Numerical modelization as a way to infer water content in Martian landslides

Introduction: Mass wasting is a geological process reported in almost every rocky body of our solar sys-tem [1], although the debate regarding the dynamics and the presence of water in martian landslides is still ongoing. The low percentage of well character-ized landslides on Mars (especially from the mechan-ical point of view) contributes to the general lack of useful data to solve the issue. For this reason, we propose a combination of geomorphological analyses and numerical modeling performed on two different models to prove the mechanical properties of a minor landslide (volumes < 1010 m3) and estimate the water content involved in the landslide’s dynamics. Crater counting technique was also used to give an age to the landslide and temporally constrain its rheological properties (and environmental conditions) to a certain period of Mars history. Methodologies: The dataset used in this research combines CTX (Context Camera, 5,99 m/px) dataset with a single HRSC (High Resolution Stereo Cam-era) derived DEM (Digital Elevation Model, up to 10m/px of vertical accuracy) covering the entire im-pact crater. This data combination allowed us to per-form detailed measurements on DEM file (total runout distance “ΔL”, total drop height “H”, width of the deposits “W”, levee height “h”), to estimate the volume of the deposits in the accumulation zone (useful for balancing the topographic reconstructions and ensuring mass conservation during simulations), and to make accurate observations regarding the morphologies and surface features of the landslide. Some of these physical measurements were used for calculating parameters to use in simulations [2, 3]. Topographic profiles of the basal and pre-event topography were reconstructed and used as input for the numerical simulations. For the sake of objectivi-ty, we performed simulations with two differend models: our in-house model [3, 4, 5] and the com-mertial software MADflow [6]. The aim was to ob-tain a good match between the simulation results and the observed extension, shape and thickness of the actual landslide. In this way, we may infer the dy-namics of failure and the water content involved in such a mass movement. The modeling phase consisted of a systematic parametric analysis. The range of the parameters used in the simulations was extracted from experimental studies [7, 8] and bibliography [8]. Case study: We chose to apply these methods to a martian landslide located in an unnamed impact crater of the Noachian highlands, close to the basin of Utopia Planitia (Fig.1). Among the reasons that drew our selection to this specific site, the most im-portant are (i) the characteristic long runout and (ii) the lobate front of the landslide’s deposits, some ge-omorphological clues that may favor a possibly vis-co-cohesive material (i.e. suggesting water in-volvment). Moreover, in the surrounding of the crater, there are outcrops of clay minerals and a sec-ondary impact with evindences of seasonal seepages (RSL) on top of the landslide’s scar. This may be a good candidate on which to test our hypothesis through numerical simulations and rheologies that include both frictional and cohesive parameters. Results: The best match obtained with our in-house model corresponds to a Bingham material, whereas with MADflow the best match is obtained for a Coulomb-viscous fluid. It is worth noting that, despite the differences in the model formulations, both seem to favor the hypothesis of a visco-cohesive material, suggesting the possible presence of water in the interstitial pore space. The shape of the deposits, as well as their thickness and lateral extent, is in good accordance with the observations, with tolera-ble discrepancies. The values of density, viscosity and yield stress used in the simulations allow us to estimate the water content in the mixture. The numerical results, combined with geomor-phological features of the landslide and the surround-ing environment, suggest that water may have been involved in this geological process. Crater counting results gave an age of approxi-mately 220 Ma, highlighting the recent nature of the studied landslide and subsequently the hypothesized presence of water in this site in a recent martian time (Late Amazonian).