Deriving focal mechanism solutions of small to moderate earthquakes in Sichuan, China via a deep learning method

Abstract

As one of the most seismically active regions, Sichuan basin is a key area of seismological studies in China. This study applies a neural network model with attention mechanisms, simultaneously picking the P-wave arrival times and determining the first-motion polarity. The polarity information is subsequently used to derive source focal mechanisms. The model is trained and tested using small to moderate earthquake data from June to December 2019 in Sichuan. We apply the trained model to predict first-motion polarity directions of earthquake recordings in Sichuan from January to May 2019, and then derive focal mechanism solutions using HASH algorithm with predicted results. Compared with the source mechanism solutions obtained by manual processing, the deep learning method picks more polarities from smaller events, resulting in more focal mechanism solutions. The catalog documents focal mechanism solutions of 22 events (M_L 2.6-4.8) from analysts during this period, whereas we obtain focal mechanism solutions of 53 events (M_L 1.9-4.8) through the deep learning method. The derived focal mechanism solutions for the same events are consistent with the manual solutions. This method provides an efficient way for the source mechanism inversion of small to moderate earthquakes in Sichuan region, with high stability and reliability.

Keywords

Deep learning / Focal mechanism solutions / Small-to-moderate earthquake / First-motion polarity / Attention mechanism / Sichuan

CRediT authorship contribution statement

Chen Zhang: Writing - original draft, Software, Methodology, Investigation, Formal analysis, Data curation. Ji Zhang: Methodology. Jie Zhang: Writing - review & editing, Supervision.

Data and resources

The data used in this study, including training and testing dataset, and the earthquake data in the Sichuan and the surrounding areas of China in 2019 were provided by China Seismic Experimental Site (CSES), available at the website (http://www.cses.ac.cn/sycdt/index.shtml, in Chinese). The source code of the neural network model can be obtained from the website (https://github.com/LolitaZJ/APP_MASTER). The HASH method to compute earthquake first-motion focal mechanism solutions was downloaded in the website (http://quake.wr.usgs.gov/research/software/#HASH).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Author agreement and Acknowledgments

I would like to declare on behalf of my co-authors that the work described was original research that has not been published previously and is not under consideration for publication elsewhere, in whole or in part. All the authors listed have approved the manuscript that is enclosed. We thank the National Key R&D Program of China (2021YFC3000701) for the financial support. We also appreciate China Seismic Experimental Site in Sichuan-Yunnan (CSES-SY) for offering the data in this study. We would like to extend our sincere gratitude to the editor, Prof. Zhigang Peng, and the two anonymous reviewers for their constructive comments.

References

[1] Allen R., 1982. Automatic phase pickers: their present use and future prospects. Bull. Seismol. Soc. Am. 72 (6B), S225-S242.

[2] Baker B., Holt M.M., Pankow K.L., Koper K.D., Farrell J., 2021. Monitoring the 2020 Magna, Utah, earthquake sequence with nodal seismometers and machine learning. Seismol Res. Lett. 92 (2), 787-801.

[3] Beaucé E., Frank W.B., Romanenko A., 2017. Fast matched filter (FMF): an efficient seismic matched-filter search for both CPU and GPU architectures. Seismol Res. Lett. 89 (1), 165-172.

[4] Brown J.R., Beroza G.C., Shelly D.R., 2008. An autocorrelation method to detect low frequency earthquakes within tremor. Geophys. Res. Lett. 35 (16), L16305.

[5] Chen C., Holland A.A., 2016. PhasePApy: a robust pure Python package for automatic identification of seismic phases. Seismol Res. Lett. 87 (6), 1384-1396.

[6] Chen Y., Saad O.M., Savvaidis A., Zhang F., Chen Y., Huang D., Li H., Aziz Zanjani F., 2024. Deep learning for p-wave first-motion polarity determination and its application in focal mechanism inversion. IEEE Trans. Geosci. Remote Sens. 62, 1-11.

[7] Cheng Y., Allen R.M., Taira T., 2023a. A new focal mechanism calculation algorithm (REFOC) using inter-event relative radiation patterns: application to the earthquakes in the Parkfield area. J. Geophys. Res. Solid Earth 128 (3), e2020JB019577.

[8] Cheng Y., Hauksson E., Ben-Zon Y., 2023b. Refined earthquake focal mechanism catalog for southern California derived with deep learning algorithms. J. Geophys. Res. Solid Earth 128 (2), e2022JB025975.

