area, as indicated by the purple arrow in Figure 4(a). At this moment, above the brightening, the CE appears as arch-shaped filament threads, outlined by gray arrows in the image, while the LF is indicated by black arrows (Figure 4(b)). By 17:03:30 UT, the loop structure of the CE can be seen clearly in both the AIA and EUVI-A observations, as outlined by gray arrows in Figures 4(l), (q), showing an arch-shaped configuration with a twisted morphology. The CE reaches its maximum height at around 17:08 UT, coincident with the maximum of the precursor activity shown in Figure 1. From the EUVI-A observation (Figure 4(d)), after the rising CE plasma drains back to the solar surface and the bright arch-shaped structure has disappeared. As shown in the 131 Å images in Figure 4, RHESSI contours (yellow for 3–10 keV, red for 10–15 keV, and blue for 15–25 keV) represent the RHESSI soft and HXR source coincident with the CE. The contour levels are 60% and 85% of the corresponding peak flux. It shows the possible presence of nonthermal HXR emission coincident with the CE.

A notable feature of the CE, concurrent with the evolution described above, is a plasma flow from east to west along the UF. It appears at around 17:03 UT. As indicated by the yellow arrow in the third row in Figure 4 (note change of scale), the plasma flow moves across the visible end of the UF and along the PIL. Its counterpart also can be identified in the EUVI-A image, although it is less visible. As shown in Figures 4(i), (n), and (s) at ∼17:09:30 UT, the plasma flows to the southwest along the PIL to form the new ARF. From the labeling of the end of the UF and the newly formed west footpoint of the ARF, it can be seen that the average flow speed during this process is about 80 km s⁻¹. In the 131 Å images, a brightening loop structure that first appears around 16:57 UT, appeared again, as indicated by the purple arrows in Figures 4(s), (t). The plasma then flows under the loop structures (Figure 4(t)) and, by 17:15 UT, the new ARF has formed. Dhakal et al. (2018), who interpreted this event in terms of a "double-decker" eruption, identified the upper filament in their scenario (filament A in their Figure 8) with what we identify here as a newly formed ARF. In their scenario, the upper filament is a preexisting structure, whereas we have shown in Figure 4 and the corresponding movies that it forms between 16:45 and 17:15 UT. While Figure 4 and the corresponding movies with multiwavelength observations in the present work show clearly the process of the filament formation. Furthermore, if filament A is preexisting, it will undergo a slow rising to reach the position that is shown to be around 17:10 UT. But from both the EUVI-A and AIA observations, the filament appears little by little, and its evolution morphology can be seen clearly and continuously in all AIA EUV wavelengths. All the observations indicate that the filament cannot exist about 1 hr before the eruption as presented in Dhakal et al. (2018).

It is worth noting that the area along the PIL outlined by the green box in Figure 2 is the region where numerous small transients such as EUV brightenings are seen before the eruption. The EUV brightening in the 171 Å image in Figure 2(b) shows an example of such transient brightening, as indicated by the blue arrow. This small EUV brightening can be identified in all wavelengths and fades about 1 minute later. By checking the evolution of the line-of-sight magnetic field, we found that the dominant features of the magnetic field evolution are small flux cancellation events, indicative of magnetic reconnection in the low solar atmosphere along the PIL and identified with the global instability leading to an eruption (Zhang et al. 2001). Figures 5(a), (b) show two line-of-sight magnetic fields recorded at 12:00 and 16:45 UT as examples to show the variation in the magnetic field before the eruption of the filament. Red curves outline the profile of the outer boundary of F1 and F2, as shown in Figure 2(d). Blue line circles are opposite polarities that located along the PIL and can be separated from neighboring polarities. The positive and negative magnetic flux evolutions during about 6 hr before the eruption within the blue area is shown in Figure 5(c). For the positive (negative) magnetic flux, it declines by about one-third (a quarter). The magnetic field evolution can be seen more clearly in the animation of Figure 5. Recent studies (Joshi et al. 2017; Dhara et al. 2017) of flux-rope eruptions preceded by a decrease in the photospheric magnetic flux in the core flaring region suggest a tether-cutting reconnection may be responsible for triggering the eruption.

