东祁连山北缘断裂带基于深度学习的密集台阵地震事件快速检测与定位研究

杨少博; 王炳文; 高级; 张海江

doi:10.11899/zzfy20220104

东祁连山北缘断裂带基于深度学习的密集台阵地震事件快速检测与定位研究

doi: 10.11899/zzfy20220104

杨少博^1,,
王炳文¹,
高级^{1, 2, ,},
张海江^{1, 2}

1.
中国科学技术大学地球和空间科学学院, 合肥 230026
2.
安徽蒙城地球物理国家野外科学观测研究站, 合肥 230026

基金项目: 国家重点研发计划（2018YFC1504102）

详细信息

作者简介:
杨少博，男，生于1996年。博士。主要从事人工智能在地震学中的应用研究工作。E-mail：yang0123@mail.ustc.edu.cn

通讯作者:
高级，男，生于1983年。副研究员，硕士生导师。主要从事近地表地球物理成像及联合反演研究。E-mail：gaoji617@ustc.edu.cn

计量
- 文章访问数: 159
- HTML全文浏览量: 64
- PDF下载量: 39
- 被引次数: 0
出版历程
- 收稿日期: 2022-02-06
- 网络出版日期: 2022-05-31
- 刊出日期: 2022-03-31

Rapid Earthquake Detection and Location for Dense Array Data in the Fault Zone of the Northern Margin of the East Qilian Mountains Based on Deep Learning

Yang Shaobo^1
,,
Wang Bingwen¹,
Gao Ji^{1, 2
, ,},
Zhang Haijiang^{1, 2}

1.
School of Earth and Space Sciences, University of Science and Technology of China, Hefei 230026, China
2.
Mengcheng National Geophysical Observatory, University of Science and Technology of China, Hefei 230026, China

摘要

摘要: 为监测东祁连山北缘断裂带附近的地震活动性，布设包含240台短周期地震仪的面状密集台阵，进行约30 d的连续观测。首先使用基于深度学习的多台站地震事件检测算法（CNNDetector）进行地震事件检测，然后使用震相拾取网络（PhaseNet）对地震事件进行P波和S波到时拾取，其次使用震相关联算法（REAL）进行震相关联及初定位，最后使用双差定位（hypoDD）进行地震重定位，最终的精定位地震目录中共有517个地震。在密集台阵观测期间，中国地震台网正式地震目录中共有39个位于台阵内的地震事件，相比而言，密集台阵检测到大量小于0级的地震。因此通过布设密集台阵，可提高活动断裂微地震活动性的监测能力。与历史地震空间分布相比，密集台阵地震精定位分布具有较好的一致性，表现出更明显的线性分布特征。基于地震分布，发现研究区域存在与地表断层迹线走向不同的隐伏活跃断裂。
- 密集台阵 /
- 深度学习 /
- 地震定位 /
- 活动断裂
Abstract: For monitoring the earthquake activity of the northern margin of east Qilianshan fault belt, we have deployed a dense seismic array consisting of 240 short-period seismic instruments for about one month observation. We first used the deep learning based multi-station event detector (CNNDetector) algorithm to detect seismic events. Then we used PhaseNet to pick first P and S arrivals for these events. To remove unreliable picks, we adopted the REAL algorithm for phase association and coarse event location. For the final step, we relocated 517 events by the double-difference seismic location method (hypoDD). For the same monitoring period, the China Earthquake Administration (CEA) catalog monitored 39 seismic events within the array. In comparison, our dense seismic array detected more weak events with magnitudes lower than zero, thus having higher monitoring ability for earthquake activity. Compared to historical earthquakes, seismic events detected by our dense array have consistent spatial distributions but more linear features. Based on the earthquake locations, we find there are active blind faults that have different strikes with surface fault traces.
- Dense array /
- Deep learning /
- Earthquake location /
- Active fracture

HTML全文

图 1 研究区域台站及历史地震事件分布

Figure 1. Distribution of stations and historical earthquakes in the study area

下载: 全尺寸图片幻灯片

图 2 CNNDetector检测出的2个地震事件波形

Figure 2. Waveforms of two seismic events detected by CNNDetector

下载: 全尺寸图片幻灯片

图 3 PhaseNet到时拾取结果和不同台站接收到的事件317的波形图

Figure 3. Examples of PhaseNet arrival time picking results and waveforms of event 317 received at different stations

