Multi-year measurements of ripple and dune migration on Mars: Implications for the wind regime and sand transport

Abstract

Aeolian sand dunes are observed across the Martian surface. The arrival of the HiRISE camera on Mars Reconnaissance Orbiter at Mars in 2006 enabled detection of modern-day movement of dunes and ripples from orbit for the first time. Since 2006, HiRISE collected a long timeseries of repeat imagery at a few Martian dune fields. We analyze this timeseries of imagery at two of these dune fields, using COSI-Corr for image registration and correlation, to study the movement and dynamics of dunes and meter-scale ripples at the Nili Patera and Meroe Patera barchan dune fields. We present measurements of whole-dune translational sand fluxes extracted at both dune fields via manual tracking of dune crestlines and slipfaces in HiRISE images. We also present a multi-Mars year timeseries of ripple flux measurements. Ripple migration shows a consistent pattern of seasonal variation, with maxima in flux during northern-hemisphere autumn and winter at both dune fields. Ripple migration is also observed to decrease away from the upwind margins of dune fields. We compare our observations with predicted sand transport using winds output from the MarsWRF atmospheric circulation model and theories of sand motion. The model predicts half-hourly, mesoscale winds, from which we estimate the 1 Hz, local-scale winds by assuming a Weibull distribution of wind speed, with parameters chosen based on landed wind data. This approach uses remote sensing observations of bedform migration, and comparisons with model output, to place constraints on the wind regime. Our measurements of the seasonal pattern of sand flux variation agree, to first order, with predictions based on modeled wind speeds. Comparison of the magnitudes of predicted and observed sand fluxes is not feasible due to the high uncertainties in our calculated sand fluxes caused by uncertainties in input parameters, most importantly the assumed fluid threshold for sand transport. However, we note that model predictions fit our observed sand fluxes best when relatively low values of the fluid threshold shear velocity of ~0.6–0.8 m/s (or shear stresses of O(10−3) Pa) are assumed.