[9] Chu R., Zhu L., Helmberger D.V., 2009. Determination of earthquake focal depths and source time functions in central Asia using teleseismic P waveforms. Geophys. Res. Lett. 36 (17), L17317.

[10] Dang P., Liu Q., Song J.S., 2020. Simulation of the Jiuzhaigou, China, earthquake by stochastic finite-fault method based on variable stress drop. Nat. Hazards 103, 2295-2321.

[11] Dreger D.S., Helmberger D.V., 1993. Determination of source parameters at regional distances with three-component sparse network data. J. Geophys. Res. Solid Earth 98 (B5), 8107-8125.

[12] Du G.B., Wu Q.J., Zhang X.M., Zhang R.Q., 2021. Activity characteristics of moderate and small earthquakes and fine crustal feature beneath Weiyuan, Sichuan and adjacent areas. Chin. J. Geophys. 64 (11), 3983-3996 (in Chinese).

[13] Feng T., Mohanna S., Meng L., 2022. EdgePhase: a deep learning model for multistation seismic phase picking. Geochem. Geophy. Geosy. 23 (1), e2022GC01045.

[14] Hara S., Fukahata Y., Iio Y., 2019. P-wave first-motion polarity determination of waveform data in western Japan using deep learning. Earth Planets Space 71 (1), 127.

[15] Hardebeck J.L., Shearer P.M., 2002. A new method for determining first-motion focal mechanisms. Bull. Seismol. Soc. Am. 92 (6), 2264-2276.

[16] Hardebeck J.L., Shearer P.M., 2003. Using S/P amplitude ratios to constrain the focal mechanisms of small earthquakes. Bull. Seismol. Soc. Am. 93 (6), 2434-2444.

[17] Hu X.H., Sheng S.Z., Wan Y.G., Wang G.J., Arno Z., Shi B.X., Jiang D.W., 2020. Study on focal mechanism and post-seismic tectonic stress field of the Changning, Sichuan, earthquake sequence on June 17th 2019. Prog. Geophys. 35 (5), 1675-1681 (in Chinese).

[18] Imanishi K., Ando R., Kuwahara Y., 2012. Unusual shallow normal-faulting earthquake sequence in compressional northeast Japan activated after the 2011 off the Pacific coast of Tohoku earthquake. Geophys. Res. Lett. 39 (9), L09306.

[19] Jia K., Zhou S., Zhuang J., Jiang C., Guo Y., Gao Z., Gao S., Ogata Y., Song X., 2020. Nonstationary background seismicity rate and evolution of stress changes in the Changning salt mining and shale-gas hydraulic fracturing region, Sichuan Basin, China. Seismol Res. Lett. 91 (4), 2170-2181.

[20] Jin Z.T., Wan Y.G., Liu Z.C., Huang J.C., Li Y., Yang F., 2019. The static stress triggering influences of the 2017M_S 7.0 Jiuzhaigou earthquake on neighboring areas. Chin. J. Geophys. 62 (4), 1282-1299 (in Chinese).

[21] Kagan Y.Y., 1991. 3-D rotation of double-couple earthquake sources. Geophys. J. Int. 106 (3), 709-716.

[22] Kingma D.P., Ba J., 2014. Adam: a method for stochastic optimization. https://arXiv.org/abs/1412.6980.

[23] Kuang W., Yuan C., Zhang J., 2021. Real-time determination of earthquake focal mechanism via deep learning. Nat. Commun. 12 (1), 1432.

[24] Lei J., Zhao D., 2009. Structural heterogeneity of the Longmenshan fault zone and the mechanism of the 2008 Wenchuan earthquake (Ms 8.0). Geochem. Geophys. Geosy. 10 (10), Q10010.

[25] Lei X.L., Wang Z.W., Su J.R., 2019. Possible link between long-term and short-term water injections and earthquakes in salt mine and shale gas site in Changning, south Sichuan Basin, China. Earth Planet. Phys. 3 (6), 510-525.

[26] Long F., Zhang Y.J., Wen X.Z., Ni S.D., Zhang Z.W., 2010. Focal mechanism solutions of M_L≥4.0 events in the M_S 6.1 Panzhihua-Huili earthquake sequence of Aug 30, 2008. Chin. J. Geophys. 53 (12), 2852-2860 (in Chinese).