下载: 全尺寸图片幻灯片

图 4 地震定位使用的一维速度模型

Figure 4. 1D velocity models used for earthquake location

下载: 全尺寸图片幻灯片

图 5 REAL震相关联及初定位结果

Figure 5. Seismic phase association and earthquake location results by the REAL algorithm

下载: 全尺寸图片幻灯片

图 6 HypoDD定位结果

Figure 6. HypoDD location results

下载: 全尺寸图片幻灯片

图 7 HypoDD 迭代过程及走时残差分布

Figure 7. HypoDD iterative process and residual distribution

下载: 全尺寸图片幻灯片

图 8 震级分布与对比

Figure 8. Magnitude distribution and comparison

下载: 全尺寸图片幻灯片

参考文献(26)

[1]	郭鹏, 韩竹军, 安艳芬等, 2017. 冷龙岭断裂系活动性与2016年门源6.4级地震构造研究. 中国科学: 地球科学, 60（5）: 929—942. Guo P., Han Z. J., An Y. F., et al., 2017. Activity of the Lenglongling fault system and seismotectonics of the 2016 M_S 6.4 Menyuan earthquake. Science China Earth Sciences, 60（5）: 929—942. (in Chinese)
[2]	郭鹏, 2019. 北祁连山冷龙岭断裂大震复发行为与危险性研究. 北京: 中国地震局地质研究所. Guo P., 2019. Earthquake recurrence behavior and seismic hazards of the Lenglongling Fault, northern Qilian Shan. Beijing: Institute of Geology, China Earthquake Administrator. (in Chinese)
[3]	侯康明, 1998. 皇城-双塔断裂带的几何分段及运动学特征. 华南地震, 18（3）: 28—34 Hou K. M. , 1998. The geometric segmentation and kinematics characteristics of Huangcheng-Shuangta fault zone. South China Journal of Seismology, 18(3): 28—34. (in Chinese)
[4]	戚帮申, 2014. 祁连山新生代构造地貌演化特征. 北京: 中国地质大学（北京）. Qi B. S., 2014. Tectonic and geomorphological evolution features in Qilian Mountains during the Cenozoic. Beijing: China University of Geosciences Beijing. (in Chinese)
[5]	莘海亮, 2020. 中国大陆岩石圈地震体波三维走时速度成像与地震定位研究. 合肥: 中国科学技术大学. Xin H. L., 2020. Three-dimensional seismic travel time tomography of lithosphere of continental China and earthquake relocations. Hefei:University of Science and Technology of China. (in Chinese)
[6]	张培震, 沈正康, 王敏等, 2004. 青藏高原及周边现今构造变形的运动学. 地震地质, 26（3）: 367—377 doi: 10.3969/j.issn.0253-4967.2004.03.002 Zhang P. Z. , Shen Z. K. , Wang M. , et al. , 2004. Kinematics of present-day tectonic deformation of the Tibetan Plateau and its vicinities. Seismology and Geology, 26(3): 367—377. (in Chinese) doi: 10.3969/j.issn.0253-4967.2004.03.002
[7]	赵明, 陈石, 2021a. 基于深度学习的地震检测模型在区域台网的泛化性研究. 地震, 41（1）: 166—179 Zhao M. , Chen S. , 2021a. The Generalization ability research of deep learning algorithm in seismic phase detection of regional seismic network. Earthquake, 41(1): 166—179. (in Chinese)
[8]	赵明, 唐淋, 陈石等, 2021b. 基于深度学习到时拾取自动构建长宁地震前震目录. 地球物理学报, 64（1）: 54—66 Zhao M. , Tang L. , Chen S. , et al. , 2021b. Machine learning based automatic foreshock catalog building for the 2019 M_S 6.0 Changning, Sichuan earthquake. Chinese Journal of Geophysics, 64(1): 54—66. (in Chinese)
[9]	Han S. C. , Zhang H. J. , Xin H. L. , et al. , 2022. USTClitho2.0: updated unified seismic tomography models for continental china lithosphere from joint inversion of body‐wave arrival times and surface‐wave dispersion data. Seismological Research Letters, 93(1): 201—215. doi: 10.1785/0220210122
[10]	Kissling E. , Ellsworth W. L. , Eberhart‐Phillips D. , et al. , 1994. Initial reference models in local earthquake tomography. Journal of Geophysical Research: Solid Earth, 99(B10): 19635—19646. doi: 10.1029/93JB03138
