Caractérisation des surfaces glacées de Mars par imagerie hyperspectrale : inversion du transfert radiatif

Abstract : Mars has a complex climate, characterized by carbon dioxide and water cycles, and dust transport at all scales. These cycles are mainly controlled by the seasonal condensation of CO₂ and water ice deposits at high latitudes during the polar night and their sublimation during the local spring. There are a lot of interactions between the CO₂ , water and dust cycles on Mars and they influence each other. An active seasonal process illustrates particularly well these links: the cryoventing, cold CO₂ gas jets triggered by seasonal sublimation of CO₂ deposits, which can put dust from the regolith in suspension into the atmosphere durably, and whose activity seems to be modulated by the exchange of water at the surface.The purpose of this thesis is to allow the use of the available hyperspectral imaging data to their full potential, to bring new constraints on seasonal exchanges between surface and atmosphere and the interactions between the different cycles (CO₂ , water , dust), focusing on cold gas jets. To achieve this, a semi-analytical radiative transfer model in compact ices and an effective inversion method were developed and validated.The radiative transfer model describes the interaction of light with a surface quantitatively, using the following parameters: thickness of the layer, volume proportions and grain-sizes of impurities, surface roughness. It is based on several key assumptions: geometrical optics, piecewise-continuous media quasi-spherical inclusions. The two-stream approximation is used for the radiative transfer inside the layer, but the surface specular reflectance is estimated taking into account the variability of the facets orientations from the surface roughness. This model was validated both numerically and on laboratory data.The inversion method consists in exploring synthetic databases generated by the radiative transfer model and determining the most likely sets of parameters to reproduce a given measure. The inversion is based on the Bayesian formalism: the manipulated variables are described by probability density functions. This allows to take into account realistic uncertainties on the data and enables to calculate a posteriori uncertainties on the result of the inversion.A focused study was conducted on a area of particular interest, to test and prove the applicability of this approach to the massive inversion of spectro-imaging data. The dune-field of Richardson Crater (72°S, 180°W) was chosen because it shows strong interactions between the water and CO₂ cycles, a major seasonal cryoventing activity, but also a large amount of data available and a high quality temporal monitoring. The monitoring of surface characteristics on this site shows a decrease in the thickness of the ice during the spring consistently with climate models simulations. We were able to estimate and monitor the content of water and dust in order to discuss the formation scenario of cold jets.
Document type :
Complete list of metadatas
Contributor : Abes Star <>
Submitted on : Wednesday, June 8, 2016 - 2:43:14 PM
Last modification on : Tuesday, May 21, 2019 - 3:17:13 AM


Version validated by the jury (STAR)


  • HAL Id : tel-01328824, version 1



François Andrieu. Caractérisation des surfaces glacées de Mars par imagerie hyperspectrale : inversion du transfert radiatif. Matière Condensée [cond-mat]. Université Paris-Saclay, 2015. Français. ⟨NNT : 2015SACLS178⟩. ⟨tel-01328824⟩



Record views


Files downloads