Ruizhu Chen and Junwei Zhao
W. W. Hansen Experimental Physics Laboratory, Stanford University, Stanford, CA 94305-4085
Sunquakes are helioseismic power enhancements initiated by solar flares, exhibiting as photospheric helioseismic waves expanding from the flaring sites1 visible about 10-30 min after the flare onset. But the occurrence of sunquakes is relatively rare — only a fraction of strong solar flares are accompanied by detectable sunquakes, and weak flares cause sunquakes with an even lower rate2. It is curious why some flares cause sunquakes while others do not. Mechanisms that are proposed to trigger sunquakes include sudden photospheric pressure perturbations caused by shock waves, photospheric heating, or Alfvén-wave heating; or by Lorentz force changes in the photosphere due to permanent changes in the photospheric magnetic field. However, none of these proposed mechanisms satisfactorily explain why only a small fraction of flares cause sunquakes.
Here we propose a hypothesis to explain the disproportionate occurrence of sunquakes considering the phosphoric background oscillations as a possible selection rule: during a flare’s impulsive phase when the flare’s impact acts upon the photosphere, delivered by shock waves, energetic particles from higher atmosphere, or by downward Lorentz Force, a sunquake tends to occur if the background oscillation at the flare footpoints happens to oscillate downward, in the same direction with the impact from above. To verify this hypothesis, we perform a systematic survey on 60 strong flares in Solar Cycle 24 observed by SDO/HMI, and examine the oscillatory velocity at the sunquake sources during the flares’ impulsive phases. Because the direct Doppler-velocity observations in the flaring regions are often corrupted during the flares’ impulsive phases due to the high temperature and violent dynamics, we reconstruct velocity fields at the flaring sites during the flares by utilizing a helioseismic holography method3. In particular, we use an observation-based Green’s function derived by the time-distance heliosiesmic method for a more precise temporal determination of the sunquake events.
Figure 1| Egression power maps of the reconstructed oscillatory velocities for the X1.8 flare on 2012 October 23, in (a) 3 − 5 mHz and (b) 5 − 7 mHz. Vertical colorbars show absolute values of the power, and horizontal colorbars show the ratios to the quiet-Sun mean values. Red and yellow contours show the 3σ levels of background oscillations before the flare in quiet-sun regions and active regions, respectively. White dashed contours show the 1σ level of quiet-Sun background.
A sunquake is recognized by a certain criterion on the egression power P, which is the oscillation power of the reconstructed (radial) velocity field. Figure 1 shows an example of such power maps for the X1.8 flare on 2012 October 23, in the low-frequency (3-5mHz) and high-frequency (5-7 mHz) bands separately. A flare is labeled as sunquake active if there are oscillations detected with power larger than 3σ (99.7 percentile) of background levels, sizes larger than 25 pixels, and close to or overlapping a flare ribbon during the flaring time, in at least one of the frequency bands. All the sunquakes that are detected in this study are also confirmed by time—space diagrams. For flares whose reconstructed oscillatory power is not compact, multiple sunquakes are counted if two or more oscillation kernels, clearly apart from each other, are associated with different flare ribbons or correspond to different (visible) sunquake ripples. A total of 24 flares, out of 60 flares surveyed, are found to be helioseismic active, giving a total of 41 sunquakes. These sunquakes are highlighted on their respective line-of-sight magnetic field maps shown in Figure 2.
Figure 2| Line-of-sight magnetic field maps of all the 24 active regions that host sunquake- generating flares. Green and blue contours show the 3σ levels of the sunquake egression power in active regions in 3 − 5 mHz and 5 − 7 mHz, respectively. The ‘+’ marks denote the centroids of oscillation cores for each sunquake. A number is marked beside the ‘+’ sign when more than one sunquake source is identified in the active region.
To examine the temporal relation between the sunquakes and the flares, we compare the temporal evolutions of the reconstructed oscillatory velocities, oscillatory power, and line-core intensity (as signatures of the flares), all of which are obtained from averaging these quantities inside the sunquake oscillation cores. Figure 3 shows these time curves for the above-mentioned sunquake example. Significant oscillation signals are recovered around the line-core intensity peak time; however, the sunquake initial impulse is not resolved, limited by the temporal resolution due to the finite frequency bandwidth. But we can still examine the direction of the sunquake velocity during the flare’s impulsive phase, defined as the time when the locally averaged line-core intensity curve is between 1/e and 1 time of its peak value. In Figure 3, the oscillatory velocities for both frequencies are positive (downward) during the flare’s impulsive phases (indicated by magenta dashed lines). Throughout the survey, it is found that in the 3−5 mHz frequency band, 25 out of 31 (80.6%) sunquakes have net downward (positive) velocities; in the 5 − 7 mHz frequency band, 33 out of 38 (86.8%) sunquakes have net downward velocities. These roughly agree with the hypothesis proposed in this study.
Figure 3| Temporal evolutions of the oscillatory velocity, oscillatory power P, and line-core intensity integrated in the oscillation cores of the sunquake following the X1.8 flare on 2012 October 23. The curves are for (a) 3 − 5 mHz, (b) 3 − 5 mHz zoomed in, (c) 5 − 7 mHz, and (d) 5 − 7 mHz zoomed in. Between magenta dashed lines in each panel is the flare’s impulsive phase, defined as the time when the locally averaged line-core intensity curve is between 1/e and 1 time of its peak value. The oscillatory velocities of this sunquake for both frequencies are positive (downward) during the flare’s impulsive phase.
We also examine the spatial relation between the sunquakes and flares, as well as the relationship between the occurrence of sunquakes, Doppler transients, and HMI continuum-intensity. A majority of sunquakes occur on or near the polarity inversion lines; and the sunquake sources also tend to be located between or at the edge of the flare ribbons denoted by the HMI line-core intensity enhancements. All the sunquake-generating flares also have both continuum-intensity enhancements and Doppler transients in our survey, but flares with Doppler transients or continuum emissions do not necessarily generate sunquakes.
For more details of this study, please refer to .
 Kosovichev, A. G., & Zharkova, V. V. 1998, Nature, 393, 317
 Sharykin, I. N., & Kosovichev, A. G. 2020, ApJ, 895, 76
 Lindsey, C., & Braun, D. C. 2000, SoPh, 192, 261
 Chen, R. and Zhao, J. 2021, ApJ, 908, 182