Fluctuations and Flares in the Ultraviolet Line Emission of Cool Stars: Implications for Exoplanet Transit Observations

Abstract

Variations in stellar flux can potentially overwhelm the photometric signal of a transiting planet. Such variability has not previously been well-characterized in the ultraviolet lines used to probe the inflated atmospheres surrounding hot Jupiters. Therefore, we surveyed 38 F-M stars for intensity variations in four narrow spectroscopic bands: two enclosing strong lines from species known to inhabit hot Jupiter atmospheres, CII $\lambda\lambda$1334,1335 and SiIII $\lambda$1206; one enclosing SiIV $\lambda\lambda$1393,1402; and 36.5 angstroms of interspersed continuum. For each star/band combination, we generated 60 s cadence lightcurves from archival HST COS and STIS time-tagged photon data. Within these lightcurves, we characterized flares and stochastic fluctuations as separate forms of variability. Flares: We used a cross-correlation approach to detect 116 flares. These events occur in the time-series an average of once per 2.5 h, over 50% last 4 min or less, and most produce the strongest response in SiIV. If the flare occurred during a transit measurement integrated for 60 min, 90/116 would destroy the signal of an Earth, 27/116 Neptune, and 7/116 Jupiter, with the upward bias in flux ranging from 1-109% of quiescent levels. Fluctuations: Photon noise and underlying stellar fluctuations produce scatter in the quiescent data. We model the stellar fluctuations as Gaussian white noise with standard deviation $\sigma_x$. Maximum likelihood values of $\sigma_x$ range from 1-41% for 60 s measurements. These values suggest that many cool stars will only permit a transit detection to high confidence in ultraviolet resonance lines if the radius of the occulting disk is $\gtrsim$ 1 R$_J$. However, for some M dwarfs this limit can be as low as several $R_{Earth}$.