Abstract
We investigate the formations and evolutions of two successive solar tornadoes in/near AR 12297 during 2015 March 19–20. Recurrent EUV jets close to two filaments were detected along a large-scale coronal loop prior to the appearances of the tornadoes. Under the disturbances from the activities, the filaments continually ascended and finally interacted with the loops tracked by the jets. Subsequently, the structures of the filaments and the loop were merged together, probably via magnetic reconnections, and formed tornado-like structures with a long spiral arm. Our observations suggest that solar tornadoes can be triggered by the interaction between filaments and nearby coronal jets, which has rarely been reported before. At the earlier development phase of the first tornado, about 30 small-scale sub-jets appeared in the tornado's arm, accompanied by local EUV brightenings. They have an ejection direction approximately vertical to the axis of the arm and a typical maximum speed of ∼280 km s−1. During the ruinations of the two tornadoes, fast plasma outflows from the strong EUV brightenings inside tornadoes are observed, in company with the untangling or unwinding of the highly twisted tornado structures. These observational features indicate that self reconnections probably occurred between the tangled magnetic fields of the tornadoes and resulted in the rapid disintegrations and disappearances of the tornadoes. According to the reconnection theory, we also derive the field strength of the tornado core to be ∼8 G.
Export citation and abstract BibTeX RIS
1. Introduction
Solar tornadoes are referred to as rapidly rotating magnetic structures in the solar atmosphere. Due to the magnetohydrodynamic (MHD) environment on the Sun, the physics of these tornado-like structures are very distinct from terrestrial tornadoes. Solar tornadoes are often found to have a close relation with prominence or filament activities (e.g., Li et al. 2012; Wedemeyer et al. 2013; Mghebrishvili et al. 2015; Su et al. 2015; Wang et al. 2017). In the early phase of the last century, based on the spectroheliograph observations from the Yerkes Observatory (Hale & Ellerman 1903), Pettit (1925) discovered that some prominences appeared like closely twisted rope or a fine screw and called them "tornado" or spiral prominences. Recently, using the EUV imaging data from the Atmospheric Imaging Assembly on board the Solar Dynamics Observatory (AIA, SDO; Lemen et al. 2012), Su et al. (2012) stated that filament barbs are actually projections of the tornado funnels and may play a role in the supply of mass and twists to filaments, confirmed by Wedemeyer et al. (2013). Relying on spectroscopic observations, some authors focused their studies on rotational motions (e.g., Su et al. 2014; Schmieder et al. 2017) or plasma diagnostics (e.g., Levens et al. 2015) of tornado-like prominences. There are also other reports about solar tornadoes, which seem to be irrelevant to filaments. Pike & Mason (1998) argued that rotating macrospicule-like features are a sort of solar tornadoes. Zhang & Liu (2011) reported a special kind of EUV rotating structures and named them "solar cyclones" (also see Yu et al. 2014). The chromospheric swirls detected by Wedemeyer-Böhm et al. (2012) have a smaller scale and a much shorter lifetime compared with the tornado-like prominences.
The origin of solar tornadoes has not yet been fully understood so far. Zhang & Liu (2011) found that the EUV cyclones are rooted in the rotating network magnetic fields (RNFs) and are probably driven by the rotation of the RNFs, which is supported by the observations of Yang et al. (2015). The MHD simulations performed by Wedemeyer-Böhm et al. (2012) provided evidence that solar tornadoes are propelled by photospheric vortex flows at their footpoints. Li et al. (2012) suggested that the cyclonic appearance and overall evolution of a tornado can be interpreted as the expansion of helical prominence structures into the cavity. Panesar et al. (2013) investigated the same tornado event as Li et al. (2012), and pointed out that the flares in the neighboring active region may have affected the cavity-prominence system before the prominence field inflates to the cavity. In this study, using the high-resolution data from the Interface Region Imaging Spectrometer (IRIS; De Pontieu et al. 2014) and SDO (Pesnell et al. 2012), we first report that solar tornadoes can be triggered by interactions between filaments and coronal jets. In addition, our observations reveal that self reconnections between the tangled magnetic fields inside solar tornadoes can lead to rapid disintegrations and disappearances of the tornadoes. In the next section, we describe the observational data we use. This is followed by a detailed investigation of the formation and development of two solar tornadoes. Finally, we summarize and discuss the results.
