G. J. D. Petrie
National Solar Observatory, Tucson, AZ, 85719, USA
The eruption and flaring of intense solar magnetic fields has been a central topic of solar physics since Carrington’s first reported white-light observation using a small telescope in 1859, and G.E. Hale established that sunspots are magnetic in 1908. Flares have long been known to be capable of causing havoc for space missions and, as in Carrington’s example, can have painful and expensive consequences on Earth. In recent years evidence has mounted indicating that flares often restructure the solar surface magnetic field. Hudson and coworkers conjectured that the exit of energy from the flaring region in the form of motion, light, heat and magnetic flux, should result in a reduction of magnetic energy in the region due to the flare, which should cause the loop structures to collapse like deflated balloons2. They estimated that the effects of this collapse should be seen in the surface field<sup2. But the great density of the solar surface fluid persuaded many people that even a powerful flare could not move it significantly, some comparing this scenario to a tail wagging a dog.
HMI surface vector field measurements offer a unique opportunity to revisit this topic, and to understand flares better. HMI samples the surface vector magnetic field every 12 minutes, allowing a direct comparison of the field structure before and after a flare. We can also see whether these changes stand out from the ongoing turbulent convection that is always happening.
Fig 1 shows the surface vector magnetic field immediately before the famous ‘Valentines’s Day’ X-class flare (02/15/2011). The box labelled ‘NL’ indicates the main neutral line, the boundary between positive and negative vertical field, where the most interesting action occurred. The atmospheric loop fields that cross the neutral line must have been highly twisted or sheared as indicated by the arrows pointing almost parallel to the neutral line. At each end is a sunspot, a positive (‘PS’) and a negative (’NS’) one, with clockwise and anti-clockwise field circulation, respectively. Observations have shown that a twisted flux tube crossed the neutral line, connecting the two sunspots.
Fig 1 | Surface vector magnetic field before the flare. The vertical field component is indicated by the color scale and the horizontal by the arrows. Red/blue represents positive/negative vertical field. The rectangle ’NL’ marks the main neutral line. Solid and dotted contours indicate strong (>1000 G) and quite strong (>100 G) fields, respectively.‘PS’ and ’NS’ contours indicate sunspots.
Fig 2 shows the surface vector field changes, the difference between the pre-flare and post-flare fields. The two sides of the neutral line have swapped colors compared to Fig 1 implying decreased vertical field strength on both sides. The horizontal changes at the neutral line represent an increase in horizontal field strength because the arrows in Fig 2 point in roughly the same direction as the arrows in Fig 1. The twist in the two sunspots decreased during the flare: their arrows in Fig 2 have opposite circulation to the corresponding arrows in Fig 1.
Fig 2 | Vector magnetic field changes during the flare. The vertical field change component is indicated by the color scale and the horizontal component by the arrows. Red/blue coloring represents positive/negative vertical field changes. The rectangle, contours and labels are as in Fig 1.
Major flare-related Lorentz (magnetic) force changes can also be estimated from surface measurements1,2,5. The surface Lorentz force changes are shown in Fig 3. Here the domination of downward vertical Lorentz forces is consistent with the loop-collapse scenario 2, and the horizontal changes at the neutral line suggest contraction: the positive (lower in the plot) side is pulled left along the horizontal field direction and the negative (upper) side is pulled right towards the positive side. The horizontal force changes circulate against the field circulation in the positive sunspot and with the field circulation in the negative sunspot, consistent with a relaxation of twist.
Fig 3. Lorentz force vector changes during each flare. The vertical force change component is indicated by the color scale and the horizontal components by the arrows. Red/blue coloring represents positive/negative (upward/downward) Lorentz force change. The rectangle, contours and labels are as in Fig 1.
The field and force changes form a coherent pattern of a collapsing, untwisting flux tube, and they stand out clearly against the background field evolution3,4. (See also the related work by Haimin Wang’s group at NJIT.) Not only can the tail wag the dog, but analysis of HMI data helps us to understand how.
References:
[1] Fisher, G.H., Bercik, D.J., Welsch, B.T., Hudson, H.S. 2012, Solar Phys. 277, 59
[2] Hudson, H.S., Fisher, G.H., Welsch, B.T. 2008, ASP Conference Series, 383, 221
[3] Petrie, G.J.D. 2012, ApJ 759, 50
[4] Petrie, G.J.D. 2013, Solar Phys. 287, 415
[5] Petrie, G.J.D. 2014, Solar Phys. submitted, http://adsabs.harvard.edu/abs/2014arXiv1403.6156P
Gordon, nice work! I am wondering whether you checked, during the period of no flares, if vertical and horizontal magnetic fields had changes of similar magnitude? That is, how can we be sure that these observed magnetic field changes are truly associated with the flare instead of regular changes an evolving active region often has?
Thanks! The field is changing all the time, of course, responding to the photospheric flows, so how is a flare-related change identified? A case can be made if it occurred at the same place and at the same time as the flare, if the change was significantly larger than the background variations, and if it occurred abruptly and left a permanent change in the field components. Here “abrupt” and “permanent” mean that the field value transitioned in a short time during the flare, say 1/2 hour or less, and the change lasted several hours or more. In past work with GONG data we fitted parametrized arctan step functions to individual pixel light curves to pick out the permanent changes and study them. With HMI we can integrate field components across areas of interest and plot the changes in time. If one does this, the flare-related changes really stand out from the background variations, appearing as steps. At times and places of no flaring, we do not find such stepwise changes. Examples are given in the references.
Thanks for your detailed explanation.