NICMOS transmission spectroscopy of HD 189733b: controversy becomes confirmation

TL;DR

This paper defends the original detection of methane and water features in HD 189733b's transmission spectrum, showing that reanalysis by Gibson et al. was flawed due to instrument modeling issues, thus confirming the initial findings.

Contribution

The study demonstrates that the original spectral features are valid and clarifies the limitations of the Gibson reanalysis method, reaffirming the molecular detections.

Findings

Gibson et al.'s reanalysis agrees with original results for HD 189733b.

Poor instrument modeling causes uncertainties and instability in spectral analysis.

Gibson's method is flawed for XO-1b due to omitted parameters.

Abstract

Spectral features corresponding to methane and water opacity were reported based on spectroscopic observations of HD 189733b with Hubble/NICMOS. Recently, these data, and other NICMOS exoplanet spectroscopy measurements, have been reexamined in Gibson et al. 2010, who claim that the features in the transmission spectra are due to uncorrected systematic errors and not molecular opacities. We examine the methods used by the Gibson team and show that, contrary to their claim, their results for the transmission spectrum of HD 189733b are in fact in agreement with the original results. In the case of HD 189733b, the most significant problem with the Gibson approach is a poorly determined instrument model, which causes (1) an increase in the formal uncertainty and (2) instability in the minimization process; although Gibson et al. do recover the correct spectrum, they cannot identify it due to the problems caused by a poorly determined instrument model. In the case of XO-1b, the Gibson method is fundamentally flawed because they omit the most important parameters from the instrument model. For HD 189733b, the Gibson team did not omit these parameters, which explains why they are able to reproduce previous results in this case, although with poor SNR.