Precise near-infrared photometry, accounting for precipitable water vapour at SPECULOOS Southern Observatory

Abstract
The variability induced by precipitable water vapour (PWV) can heavily affect the accuracy of time-series photometric measurements gathered from the ground, especially in the near-infrared. We present here a novel method of modelling and mitigating this variability, as well as open-sourcing the developed tool – Umbrella. In this study, we evaluate the extent to which the photometry in three common bandpasses (r′, i′, z′), and SPECULOOS’ primary bandpass (I+z′), are photometrically affected by PWV variability. In this selection of bandpasses, the I+z′ bandpass was found to be most sensitive to PWV variability, followed by z′, i′, and r′. The correction was evaluated on global light curves of nearby late M- and L-type stars observed by SPECULOOS’ Southern Observatory (SSO) with the I+z′ bandpass, using PWV measurements from the LHATPRO and local temperature/humidity sensors. A median reduction in RMS of 1.1 per cent was observed for variability shorter than the expected transit duration for SSO’s targets. On timescales longer than the expected transit duration, where long-term variability may be induced, a median reduction in RMS of 53.8 per cent was observed for the same method of correction.