Figures 3 and 4 show that the new ARF formed as a result of the evolution and interaction of F1, F2, and the CE described in the previous subsection. The ARF began a slow rise, estimated to be 73 km s⁻¹ between 17:17 and 17:21 UT as viewed from STA. Figure 6 shows the dynamic eruption of the newly formed ARF. Both the running difference and the original images are used to highlight the morphology at different times. The twisted structure of the filament can be identified clearly during the rise and expansion of the filament in both the AIA and EUVI-A images. The first row in Figure 6, obtained at around 17:18 UT, shows an example of the twist. A narrow, bright feature (indicated by yellow arrows) is entwined around the dark filament body from one footpoint to the other. This twisted structure can be easily identified in the 335 Å image along the length of the filament. Using yellow arrows as guides we find that the twist of the whole filament is about four turns (8π) but it is not uniform along the filament. The twisting is greater along the newly formed part of the ARF to the west (three turns) than is the portion to the east (one turn).

Zoom In Zoom Out Reset image size

In the EUVI-A 171 Å image, the filament shows a spiral structure (yellow arrows) corresponding to the western portion of the ARF (yellow arrows in the AIA observations). A time difference EUVI-A image (Figure 6(q)), using images at 17:21:00 and 17:19:45 UT, shows the twisted morphology quite clearly: a bright feature entwined with a dark feature with four turns (indicated by yellow arrows). The EUVI-A images subsequently show the clear transition of the filament between its slow (first three rows) and its rapid rising phases (bottom four rows): the morphology of the erupting filament rapidly transforms from twist to writhe at the beginning of the fast rising phase. As shown in Figure 6(r), the EUVI-A image obtained at 17:22:15 UT shows a clear reversed-γ shape, which then expands outward. In the 335 and 131 Å images the morphology of the erupting filament and flux rope evolve continuously at the beginning of the fast rising phase. As shown in Figure 6(e), the reversed-γ shape appears in the filament body around 17:23 UT in the 335 Å image.

An MFR is usually defined as a coherent helical magnetic structure with all field lines wrapping around the central axis at least one turn (Cheng et al. 2017). From an observational point of view, it is an expanding structure with a twisted morphology. In the present work, a typical MFR structure with four turns is identified during the eruption of the ARF as shown in Figure 6. In AIA high temperature wavelength bandpasses, it appeared as thread-like structures (as indicated by red arrows in Figures 6(i)–(l)), which are entwined around the dark filament body and linked both footpoints of the filament. Here, the thread-like structures and the filament can be regarded as the tracer of the helical magnetic structure and the central axis, respectively. In order to help us to track back the MFR, we draw a white line in Figure 6(i). It crosses the thread-like structures and the newly formed dark filament body. The same line is drawn in the 131 Å image in Figure 6(h) and Figure 4(t). Using the white line as a reference location, the evolution of the MFR can be reconstructed. As shown in Figures 6(h) and (i), the MFR clearly shows a helical bright structure entwined with the dark filament. Going back to Figure 4, the bright feature is identified as co-parallel brightening loops that cross the PIL above the ARF. This temporal evolution of the MFR is consistent with the simple sketches of MFR formation proposed by Low (2001).

In order to illustrate the points made in the previous paragraph another way, we show position–time stack plots in Figure 7 from a series of AIA images along the white reference line. The brightening loops first appear in both the 131 and 94 Å images at around 16:57 UT. Their morphology at that time is shown in the lower right corner in Figure 7, with the reference position again indicated by a white line. The parallel loop structure is obvious in the 131 Å image, is weak in the 94 Å image, and has no counterpart at all in the 335 Å image. In Figure 7, several discrete times are indicated. The temporal evolution of the brightening loop shows features typical of an evolving helical structure. A feature changes its location slowly along the reference slit in the position–time plots, with the helical structure moving or expanding when the feature is on the side toward the observer. Then it disappears suddenly when it moves from the side toward observer to the side backwards of the observer. Around the same time, another similar feature may appears on the other side. As shown by the position–time stack, the signature of the helical structure first appeared at around 16:57 UT. At around 17:07 UT, two bright features can be identified in the 94 Å position–time plot, with one going to disappear and the other just appearing at almost the same place as the disappearing one's first appearance. The newly appearing bright feature also moves from left to right and then disappears. Figure 7 shows that this process repeated four times. In the 335 Å image the plasma flow, which forms the ARF after the interaction between F1 and F2, first passes through the slit at around 17:10 UT. The brightening feature then appears in the 335 Å image associated with the filament formation here. It shows clearly that the bright feature and the filament body evolve simultaneously. The red lines indicate the rise of the outer edge of the MFR as observed in the 131 and 94 Å images. During the first three times, the MFR rises slowly with a projected speed of about 2 km s⁻¹. Associated with the filament eruption, the MFR also experiences a slow and fast moving phase with the projected velocity of and 26 and 152 km s⁻¹. The observational evidence shows that the pre-activities configuration of F2 is a helical MFR. Part of the MFR is emerged into the low corona. It contains plasma materials, and appears as F2. Others (west part) are in their emerging phase and appear as arch-shaped brightening loops that cross the PIL. The plasma flows along the hollow helical MFR from one footpoint to another, after the interaction between F1 and F2, forming the ARF. This observation shows that the twisted MFR can preexist along the FC and once the coronal plasma gathers here, the filament forms.

