Model parameters
The joint inversion program, joint_inv, samples the posterior probability distributions of model parameters regarding 1-D velocity structure. The model parameters are:
- Number of layers \(k\)
- Depth of each layer bottom \(z_i (i=1,\cdots,k)\)
- S-wave velocity of each layer \(\beta_i (i=1,\cdots,k)\)
- P-wave velocity of each layer (optional) \(\alpha_i (i=1,\cdots,k)\)
- Standard deviation of data noise (optional) \(\sigma_i (i=1,\cdots,N_{data})\)
Alternatively, one can solve for velocity anomalies centered on the given reference velocity, as is done by Akuhara et al. (2020). This mode is activated by setting solve_anomaly to be 'T' in the main parameter file.

Note
- The ocean and bottom half-space layers are not counted for \(k\).
- The properties of the ocean layer are fixed during the inversion (P-wave velocity: 1.5 km/s and density: 1.0 g/cm^3, by default). These default values can be changed through the parameters vp_ocean and rho_ocean.
- The properties of the bottom half-sapce are also fixed during the inversion (P-wave velocity: 8.1 km/s, S-wave velocity: 4.6 km/s and density: 3.3 g/cm^3, by default). These default values can be changed through the parameters vp_bottom, vs_bottom, and rho_bottom. If you want to solve for the velocities of this layer, set vp_bottom and vs_bottom to be negative (NOT RECOMMENDED WHEN INVERTING DISPERSION CURVES because low velocity at the half space often results in the failure of dispersion computation.)
- The density of is calculated from the P-wave velocity using the empirical relationship by Brocher (2005):