mtuq

MTUQ provides moment tensor estimates and uncertainty quantification from broadband seismic data.

Use cases include full moment tensor anaysis, constrained moment tensor analysis, and many other possibilities involving moment tensor, force, hypocenter, depth, and wavelet parameters.

To provide these capbilities, MTUQ draws on ObsPy, instaseis, and pandas data structures.

Misfit evaluation

Waveform difference and cross-correlation time-shift misfit evaluation on body-wave and surface-wave windows is implemented in C-accelerated Python.

These misfit functions can be used with mtuq.grid_search, which automatically partitions the grid over multiple MPI processes if invoked from an MPI environment. For efficient and unbiased uncertainty quantification, uniform grids can be used for the grid search, drawing from Tape2015.

Alternatively, MTUQ misfit functions can be used as a starting point for Bayesian uncertainty quantification using pymc or other MCMC libraries.

Solver interfaces

Data I/O functions are included for reading AxiSEM, SPECFEM3D, and FK Green's functions as well as downloading Green's functions from remote syngine databases.

Visualization

Visualization utilities are included for both the eigenvalue lune and v-w rectangle, with matplotlib and Generic Mapping Tools graphics backends.

Testing

The package has been tested against legacy Perl/C codes as well as published studies.

Getting started

Installation

Quick start

Documentation

Acquiring seismic data

Acquiring Green's functions

Data processing

Visualization galleries

Library reference