The abundance, vertical distribution and origin of H2O in Titan’s atmosphere: Herschel observations and photochemical modelling

Disk-averaged observations of water vapor in Titan’s atmosphere acquired with the Herschel satellite are reported. We use a combination of unresolved measurements of three H2O rotational lines at 66.4, 75.4 and 108.0 μm with the PACS instrument, and spectrally-resolved observations of two other transitions at 557 GHz (538 μm) and 1097 GHz (273 μm) with the HIFI instrument, to infer the vertical profile of H2O over the 100–450 km altitude range. Monitoring of the 66.4 μm line indicates no variation between Titan leading and trailing sides, nor variation over a ∼1 year interval. Both the narrow (2–4 MHz) widths of the HIFI-observed lines, and the relative contrasts of the five H2O lines indicate that the H2O mole fraction strongly increases with altitude, with a best fit mole fraction of q0 = (2.3 ± 0.6) × 10−11 at a pressure p = 12.1 mbar, a slope −d(ln q)/d(ln p) = 0.49 ± 0.07, and a H2O column density of (1.2+/−0.2) × 1014 cm−2. This H2O profile also matches the original ISO observations of Titan H2O. Water vertical profiles previously proposed on the basis of 1-D photochemical models are too water-rich, and none of them have the adequate slope; in particular, the water profiles of Lara et al. (Lara, L.M., Lellouch, E., López-Moreno, J.J., Rodrigo, R. [1996]. J. Geophys. Res. E 101, 23261–23283) and Hörst et al. (Hörst, S.M., Vuitton, V., Yelle, R.V. [2008]. J. Geophys. Res. E 113, E10006) are too steep and too shallow, respectively, in the lower stratosphere. Photochemical models of oxygen species in Titan’s atmosphere are reconsidered, updating the Lara et al. model for oxygen chemistry, and adjusting the eddy diffusion coefficient in order to match both our H2O observations and the C2H6 and C2H2 vertical profiles determined from Cassini/CIRS. We find that the H2O profile can be reproduced by invoking a OH/H2O influx of (2.7–3.4) × 105 mol cm−2 s−1, referred to the surface. This is essentially one order of magnitude lower than invoked by previous modellers, and also a factor of ∼10 less than required to match the observed CO2 mole fraction. As H2O has a more shorter atmospheric lifetime than CO2 (∼9 years vs ∼450 years), we suggest that this reflects a temporal change in the oxygen influx into Titan, that could be currently much smaller than averaged over the past centuries. Both interplanetary dust particles and Enceladus’ activity appear to provide sufficient supply for the current Titan H2O. We tentatively favor the latter source as potentially more prone to time variability.