/mtbench

moment tensor inversion benchmarks

Primary LanguagePython

Instructions
------------

1. Install mtuq within a conda virtual environment as follows

    https://uafgeotools.github.io/mtuq/install/index.html


2  Within the same virtual environment install mtbench

   >> cd mtbench/
   >> pip install -e .


3. Download waveform data

  >> bash WAVEFORMS/download.bash


4. Run scripts

  >> cd scripts/
  >> python run_Silwal2016_syngine
  >> python run_Alvizuri2018_syngine


5. Optionally, compare your results with the expected output

    https://github.com/rmodrak/mtbench/tree/FiguresSilwal2016



Notes
-----

 - AxiSEM and FK scripts closely reproduce published figures, but require local Green's function databases (available at University of Alaska Fairbanks)

 - Since AK135F is used in place of regional velocity models, syngine scripts do NOT closely reproduce published figures

- syngine scripts download Green's functions from a remote server, which can take a very long time. Afterward, Green's functions are locally cached, so any subsequent runs will be much faster.



References
----------

[`Alvizuri2018`] C. Alvizuri and C. Tape (2018), Full Moment Tensor Analysis of Nuclear Explosions in North Korea, SRL, https://doi.org/10.1785/0220180158
 

[`Silwal2016`] V. Silwal and C. Tape (2016), Seismic moment tensors and estimated uncertainties in southern Alaska, JGR, https://doi.org/10.1002/2015JB012588