rapid sunspot rotation associated with the x2.2 flare on ... - IOPscience

The Astrophysical Journal, 744:50 (6pp), 2012 January 1 C 2012.

doi:10.1088/0004-637X/744/1/50

The American Astronomical Society. All rights reserved. Printed in the U.S.A.

RAPID SUNSPOT ROTATION ASSOCIATED WITH THE X2.2 FLARE ON 2011 FEBRUARY 15 Yunchun Jiang1 , Ruisheng Zheng1,2 , Jiayan Yang1,3 , Junchao Hong1,2 , Bi Yi1,2 , and Dan Yang1,2 1

National Astronomical Observatory/Yunnan Astronomical Observatory, Chinese Academy of Sciences, P.O. Box 110, Kunming 650011, China; [email protected] 2 Graduate School of Chinese Academy of Sciences, Beijing 100049, China 3 Key Laboratory for the Structure and Evolution of Celestial Objects, Chinese Academy of Sciences, Kunming 650011, China Received 2011 May 28; accepted 2011 September 19; published 2011 December 13

ABSTRACT We present observations of sunspot evolution associated with the first X-class flare of the present solar cycle 24, which occurred in AR 11158 on 2011 February 15. The active region consisted of four emerging bipoles that showed complicated sunspot motion. The preceding spot of a bipole underwent the fastest movement. It not only passed through the following end of another bipole, thus causing a shearing motion, but also merged with the same-polarity spots and formed a single, larger umbra. This led to the formation of a δ configuration with an S-shaped neutral line, above which an extreme ultraviolet filament channel and a sigmoid formed and erupted to produce the flare. Along with the development of a clockwise (CW) spiral penumbra-filament pattern, the merged spot started rapid CW rotation around its umbral center 20 hr before the flare. The rotation persisted throughout the flare but stopped sharply about 1 hr after the flare ended, maintaining the twisted penumbra-filament pattern. The moving spot also caused continuous flux cancellation; in particular, its outer penumbra directly collided with small opposite-polarity spots only 100 minutes before the flare. When the shearing and rotational motions are main contributors to the energy buildup and helicity injection for the flare, the cancellation and collision might act as a trigger. Our observations support the idea that the rotation can be attributed to the emergence of twisted magnetic fields, as proposed in recent theories. Finally, the cause of its sudden halt is discussed. Key words: Sun: activity – Sun: flares – sunspots – Sun: surface magnetism On 2011 February 15, an X2.2 flare occurred in NOAA AR 11158 (S21◦ , W21◦ ). Rapidly shearing and rotational motions of a sunspot were observed by the Hinode satellite (Kosugi et al. 2007) with high spatial resolution and by the Helioseismic and Magnetic Imager (HMI; Wachter et al. 2011) on board the Solar Dynamics Observatory (SDO; Schwer et al. 2002) with high cadence. This provided us with a good opportunity to detail their origin and correlation with the flare.

1. INTRODUCTION It is generally believed that solar flares and coronal mass ejections (CMEs) are powered by the free energy stored in stressed magnetic fields, and the evolution of active regions (ARs) reflects the transport of magnetic energy and helicity from the subsurface into the corona. After decades of thorough research it is well established that some kinds of AR configurations and changes, such as δ-sunspots, soft X-ray (SXR) sigmoids, flux emergence, sunspot collision, and shearing and rotational motions, yield a high probability of producing flares (Zirin & Leggett 1987; Canfield et al. 1999; Wang 2001). In particular, sunspot rotation, first observed by Evershed (1910) a century ago, has been considered as an attractive flare energy buildup mechanism (Stenflo 1969; Barnes & Sturrock 1972; Amari et al. 1996; Tokman & Bellan 2002; Török & Kliem 2003). The improvement of observational capability makes it easier to identify rotating sunspots and more accurately measure them, and thus they have recently received considerable attention. It is found that some sunspots rotate as much as several hundred degrees around their umbral centers over a few days (Brown et al. 2003; Zhang et al. 2007; Min & Chae 2009), and both idealized analysis and simulation models suggested that sunspot rotation can be caused by the emergence of a twisted magnetic flux tube at its footpoints (Longcope & Welsch 2000; Gibson et al. 2004; Magara 2006; Fan 2009). Consistent with the results of magnetohydrodynamic (MHD) numerical simulations (Gerrard et al. 2003; DeVore & Antiochos 2008; Titov et al. 2008; Aulanier et al. 2010), rotating sunspots were also observed to be temporally and spatially associated with later flares and CMEs in a few events (Régnier & Canfield 2006; Tian & Alexander 2006; Zhang et al 2008; Srivastava et al. 2010), although the temporal relationship might not be strong (Kazachenko et al. 2009). However, the cause and effect of rotating sunspots are not yet very well understood.

