2022MNRAS.516.3701C

The exomoon candidate Kepler-1708 b-i was recently reported using two transits of Kepler data. Supported by a 1 per cent false-positive probability, the candidate is promising but requires follow-up observations to confirm/reject its validity. In this paper, we consider the detectability of the exomoon candidate's transit, most specifically in the next window (2023 March) using the WFC3 instrument aboard the Hubble Space Telescope (HST). Using realistic noise estimates, accounting for the visit-long trends, and propagating the model posteriors derived using the Kepler data, we perform 75 injection-recovery trials with Bayesian model selection. Defining a successful detection as one that meets thresholds of the Bayes factor, Akaike information criterion, and error of the retrieved parameters, only 7 of our 75 injections were recovered when considering HST data alone. This implies a true-positive probability of 10 ± 3 per cent. Despite HST's superior aperture to Kepler, both instrumental systematics and the compactness of the candidate exomoon's orbit typically obfuscate a strong detection. Although the noise properties of the James Webb Space Telescope have not yet been characterized in flight, we estimate the signal would be easily recovered using NIRSpec operating in its Bright Object Time Series mode.