Can EUV Power-Spectral Indices Reveal Imminent Solar Flares?

Nugget
Number:	521
1st Author:	Sihui ZHONG
2nd Author:	Dmitrii KOLOTKOV and Valery M. NAKARIAKOV
Published:	April 6, 2026?
Next Nugget:	TBD
Previous Nugget:	How energetic can solar flares become?

Introduction

As the closest star to Earth, the Sun provides the light and heat necessary for life, but it also poses hazards through solar flares and coronal mass ejections Despite decades of research, existing solar flare forecasting methods, whether physical or empirical, still have limited predictive reliability (Ref. [1]).

Producing a solar flare requires both stored magnetic energy in non-potential plasma structures, such as an unstable flux rope. An instability induced by a trigger mechanism then can produce a flare. To tell whether an active region is about to erupt, we need to track the evolving coronal dynamics within it rather than rely only on the presence of flare-favourable conditions. Tiny bubbles in a kettle precede the eruption of boiling water, for example.

The dynamics of the solar atmosphere can be probed via the time variability of radiation. Across the chromosphere to the inner heliosphere, intensity fluctuations often follow power-law spectra in frequency, S ~ f^-α. The power laws can be single-slope or segmented into a broken power law, with low-frequency α_lf and high-frequency α_hf values. The observed values of α carry information about local coronal dynamics that may be relevant to the flare process (Ref. [2]). Building on these findings, our recent paper (Ref. [3]) investigates the spatial and temporal evolution of the power-law indices of EUV intensity power spectra in the solar corona, to determine whether changes in these indices can serve as flare precursors.

Results

We construct "α maps" from a sequence of full-disk images and study their relation to plasma structures and their dynamics. As shown in Figure 1(a-b), the α_hf map closely resembles the AIA intensity image, with high values appearing in the high-contrast plasma structures, clearly outlining the active region (AR) belts.

Figure 1: Power-law index map (b) for a full-disk image (a) sequence from 08:00 UT to 16:00 UT on 2023 March 29th and time evolution of intensity and α values in three selected regions. The grey shading indicates the time interval of an M1.3 flare.

The temporal evolution of α_lf varies in different regions. It remains nearly steady in non-flaring regions (Figure 1(c-d)), while strong variability is seen in the flaring area (e). To explore whether this behaviour is a common preflare trend, we further analyse 13 flare events and one additional quiescent AR for comparison. We find that the two power-law indices have typical values, α_lf > 1 and α_hf ~ 0, but a reversal of the two indices appears in regions with short-lived brightenings before the flare onset. In addition, some regions show a continuous increase of α_lf before the flare. Based on these signatures, we quantified the temporal and spatial evolution of indices using three proxies: (1) the minimum of α_lf - α_hf defining a reversal when less than -1; (2) Δα_lf, the temporal change of α_lf and(3) the growth rate of α_lf, Δα_lf/Δt.

Figure 2 shows AIA images and proxy maps for two representative cases. In the quiescent AR (top), reversals are almost not seen (except a sporadic jet near one loop footpoint), and both of the proxy maps appear spatially random. In flaring ARs (before flares), however, areas where the proxies exceed a pre-defined threshold form coherent patterns near the flare site. The Δα_lf map resembles the EUV image with distinct loop-like and open-field structures, while the growth-rate map is less obvious due to temporal averaging.

Figure 2: Three power-law index proxies as potential flare precursors in ARs of interest: a nonflaring AR (top row) and a representative C-class flaring AR (bottom row).

We further examined the temporal distribution of two other proxies, min(α_lf-α_hf) and Δα_lf. Because the first of these is generally related to short-lived brightenings in reconnection-indicative morphologies, it can be an indicator of small-scale magnetic reconnection. If unrelated to flares, such events should occur stochastically (Figure 3a, top). Instead, in several cases, they cluster before flare onset, forming distinct peaks at about 30-90 minutes before flares (Figure 3a, bottom) or the occurrence intensifies at a certain time before flares (Figure 3a, middle). In Figure 3(b), instances when Δα_lf exceeds a threshold form a peak near the flare onset, often within the last few minutes. These behaviours were derived from the analysis of 14 flare events. If confirmed in a larger flare sample, these variations may serve as quantitative precursors of flare activity.

&\Delta;α_lf > 0.6 of its maximum occur in the selected events of analysis.

Conclusion

This work investigates the spatial and temporal evolution of power-law indices in coronal EUV intensity power spectra in flare-hosting ARs. The spatial distribution of the power-law index in the low-frequency domain closely mirrors EUV intensity images, indicating that it can reveal the dynamics of coronal plasma structures. Temporally, the low-frequency power law remains stable in quiescent regions, but shows clear variability before flares. Across 14 flare events, notable deviations of α_lf beyond a defined threshold consistently occurred at the flare site within a few minutes before the flare. In some cases, the change in the value of the slope difference was detected within 30-90 minutes before the flare.

This proof-of-concept study suggests that the temporal changes in power-law indices may provide short-term precursors of solar flares, although confirmation with a larger dataset is required.

References

[1] "A Comparison of Flare Forecasting Methods"

[2] "Coronal Fourier Power Spectra: Implications for Coronal Seismology and Coronal Heating"

[3] "Power-law Indices of EUV Intensity Power Spectrum in Flaring Coronal Active Regions"