2. OBSERVATIONS The AR and the flare were observed by the Solar Optical Telescope on board Hinode in the near-photospheric G-band (4300 Å) and in the chromospheric Ca ii H (3970 Å) channel with a pixel size of 0. 108. Unfortunately, the high-resolution G-band observations only have a low cadence of 2 hr. To detail the AR evolution, we adopt photospheric full-disk continuum intensity images and line-of-sight magnetograms observed by SDO/HMI in the Fe i 6173 Å absorption line, with a spatial sampling of 0. 5 pixel−1 and a higher cadence of 45 s. The associated chromospheric and coronal structures are examined using extreme ultraviolet (EUV) and SXR data from the Atmospheric Imaging Assembly (AIA; Boerner et al. 2011) and the X-Ray Telescope (XRT) on board SDO and Hinode, respectively. AIA takes full-disk multi-wavelength images with a pixel size of 0. 6 and a cadence of 12 s. XRT followed the AR with a 2 minute cadence, and its data cover a wide range of temperature that reveals all of the coronal features. All of the above data are then differentially rotated to a reference time close to the flare (2011 February 15 02:00 UT) and cropped to follow the AR. The associated CME is identified using coronagraph data from SOHO/LASCO, and the long-term and back-disk evolution of the AR is also examined using EUV movies from STEREO/EUVI. Finally, we used the SXR light curves 1

The Astrophysical Journal, 744:50 (6pp), 2012 January 1

Jiang et al.

whereas Fb only underwent minor motion at an average speed of about 0.03 km s−1 (panel (a6)). Thus, the shearing was mainly due to the northwestern motion of p2. Moreover, Fb and the merged p2 overlapped into a common penumbra, thus forming a δ configuration with an S-shaped neutral line (NL) between them (panels (a3–a6) and (b)). The flare took place in the course of the merged p2 moving away from Fb, and the Hinode Ca ii H image at the flare peak (panel (c)) showed that there were two ribbons at both sides of the NL, implying that it was closely associated with the shearing motion between Fb and the merged p2. By examining AIA and XRT observations, it was found that complex bright and dark features appeared that changed quickly along the NL, but an S-shaped EUV filament channel and an SXR sigmoid formed over it just before the flare. This is also shown by the close-up view of the NL in AIA 304 Å and XRT Be-thin filter images in Figure 2 (panels (d1–d2) and (e1–e2)). These S-shaped structures can be tracked until the flare starts (panels (d2) and (e2)), and it is very likely that the flare was produced only by their eruptions, although the eruption process was too rapid to be traced. These observations suggest that the fast motion of p2 and its interaction and mergence with Fb and p1 might contain the key elements that are needed to produce the flare. In Figure 2, the merged p2 showed two distinct characteristics. One is its rapid rotation around the umbral center, which can be seen by comparing panels (a3) and (a5). The other is its cancellation and collision with small opposite-polarity sunspots, as indicated by the black arrows. The rotation started at about 06:00 UT on February 14. Using the method described by Brown et al. (2003), we first obtained the changing center of p2’s umbra over time for both Hinode G-band and HMI intensity images. Because the merged p2 had a sole umbra that showed obvious motion (see panels (a3–a5)), a shifting small window covering p2 only is manually selected for each image, and then the umbral center is determined by a center-of-mass calculation for any pixel with intensity below 0.22/0.35 times that of the solar surface in G-band/HMI images, respectively. Figure 3 presents high-resolution G-band images centered at p2’s varying center. As indicated by the dashed lines, three penumbral features, “A,” “B,” and “C,” can be identified to undergo clearly clockwise (CW) rotation. Before the flare (panels (a–c)), C and B rotated slowly, while A had fast rotation that overtook and coalesced with B (panel (c)). After the flare (panel (d)), however, these features only showed minor or negligible rotation. It is also clear that the rotation was accompanied by the development of the CW spiral pattern of p2’s penumbral filaments, meaning that the CW rotation was responsible for right-handed twist and positive magnetic helicity. Consistent with the formation of the S-shaped NL, EUV filament channel, and sigmoid, such behavior obeys the chirality rule in the southern hemisphere (Zirker et al. 1997). As a key point of the event, it is noteworthy that the highly twisted filament pattern still persisted after the flare (panel (d)). To obtain information on the angular speed of the rotation, HMI intensity images are then unwrapped anticlockwise to a polar r–θ frame from the initial Cartesian x–y frame, starting from a westward-pointing chord (see panel (d) in Figure 3). According to Brown et al. (2003), the penumbra is the part of the sunspot that rotates the fastest and the feature-rich penumbral measurements are more reliable than umbral ones. A time slice is thus taken at a constant radius of 10. 8 from p2’s center, which includes A–C (indicated by the dashed circles in Figure 3). The result is displayed in Figure 4, with a time span of 41 hr and a

