Detecting molecules in Ariel Tier 1 transmission spectra

Bocchieri, Andrea, Mugnai, Lorenzo V. ORCID: https://orcid.org/0000-0002-9007-9802, Pascale, Enzo, Changeat, Quentin and Tinetti, Giovanna 2022. Detecting molecules in Ariel Tier 1 transmission spectra. Presented at: Europlanet Science Congress 2022, 18–23 Sep 2022. EPSC Abstracts, Europlanet Science Congress 2022. , vol.16 10.5194/epsc2022-416

Full text not available from this repository.

Abstract

The Ariel Space Mission will observe a large and diverse sample of exoplanetary atmospheres in the 0.5 to 7.8-micron range of the electromagnetic spectrum. As part of the Ariel observing programme, a shallow Reconnaissance survey (Tier 1) will provide transiting and eclipse spectroscopy on about 1000 targets, with low spectral resolution but sufficient SNR to identify the signature of molecular species. The wealth of information provided by this survey will be the basis for promoting targets for re-observation to reach sufficient SNR at higher spectral resolution. At the same time, these low spectral resolution observations are not suitable for estimating molecular abundances with an appropriate confidence level. Therefore, it is paramount to develop special data analysis techniques to extract their information content. This work investigates using the abundance posteriors from spectral retrieval as an unbiased metric to assess the presence of a molecule up to a certain threshold. The experimental dataset comprises simulated Tier 1 transmission spectra for about 300 targets from the Ariel Mission Reference Sample produced using the Alfnoor software. We use the TauRex 3 retrieval framework to run spectral retrievals on each “observed” spectrum, and we compute the probability that the spectra bear a molecule by integrating the posteriors above a specified threshold of molecular concentration. We find that the retrieved probabilities correlate with the abundances in the forward models and that this method is statistically reliable and has considerable predictive power and diagnostic ability. The predictive power is not significantly affected by adding molecules in the fitted composition that are not present in the forward models, while omitting molecules should be discouraged as it can lead to biased results.

Item Type:	Conference or Workshop Item (Paper)
Status:	Published
Schools:	Physics and Astronomy
Last Modified:	01 Jun 2023 15:15
URI:	https://orca.cardiff.ac.uk/id/eprint/159655

Actions (repository staff only)

Edit Item