CIRCULAR RIBBON FLARES AND HOMOLOGOUS JETS

The active region of interest NOAA AR 6545 emerged on the east limb on 1991 March 10. It appears in the βγδ magnetic configuration till March 16 when it rotated to the disk center. The region is the source region for five X-class and nine M-class flares. According to the USAF/NOAA report, the region showed slight decay and possible magnetic simplification during March 17–18, while there is still clear evidence of a δ configuration in the bottom part of the leading spot (cf. Figures 2(a) and (c)). As an obvious property of magnetic field structure of this region, the central negative field seems to be encircled by the surrounding positive field, which forms a quasi-circular PIL (Figure 2(a)). The circular shape of the source region and a filament lying along the PIL are also clearly visible in Hα line center (Figure 2(b)).

From ∼17:10 UT on March 17 before the event occurrence, a dark preflare surge (PS) apparently rooted at the northern side of the central negative field gradually extends upward to form a long and slim structure (see Figure 2(e)). This outward ejection as observed in Hα blue wing seems to subside after ∼17:32 UT, while it picks up again from ∼17:37 UT till the event onset around 17:42 UT (Figure 3(a)). The M2 flare starts with rapid brightenings of the kernels k1 and k2 located in the negative and positive field regions, respectively (cf. Figures 3(b) and 2(a)). As the flare progresses, we note the following dynamic activities. First, k2 shows sequential brightening in the clockwise direction, moving from the top to the bottom of the circular region from ∼17:45 to 17:51 UT with an average speed of ∼100 km s⁻¹. This motion also seems to be apparently accompanied by the dark surge ejection (Figures 3(c)–(k)). Weaker ribbon brightenings in the south are discernible from ∼17:48 UT, so that the entire outer ribbon appears as a circle (Figure 3(e)). The circular shape of the flare ribbon and the progressive brightening along the rim of the circle are directly reminiscent of the event of M09. Meanwhile, the clockwise drifting of the surge axis seems in line with the model of P10. Second, k1 first moves northeastward from ∼17:46 UT at ∼40 km s⁻¹, and after k1 reaches the strongest negative field region at ∼17:49 UT (Figure 3(f)), intense surge associated with the circular ribbon in the west begins to eject outward. Subsequently, k1 turns back to propagate southwestward at ∼25 km s⁻¹ during ∼17:52–17:58 UT. There is a separate episode of ejection possibly due to the partial eruption of the circular filament (Figures 3(l)–(n)). k1 thus evolves from compact to elongated shape and undertakes an intriguing "round-trip" motion (illustrated in Figure 3(b)).

To track the flare ribbon quantitatively, we use the intensity-based binary masks method (e.g., Liu et al. 2007; Liu & Wang 2009) and depict the aforementioned motion of k1 and k2 from ∼17:42 to 18:01 as color-coded areas overplotted on the co-aligned NSO/KP magnetogram in Figure 2(d). In the standard 2D model, the flare ribbon motion is a direct mapping of energy release via coronal magnetic reconnection, the rate of which can be evaluated using the magnetic flux change rate of the chromospheric ribbon area as $\dot{\phi }= (\partial / \partial t) \int B\,da$, where B is the vertical magnetic field component of the surface ribbon element da. For this event near the disk center, we take the LOS magnetic field measurement and in order to estimate uncertainties of the ribbon selection, we vary the intensity cutoff value (about 160) within ±10% to obtain the error bars. The resulting magnetic flux change rate in both the positive ($\dot{\phi }_+$) and negative ($\dot{\phi }_-$) fields as a function of time are plotted in Figure 2(f). It can be seen that $\dot{\phi }_+$ and $\dot{\phi }_-$ are temporally correlated with the flare non-thermal emission (approximated using the time derivative of the SXR light curve), which is theoretically expected (Priest & Forbes 2002) and may demonstrate the essential role of magnetic reconnection in producing this event. We note that the intensity threshold used in selecting ribbons does not affect the overall timing (e.g., the peak time) of the derived $\dot{\phi }$ (Liu et al. 2007). However, the results of $\dot{\phi }_{+(-)}$ also obviously show that the magnetic flux change in two polarities are grossly unbalanced (cf. Qiu & Yurchyshyn 2005; Liu et al. 2007). We are cautious that (1) the calculation may be affected by the limitation of the quality of the digitized film data (such as saturation seen at the center of flare ribbons), (2) the measurement does not take into account the reconnected flux associated with any open fields, and (3) most importantly, our method is only applicable if the event configuration can be approximated by a 2D-like picture.

On the other hand, it is noteworthy that the event exhibits no clear evidence of the eruption of the PIL filament (Figure 2(b)) preceding the event initiation, while possesses many properties (e.g., circular ribbon and drifting surge) that are very similar to the observations and models of M09 and P10. This strongly implies that a 3D reconnection possibly associated with a coronal null point, rather than the standard 2D scenario, could be the flare mechanism of this event. We make such considerations further in the following.

