ejhgeo/BayGrav3D

BayGrav3D is a Bayesian linear gravity inverse modeling program that inverts gravity data to determine the best-fitting densities of spatially discretized 3D subsurface prisms in a least-squares sense. We use a Bayesian approach to incorporate both data uncertainty and prior geophysical constraints, such as seismic data. Gaussian priors are appl...

BayGrav3D is a Bayesian linear gravity inverse modeling program that inverts gravity data to determine the best-fitting densities of spatially discretized 3D subsurface prisms in a least-squares sense. We use a Bayesian approach to incorporate both data uncertainty and prior geophysical constraints, such as seismic data. Gaussian priors are applied to the model parameters as absolute equality constraints. BayGrav3D utilizes Tikhonov regularization as a relative equality constraint that smooths and stabilizes the inversion solution. Given gravity data and a set of priors, the inversion produces a solution to the model parameters (i.e. density) and the full covariance and resolution matrices to quantify the error on the solution. BayGrav3D is capable of working on both local and regional scales and with both simple and complex subsurface geometries. BayGrav3D is written is Matlab, and the release contains an example project with data with which users may test the scripts. A python release is planned sometime in the coming year or so. For more information on the methods, please refer to our paper - Hightower et al. (2020), A Bayesian 3-D linear gravity inversion for complex density distributions: application to the Puysegur subduction system, GJI, v. 223, p. 1899-1918. For using the program, please refer to the documentation (soon to be available on ReadTheDocs) and the commentary within the scripts.