Back
Science

Drilleau et al. (2024) model selected as seismic velocity model for InSight landing site

View source

Drilleau et al. (2024) Model Selected for Mars InSight Landing Site

A new seismic velocity model, developed by Drilleau et al. (2024), has been adopted for the InSight landing site region. The model was derived from a joint inversion of P- and S-wave travel times and surface-wave dispersion curves from meteoroid impacts.

A probabilistic Markov chain Monte Carlo framework was used, incorporating prior constraints on crustal structure and refined by consistency with receiver function data.

The model avoids velocity averaging from distant observations and is consistent with multiple independent datasets.

Geochemical Database and Sample Selection

A comprehensive database of 883 Martian rock samples was compiled for the analysis.

  • Mafic samples were defined as those with 45–52 wt% SiO₂ (basaltic composition).
  • Ultramafic samples had less than 45 wt% SiO₂.
  • Additional filters were applied to minimize surface alteration effects.

Phase Equilibrium Modelling and Forward Seismic Modelling

Stable phase assemblages and elastic properties were computed using MAGEMin software with composition-dependent equations of state.

Modelling parameters:

  • Depth range: 15–38 km
  • Early Mars areotherm: 16°C/km
  • Present-day Mars areotherm: 10°C/km

Seismic velocities (VP and VS) were calculated from thermodynamic data and empirical relations, using Voigt-Reuss-Hill averaging.

Bayesian Classification

Posterior probabilities for mafic versus ultramafic compositions in each layer were computed using Bayesian model selection. The marginal likelihood was approximated by summing likelihoods over sampled compositions.

Layer 4 showed higher likelihood under the ultramafic model, while layer 3 showed the opposite.

Uncertainty Assessment

A Latin Hypercube Sampling approach with 1,000 parameter sets was used to assess how uncertainties in areotherm, water content, and Fe³⁺ affect posterior probabilities.

Thermal Modelling

The temperature profile was calculated using internal heating from radiodecay and basal heat flux.