Estimating dayside effective temperatures of hot Jupiters and associated uncertainties through Gaussian process regression

In this work, we outline a new method for estimating dayside effective temperatures of exoplanets and associated uncertainties using Gaussian process (GP) regression. By applying our method to simulated observations, we show that the GP method estimates uncertainty more robustly than other model-independent approaches. We find that unbiased estimates of effective temperatures can be made using as few as three broad-band measurements (white-light HST WFC3 and the two warm Spitzer IRAC channels), although we caution that estimates made using only IRAC can be significantly biased. We then apply our GP method to the twelve hot Jupiters in the literature whose secondary eclipse depths have been measured by WFC3 and IRAC channels 1 and 2: CoRoT-2 b; HAT-P-7 b; HD 189733 b; HD 209458 b; Kepler-13A b; TrES-3 b; WASP-4 b; WASP-12 b; WASP-18 b; WASP-33 b; WASP-43 b; and WASP-103 b. We present model-independent dayside effective temperatures for these planets, with uncertainty estimates that range from $\pm$66 K to $\pm$136 K.
Comment: 11 pages, 6 figures, accepted for publication in MNRAS