Sub-second cadence structure of optical flares on AD Leo

Abstract

Stellar flares are sudden brightenings caused by magnetic reconnection and are frequently observed on late-type stars. High-cadence photometry of flares provides valuable insights into the mechanisms of these events, yet such observations remain scarce. We seek to explore the sub-second fine structure of stellar flares and assess the information content in high-speed photometry. New 0.3 s-cadence photometry from a six-year-long observing campaign of the active M-dwarf AD Leo is presented. We use time--frequency analysis to detect quasi-periodic pulsations in the decay phase of flares. We explore statistical measures of time series complexity of the detected flares to quantify the information gain achievable with high-cadence photometry. We detect 42 flares in 211 hours of observations. The flare frequency distribution is consistent with the previous literature. We find no quasi-periodic pulsations with periods below a few seconds, and identify two candidate signals with periods around 1 and 3 min. Using different measures of complexity on the binned flare light curves we confirm the advantages of high observing cadence. However, we also find a plateau up to a binning of ~4--5 s for a few complex flares, suggesting that an exposure time of a few seconds is usually enough to retain most of the information carried by a single-filter observation. New photometric observations of AD Leo revealed sub-structures of flare light curves on the timescale of a few seconds, but we found no features on timescales below that.

Figures (8)

• Figure 1: Complexity assessment with generalized additive models (GAMs) for two flares. Top: generalized cross-validation score as a function of the number of spline terms, with the red point marking the minimum. Bottom: flares with the best-fitting curve corresponding to this minimum; associated spline values are indicated in the panels.
• Figure 2: The flare frequency distribution derived in this work (grey data points and corresponding best-fit line), compared to fitted trends reported in the literature.
• Figure 3: Light curves of flares exhibiting complex temporal structures are shown as grey points, with the corresponding smoothed curves indicated by red lines. The first column shows the full flare profiles, while the second column provides magnified views of the peaks, illustrating the two fitted flare components with dashed lines. The HJD indicated in each panel of the first column corresponds to t = 0 on the time axis.
• Figure 4: Time--frequency analysis of the flare at $\mathrm{HJD}=2458564.51777$. Red line in the upper right panel shows the smoothed light curve that was subtracted before the application of the windowed Lomb--Scargle periodogram. Possible quasi-periodic pulsation is present with a period around 1 min (around $t=6-8$ min).
• Figure 5: Same as Fig. \ref{['fig:stft_1']} for the flare at $\mathrm{HJD}=2458567.44626$. Possible quasi-periodic pulsation is present with a period around 3 min.