Repeating earthquake discovery using cross-correlation and differential arrival times
Project description
FINDRES
A Python code for detecting true Repeating Earthquakes from Self-similar Waveforms (FINDRES)
Abstract
The software package FINDRES is an open-source seismological software. FINDRES is designed to discriminate repeating earthquakes starting from a family of candidate repeating earthquakes, based on the cross-correlation values and S-P time difference between pairs of earthquakes (estimated using cross-spectrum).
Motivation and significance
The code FINDRES is inspired to previously published methods that combine both seismic waveform similarity, using cross-correlation function, and differential S-P travel time measured at each seismic station (Chen et al., 2008 and Shakibay Senobari and Funning, 2019). The code is versatile and works with and without P and S-wave phase pickings. At the moment the reading of phases in Hypoellipse (Lahr 1999), Hypoinverse (Klein, 2002), NonlinLoc (Lomax et al. 2000), and QuakeML (https://quake.ethz.ch/quakeml) format are implemented. The code has been tested using synthetic and real data, providing accurate results. It contributes to the implementations of open-source Python packages in seismology aiming to support the activities of researchers and the reproducibility of scientific results.
Requirements
Python installed to run the program (version 3.8 or more) is required [http://python.org].
How to use it
The package has few dependencies; the recommended way of installing them is via the Conda package manager. You can create a test environment using
conda create -n findres-test python=3.8 numpy=1.21 scipy scikit-learn pandas tqdm pyyaml obspy mtspec -c conda-forge
Remember to activate the environment with
conda activate findres-test
The package is registered on PyPi, you can install it using pip
pip install findres
After that, you'll have the findres
script in your path.
# findres --help
usage: ./bin/findres [-h] [--phase_file PHASE_FILE] [--phase_type {hypoinv,nll,quakeml,hypoel}] [--taup_model TAUP_MODEL] [--rebuild_model]
[--graphics_dir GRAPHICS_DIR] [--graphics_format GRAPHICS_FORMAT] [--hypodd] [--stop] [--log LOG] [--progress]
catalogue inventory parameters output
positional arguments:
catalogue Modified ZMAP catalogue containing repeater candidates, their MSEED location, and names
inventory Inventory of the stations data
parameters Numerical parameters file
output Output YAML file
optional arguments:
-h, --help show this help message and exit
--phase_file PHASE_FILE
Catalogue containing the phase picking and other information if available (default: None)
--phase_type {hypoinv,nll,quakeml,hypoel}
Type of PHASE_FILE (default: None)
--taup_model TAUP_MODEL
Velocity model file without extension (assumed to be .tvel) if available (default: None)
--rebuild_model Force the rebuild of the velocity model (regenerate ObsPy local cache) (default: False)
--graphics_dir GRAPHICS_DIR
Where to put the graphics relative to OUTPUT_DIR, if not specified the graphics is not generated (default: None)
--graphics_format GRAPHICS_FORMAT
Graphics format, must be one of the extensions recognized by matplotlib (default: pdf)
--hypodd Whether to output hypodd input files (default: False)
--stop Stop if exceptions are raised during the analysis, otherwise skip to the next event station (default: False)
--log LOG Log level (default: info)
--progress Show progress bar (default: False)
You can run a test using the data provided in this repository and accessing the following folder
cd data/california
For example the command to analyse the data and produce graphics while showing a progress bar is
findres cre.zmap inventory.xml parameters.yaml results --phase_file=phases_hypoinv.txt --phase_type=hypoinv --taup_model=ncmodel --graphics_dir=figures --progress
The command to analyse the data without graphics (speeding up the time computation) while showing a progress bar is
findres cre.zmap inventory.xml parameters.yaml results --phase_file=phases_hypoinv.txt --phase_type=hypoinv --taup_model=ncmodel --progress
The command to analyse the data without graphics (speeding up the time computation) and producing the HypodDD output file while showing a progress bar is
findres cre.zmap inventory.xml parameters.yaml results --phase_file=phases_hypoinv.txt --phase_type=hypoinv --taup_model=ncmodel --progress --hypodd
The Hypodd output file can be used to locate the events. HypoDD software (Waldhauser, F., and W.L. Ellsworth 2000) must be installed. To run an exmaple you can accesss the following directory:
cd /data/california/Hypodd_svd_1_RES
and type:
HypoDD hypoDD.inp
The numerical parameters are set using the parameters.yaml
file. The name of the fields are self-explicative and more
extensive information can be found in the provided example.
References
Sugan, M., Campanella, S., Vuan, A., and Shakibay Senobari, N., (2022). A Python Code for detecting true Repeating Earthquakes from Self-similar Waveforms (FINDRES). Submitted
Shakibay Senobari, N., and Funning, G. J., (2019). Widespread Fault Creep in the Northern San Francisco Bay Area Revealed by Multistation Cluster Detection of Repeating Earthquakes, Geophysical Research Letters, 46(12), 6425-6434, https://doi.org/10.1029/2019GL082766.
Chen, K. H., Nadeau, R. M, and Rau, R.-J. (2008). Characteristic repeating microearthquakes on an arc-continent collision boundary – the Chihshang fault of eastern Taiwan, Earth Planet. Sci. Lett., 276, 262–272, https://doi.org/10.1016/j.epsl.2008.09.021.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.