First, the presence of a coronal null point in this event could already be well indicated by the circular shape of the outer flare ribbon k2 (M09). The fan–spine magnetic topology may be envisioned by the result of potential field extrapolation performed using the NSO/KP magnetogram at March 17 15:20 UT (see Figure 4(a)), in which red (respectively blue) lines denote closed (respectively open) fields. The expected dome-shaped magnetic structure is also discernible by the black surges observed in Figure 3(o). Similar to M09, we thus link the circular ribbon k2 and the inner ribbon k1 to the chromospheric mapping of the fan and the inner spine field lines (also cf. Figure 1), respectively. Another notable observation is the slim surge right before this event (Figures 2(e) and 3(a)). Its root is close to the northeast corner of the PIL, and the entire surge seems to lie directly above the circular source region (cf. Figures 2(b) and (e)). These are obviously distinct from the later surge that are cospatial with the circular ribbon (Figures 3(c)–(k)). We speculate that (1) this slim surge could be the precursor of this event tracing the outer spine field lines, and that (2) the coronal null point might be near the northeast resulting in a non-axisymmetric fan–spine magnetic configuration. The latter may be implied by the asymmetric surface field distribution within the circle (Figure 2(a)), and has been shown to be a favorable condition for the 3D reconnection (M09; P10).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Second, the clearly observed sequential brightening of the fan ribbon k2 and the elongated shape of the inner spine ribbon k1 resemble the event of M09 well, and would strongly suggest the existence of extended QSLs, in which the null point and the related fan–spine structure are embedded. In the model of M09, field lines within the QSLs first undergo slipping and slip-running reconnection toward the null point before the occurrence of the null-point reconnection. In a second phase, field lines encounter slip-running and slipping reconnection as they slip away from the null. These modeled results agree with the ordered motion of the fan and spine ribbons in their event observed in TRACE 1600 Å. We note that such UV emission has a slow decay due to slow cooling or gradual heating not related to the non-thermal electrons (Qiu et al. 2010). In contrast, blue wing Hα images may better represent the precipitation of energetic electrons in the lower atmosphere (e.g., Lee et al. 2006) and hence may better trace the motion of flare ribbons/kernels. Besides the clearly observed sequential brightening of k2, a prominent finding of our observations is thus that the inner spine ribbon/kernel k1 first moves northeast presumably toward the null point and then reverts back. These observations may then well reflect the different phases of the QSL reconnection involving the succession of the slipping and slip-running reconnection. It is pertinent to point out that they are not the same field lines that slip reconnect through the QSL before and after the null-point reconnection (M09). Moreover, the intense surge, which should be resulted from the null-point reconnection involving open outer spine field line (P10), occurs after k1 reaches the northern turning point. This could be consistent with the view that the eruptive null-point reconnection can only be triggered when the system has stored a sufficient amount of energy (M09). The measured average velocities of k1 and k2 are comparable with or larger than the estimated slippage velocity of 30 km s⁻¹ in the model of M09. It is also worthwhile to mention that different from the confined event of M09 where the null-point reconnection is related to remote brightenings, our eruptive event exhibits jet activities as the signature of null-point reconnection as in P10.

This δ region in a circular shape produced three other flares in about one day, including an X1 (event 2), an M5 (event 3), and an M1 (event 4) flare. These three events show similar characteristics in the following aspects: (1) they all have a circular ribbon that encompasses another smaller elongated ribbon, and the ribbons evolve in nearly the identical regions. (2) In all the events, the circular ribbon brightens sequentially in a clockwise fashion, in tandem with the cospatial surge (except event 4); meanwhile, the inner ribbon shows a similar "round-trip" motion as elaborated in Section 3.1. It is however noticeable that these ribbon dynamics are most pronouncedly observed in event 1 with the longest flare duration (see Table 1). These properties thus connote that events 1–4 constitute homologous flares and surges (while event 4 shows insignificant surge activity probably consistent with its lowest flare magnitude), and that they could be associated with similar magnetic reconnection processes.

