@article{81a3e9ca67dd4dffbaf9bed0a12df161,
title = "Deep Learning for Picking Seismic Arrival Times",
abstract = "Arrival times of seismic phases contribute substantially to the study of the inner working of the Earth. Despite great advances in seismic data collection, the usage of seismic arrival times is still insufficient because of the overload manual picking tasks for human experts. In this work we employ a deep-learning method (PickNet) to automatically pick much more P and S wave arrival times of local earthquakes with a picking accuracy close to that by human experts, which can be used directly to determine seismic tomography. A large number of high-quality seismic arrival times obtained with the deep-learning model may contribute greatly to improve our understanding of the Earth's interior structure.",
keywords = "arrival times, deep learning, seismic tomography",
author = "Jian Wang and Zhuowei Xiao and Chang Liu and Dapeng Zhao and Zhenxing Yao",
note = "Funding Information: High-quality arrival-time data sets for this study were provided by the data centers of Tohoku University (http://www.aob.tohoku.ac.jp), the JMA Unified Earthquake Catalogue (https://hinetwww11.bosai.go.jp/auth/JMA/), the China Earthquake Administration (https://www.cea.gov.cn/), and the International Seismological Centre (http://www.isc.ac.uk). Waveform data for this study were provided by the Hi-net (http://www.hinet.bosai.go.jp), IRIS (https://www.iris.edu), the International Federation of Digital Seismograph Networks (https://www.fdsn.org), the Southern California Earthquake Data Center (http://scedc.caltech.edu/), and Data Management Centre of the China National Seismic Network at Institute of Geophysics, China Earthquake Administration (http://www.seisdmc.ac.cn). The TensorFlow libraries (Abadi et al.,; https://www.tensorflow.org) were used to train and build our deep-learning model. The Obspy (Beyreuther et al.,; Krischer et al.,; Megies et al.,) was used for reading, writing, and slicing the seismograms. Figures were generated using the free and open software GMT 4.5.3 (Wessel & Smith,) and Matplotlib (Hunter,). This study was financially supported by the National Key R&D Program of China (grant 2017YFC0601206), the National Natural Science Foundation of China (grants 41474043 and 41274089), and the Youth Innovation Promotion Association of CAS (2014058). Yehuda Ben-Zion (editor), Zachary Ross, and Qingkai Kong provided constructive review comments and suggestions that have improved the manuscript. The PickNet codes and the trained models associated with this paper will be available online (https://www.researchgate.net/profile/Jian_Wang78/publications). Publisher Copyright: {\textcopyright}2019. American Geophysical Union. All Rights Reserved.",
year = "2019",
doi = "10.1029/2019JB017536",
language = "English",
volume = "124",
pages = "6612--6624",
journal = "Journal of Geophysical Research: Solid Earth",
issn = "2169-9313",
publisher = "Wiley-Blackwell",
number = "7",
}