5 Conclusions

We examined the structure and dynamics of the atmospheres of Pegasi planets$\,\,$and concluded that strong day-night temperature contrasts ($300\rm\,K$ or more) are likely to occur near the level where radiation is emitted to space ($\sim$1 bar). These temperature variations should drive a rapid circulation with peak winds of 1 km s^-1 or more. Force-balance arguments suggest that the mean midlatitude winds are eastward, but the equatorial winds could blow either east or west. Depending on the dynamical regime, the cloud coverage and consequently the radiative absorption of the incoming stellar flux will be very different.

Preliminary numerical simulations show that, while the dayside is generally hotter than the nightside, the specific temperature distribution depends on the dynamics. In our simulations, a broad superrotating jet develops that sweeps the high-temperature regions downwind by about 1 radian. This implies that the greatest infrared flux would reach Earth $\sim$10-15 hours before the occultation of the planet behind the star. On the other hand, if a subrotating jet existed, the greatest infrared flux would reach Earth after the occultation. Measurements of the infrared lightcurve of Pegasi planets$\,\,$will therefore constrain the direction and magnitude of the wind.

The simulations also produce a downward flux of kinetic energy across the $\sim$100 bar surface equal to $\sim$1% of the absorbed stellar flux. Although the simulation did not explicitly include the convective interior, we surmise that a substantial fraction of this kinetic energy would be converted to thermal energy by Kelvin-Helmholtz instabilities and other processes. As discussed in Paper I, "standard'' evolution models explain HD 209458b's radius only when the atmosphere is assumed to be unrealistically hot, but addition of an internal energy source allows a more realistic model to match the observed radius. The downward transport and subsequent dissipation of kinetic energy described here is a promising candidate. Bodenheimer et al. (2001) suggested an alternative - that the internal heating could be provided by tidal circularization of an initially eccentric orbit. The difficulty, as Bodenheimer et al. were careful to point out, is that the tidal heating is a transient process in the absence of a detected close, massive companion capable of exciting the planet's eccentricity. In contrast, the mechanism proposed here can last throughout the star's lifetime. Nevertheless, to fully determine the feasibility of the mechanism, more detailed numerical simulations are required (in particular, to test the dependence of the energetics on the possible existence of winds in the interior). We will present such simulations in a future paper.

Upcoming observations are likely to provide key information within the decade. Several spacecraft missions (either proposed or accepted) and dedicated ground programs will observe extrasolar planets. Measurement of starlight reflected from these planets may allow the albedo to be estimated. Because the star-planet-Earth angle changes throughout the planet's orbit, crude information on the scattering properties of the atmosphere (e.g., isotropic versus forward scattering) may be obtainable. Asymmetries in the reflected flux as the planet approaches and recedes from the transit could give information on the differences of albedo near the leading and trailing terminators, which would help constrain the dynamics. Finally, transit observations of Pegasi planets$\,\,$using high resolution spectroscopy should in the near future yield constraints on the atmospheric temperature, cloud/haze abundance, and winds (Brown 2001; see also Seager & Sasselov 2000; Hubbard et al. 2001). If these measurements are possible during the ingress and egress, i.e., the phases during which the planets enters and leaves the stellar limb, respectively, asymmetries of the planetary signal should be expected and would indicate zonal heat advection at the terminator. The duration of these phases being limited to less than 10 min, it is not clear that this effect is observable with current instruments.

Acknowledgements

We wish to thank F. Allard, P. Bodenheimer, H. Houben, S. Peale, D. Saumon, D. J. Stevenson, R. E. Young and K. Zahnle for a variety of useful contributions, and T. Barman for sharing results in advance of publication. This research was supported by the French Programme National de Planétologie, Institute of Theoretical Physics (NSF PH94-07194), and National Research Council of the United States.