Near-Field Seismic Propagation and Coupling Through Mars’ Regolith: Implications for the InSight Mission

NASA’s InSight Mission will deploy two three-component seismometers on Mars in 2018. These short period and very broadband seismometers will be mounted on a three-legged levelling system, which will sit directly on the sandy regolith some 2–3 meters from the lander. Although the deployment will be covered by a wind and thermal shield, atmospheric noise is still expected to couple to the seismometers through the regolith. Seismic activity on Mars is expected to be significantly lower than on Earth, so a characterisation of the extent of coupling to noise and seismic signals is an important step towards maximising scientific return. In this study, we conduct field testing on a simplified model of the seismometer assembly. We constrain the transfer function between the wind and thermal shield and tripod-mounted seismometers over a range of frequencies (1–40 Hz) relevant to the deployment on Mars. At 1–20 Hz the displacement amplitude ratio is approximately constant, with a value that depends on the site (0.03–0.06). The value of the ratio in this range is 25–50% of the value expected from the deformation of a homogeneous isotropic elastic halfspace. At 20–40 Hz, the ratio increases as a result of resonance between the tripod mass and regolith. We predict that mounting the InSight instruments on a tripod will not adversely affect the recorded amplitudes of vertical seismic energy, although particle motions will be more complex than observed in recordings generated by more conventional buried deployments. Higher frequency signals will be amplified by tripod-regolith resonance, probably reaching peak-amplification at $\sim 50$ Hz. The tripod deployment will lose sensitivity at frequencies $>50$ Hz as a result of the tripod mass and compliant regolith. We also investigate the attenuation of seismic energy within the shallow regolith covering the range of seismometer deployment distances. The amplitude of surface displacement decays as $r^{-n}$ , where $1.5 < n < 2$ . This exceeds the value expected for a homogeneous isotropic elastic halfspace ( $n \sim 1$ ), and reflects an increase in Young’s modulus with depth. We present an updated model of lander noise which takes this enhanced attenuation into account.