Longitudinally Resolved Spectral Retrieval (ReSpect) of WASP-43b

• Published in: The Astrophysical journal Vol. 915; no. 1; pp. 45 - 61
• Main Authors: Cubillos, Patricio E., Keating, Dylan, Cowan, Nicolas B., Vos, Johanna M., Burningham, Ben, Ygouf, Marie, Karalidi, Theodora, Zhou, Yifan, Gonzales, Eileen C.
• Format: Journal Article
• Language: English
• Published: Philadelphia The American Astronomical Society 01.07.2021; IOP Publishing
• Subjects: Astrophysics; Bayesian statistics; Complexity; Emission spectra; Exoplanet atmospheres; Extrasolar planets; Hubble Space Telescope; James Webb Space Telescope; Radiative transfer; Retrieval; Solar system; Space telescopes; Spectroscopy; Temperature gradients
• ISSN: 0004-637X; 1538-4357
• DOI: 10.3847/1538-4357/abfe14

Abstract Thermal phase variations of short-period planets indicate that they are not spherical cows: day-to-night temperature contrasts range from hundreds to thousands of degrees, rivaling their vertical temperature contrasts. Nonetheless, the emergent spectra of short-period planets have typically been fit using one-dimensional (1D) spectral retrieval codes that only account for vertical temperature gradients. The popularity of 1D spectral retrieval codes is easy to understand: they are robust and have a rich legacy in solar system atmospheric studies. Exoplanet researchers have recently introduced multidimensional retrieval schemes to interpret the spectra of short-period planets, but these codes are necessarily more complex and computationally expensive than their 1D counterparts. In this paper we present an alternative: phase-dependent spectral observations are inverted to produce longitudinally resolved spectra that can then be fit using standard 1D spectral retrieval codes. We test this scheme on the iconic phase-resolved spectra of WASP-43b and on simulated observations for the James Webb Space Telescope (JWST) using the open-source Pyrat Bay 1D spectral retrieval framework. Notably, we take the model complexity of the simulations one step further from previous studies by allowing for longitudinal variations in composition in addition to temperature. We show that performing 1D spectral retrieval on longitudinally resolved spectra is more accurate than applying 1D spectral retrieval codes to disk-integrated emission spectra, even though this is identical in terms of computational load. We find that for the extant Hubble and Spitzer observations of WASP-43b, the difference between the two approaches is negligible, but JWST phase measurements should be treated with longitudinally re solved spect ral retrieval (ReSpect).