ATMOSPHERIC HEAT REDISTRIBUTION ON HOT JUPITERS.

Infrared light curves of transiting hot Jupiters present a trend in which the atmospheres of the hottest planets are less efficient at redistributing the stellar energy absorbed on their daysides—and thus have a larger day-night temperature contrast—than colder planets. To this day, no predictive atmospheric model has been published that identifies which dynamical mechanisms determine the atmospheric heat redistribution efficiency on tidally locked exoplanets. Here we present a shallow-water model of the atmospheric dynamics on synchronously rotating planets that explains why heat redistribution efficiency drops as stellar insolation rises. Our model shows that planets with weak friction and weak irradiation exhibit a banded zonal flow with minimal day-night temperature differences, while models with strong irradiation and/or strong friction exhibit a day-night flow pattern with order-unity fractional day-night temperature differences. To interpret the model, we develop a scaling theory which shows that the timescale for gravity waves to propagate horizontally over planetary scales, τwave, plays a dominant role in controlling the transition from small to large temperature contrasts. This implies that heat redistribution is governed by a wave-like process, similar to the one responsible for the weak temperature gradients in the Earth's tropics. When atmospheric drag can be neglected, the transition from small to large day-night temperature contrasts occurs when , where τrad is the radiative relaxation time and Ω is the planetary rotation frequency. Alternatively, this transition criterion can be expressed as τrad ∼ τvert, where τvert is the timescale for a fluid parcel to move vertically over the difference in day-night thickness. These results subsume the more widely used timescale comparison for estimating heat redistribution efficiency between τrad and the horizontal day-night advection timescale, τadv. Only because τadv ∼ τvert for hot Jupiters does the commonly assumed timescale comparison between τrad and τadv yield approximately correct predictions for the heat redistribution efficiency. [ABSTRACT FROM AUTHOR]