MULTIWAVELENGTH CONSTRAINTS ON THE DAY-NIGHT CIRCULATION PATTERNS OF HD 189733b

We present new Spitzer observations of the phase variation of the hot Jupiter HD 189733b in the MIPS 24 mm bandpass, spanning the same part of the planet's orbit as our previous observations in the IRAC 8 mm bandpass (Knutson et al. 2007). We find that the minimum hemisphere-averaged flux from the planet in this bandpass is 76% +- 3% of the maximum flux; this corresponds to minimum and maximum hemisphere-averaged brightness temperatures of 984 +- 48 K and 1220 +- 47 K, respectively. The planet reaches its maximum flux at an orbital phase of 0.396 +- 0.022, corresponding to a hot region shifted 20deg-30deg east of the substellar point. Because tidally locked hot Jupiters would have enormous day-night temperature differences in the absence of winds, the small amplitude of the observed phase variation indicates that the planet's atmosphere efficiently transports thermal energy from the day side to the night side at the 24 mm photosphere, leading to modest day-night temperature differences. The similarities between the 8 and 24 mm phase curves for HD 189733b lead us to conclude that the circulation on this planet behaves in a fundamentally similar fashion across the range of pressures sensed by these two wavelengths. One-dimensional radiative transfer models indicate that the 8 mm band should probe pressures 2-3 times greater than at 24 mm, although the uncertain methane abundance complicates the interpretation. If these two bandpasses do probe different pressures, it would indicate that the temperature varies only weakly between the two sensed depths, and hence that the atmosphere is not convective at these altitudes. We also present an analysis of the possible contribution of star spots to the time series at both 8 and 24 mm based on near-simultaneous ground-based observations and additional Spitzer observations. Accounting for the effects of these spots results in a slightly warmer night-side temperature for the planet in both bandpasses, but does not otherwise affect our conclusions.