MTid is a Python3 routine for assessing the likelihood that a seismic moment tensor belongs to any number of predefined event populations. For example, it can determine the probability that an event was an explosion as opposed to an earthquake. It does so by reading in the six independent components of the seismic moment tensor, normalising them, projecting them onto a hypersphere, and evaluating any number of probability distribution functions that have previously been defined on that hypersphere. It follows the methodology outlined in Hoggard et al. (2024), which builds on work by Ford et al. (2020).
The file example_MTs.csv contains moment tensor solutions for nine seismic events that were published in Alvizuri & Tape (2018). The first six are for nuclear tests carried out by the Democratic People's Republic of Korea from 2006 to 2017, the seventh is believed to be a "cavity collapse" that immediately followed the 2017 test, and the final two are earthquakes that ocurred in South Korea in 2016 and 2017.
The six components of the moment tensor must be comma-separated, can be in any unit (provided that it is consistent for all six components), and must be in a specific order: Mxx is in the first column, followed by Mxy, Mxz, Myy, Myz and Mzz (x represents east, y is north, and z is the vertical/radial component). The program can be run on the command line using
python MTid.py example_MTs.csv
which will state that only the first six events belong to the explosion population and will also output two files. The first is prob.csv and contains a comma-separated list of the probabilities that each moment tensor in the input file belongs to each of the possible event populations. The second is lune_coords.csv and contains a comma-separated list of the coordinates (longitude, latitude) of the moment tensors when projected onto the fundamental lune of Tape & Tape (2012), which can be useful for plotting purposes. A help message can be accessed using
python MTid.py -h
The most likely modification that a user may wish to make is to alter the prior probability that an event was an explosion from the default values of 0.9999 for explosions and 0.0001 for earthquakes (see Section 3.2 of Hoggard et al. 2024). This modification can be acheived by locating the library populations.py and updating the prior = variables, staying faithful to the fact that the sum of all priors must be equal to 1. It is also within this same library that a user may wish to update parameters that control the shape of the probability distribution functions or add new population descriptions for other types of seismic event.
- Hoggard, M.J., J.L. Scealy & B.G. Delbridge (2024). Seismic moment tensor classification using elliptical distribution functions on the hypersphere. Geophysical Journal International, 237, 1–13.
- Ford, S.R., G.D. Kraft & G.A. Ichinose (2020). Seismic moment tensor event screening. Geophysical Journal International, 221, 77–88.
- Alvizuri, C. & C. Tape (2018). Full moment tensor analysis of nuclear explosions in North Korea. Seismological Research Letters, 89, 2139–2151.
- Tape, W. & C. Tape (2012). A geometric setting for moment tensors. Geophysical Journal International, 190, 476–498.