Nevertheless, different from event 1, a conspicuous common feature of events 2–4 is the existence of a third elongated flare ribbon brightening in the eastern plage region of negative polarity (same as the central parasitic field), which is about 120'' apart from the main circular flaring site (see Figures 4(c)–(h) and Figures 5(a) and (c)). Remote brightenings are often seen in large eruptive flares, and have been understood as the result of the direct heating by hot electrons at relativistic speed traveling along large-scale, closed magnetic loops connecting the main flare to remote sites (e.g., Liu et al. 2006 and references therein). This is most likely the mechanism for our event, in which the remote flare ribbon brightened almost simultaneously, with a minor trend of drifting motion toward the south. It is unlikely that these brightening patches are the sequential chromospheric brightenings observed to propagate away from the flare site (Balasubramaniam et al. 2005). The present remote ribbons also do not bear resemblance to those described in Wang (2005), where the remote brightenings were interpreted as the interaction between the erupting fields and the nearby larger scale fields. In view of the analogous study of M09 and P10, it is therefore appealing to conjecture that some outer spine fields dip down to the remote brightening region, while some open upward related to the surges, and that both the remote brightening and the surge are due to the null-point reconnection involving the outer spine field lines. The corresponding magnetic structure as sketched in Figure 1 based on our observations, which can also be envisioned from the extrapolated fields (Figure 4(a)), is hence a combination of those depicted in M09 and P10.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Another persuasive argument for the close linkage between the remote flare ribbon and the main circular flare lies in the fact that the remote brightenings in events 2–4, as manifested by the peaks of $\dot{\phi }_{-}$, are all co-temporal with a second phase of flare energy release signaled by a separate peak in the time derivative of the SXR flux. This is clearly evidenced in Figures 4(b) and 5(b) and (d). We are convinced that the peaks of $\dot{\phi }_{-}$ are caused by the remote brightenings, as those peaks would disappear if only the inner spine ribbons in the negative field region are included. We also note that there is a ∼1 minute lag between the peak of the remote brightenings and the main energy release of flares, the latter of which is reflected by the earlier peaks in the SXR flux derivative and $\dot{\phi }$. As implied by M09 and elaborated in Reid et al. (2012), we tend to interpret this time delay as the time related to the slipping of the magnetic fields within the QSLs toward the null point. Other possibilities for the delay include that (1) it is the time required to store enough energy in the system to trigger an impulsive surge ejection, and (2) since the outer spine could be initially open, which is noticeable from the magnetic field extrapolation of the active region (Figure 4(a)), the delay might be due to the time required for the outer spine to be close down to the photosphere to produce remote brightenings. A more intricate issue is posed by the presence of both the remote brightenings and surges in events 2 and 3, in which the peak of the remote brightenings appears to coincide with the start of surges as seen in the time-lapse movies. We surmise that this may indicate a change of magnetic topology quasi-instantaneously. Specifically, our observations could be the first ones that can be interpreted as the signature of the opening and closing of the outer spine in a null-point topology.

This flare is associated with the parasitic negative field in the leading positive plage region of NOAA AR 6509 on 1991 February 25 (NSO/KP magnetogram was not available for this day). The findings of the filament material lying along the circular PIL and the overall sequential brightening of the fan-related circular flare ribbon, as shown in the Figure 6 (upper panels) and the movie, are similar to those of events 1–4. It is nevertheless obvious that the inner spine ribbon remains compact throughout the event period. According to Hα images, this circular ribbon appears more circular than those in the other events, which suggests that the fan–spine topology is more symmetric in this event. Therefore, the QSL around the inner spine is expected to be less squashed, which implies that the inner spine ribbon should show a more compact form. Another notable feature of this event is the well-separated phases of the circular ribbon flare (from ∼16:49–17:10 UT) and the surge (after ∼17:10 UT). This may provide another example in which the slipping/slip-running reconnection and the null-point reconnection may occur successively in a single event (M09). As a side note, an additional brightening is discernible from ∼17:17:29 (see the animation) during the phase of surge in a region close to the main sunspots. However, the timing of the brightening seems not to be related to significant peaks in X-rays (cf. events 2–4).

Zoom In Zoom Out Reset image size

We use the available BBSO VMG observations to study the associated magnetic field evolution. From Figure 6 (lower left panels), one can see that the central negative field is broken into several fragments, which move toward the PIL and cancel with the nearby positive field. Since the positive flux covers a much larger area, it is hard to calculate the small amount of change due to flux cancellation. Instead, we plot the total negative magnetic flux as a function of time in Figure 6 (lower right panel). The result shows that from 16:10 to 17:40 UT during which event 5 occurred, magnetic flux decreases rapidly by an amount of ∼10²⁰ Mx. Assuming a depth of magnetic energy dissipation d = 1000 km and the mean flux density in the calculation box B = 500 G, the magnetic energy loss E = BFd/4π is estimated to be on the order of 5 × 10²⁹ erg. The flux cancellation across the circular PIL could be closely related to the triggering and progress of this flare related to the 3D null-point reconnection. Using a topological model, Fletcher et al. (2001) demonstrated that such a flux cancellation may cause rapid coronal field restructuring by reconnection at the null, which produces heating and subsequent material ejection from multiple locations along the footprint of the fan in the lower atmosphere. Our event showing a dynamic nature of brightenings and jetting activities may also fit into this picture. However, the QSL reconnection should still play a central role in explaining the sequential brightening along the circular ribbon.