observed by the Geostationary Operational Environmental Satellite (GOES) to track the flare time. 3. RESULTS The flare attracted our attention because it was the first X-class flare of the current solar cycle 24. It had start, peak, and end times around 01:44, 01:56, and 02:06 UT, respectively, and was associated with an EUV Imaging Telescope wave event and a halo CME observed by AIA and LASCO (Schrijver et al. 2011). AR 11158, an emerging one, underwent a rapid growth and complex sunspot motion and interaction during its disk passage. Figure 1 shows its general evolution in HMI magnetograms and intensity images. The AR crudely consisted of four emerging bipoles, “Pa-Fa,” “Pb-Fb,” “p1-f1,” and “p2-f2.” The preceding (positive) sunspots are marked with P and p, while the following (negative) ones are marked with F and f. HMI magnetogram movies showed that Pa-Fa and p1-f1 first appeared at about 04:00 and 21:00 UT, respectively, on February 10 and then displayed separate motion without interaction between them. At 14:11 UT on February 12 (panels (a1) and (b1)), the two bipoles were clearly seen. Pb-Fb and p2-f2 begun to emerge inside Pa-Fa and p1-f1 after around 14:30 and 23:30 UT, respectively, on February 12. By 02:35 UT on February 13 (panels (a2) and (b2)), they can be clearly identified in the interior of Pa-Fa and p1-f1. Note that Pa-Fa and Pb-Fb had east–west axis orientations, while p1-f1 and p2-f2 had southeast–northwest ones. During the growth of the bipoles, their polarities, except that of p2, were highly fragmentized into some small pores or spots. The whole AR expanded rapidly as a result of the usual spreading action, accompanied by interaction, collision, and mergence of some polarities of the bipoles. By 14:11 UT on February 15 (panels (a4) and (b5)), mergence of the polarities of the four bipoles can be identified clearly, that is, Pb merged with Pa, f2 merged with f1, Fb merged with Fa, and p2 merged with p1 (labeled as p2). Therefore, it is difficult to determine exactly the velocity for each polarity except p2. The trajectory of p2’s center is easily traced and plotted as black curves in panels (a2) and (a4). In order to show approximately the motions of other polarities, their crude centers in panels (a1) and (a2) are simply connected to those in panel (a4) by the black/white arrows. Generally, the preceding polarities moved faster forward, whereas the following polarities moved more slowly backward. The average speeds were less than 0.1 km s−1 for the following polarities, but they fell in a range of 0.07–0.19 km s−1 for the preceding polarities pa, p1, and pb. In particular, p2 traveled along a curved trajectory from the following ends of Pa-Fa and Pb-Fb to their middle part, thus showing the largest displacement and the fastest separate motion from f2 since its first appearance at about 01:30 UT on February 13. The average speed of p2 reached 0.92 km s−1 , which was much higher than those of any other polarities. These results were consistent with the observations given in Figure 1. Panels (a3–a4) and (b3–b4) in Figure 1 show that the fast-moving p2 interacted with p1 and Fb. To reveal such a process clearly, Figure 2 presents a close-up view of the AR in G-band images with limited field of view (FOV). We see that p2 quickly moved toward and interacted with the growing Fb (panels (a1–a2)); overtook and merged with p1 as a larger, single spot (panel (a3)); and then diverged from Fb (panels (a4–a6)). Therefore, there was obvious shearing motion between Fb and the merged p2. After the interaction and mergence with Fb and p1, the average speed of the merged p2 decreased from 0.92 to 0.73 km s−1 and persisted at least until the end of February 15, 2


