FIGURE 3. Simulated emission spectra of HD 189733b in the infrared.

We show here different simulated emission spectra, all generated at the same spectral resolution and corresponding to the same atmospheric composition but different temperature–pressure (T–P) profiles. Accordingly, we obtain very different spectral responses: hot isothermal profile (top, dotted line), isothermal profile and a cloud at 10–100 mbar altitude (bottom, dashed), and temperature profile decreasing with altitude (middle, solid). Although present and with the same abundances in all the three scenarios, water is detectable through emission spectra only in the last case, showing up as a steep gradient between 8 and 10 m when compared to the 10–14 m region. An isothermal profile in the upper part^{18, 28} and a cloud in the lower part of the atmosphere is sufficient to explain the non-detection of water⁴ even if water were present in high abundance. HD 189733b is presumably tidally locked, therefore the thermal profiles and the condensate dynamics might be very different on the two sides of the planet. The same reasoning applies to HD 209458b and the most recent observations in secondary transit of that planet³.