Zoom In Zoom Out Reset image size

As shown in the EUVI-A images in the bottom four rows of Figure 6, the MFR and filament erupt into overlying coronal loops (indicated by the red arrow in Figure 6(r)), which expand upward with the MFR and filament and move out of the EUVI FOV by 17:36 UT, with the average velocity of about 274 km s⁻¹. Associated with loops moving out of the EUVI FOV is the appearance of the brightening front of a CME. First appearing in COR1 at around 17:35:00 UT as faint brightening, the semicircular bright front is clear by 17:37:30 UT, as shown in Figure 8(a). The filament, which moves behind the coronal loops, evolves as the coiled structure within the associated CME. Figure 8(b) shows that by around 17:42:30 UT, the bright front has separated from the coiled filament and the classical three-part CME moves outward at 1400 km s⁻¹. During this phase, the morphology of AR 11429 is typical of an eruptive flare, showing two flare ribbons separating and expanding, spanned by post-flare loops. It is primarily during the post-eruption phase that is observed by the Jansky VLA, producing a variety of coherent decimetric radio bursts. These will be described in a separate publication.

Zoom In Zoom Out Reset image size

The difference evolutionary phase and corresponding main observational features are listed in Table 1. Here, we briefly summarized the dynamical evolution of the whole event as follows:

• 1.  
  Formation of the ARF: Prior to the formation of the ARF, two filaments (F1 and F2) are identified in close proximity in the FC. At that time, there is no identifiable filament along the western part of the FC. An EUV brightening near the point where F1 and F2 cross appears at about 16:46 UT. Soon afterwards, at around 16:48 UT, an arch-shaped structure L1 appears in the 1600 Å image that evolves continuously with the 1600 Å brightening co-spatial with the EUV brightening. F1 and F2 reconfigure at around 16:53 UT, with the formation of UF and LF with L1 located in between. The high temperature bandpass of 131 Å shows the appearance of a reversed-ω-shaped structure at about 16:54 UT, and the sunken structure spatially coincident with the EUV brightening and the location where F1 and F2 cross. The UF experiences a partial and confined eruption, peaking at about 17:08 UT. RHESSI showed the possible nonthermal HXR emission that was coincident with the CE. Meanwhile, a plasma flow, which originates from the east footpoint of F1 and moves toward west along the PIL continuously after the reconfiguration to the UF and LF, extends the visible end of UF to the west and forms a new ARF. Magnetic flux in the area, where numerous EUV brightening occurred frequently, decrease associated with a small flux cancellation.
• 2.  
  Eruption of the ARF and MFR: The newly formed ARF begins to rise and expand about 2 minutes after its formation. During its slow rise phase, low-lying bright loops, which initially cover the western part of the PIL, are entwined with the ARF, rising and expanding with the ARF. From both AIA and EUVI-A perspectives it is possible to show that the bright feature is twisted around the ARF body by four turns, which we identify as an MFR. The configuration of the erupting filament and MFR changes from a spiral-arch shape to a reversed-γ shape within 75 s at the beginning of the fast rising phase, which represents the transformation from strong twist to writhe. Then the reversed-γ-shaped filament expands and rises, moving out of the EUVI-A FOV within 15 minutes.
• 3.  
  Post-eruption: After the filament and MFR eruptions, the evolution event in both the AIA and EUVI-A images of AR 11429 show flare ribbons and the formation of post-flare loops. The associated CME appears in COR1 at around 17:35 UT.