Jiang et al.

Figure 1. HMI magnetograms (a1–a4) and intensity images (b1–b5) showing the evolution of AR11158. The AR consisted of four bipoles, “Pa-Fa,” “Pb-Fb,” “p1-f1,” and “p2-f2.” The vertical bars indicate the moving p2, and the black curves mark the trajectory of its center from 01:30 UT of February 13 to the end of February 15, showing five times with a 12 hr interval being indicated. The arrows in panels (a1) and (a2) indicate the spread motions of the other poles of the bipoles. The FOV is 207 × 126 , and the boxes indicate the FOV for Hinode images in Figure 2.

Figure 2. Close-up view of AR11158 and the X2.2 flare in Hinode G-band images (a1–a6), HMI magnetogram (b) and Hinode Ca ii H image (c) at the flare peak, and AIA 304 Å (d1–d2) and XRT SXR (e1–e2) images before the flare start. The S-shaped neutral line, “NL,” determined from the 01:56 UT HMI magnetogram, is superimposed as dashed curves, the umbral outlines of Fb and p2 determined from the corresponding HMI intensity image are overplotted in (e1), and the white curve in (a6) shows the trajectories of p2’s center from 18:00 UT of February 13 to the end of February 15, along with the thin white arrow indicating its direction of motion and four times with a 12 hr interval being marked. The thick white arrow indicates the EUV filament channel, and the thick black arrows indicate the cancellation and collision of p2 with small opposite-polarity sunspots on its way. The FOV for AIA and XRT images is 108 × 66 , and it is 98. 4 × 60 for the others (indicated by the box in (d1)). The solid/dashed box in (a4)/(b) indicates the FOV in Figure 3/5.

3


Jiang et al.

Figure 5. Close-up view of the cancellation and collision of p2 with small opposite-polarity sunspots, “1” and “2,” in HMI magnetograms (a1–a3) and Hinode G-band images (b1–b3), with an FOV of 24. 6 × 29. 4. The black curve marks the trajectory of p2’s center from 23:17 UT of February 14 to 23:00 UT of February 15, and three times on February 15 with a 6 hr interval are indicated. The white arrows indicate the collision location, and the solid line, “L,” pointing from SE to NW, marks the slit position of the time slices shown in Figure 6.

Figure 3. Hinode G-band images centered at p2’s varying center. Three penumbral features, “A,” “B,” and “C,” show clear CW rotation, along with the development of CW spiral pattern of p2’s penumbral filaments. The dashed circles, with a radius of 10. 8 from p2’s center, indicate the location of the time slice shown in Figure 4. The r–θ polar coordinate system for uncurling the rotating p2 is shown in (d). The FOV is 30 × 30 .