2. Observations
During 2015 March 19–20, two tornadoes (T1 and T2) successively formed and developed near active region AR 12297 (∼S16W79). The AIA which provides full-disk images up to 0.5 
above the solar limb with 0
6 pixel size and 12 s cadence in 10 wavelengths, covered this event well. We mainly used the data (Level 1.5 images) at 7 EUV channels centered at 304 Å (He ii, 0.05 MK), 171 Å (Fe ix, 0.6 MK), 193 Å (Fe xii, 1.3 MK and Fe xxiv, 20 MK), 211 Å (Fe xiv, 2 MK), 335 Å (Fe xvi, 2.5 MK), 94 Å (Fe xviii, 7 MK), and 131 Å (Fe viii, 0.4 MK and Fe xxi, 11 MK), respectively. This event was captured by IRIS slit jaw imager (SJI) in 1330 Å during three periods (09:09–14:07 UT; 16:18–21:16 UT; 21:26–03:26 UT) from March 19 to 20, using a spatial scale of 0332 and a cadence of 9.3 s. The longitudinal magnetograms and continuum intensity images with a 05 plate scale and 45 s cadence from the Helioseismic and Magnetic Imager (HMI; Schou et al. 2012) on board SDO are also utilized for this study.
3. Results
3.1. Formation and Development of T1
To explore the initiation mechanism of T1 and T2, we checked the AIA and HMI data on March 18. From Figures 1(a)–(c), we can see that two associated filaments (F1 and F2) were orderly aligned along the magnetic neutral lines of AR 12297 and the ambient quiet region. The animation of Figure 1 shows the evolutions of AR 12297 and the two filaments on that day. It can be found that some bright and dark plasma were intermittently ejected from the joint of F1 and F2, and moved along a large-scale coronal loop (also see the bottom panels of Figure 1). Disturbed by these recurrent EUV jets, F1 began to oscillate and continually rose. At around 07:41 UT on March 19, an obvious brightening appeared at the top of F1 (Figure 2(b)). Then, the dark and bright heated plasma in F1 started to rotate rapidly and move along the trajectory of the jets to the far end of the coronal loop. These phenomena suggest that there existed interaction between the magnetic fields of F1 and the large-scale magnetic loop. As a result of this kind of interaction, a tornado-like structure (T1) with a long spiral arm came into being, which is outlined by the helical dotted line in Figure 2(d). During the following development, more explosive events seem to take place inside T1. We can see that some bright blobs suddenly appeared and moved round the arm of T1 (Figure 2(e)) and some thread-like structures of the tornado's arm quickly spun around (Figure 2(f)). Filament F2 was not affected much by these activities.
Figure 1. (a) HMI continuum intensity map; (b) AIA 171 Å image displays the filaments F1 and F2 in/near AR 12297; (c) HMI longitudinal magnetogram; ((d)–(f)) AIA 171 Å images show three examples of the recurrent EUV jets close to F1 and F2 on March 18 (also see the animation). The red box in panel (a) indicates the field of view (FOV) of panels (b) and (c). The green and turquoise dotted curves outline F1 and F2, respectively; the blue arrows denote the ejection direction of the jets. The images in panels (b)–(f) have been rotated counterclockwise by 120°, which is the same for all AIA and IRIS images in the following figures. The center of panels (d)–(f) is at solar (x, y) = (886'', –338'') and the FOV is 210'' × 210''.
(An animation of this figure is available.)
