PrISM.jl (Probabilistic Inference of Subsurface Models) is a performant package written in Julia for doing forward and inverse modeling in geophysics. We hope to write the code structure such that any other geophysical survey can also be used, and we can tend towards a joint forward and inverse modeling library.
While forward modeling typically requires solving a PDE obtained using the quasi-static approximation, in 1D, we are fortunate to have the solutions in an analytical form. Currently, this is what is supported.
Supported methods:
- 1D Magnetotellurics (MT)
- 1D Rayleigh waves
- 1D Love waves
- 1D Direct Current (DC) Resistivity
No surprises here that we are almost always trying to solve for an under-determined system.
Deterministic schemes supported:
- Occam
- Nonlinear schemes using
NonlinearSolve.jl - Nonlinear schemes using
Optimization.jl
Probabilistic schemes supported:
- MCMC with fixed grids
- MCMC with flexible grids
- RTO-TKO