θ range of 120◦ –360◦ that completely covers the rotating A–C as diagonal dark streaks. Beginning at about 06:00 UT on February 14, A (C) rotated about 107◦ (38◦ ) in a period of 21 hr. B did not show obvious rotation until 17:30 UT on February 14, but then it rotated about 43◦ in 9.5 hr and eventually merged with A. The rotating A–C showed up as nearly straight diagonal streaks, suggesting that they had almost uniform angular speed. Applying linear fitting to the diagonal streaks gives average rotational speeds of 5.◦ 1, 4.◦ 5, and 1.◦ 8 hr−1 for A–C, respectively. Compared with previous results (Brown et al. 2003; Zhang et al 2008), the rotation of A and B was very strong. Consistent with what we deduce from G-band images and with the previous results (Tian & Alexander 2006, and references therein), it is clear that p2’s rotation started about 20 hr before the flare and persisted throughout the flare duration. However, the most important characteristic revealed by the time slice is the sharp stop of A–C rotation at about 03:00 UT, within 1 hr after the flare ended at 02:06 UT. Although the decrease of angular velocity in rotating sunspots after a major flare had been reported previously (Zhang et al. 2007; Min & Chae 2009), this sudden halt of sunspot rotation was observed for the first time. As mentioned above, the rotating p2 canceled and collided with small sunspots when it moved toward the northwest. This is illustrated by the close-up view of HMI magnetograms and G-band images in Figure 5. Two small negative-polarity sunspots, “1” and “2,” first stood in p2’s way (panel (b1)). The moving p2 then pushed 2 to approach 1. At 00:13 UT before the flare, p2’s outer penumbra had collided with 2. By 07:58 UT after the flare, 1 and 2 were squeezed together and connected with p2’s penumbra. In this process, flux cancellation was manifested by a continuous decrease in area of the negative flux. To display the cancellation and collision clearly, time slices along a line, “L,” which passes the collision location and the two small spots, are taken from both HMI intensity images and magnetograms, and changes of the negative flux in the window of Figure 5 are measured. Note that the slit L is not along the trajectory of p2’s center. The results are given in Figure 6 with

Figure 4. Time slice of a band in p2’s penumbra at a fixed radius (see Figure 3) made from HMI intensity images. Time profiles of the GOES-14 1–8 Å SXR flux are overplotted, and the vertical black bar indicates the flare end time. The dashed lines mark the linear fittings to the diagonal dark streaks A–C, and the vertical white line indicates the sudden halt of the rotation. The white arrow indicates that B showed almost no rotation at the beginning, and the black arrow indicates the mergence of A and B.

4


Jiang et al.

Our observations reveal some important aspects about the origin and effect of sunspot rotation. (1) Consistent with the idea suggested by some researchers (Brown et al. 2003; Tian & Alexander 2006; Min & Chae 2009), the rotation might result from the emergence of pre-twisted magnetic fields, given that the whole AR was in its growth phase and, in particular, p2 was formed by the collection of the preceding ends of two emerging bipoles. (2) Accompanied by the development of the whorl pattern of p2’s penumbral filaments, it is not surprising that p2 started rapid rotation about 20 hr prior to the flare. This process might lead to steadily coronal freeenergy buildup that is later released by the flare (Stenflo 1969; Barnes & Sturrock 1972, and references therein), as well as helicity accumulation that was removed from the Sun through the associated halo CME (Zhang & Low 2005; Zhang et al 2008). (3) The rotation persisted throughout the flare and lasted about 1 hr more after the flare ended. This further indicates that it might be produced by a primary interior driver that can inject twist into the atmosphere after the coronal energy release and helicity removal. (4) However, a natural question is: what caused the sharp brake of the rotation? In her threedimensional MHD numerical simulations, Fan (2009) showed that sunspot rotation can be a result of propagation of nonlinear torsional Alfvén waves along an emerging pre-twisted magnetic flux tube, which is driven by a gradient in the rate of twist and transports significant twist from the tube’s interior highly twisted portion to the expanded coronal portion (Longcope & Welsch 2000). Whereas this is supported by previous studies showing that subsurface horizontal vortical flows and magnetic twists also exist beneath the photosphere and can be related to solar eruptions (Zhao & Kosovichev 2003; Mason et al. 2006), more recently Gosling et al. (2010) presented evidence for a torsional Alfvén wave embedded within a small magnetic flux rope in a solar wind event. Fan (2009) further suggested that sunspot rotation would stop when a final equilibrium was reached in the interior and coronal twist. Owing to the high cadence of HMI data, two key factors lead us to believe that direct photospheric reflections of such a final equilibrium status were observed in our event. One is the sudden halt of p2’s rotation, and the other is the maintenance of its highly twisted filament pattern even after the halt. Moreover, it is found that no major eruption occurred in the AR since the X2.2 flare by tracking its evolution using observations from AIA and STEREO/EUVI. Finally, we would like to point out that since sunspot rotation is not an uncommon phenomenon in ARs (Yan et al. 2008), HMI observations with high time cadence and spatial resolution will be greatly helpful in further understanding their real nature.

