Estimations of the Seismic Pressure Noise on Mars Determined from Large Eddy Simulations and Demonstration of Pressure Decorrelation Techniques for the Insight Mission

Abstract

International audienceThe atmospheric pressure fluctuations on Mars induce an elastic response in the ground that creates a ground tilt, detectable as a seismic signal on the InSight seismometer SEIS. The seismic pressure noise is modeled using Large Eddy Simulations (LES) of the wind and surface pressure at the InSight landing site and a Green’s function ground deformation approach that is subsequently validated via a detailed comparison with two other methods: a spectral approach, and an approach based on Sorrells’ theory (Sorrells,Geophys. J. Int. 26:71–82, 1971; Sorrells et al., Nat. Phys. Sci. 229:14–16, 1971). The horizontal accelerations as a result of the ground tilt due to the LES turbulence-induced pressure fluctuations are found to be typically ∼ 2–40 nm/s2 in amplitude, whereas the direct horizontal acceleration is two orders of magnitude smaller and is thus negligible in comparison. The vertical accelerations are found to be ∼ 0.1–6 nm/s2 in amplitude. These are expected to be worst-case estimates for the seismic noise as we use a half-space approximation; the presence at some (shallow) depth of a harder layer would significantly reduce quasi-static displacement and tilt effects. We show that under calm conditions, a single-pressure measurement is representative of the large-scale pressure field (to a distance of several kilometers), particularly in the prevailing wind direction. However, during windy conditions, small-scale turbulence results in a reduced correlation between the pressure signals, and the single-pressure measurement becomes less representative of the pressure field. The correlation between the seismic signal and the pressure signal is found to be higher for the windiest period because the seismic pressure noise reflects the atmospheric structure close to the seismometer. In the same way that we reduce the atmospheric seismic signal by making use of a pressure sensor that is part of the InSight Auxiliary Payload Sensor Suite, we also the use the synthetic noise data obtained from the LES pressure field to demonstrate a decorrelation strategy. We show that our decorrelation approach is efficient, resulting in a reduction by a factor of ∼ 5 in the observed horizontal tilt noise (in the wind direction) and the vertical noise. This technique can, therefore, be used to remove the pressure signal from the seismic data obtained on Mars during the InSight mission

The Noise Model of the SEIS Seismometer of the InSight Mission to Mars

“…In order to make a sanity check, we have also developed a quasi-static model using Green's function approach: a pressure field applies in a static way on a homogeneous half-plane. This simpler model gives very similar results, as described in Murdoch et al (2016a). The advantage of this simpler model is that it can be considered as validated on Earth: Lorenz et al (2015) has shown that the size and amplitude of seismic signal generated on a seismometer by a dust devil (which can be assimilated to a pressure point source) can be measured, and is consistent with the Greens' function approximation used.…”

“…The correlation with a local measurement of the pressure variation can, therefore, be questioned. Murdoch et al (2016a) demonstrate that the local tilt is well correlated with the local pressure measurement up to a distance of 1 to 4 km depending on the wind conditions (windier conditions induce a better correlation). By using the synthetic noise data obtained from the LES pressure field, we were able to demonstrate that our decorrelation approach is efficient, resulting in a reduction by a factor of ∼5 observed in the horizontal tilt noise (in the wind direction) and the vertical noise.…”

“…These local effects include ground tilt due to the propagation of pressure wavessee e.g. Murdoch et al (2016a) or wind acting on local trees coupling into seismic noise. Local topography or complex local geology can cause incident waves to refract and reflect, adding complexity to the seismograms.…”

“…5.9 and 5.7) in the noise budget. The theoretical background for pressure decorrelation can be found in Murdoch et al (2016a).…”

“…In order to make this contribution precise in our case, we have to evaluate the pressure field around the InSight landing site, then model the deformation of the ground at the seismometer location before deriving the induced noise in the seismic bandwidth. This section summarizes the approach we have followed, but the whole process is described in detail in the companion paper Murdoch et al (2016a). …”