Download figure:
Video Standard image High-resolution imageDownload figure:
Video Standard image High-resolution imageFigure 2. ((a)–(f)) AIA 171 Å images present the formation and development of the first tornado T1 (also see animation (a)); ((g)–(i)) IRIS 1330 Å SJI images exhibit some sub-jet structures in T1 (also see animation (b)). The blue dotted curves in panel (f) outline some bright thread structures in the tornado's arm. The circles and thick arrows in panels (e) and (g)–(i) indicate the bright blobs or sub-jets and their ejection directions, respectively. The blue rectangle denotes the FOV of panels (g)–(i). The center of panels (a)–(f) is at solar (x, y) = (922'', –341'') and the FOV is 120'' × 108''.
Download figure:
Video Standard image High-resolution imageThe IRIS observations on this event began from 09:09:11 UT on March 19. Thus, the formation process of T1 was not captured by IRIS. The bottom panels of Figure 2 present the main part of T1 in the IRIS 1330 Å channel during some internal explosive activities. It can be seen that accompanied by the local brightenings, some small-scale jet-like structures formed at different locations of the tornado's arm. Their ejection directions are approximately vertical to the axis of the arm. From 09:21 UT to 09:39 UT, about 30 such sub-jet structures appeared in total, with a typical maximum speed of ∼280 km s−1. A local correlation tracking (LCT) method (Welsch et al. 2004) is taken to compute the velocity field in the plane of the sky by employing two SJI images separated by 9 s. The top panels in Figure 4 show the velocity map of T1 at 09:31 UT and 09:52 UT, when the tornado was in active and relative stable stage, respectively. Probably correlated with local magnetic reconnection, the sub-jets (indicated by the circle in Figure 4(a)) obviously have much larger velocities of ∼300 km s−1 (a mean value of the velocities of all plasma in the sub-jets) than the other plasma. After 09:50 UT, T1 evolved into a gentle stage without significant brightening events. In Figure 4(b), we can clearly see the vortical flows at the tornado core region, which had a median angular speed of ∼3
6 min−1.
Figure 3. ((a)–(d)) IRIS 1330 Å SJI and ((e)–(h)) AIA 131 Å images display the disintegration of T1 (also see the animation); The curved arrows in panels (c) and (d) denote the movement directions of the ejected plasma from the self-reconnection region of T1. The arrows in panels (f) and (g) point to the hot tangled structures of T1 during the self-reconnection. The yellow box indicates the FOV of panels (a)–(d). The center of panels (e)–(h) is at solar (x, y) = (925'', –336'') and the FOV is 120'' × 96''.
(An animation of this figure is available.)
Download figure:
Video Standard image High-resolution image
Figure 4. IRIS 1330 Å SJI image taken at 09:31:14 UT (a) and 09:52:40 UT (b) on March 19, overplotted with the velocity fields calculated by the LCT method. The time periods of the velocity fields are 9 s from 09:31:14 UT to 09:31:23 UT and from 09:52:40 UT to 09:52:49 UT, respectively; (c) Temperature and (d) EM map of T1 at 12:26:56 UT on March 19. The arrows in panels (c) and (d) point to the self-reconnection region of T1. The FOV of panels (a) and (b) is the same as that of Figures 2(g)–(i). The center of panels (c) and (d) is at solar (x, y) = (927'', –344'') and the FOV is 120'' × 108''.
Download figure:
Standard image High-resolution image
3.2. Disintegration of T1
Figure 3 displays the final evolution stage of T1. After a relatively stable development of 2 hr, a new EUV brightening arose at the top of the tornado core at ∼11:58 UT on March 19. Subsequently, bright plasma blobs spouted out of the brightening region at a velocity of ∼140 km s−1. The arrows in Figures 3(c) and (d) point out the motion directions of the blobs. We can see that the ejected plasma initially moved along the spiral tornado arm (Figure 3(c)), but directly ran to the far end of the loop from the source region finally (Figures 3(d) and 1(e)). This phenomenon reflects that the complex tangled structures of T1 experienced an untangling and disintegration process, which was highly likely caused by the self-reconnection between the tangled magnetic fields inside T1. The self-reconnection and rapid mass ejection processes lasted for about 20 minutes and T1 completely disappeared after 13:00 UT.