Figure 6. Time slices from HMI intensity images (a) and magnetograms (b) for the slit L shown in Figure 5. Time profiles of the GOES-14 1–8 Å SXR flux and changes in negative magnetic flux in the window of Figure 5 are overplotted in (b). For clarity, the absolute values for the negative flux are plotted. The arrows indicate the collision of p2’s penumbra with 2 at 00:06 UT of February 15. The dotted and dashed lines in (a), the linear fittings to the outer edge of p2’s umbra, indicate two distinct motion stages before and after 01:12 UT. The horizontal bars indicate the flare duration, and the flare start time of 01:44 UT is also indicated.

a time span of 14 hr. We can see coalescence and flux loss of 1 and 2 from the magnetogram time slice, and their flux continuously decreased from 6.9 to 3.9 × 1020 Mx within 14 hr. In particular, p2’s umbra and penumbra can be clearly identified in the time slice of intensity images. This leads to the findings that p2’s penumbra began to collide with 2 at about 00:06 UT on February 15 and then the average speed at the outer boundary of its umbra decreased from 0.24 to 0.06 km s−1 at 01:12 UT. Such a decrease may represent a response of p2’s motion to the collision with 2. We tried but failed to find a similar decrease in the average speed of p2’s center. The flux cancellation may play a role in triggering of the eruption by modifying the line-tying condition for the EUV filament channel and sigmoid (Wang & Shi 1993; Jiang et al. 2007), which was further evidenced by the high temporal closeness between the collision and the flare. 4. CONCLUSIONS AND DISCUSSION

We thank an anonymous referee for many constructive suggestions and thoughtful comments that improved the quality of this paper. We thank the Hinode, AIA, HMI, GOES, SOHO/LASCO, and STEREO/EUVI teams for data support. Hinode is a Japanese mission developed and launched by ISAS/JAXA, with NAOJ as domestic partner and NASA and STFC (UK) as international partners. It is operated by these agencies in cooperation with ESA and NSC (Norway). This work is supported by the 973 Program (2011CB811403) and by the Natural Science Foundation of China under grants 10973038 and 11173058.

The AR was in a favorable condition to produce flares, such as the emerging flux, the formation of a δ configuration with the EUV filament channel and sigmoid over the S-shaped NL, the shearing and rotational motions of p2 and the resultant twist development, and the cancellation and collision of p2 with small opposite-polarity sunspots. All of these factors might make more or fewer contributions to the flare’s occurrence, but the shearing and rotation are probably the vital elements for the energy buildup and helicity injection owing to their close spatial relationship with the flare, while the cancellation and collision might be a trigger for the energy release. Although beyond the scope of this paper, it is clear that quantitative analysis, as done by Kazachenko et al. (2010), is required to determine the relative importance of the shearing and rotation in the flare.

