Optimising spectroscopic observations of transiting exoplanets

Abstract

When observing the atmospheres of transiting exoplanets using high-resolution spectroscopy, one aims to detect well-resolved spectral features with high signal-to-noise ratios (SNR) as is possible today with modern spectrographs. However, obtaining such high-quality observations comes with a trade-off: a lower cadence of fewer, longer exposures across the transit collects more photons thanks to reduced overheads, enhancing the SNR of each observation, while a higher cadence of several, shorter exposures minimises spectral feature smearing due to the continuously changing radial velocity of the planet. Considering that maximising SNR and minimising smearing are both beneficial to analysis, there is a need to establish where the optimal compromise lies. In this work, we model real transit events based on targets as they would be observed with VLT/CRIRES+ at Paranal Observatory. Creating four hypothetical scenarios, we simulate each observation across 100 realisations of the same transit event in order to vary the time resolution only. We remove telluric and stellar lines using the SYSREM algorithm and analyse them through cross-correlation with model templates, measuring how successfully each time resolution and case detects the planetary signal. We demonstrate that there is a continuous change in the detection significance based on time resolutions, and that the function of this significance has clear maxima. The strength and location of this maxima varies on e.g. planet system parameters, instrumentation, and no. of removal iterations. We discuss why observers should therefore take several factors into account, using a strategy akin to the 'exposure triangle' from traditional photography where a balance must be struck by considering the full context of the observation. Our method is robust and may be employed by observers to estimate best observational strategies for other targets.