[27] Li X., Gao K., Feng Y., Zhang C., 2022. 3D geomechanical modeling of the Xianshuihe fault zone, SE Tibetan Plateau: implications for seismic hazard assessment. Tectonophysics 839 (20), 229546.

[28] Liu S., Yang Y., Deng B., Zhong Y., Wen L., Sun W., Li Z., Jansa L., Li J., Song J., Zhang X., Peng H., 2020. Tectonic evolution of the Sichuan basin, southwest China. Earth Sci. Rev. 213, 103470.

[29] Liu Y., Yao H., Zhang H., Fang H., 2021. The community velocity model v. 1.0 of southwest China, constructed from joint body- and surface-wave travel-time tomography. Seismol Res. Lett. 92 (5), 2972-2987.

[30] Luo G., Liu M., 2010. Stress evolution and fault interactions before and after the 2008 Great Wenchuan earthquake. Tectonophysics 491 (1-4), 127-140.

[31] Mousavi S.M., Ellsworth W.L., Zhu W.Q., Chuang L.Y., Beroza G.C., 2020. Earthquake transformer-an attentive deep-learning model for simultaneous earthquake detection and phase picking. Nat. Commun. 11 (1), 3952.

[32] Mousavi S.M., Zhu W.Q., Ellsworth W., Beroza G., 2019. Unsupervised clustering of seismic signals using deep convolutional autoencoders. IEEE Geosci. Remote. Sens. Lett. 16 (11), 1693-1697.

[33] Mu X., Song J., Yang H., Huang J., Yao H., Tian B., 2023. High-resolution shallow structure along the Anninghe Fault zone, sichuan, China, constrained by active source tomography. Seismol Res. Lett. 95 (1), 408-420.

[34] Pennington C., Chen X., 2017. Coulomb stress interactions during the M_W 5.8 Pawnee sequence. Seismol Res. Lett. 88 (4), 1024-1031.

[35] Pei S., Niu F., Ben-Zon Y., Sun Q., Liu Y., Xue X., Su J., Shao Z., 2019. Seismic velocity reduction and accelerated recovery due to earthquakes on the Longmenshan fault. Nat. Geosci. 12, 387-392.

[36] Perol T., Gharbi M., Denolle M., 2018. Convolutional neural network for earthquake detection and location. Sci. Adv. 4 (2), e1700578.

[37] Pondrelli S., Salimbeni S., Ekström G., Morelli A., Gasperini P., Vannucci G., 2006. The Italian CMT dataset from 1977 to the present. Phys. Earth Planet. Inter. 159 (3-4), 286-303.

[38] Prechelt L., 1998. Early stopping-but when?. In: Neural Networks: Tricks of the Trade. Springer, Berlin, Heidelberg, pp. 55-69.

[39] Pugh D.J., White R.S., Christie P.A.F., 2016. Automatic Bayesian polarity determination. Geophys. J. Int. 206 (1), 275-291.

[40] Ross Z.E., Meier M.-A., Hauksson E., 2018. P wave arrival picking and first-motion polarity determination with deep learning. J. Geophys. Res. Solid Earth 123 (6), 5120-5129.

[41] Sokos E.N., Zahradnik J., 2008. ISOLA a Fortran code and a Matlab GUI to perform multiple-point source inversion of seismic data. Comput. Geosci. 34 (8), 967-977.

[42] Trugman D.T., Ross Z.E., Johnson P.A., 2020. Imaging stress and faulting complexity through earthquake waveform similarity. Geophy. Res. Lett. 47 (1), e2019GL085888.

[43] Uchide T., 2020. Focal mechanisms of small earthquakes beneath the Japanese islands based on first-motion polarities picked using deep learning. Geophys. J. Int. 223 (3), 1658-1671.

[44] Uchide T., Shearer P.M., Imanishi K., 2014. Stress drop variations among small earthquakes before the 2011 Tohoku-oki, Japan, earthquake and implications for the main shock. J. Geophys. Res. Solid Earth 119 (9), 7164-7174.

[45] Wang F., Wang M., Wang Y., Shen Z.K., 2015. Earthquake potential of the SichuanYunnan region, western China. J. Asian Earth Sci. 107, 232-243.

