2022MNRAS.515..749G

Exoplanet atmospheres are inherently three-dimensional systems in which thermal/chemical variation and winds can strongly influence spectra. Recently, the ultra-hot Jupiter WASP-76 b has shown evidence for condensation and asymmetric Fe absorption with time. However, it is currently unclear whether these asymmetries are driven by chemical or thermal differences between the two limbs, as precise constraints on variation in these have remained elusive due to the challenges of modelling these dynamics in a Bayesian framework. To address this, we develop a new model, HyDRA-2D, capable of simultaneously retrieving morning and evening terminators with day-night winds. We explore variations in Fe, temperature profile, winds, and opacity deck with limb and orbital phase using VLT/ESPRESSO observations of WASP-76 b. We find Fe is more prominent on the evening for the last quarter of the transit, with $\log (X_\mathrm{Fe}) = {-4.03}^{+0.28}_{-0.31}$, but the morning shows a lower abundance with a wider uncertainty, $\log (X_\mathrm{Fe}) = {-4.59}^{+0.85}_{-1.0}$, driven by degeneracy with the opacity deck and the stronger evening signal. We constrain 0.1-mbar temperatures ranging from $2950^{+111}_{-156}$ to $2615^{+266}_{-275}$ K, with a trend of higher temperatures for the more irradiated atmospheric regions. We also constrain a day-night wind speed of $9.8^{+1.2}_{-1.1}$ km s^–1 for the last quarter, higher than $5.9^{+1.5}_{-1.1}$ km s^–1 for the first, in line with general circulation models. We find our new spatially and phase-resolved treatment is statistically favoured by 4.9σ over traditional 1D-retrievals, and thus demonstrate the power of such modelling for robust constraints with current and future facilities.