In the AIA 131 Å channel, some hot coronal loop structures enwrapping the source region of the ejected plasma (denoted by the arrows in Figures 3(f) and (g)) were visible as emission. These structures seems to be tangled in the top of the tornado core at first and gradually developed into nearly potential field geometry with the disintegration and disappearance of T1 (see Figure 3(h)). Applying the differential emission measure (EM) method (Cheng et al. 2012) to the AIA simultaneous data in six EUV wavebands, we calculated the temperature (Figure 4(c)) and the EM (Figure 4(d)) of T1 at 12:26:56 UT, when the tornado was disintegrating. The bright self-reconnection region (indicated by the arrows in Figures 4(c) and (d)) and the hot loops are clear in the temperature and EM maps.
Based on our calculation, the plasma temperature can reach 10 MK and the mean EM value is ∼6.3 × 1028 cm−5 in the reconnection region. Assuming that the depth of the reconnection region along the line of sight equals to the width, an average electron density of ∼1.7 × 1010 cm−3 can be derived from these results. According to MHD theory, the plasma would be accelerated up to the Alfvén speed (
) after reconnection. Thus, we estimate the magnetic field strength (
) of the tornado core to be ∼8 G from the formula 
, where ρ is the mass density and 
approximately equals to 140 km s−1 assuming that the observed flow speed close to the reconnection site (the core) is equal to the reconnection outflow velocity. Allowing for the projection effect, 8 G is only the lower limit of the magnetic field strength. The result is consistent with that of ∼15 G measured from spectropolarimetric observations by Levens et al. (2016) (also see Martínez González et al. 2015, 2016; Schmieder et al. 2015).
3.3. Formation and Disintegration of T2
Some snapshots of the second tornado (T2) during its formation and disintegration are shown in the top and bottom panels of Figure 5, respectively. From about 15:50 UT on March 19, the filament F2 developed into a fast-rise phase. When reaching a height of ∼10 Mm, the top of F2 started to interact with the surrounding background field and produced a bright jet, which seems to pour along the trajectory of the aforementioned coronal loop (Figure 5(a)). In the IRIS 1330 Å SJI images, as F2 continually ascended, some bright threads (see the dotted curves in Figure 5(b)) were observed to writhe and spin quickly (Figure 5(b)). Finally, the rising filament plasma was ejected along the loop to form a helical tornado-like structure (T2) as outlined by the dotted curves in Figure 5(c). During the following relatively stable evolution of ∼7 hr, the inner material of T2 continuously rose and moved along the helical trajectory to the far end of the loop. From 00:12 UT on March 20, similar to the disintegration of T1, self reconnections seem to occur between the tangled fields inside T2, which heated and expelled the plasma out of the strong brightening regions. Meanwhile, the knotted structures of T2 underwent an obvious unwinding and disintegration process, which is reflected by the rapid spinning of some threads during the reconnection (Figure 5(d)) and the crossing motion of the expelled plasma in the end (Figure 5(f)). After 01:00 UT on March 20, T2 disappeared totally from the IRIS and AIA observations.
Figure 5. (a) AIA 171 Å images; IRIS 1330 Å SJI images present the formation ((b)–(c)) and disintegration ((d)–(f)) of the second tornado T2 (also see the animation); The dotted curves and arrows in panels (b) and (d) separately indicate some bright thread structures in F2 and T2 and the spinning directions of these threads. The ellipse in panel (e) marks a strong brightening produced by the self-reconnection of T2. The arrows in panels (e) and (f) denote the ejection directions of the plasma during the disintegration of T2. The blue box corresponds to the FOV of the SJI images. The center of panel (a) is at solar (x, y) = (922'', –341'') and the FOV is 120'' × 96''.
(An animation of this figure is available.)