[46] Wang X., Zhan Z., 2020. Seismotectonics and fault geometries of the 2019 Ridgecrest sequence: insight from aftershock moment tensor catalog using 3-D Green's functions. J. Geophys. Res. Solid Earth 125 (5), e2020JB019577.

[47] Xu J., Shao Z.G., Liu J., Ji L.Y., 2019. Coulomb stress evolution and future earthquake probability along the eastern boundary of Sichuan-Yunnan block. Chin. J. Geophys. 62 (11), 4189-4213 (in Chinese).

[48] Xu Y.C., Guo Y.X., 2023. Relocation of the 2022M_S 6.0 Maerkang earthquakes warm in Sichuan Province and its seismic fault analyst. Seismol. Geol. 45 (4), 1006-1024 (in Chinese).

[49] Yang J., Chen J.H., Liu H.L., Zheng J.C., 2014. Comparison of two large earthquakes in China: the 2008 sichuan wenchuan earthquake and the 2013 sichuan lushan earthquake. Nat. Hazards 73 (10), 1127-1136.

[50] Yang W.Z., Hauksson E., Shearer P.M., 2012. Computing a large refined catalog of focal mechanisms for Southern California (1981-2010): temporal stability of the style of faulting. Bull. Seismol. Soc. Am. 102 (3), 1179-1194.

[51] Yi G.X., Long F., Liang M.J., Zhao M., Wang S.W., 2020. Geometry and tectonic deformation of seismogenic structures in the Rongxian-Weiyuan-Zizhong region, Sichuan Basin: insights from focal mechanism solutions. Chin. J. Geophys. 63 (9), 3275-3291 (in Chinese).

[52] Yoon C.E., O'Reilly O., Bergen K.J., Beroza G.C., 2015. Earthquake detection through computationally efficient similarity search. Sci. Adv. 1 (11), e1501057.

[53] Yu M.X., Wu Q.J., Qiang Z.Y., Bai L.S., 2023. Focal mechanism solutions and stress field characteristics of small and moderate earthquakes in Weiyuan area, southern Sichuan basin. Acta Seismologica Sinica 45 (6), 1025-1039 (in Chinese).

[54] Zhai Q., Peng Z., Chuang L.Y., Wu Y.-M., Hsu Y.-J., Wdowinski S., 2021. Investigating the impacts of a wet typhoon on microseismicity: a case study of the 2009 Typhoon Morakot in Taiwan based on a template matching catalog. J. Geophys. Res. Solid Earth 126 (12), e2021JB023026.

[55] Zhang J., Li Z.F., Zhang J., 2023. Simultaneous seismic phase picking and polarity determination with an attention-based neural network. Seismol Res. Lett. 94 (2A), 813-828.

[56] Zhang J., Zhang H., Chen E., Zheng Y., Kuang W., Zhang X., 2014. Real-time monitoring earthquake using a search engine method. Nat. Commun. 5 (1), 5664.

[57] Zhang X., Zhang J., Yuan C., Liu S., Chen Z., Li W., 2020. Locating induced earthquakes with a network of seismic stations in Oklahoma via a deep learning method. Sci. Rep. 10, 1941.

[58] Zhao B., Gao Y., Liu J., Liang S.S., Xu Z.G., Du G.B., 2019. Focal mechanism inversion and source depth locating of moderate-major earthquakes in the Sichuan region since 2010. Chin. J. Geophys. 62 (1), 130-142 (in Chinese).

[59] Zhao L.S., Helmberger D.V., 1994. Source estimation from broadband regional seismograms. Bull. Seismol. Soc. Am. 84 (1), 91-94.

[60] Zhao M., Xiao Z., Zhang M., Yang Y., Tang L., Chen S., 2023. DiTingMotion: a deeplearning first-motion-polarity classifier and its application to focal mechanism inversion. Front. Earth Sci. 11, 1103914.

[61] Zhu L., Helmberger D.V., 1996. Advancement in source estimation techniques using broadband regional seismograms. Bull. Seismol. Soc. Am. 86 (5), 1634-1641.

[62] Zhu L., Zhou X.F., 2016. Seismic moment tensor inversion using 3D velocity model and its application to the 2013 Lushan earthquake sequence. Phys. Chem. Earth 95, 10-18.

[63] Zoback M.L., 1992. First and second order patterns of stress in the lithosphere: the world stress map project. J. Geophys. Res. 97 (B8), 11703-11728.