Robustness measures for molecular detections using high-resolution transmission spectroscopy of exoplanets

Abstract

Ground-based high-resolution transmission spectroscopy has emerged as a promising technique for detecting chemicals in transiting exoplanetary atmospheres. Despite chemical inferences in several exoplanets and previous robustness studies, a robust and consistent detrending method to remove telluric and stellar features from transmission spectra has yet to be agreed upon. In this work we investigate the robustness of metrics used to optimise PCA-based detrending for high-resolution transmission spectra of exoplanets in the near-infrared. As a case study, we consider observations of the hot Jupiter HD 189733 b obtained using the CARMENES spectrograph on the 3.5 m CAHA telescope. We confirm that optimising the detrending parameters to maximise the S/N of a cross-correlation signal in the presence of noise has the potential to bias the detection significance at the planetary velocity of optimisation. However, we find that optimisation using the difference between a signal-injected cross-correlation function and the direct cross-correlation function is more robust against over-optimisation of noise and spurious signals. We additionally examine the robustness of weighting the contribution of each order to the final CCF, and of S/N calculations. Using a prescribed robust methodology, we confirm H2O in the atmosphere of HD 189733 b (S/N = 6.1). We then investigate two further case studies, of exoplanets HD 209458 b and WASP-76 b, confirming OH in the atmosphere of WASP-76 b (S/N = 4.7), and demonstrating how non-robust methods may induce false positive or inflated detections. Our findings pave the way towards a robust framework for homogeneous characterisation of exoplanetary atmospheres using high-resolution transmission spectroscopy in the near-infrared.

JWST Transmission Spectroscopy of HD 209458b: A Supersolar Metallicity, a Very Low C/O, and No Evidence of CH4, HCN, or C2H2

“…Studies on the robustness of molecular detections by HRS of exoplanets demonstrate that some detrending methods may induce false positive or inflated detections (Cheverall et al 2023). In their case study on HD 209458b, HCN, NH 3 , and CH 4 were not observed by Cheverall et al (2023). However, a signal for CH 4 can be detected if the telluric contamination is not correctly removed.…”

A Combined Ground-based and JWST Atmospheric Retrieval Analysis: Both IGRINS and NIRSpec Agree that the Atmosphere of WASP-77A b Is Metal-poor

“…For such a more finely tuned approach to selecting the number of PCs to remove, it would be appropriate to develop a quantitative metric to indicate how many is "enough." This is not trivial; CCF signals can be spuriously overoptimized to particular models or velocities (see, e.g., Cabot et al 2019;Cheverall et al 2023), and searching for correlations between the LSVs and physical components of the data requires a comprehensive record of such components. Air mass and humidity appear to be prevalent quantities captured by the SVD in these data, but information about, e.g., the seeing or dew point was not available, yet it may still impact the data.…”

A Formally Motivated Retrieval Framework Applied to the High-resolution Transmission Spectrum of HD 189733 b

“…6 This data set has been previously used to successfully detect water vapor and to infer the hazy nature of the atmosphere of this hot Jupiter (Alonso-Floriano et al 2019;Sánchez-López et al 2019. In addition, it was used by Cheverall et al (2023) as a case study for evaluating different preparing methods for CCF analysis. We used the data from the NIR channel's fiber A (focused on the target star HD 189733), which covers the spectral range 0.96-1.71 μm in 28 echelle orders at a spectral resolution of »  80,400.…”

“…The preparing steps are often not described exhaustively, sometimes making reproducibility and comparisons difficult. This need for a consistent, reproducible, and robust preparation technique has already been raised (Cabot et al 2018;Cheverall et al 2023). Similarly, how the likelihood function terms are modified can vary significantly between authors-especially regarding the modification of the forward models.…”