Probing the Martian atmospheric boundary layer using impact-generated seismo-acoustic signals

In-situ measurements of atmospheric variables are key to the validation and improvement of current models of the Martian climate, particularly of the planetary boundary layer. This highly dynamical region, whose thickness and altitude vary, is governed by unique thermodynamical, physical and chemical exchanges between the troposphere and surface. However, only sparse data is available in this location, as information is collected mostly by a few surface probes and during occasional spacecraft descent and landing.  Recently, the seismometers of the InSight lander recorded short low-frequency, dispersed waveforms on six occasions. These signals were shown to be impact-generated guided infrasound waves. They were excited by the atmospheric entry and surface impact of bolides, and propagated  in a low-altitude atmospheric waveguide. The location of their source, i.e., the impact crater, is well characterized by orbital imaging and the origin time by joint seismic and acoustic analysis. The deformation of the ground by the propagating infrasound wave allowed their detection by InSight seismometers. Using analytical modeling, we show that the signal group velocity depends on the vertical profile of the effective sound speed in the Martian boundary layer. These impact-generated signals provide a unique opportunity to probe the atmosphere at these low altitudes. Here, we conduct a Bayesian inversion of effective sound speed up to  ~2000 m altitude using the group velocity measured for events S0981c, S0986c and S1034a. We compare these results with estimates of effective sound speed profiles provided by the Mars Climate Database based on global circulation models. We show that the differences between inverted and modeled profiles can be attributed to a local variation in wind in the impact → station direction, with amplitude smaller than 2 m/s, and that the infrasound data generally validates the Mars Climate Database results for each date and location.