Differentiability and retrievability of CO2 and H2O clouds on Mars from MRO/MCS measurements: A radiative-transfer study

Since the 1970s, it has been predicted that both CO2 and H2O clouds can form in the Martian atmosphere, and many remote-sounding instruments have directly observed layers of extinction asserted to be clouds composed of either CO2 or H2O ice on Mars. The Mars Climate Sounder, onboard the Mars Reconnaissance Orbiter (MRO/MCS), entered orbit around Mars in 2006, and has been providing near-continuous coverage of the full planet since, at wavelengths from visible through to the mid-infrared, primarily in limb-viewing geometry, making it a suitable candidate to study the parameters of these clouds. In this work, the multiple scattering radiative-transfer tool NemesisMCS has been used to create a large dataset of simulations of CO2 and H2O clouds on Mars as would be measured by MRO/MCS, using a range of atmospheric conditions as well as cloud parameters derived from literature suitable for upper atmospheric clouds, and building specifically on the work of Sefton-Nash et al. (2013). This ensemble of simulations has been used to characterise the spectral signature of plausible CO2 and H2O clouds, as well as to assess the suitability of MRO/MCS to observe, to differentiate between, and to derive properties of such clouds. It has been found, given the noise levels expected for MRO/MCS and the range of atmospheric and cloud parameters sampled in this study, that radiance signals introduced by upper atmospheric clouds having nadir optical depths greater than about 10(-5) should be distinguishable, with S/N >= 1. This corresponds to specific concentrations greater than about 105 particles/g, particle radii greater than around 0.5 mu m, and cloud depths greater than about 2 km. MRO/MCS measurements should be able to be used with confidence to differentiate between upper atmospheric cloud and dust in the lower atmosphere, and clear conditions, with high success (approximate to 100%). Lower reliability classification is accomplished for CO2 clouds, with only 60% being correctly identified as CO2, and the remainder classified instead as H2O cloud, in the case of optical depths in the expected range for upper atmospheric cloudswhich are detectable by MRO/MCS, although this result is highly dependent upon the sampled selection of optically thin and thick clouds and the atmospheric model employed. Although almost all the H2O clouds are correctly identified, the fact that such a large proportion of CO2 clouds are misclassified as H2O clouds shows that the spectral information alone from MRO/MCS is insufficient to differentiate between CO2 and H2O clouds when optically thin but detectable clouds are included in the analysis. Using a simple look-up table (LUT) scheme and simulated data, retrieval of properties of upper atmospheric clouds of sufficient opacity is possible, with preliminary estimates indicating that H2O cloud and dust parameters can be correctly reproduced between 48% and 100% of the time, and between 18% and 92% of the time for CO2 cloud test cases, although it must be noted that these values must be taken as a qualitative measure which does not capture the full range of atmospheric and cloud conditions on Mars which would be present in real MRO/MCS data. Furthermore, because of the optical properties of H2O and CO2, on a like-with-like selection, the H2O clouds always produce greater perturbations in radiance, thus biasing results to a higher success rate for H2O cloud retrievals. Application of the method to MRO/MCS data with a full-optimal estimation retrieval tool such as Nemesismcs will be the topic of a future study. (C) 2014 Elsevier Ltd. All rights reserved.