Machine-learning based toolbox to characterize seismic phases, i.e. phase detection, phase classification, and phase picking.
Currently, the quakephase toolbox can be used to largely enhance the pre-trained ML seismic phase picking models.
Install using pip (recommend)
pip install quakephaseInstall from source code
conda create -n quakephase python=3.9
conda activate quakephase
git clone https://github.com/speedshi/quakephase.git
cd quakephase
pip install .Follow the example scripts to use quakephase:
use_quakephase_example.py
use_quakephase_example.ipynb
Input parameters are explained and set in the parameter YAML file: "parameters.yaml". Feel free to explore and play with different parameters.
Note for data with distinct data sampling rates, the rescaling factor affects the performance most, choose it wisely. And if computing power is allowed, you can assemble different sets of rescaling rates to maximize the performance.
Please cite the following paper in your documents if you use quakephase in your work:
Peidong Shi, Men-Andrin Meier, Linus Villiger, Katinka Tuinstra, Paul Selvadural, Federica Lanza, Sanyi Yuan, Anne Obermann, Maria Mesimeri, Jannes Münchmeyer, Patrick Bianchi, and Stefan Wiemer. From labquakes to megathrusts: Scaling deep learning based pickers over 15 orders of magnitude. ESS Open Archive. April 12, 2024. DOI: 10.22541/essoar.171291855.54784565/v1