Download figure:
Video Standard image High-resolution image
4. Summary and Discussion
We presented detailed observations of two successive solar tornadoes formed and evolved in/near AR 12297 during 2015 March 19–20. The two tornadoes (T1 and T2) were closely associated with the evolutions of two active filaments (F1 and F2), respectively. According to our observations, F1 and F2 were obviously disturbed by the nearby recurrent jet activities before they developed into tornadoes. The fast rotation and spiral appearances of the tornadoes are probably related to the relaxations of the magnetic twist stored in the filaments (e.g., Yang et al. 2014) during the reconnections and the following mass flows along the helical fields. Similar process may also take place during the formations of untwisting jets (e.g., Shen et al. 2011; Chen et al. 2012; Zhang & Ji 2014). So far, very few cases that tornadoes are triggered by solar activities have been reported. Panesar et al. (2013) hinted that the dynamical developments of a tornado may correlate with the flares in the adjacent active region. For the first time, our observations clearly show that solar tornadoes can originate from the interaction between filaments and coronal jets.
In this work, many observational features such as the EUV brightenings; the rapid rotating and spinning of field structures; mass flows; and sub-jets, suggest the presence of reconnections. It should be noted that the filaments associated with the tornadoes studied here are near/in solar active region. Thus, their evolutions are easily influenced by the activities from the active region. In addition, the filaments are more active and their field structures may have more twist (e.g., Yan et al. 2014) than the quiescent tornado-like prominences investigated by the other authors (e.g., Li et al. 2012; Su et al. 2012). This may be the reason why so many reconnection events corresponding to the EUV brightenings are observed during the tornadoes developed. Some small-scale jet-like structures arose in the tornado's spiral arm at the earlier growth of T1, which is an interesting phenomenon and has not been mentioned before. They have a typical maximum speed of ∼280 km s−1, which is comparable with that of some coronal jets related to strong flares (e.g., Chen et al. 2015). However, the sub-jets reported here evidently have much smaller spatial and timescales than usual 
surges or EUV jets (e.g., Jiang et al. 2007; Chen et al. 2008). They were accompanied by local EUV brightenings and ejected nearly vertical to the arm's axis, indicating that magnetic reconnection and the Lorentz force may play a joint role for their formations.
All solar tornadoes will go to extinction, but the way that they collapse may be different. According to Zhang & Liu (2011), many EUV cyclones are associated with microflares at the later phase and would become extinct as the cancellations of opposite photospheric fluxes take place. Li et al. (2012) reported a gradual dimming process of a solar tornado, but the detailed mechanism is not clear yet. Taking tornadoes as part of the filament, Su et al. (2012) and Wedemeyer et al. (2013) stated that tornadoes erupt together with the filament. However, the important role of the tornadoes in triggering the eruption still awaits to be assured. The tornadoes under our investigation exhibit a very different evolution pattern at the end of their lifetimes. Strong EUV brightenings accompanied by the fast plasma outflows and the untangling or unwinding of the field structures reflect that magnetic reconnections most likely occurred between the tangled fields inside the tornadoes. This may be a consequence of vortical flows winding up the magnetic structures of the tornadoes during the periods of developments. When the augmented twist or kink reaches a threshold and the neighboring magnetic field lines with locally opposite polarity vector components come into contact, the reconnection happens. We call this process "self-reconnection" of solar tornadoes, which is likely similar to the scenarios numerically simulated by Browning et al. (2008) or Kliem et al. (2010). The following rapid disintegrations and disappearances of the tornadoes just resulted from the self reconnections. The remarkable observations presented in this study are helpful to provide new insights for the origin and dynamical evolution of solar tornado-like filament.
IRIS is a NASA small explorer mission developed and operated by LMSAL with mission operations executed at NASA Ames Research center and major contributions to downlink communications funded by ESA and the Norwegian Space Centre. The SDO data are courtesy of NASA, the SDO/AIA and HMI science teams. This work was supported by NSFC (11533008, 41331068, 41204124, 11221063, 11673034, and 11673035), and the project funded by China Postdoctoral Science Foundation (2015M571126).