Activity-induced radial velocity jitter in a flaring M dwarf

We investigate the effect of stellar activity and flares on short-term radial velocity measurements in the mid-M flare star CN Leo. Radial velocity variations are calculated from 181 UVES spectra obtained during three nights. We searched for spectral orders that contain very few atmospheric absorption lines and calibrated them against the telluric A-band from O$_2$ in the Earth’s atmosphere. One giant flare occurred during our observations, which has a very strong effect on radial velocity. The apparent radial velocity shift due to the flare is several hundred m s$^{-1}$ and clearly correlated with H$\alpha$ emission. Outside the flare, only spectral orders containing the most prominent emission lines of H, He, and Ca show a correlation to chromospheric activity together with a radial velocity jitter exceeding a few 10 m s$^{-1}$. We identify a number of spectral orders that are free of strong emission lines and show no flaring-related radial velocity jitter, although flares occurred as strong as 0.4 dex in normalized H$\alpha$ luminosity. The mean radial velocity jitter due to moderate flaring is less than 10 m s$^{-1}$. Strong flares are easily recognized directly in the spectra and should be neglected for planet searches.

💡 Research Summary

The paper presents a detailed investigation of how stellar activity, particularly flares, influences short‑term radial‑velocity (RV) measurements of the mid‑M dwarf CN Leo (Gliese 406). Using the UVES spectrograph on the VLT, the authors obtained 181 high‑resolution spectra over three nights. To isolate the stellar RV signal from instrumental drifts and telluric contamination, they identified spectral orders that contain very few atmospheric absorption features and calibrated each order against the terrestrial O₂ A‑band near 760 nm. This approach provides a robust wavelength reference that is independent of the star’s own line variability.

During the observing run a single, very energetic flare was captured. The flare produced strong emission in Hα, He I, Ca II and several other chromospheric lines. In the spectra taken during the flare the apparent RV shifted by several hundred metres per second. The authors demonstrate a clear correlation between the magnitude of the RV shift and the Hα equivalent width, indicating that the flare’s radiative output directly perturbs the measured line centroids. The physical origin of the shift is likely a combination of line‑profile asymmetries caused by rapid chromospheric heating, line‑broadening from turbulent motions, and a change in the overall spectral energy distribution that alters the weighting of the cross‑correlation function used for RV extraction.

Outside the flare, the authors examined the RV stability of each spectral order. Orders that contain the strongest emission lines (H, He, Ca) still show a modest correlation with chromospheric activity indicators, leading to RV jitter of a few tens of metres per second. In contrast, orders dominated by photospheric metal absorption lines and free of strong emission features exhibit remarkably low jitter, typically below 10 m s⁻¹, even when the star experiences moderate flaring (up to 0.4 dex increase in normalized Hα luminosity). This demonstrates that the choice of spectral orders is crucial for minimizing activity‑induced noise in RV surveys of active M dwarfs.

The authors further quantify the mean RV jitter associated with moderate flares as <10 m s⁻¹ when emission‑free orders are used. They argue that strong flares are readily identifiable in the spectra because of their conspicuous emission lines, and therefore can be excluded from planet‑search data sets without loss of valuable observing time. By applying a simple line‑mask or by discarding orders that contain variable emission, one can retain the bulk of the data while achieving RV precision comparable to that obtained for quieter, earlier‑type stars.

The study’s implications are significant for the ongoing search for low‑mass exoplanets around late‑type stars. M dwarfs are attractive targets because their low masses amplify the RV signal of Earth‑mass planets, but their high levels of magnetic activity have traditionally limited RV precision. This work shows that, with careful order selection and flare identification, it is possible to reach sub‑10 m s⁻¹ precision even on a highly active flare star. Such precision is sufficient to detect super‑Earths in short‑period orbits and, when combined with transit surveys, to confirm planetary candidates around the most common stars in the Galaxy.

In summary, the paper provides a practical methodology for mitigating activity‑induced RV jitter in M dwarfs: (1) use telluric O₂ A‑band calibration to anchor the wavelength solution, (2) select spectral orders free of strong chromospheric emission, and (3) flag and discard spectra taken during obvious flares. By following these steps, future RV programs can include active M dwarfs in their target lists without compromising the detection of low‑mass planets.