Vertical distribution of dust in the Martian atmosphere: OMEGA/MEx limb observations

The OMEGA spectrometer on board Mars Express acquired several observations of the Martian limb that are still largely unexploited. Here, we explore the information content of these data in terms of abundance and size of the Martian atmospheric dust. We approach the dust retrievals by applying a full spherical multiple scattering Monte Carlo 1D radiative transfer code to model the 0.5-2.5 mu m spectral range (VNIR and SWIR OMEGA channels), in order to derive dust effective radius and number density variation with altitude, between approx. 8 and 50 km. This is the first time this approach is applied to OMEGA limb data; therefore, we only present three study cases where water ice is below the detectability level, in order to focus on the methodological issues, assumptions, and performances before future more extensive applications. The model fully includes multiple scattering effects, which are known to be responsible for the coupling between the limb spectra taken at different altitudes and the surface. An extended three-dimensional modelling of the surface reflectance, acting as surface-shine for limb spectra, is developed. The VNIR channel is found useful in reducing the degeneracy of the radiative transfer solutions. The derived dust vertical distributions, yielding dust effective radii of the order of 0.85 +/- 0.15 mu m (corresponding to a modal radius rm of similar to 0.3 mu m) between 15 and 30 km are quite in agreement with Global Circulation Models (GCMs), but the abundances lie about one order of magnitude above model predictions, as previously reported from comparison between models and Mars Climate Sounder (MCS) data. An overall agreement is in fact found with MCS data, and in one case the OMEGA-retrieved dust is compatible with a local storm rising from the Hellas basin. Such storms are not well represented in the Mars Climate Database, which provides monthly mean statistics. Our results demonstrate the capability of OMEGA limb data to quantitatively contribute to Martian dust studies, despite the complex and slow radiative transfer computation scheme required to accurately model multiple scattering in the probed spectral range. The desirable application of the retrieval method to the whole OMEGA limb dataset, also expected to help assess the occurrence of local dust storms, requires further work aimed to include water ice aerosols and possibly thermal emission. Being the first attempt to use a Monte Carlo modelling approach to OMEGA limb data, this work represents a useful benchmark for validating faster, although less accurate, radiative transfer models.