Repeated brightening in a flare loop

Science Nugget: Nov 13, 1998

This week we examine a coronal loop associated with a flare. The flare had a GOES classification of approximately C3, and occurred near the northwest limb at 21:00 UT on 13-Nov. A plot showing the location is shown here, and a raster of images is shown below (click to see a larger version).


In particular, let's look carefully at the loop in the north part of this region, marked by an arrow in the raster image above. During the flare, this loop is seen to brighten up, then fade, then brighten again, fade again.... This can be clearly seen by making light curves of portions of the loop. The light curves shown below are made by measuring the average brightness of small groups of pixels along the loop's length.


The groups of pixels corresponding to the respective light curves are shown in the image below (click to see a larger version).


Region "F" is made very large, and with a strange shape, in order to include all the signal that has spilled out from saturated pixels at the location of the flare. Boxes A, B, C, and D are along the loop in question; box E was chosen simply for sake of comparison.

What is observed in the light curves is a fast brightening in the flare kernel (box F), followed by a quick initial decay and then slower fading. But the more interesting behavior is seen in regions A, B, and C, which can be seen to possess not a single bright peak, but two and maybe three bright humps spaced about 6 minutes apart.

It is interesting to speculate about possible causes for this repeated brightening. One obvious possibility is that there is more than a single flare event in the kernel; the lack of significant subsequent brightenings in light curve F tends to argue against this possibility, although we cannot rule it out entirely.

Other possibilities include Alfvenic/magnetic wave oscillation in the coronal loop, mirroring of trapped energetic particles, or compressive pulses propagating along the loop field lines. The idea of wave oscillation in the loop is based on the interpretation of the repeated brightenings as being due to a disturbance (magnetic or otherwise) being triggered by the flare and then propagating down the length of the loop to the footpoint near box A. The wave would then be reflected at the footpoint, return to the western footpoint (presumably near the flare site), and "bounce" back to region A again. The time between brightenings seen at position A would thus be a function of the wave propagation speed and twice the length along the loop. Since the loop length can be judged from the images to be 30-42 megameters, and the time between brightenings at position A is 340 seconds, this implies a wave speed of 180-250 kilometers per second.

This speed is far below what one expects for Alfven wave speeds in coronal loops. If (and that's a big IF ) this repeated brightening can be interpreted as a "bouncing" wave, then the speed is more consistent with those of the compressive wave trains that EIT has seen in polar plumes (cf. DeForest and Gurman 1998, ApJL, 501, L217) and in loops interconnecting active regions (Berghmans and Clette 1998, Solar Physics, in press). Those observations of brightenings can be interpreted as compressive pulses or waves with propagation speeds of 75-150 km/sec (DeForest and Gurman), and 150-185 km/sec (Berghmans and Clette). However, those observations were of waves or pulses travelling on more or less open field lines, and not resonating in closed systems. It is not clear that this kind of wave would be reflected efficiently enough at the footpoints to enable resonance in the fashion described here.

Nov. 14, 1998: D.E. McKenzie (email