REFERENCES Amari, T., Luciani, J. F., Aly, J. J., & Tagger, M. 1996, ApJ, 466, L39 Aulanier, G., Török, T., Démoulin, P., & DeLuca, E. E. 2010, ApJ, 708, 314

5


Jiang et al. Régnier, S., & Canfield, R. C. 2006, A&A, 451, 319 Schrijver, C. J., Aulanier, G., Title, A. M., Pariat, E., & Delannée, C. 2011, ApJ, 738, 167 Schwer, K., Lilly, R. B., Thompson, B. J., & Brewer, D. A. 2002, AGU Fall Meeting Abstracts, SH21C-01 Srivastava, A. K., Zaqarashvili, T. V., Kumar, P., & Khodachenko, M. L. 2010, ApJ, 715, 292 Stenflo, J. O. 1969, Sol. Phys., 8, 115 Tian, L., & Alexander, D. 2006, Sol. Phys., 233, 29 Titov, V. S., Mikic, Z., Linker, J. A., & Lionello, R. 2008, ApJ, 675, 1614 Tokman, M., & Bellan, P. M. 2002, ApJ, 567, 1202 Török, T., & Kliem, B. 2003, A&A, 406, 1043 Wachter, R., Schou, J., Rabello-Soares, M. C., et al. 2011, Sol. Phys., in press Wang, J. 2001, Space Sci. Rev., 95, 55 Wang, J., & Shi, Z. 1993, Sol. Phys., 143, 119 Yan, X. L., Qu, Z. Q., & Xu, C. L. 2008, ApJ, 682, L65 Zhang, J., Li, L., & Song, Q. 2007, ApJ, 662, L35 Zhang, M., & Low, B. C. 2005, ARA&A, 43, 103 Zhang, Y., Liu, J., & Zhang, H. 2008, Sol. Phys., 247, 39 Zhao, J., & Kosovichev, A. G. 2003, ApJ, 591, 446 Zirin, H., & Leggett, M. 1987, Sol. Phys., 113, 267 Zirker, J. B., Martin, S. F., Harvey, K., & Gaizauskas, V. 1997, Sol. Phys., 175, 27

Barnes, C. W., & Sturrock, P. A. 1972, ApJ, 174, 659 Boerner, P., Edwards, C., Lemen, J., et al. 2011, Sol. Phys., in press Brown, D. S., Nightingale, R. W., Alexander, D., et al. 2003, Sol. Phys., 216, 79 Canfield, R. C., Hudson, H. S., & McKenzie, D. E. 1999, Geophys. Res. Lett., 26, 627 DeVore, C. R., & Antiochos, S. K. 2008, ApJ, 680, 740 Evershed, J. 1910, MNRAS, 70, 217 Fan, Y. 2009, ApJ, 697, 1529 Gerrard, C. L., Brown, D. S., Mellor, C., Arber, T. D., & Hood, A. W. 2003, Sol. Phys., 213, 39 Gibson, S. E., Fan, Y., Mandrini, C., Fisher, G., & Demoulin, P. 2004, ApJ, 617, 600 Gosling, J. T., Teh, W.-L., & Eriksson, S. 2010, ApJ, 719, L36 Jiang, Y. C., Shen, Y. D., & Wang, J. X. 2007, Chin. J. Astron. Astrophys., 7, 129 Kazachenko, M. D., Canfield, R. C., Longcope, D. W., & Qiu, J. 2010, ApJ, 722, 1539 Kazachenko, M. D., Canfield, R. C., Longcope, D. W., et al. 2009, ApJ, 704, 1146 Kosugi, T., Matsuzaki, K., Sakao, T., et al. 2007, Sol. Phys., 243, 3 Longcope, D. W., & Welsch, B. T. 2000, ApJ, 545, 1089 Magara, T. 2006, ApJ, 653, 1499 Mason, D., Komm, R., Hill, F., et al. 2006, ApJ, 645, 1543 Min, S., & Chae, J. 2009, Sol. Phys., 258, 203

6