Skip to content
/ mtuq Public
forked from uafgeotools/mtuq

moment tensor (mt) estimation with uncertainty quantification (uq)

License

Notifications You must be signed in to change notification settings

rmodrak/mtuq

 
 

Repository files navigation

mtuq

Build Status SCOPED

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

Getting started

Installation

Quick start

Documentation

Acquiring seismic data

Acquiring Green's functions

Data processing

Visualization galleries

Library reference

Highlights

Common use cases include double couple moment tensor, full moment tensor, depth and hypocenter uncertainty analysis. Applications involving composite sources, force sources, constrained moment tensor sources, source-time functions, and other source parameters are also possible.

Solver interfaces

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.

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.

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.

About

moment tensor (mt) estimation with uncertainty quantification (uq)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 94.3%
  • Shell 4.4%
  • C 1.3%