In Figure 9, we present an NLFFF extrapolation (Wiegelmann 2004; Wiegelmann et al. 2006), and simultaneous 211 and 131 Å images for a reference time of 16:48 UT as seen from the top of the domain. In Fig 9(a), some selected representative field lines outline the magnetic topology and EUV counterpart just before formation of the ARF. In Fig 9(b), blue and red arrows indicate F1 and F2, which we can see more clearly in an enlarged image in Figure 2(d). Qualitatively, the selected extrapolated field lines in the whole active region show reasonable agreement with the plasma loops seen in EUVI images taken by AIA. Two low-lying magnetic structures exhibit MFR morphology and are found to be co-spatial with F1 (blue field lines) and F2 (red field lines). Especially, the structure of MFR2 extends to the west longer than that of F2. Comparing the magnetic field extrapolation and observational results, we suggest a physical scenario for the whole event as summarized schematically in Figure 10. Two low-lying MFRs are located along the FC of AR 11429 before the eruptions. One MFR (blue field lines, MFR1) contains a filament as F1. The twist of this MFR is about one turn. The other MFR (red field lines, MFR2) is still during its emerging phase and only the east part where emerged contains coronal material as F2. While the twist of it is about three turns. Reconnection occurs at the location where MFR1 and MFR2 and their associated filaments cross (Figure 10(a)) as proposed by the tether-cutting model, forming a larger complex MFR that undergoes a partial and confined eruption (Figure 10(b)) seen in the EUV and HXR bands. The larger MFR structure enables plasma to flow from the east to the west, forming a long ARF embedded in the twisted MFR (Figure 8(c)). The ARF and MRF present a twist of four turns, which is well above the critical twist needed for kink instability. Then it begins rising slowly about 2 minutes after it forms. During its slow rise, its twisted morphology can be identified clearly from both SDO and STA observations. It then erupts rapidly, transitioning from twist to writhe. The erupting MFR and filament then drive a fast CME, producing a classic two-ribbon flare and post-flare loops in its aftermath, consistent with the standard CSHKP (Carmichael 1964; Sturrock 1968; Hirayama 1974; Kopp & Pneuman 1976) flare-CME model (Figure 10(d)). The reconnection between coronal flux-rope configuration was simulated by Török et al. (2011) for corona conditions when they studied the dynamic interaction between two ARFs which occurred on 2003 November 20. The observations and their simulation show that the reconnection can occur when the two flux ropes locate close enough. Here, in our case, the two MFRs and their associated filaments share the same FC and crisscross from head to tail, indicative of a similar process to occur.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

It widely believed that flare, filament eruption, and CME are different responses of spectacular eruptions in the solar atmosphere. They originate from a subregion of the active region in the photosphere (flare), expand to cover a significant part of the solar surface in the corona (CME), and further extend all the way from the low corona to the interplanetary space. For the M8.4 flare, which we presented in the present work, it associates with a filament eruption, a CME, and also other different solar activities. In the current study, we mainly focus on the precursor phase in the low atmosphere, i.e., the formation of a filament via the interaction between two small filaments and the morphology of its hosted MFR during the eruption. So the physical scenario proposed in the present study has ignored other activities. We mainly considered the formation of the filament in the low atmosphere and simply regarded the following eruption as the eruption of an MFR-hosted filament. Another study, which presented the properties of two consecutive eruption magnetic structures and their associated "double-decker" filaments during the impulsive phase of this event, was presented by Dhakal et al. (2018). Even though we suggest that one of the filaments is newly formed during the pre-activities, their erupting scenarios are helpful in understanding double-decker eruption. The double-decker structure was first identified by Liu et al. (2012) as filament composed of two branches separated in height. Subsequently, double-decker flux rope was reported with two flux ropes intertwined with each other (Cheng et al. 2014; Mitra et al. 2020). Dhakal et al. (2018) presented a study of this event in which they interpreted the filament eruption in terms of double-decker filaments (Liu et al. 2012). In contrast, we suggest that instead of the eruption of a preexisting double-decker morphology, the precursor phase is characterized by filament formation and a confined eruption.