[11]	Li Z. F. , Peng Z. G. , Hollis D. , et al. , 2018. High-resolution seismic event detection using local similarity for Large-N arrays. Scientific Reports, 8(1): 1646. doi: 10.1038/s41598-018-19728-w
[12]	Liu M. , Zhang M. , Zhu W. Q. , et al. , 2020. Rapid characterization of the July 2019 Ridgecrest, California, earthquake sequence from raw seismic data using machine‐learning phase picker. Geophysical Research Letters, 47(4): e2019GL086189.
[13]	McBrearty I. W. , Delorey A. A. , Johnson P. A. , 2019. Pairwise association of seismic arrivals with convolutional neural networks. Seismological Research Letters, 90(2A): 503—509. doi: 10.1785/0220180326
[14]	Meng H. R. , Ben-Zion Y. , 2018. Detection of small earthquakes with dense array data: Example from the San Jacinto fault zone, southern California. Geophysical Journal International, 212(1): 442—457. doi: 10.1093/gji/ggx404
[15]	Mousavi S. M. , Ellsworth W. L. , Zhu W. Q. , et al. , 2020. Earthquake transformer—an attentive deep-learning model for simultaneous earthquake detection and phase picking. Nature Communications, 11(1): 3952. doi: 10.1038/s41467-020-17591-w
[16]	Perol T. , Gharbi M. , Denolle M. , 2018. Convolutional neural network for earthquake detection and location. Science Advances, 4(2): e1700578. doi: 10.1126/sciadv.1700578
[17]	Ross Z. E. , Meier M. A. , Hauksson E. , et al. , 2018. Generalized seismic phase detection with deep learning. Bulletin of the Seismological Society of America, 108(5A): 2894—2901. doi: 10.1785/0120180080
[18]	Ross Z. E. , Trugman D. T. , Hauksson E. , et al. , 2019a. Searching for hidden earthquakes in southern California. Science, 364(6442): 767—771. doi: 10.1126/science.aaw6888
[19]	Ross Z. E. , Yue Y. S. , Meier M. A. , et al. , 2019b. PhaseLink: a deep learning approach to seismic phase association. Journal of Geophysical Research: Solid Earth, 124(1): 856—869. doi: 10.1029/2018JB016674
[20]	Ross Z. E. , Cochran E. S. , Trugman D. T. , et al. , 2020. 3D fault architecture controls the dynamism of earthquake swarms. Science, 368(6497): 1357—1361. doi: 10.1126/science.abb0779
[21]	Waldhauser F. , Ellsworth W. L. , 2000. A double-difference earthquake location algorithm: method and application to the northern Hayward fault, California. Bulletin of the Seismological Society of America, 90(6): 1353—1368. doi: 10.1785/0120000006
[22]	Wang R. J. , Schmandt B. , Zhang M. , et al. , 2020. Injection‐induced earthquakes on complex fault zones of the Raton Basin illuminated by machine‐learning phase picker and dense nodal array. Geophysical Research Letters, 47(14): e2020GL088168.
[23]	Xiao Z. W. , Wang J. , Liu C. , et al. , 2021. Siamese earthquake transformer: a pair‐input deep‐learning model for earthquake detection and phase picking on a seismic array. Journal of Geophysical Research: Solid Earth, 126(5): e2020JB021444.
[24]	Yang S. B. , Hu J. , Zhang H. J. , et al. , 2021. Simultaneous earthquake detection on multiple stations via a convolutional neural network. Seismological Research Letters, 92(1): 246—260. doi: 10.1785/0220200137
[25]	Zhang M. , Ellsworth W. L. , Beroza G. C. , 2019. Rapid earthquake association and location. Seismological Research Letters, 90(6): 2276—2284. doi: 10.1785/0220190052
[26]	Zhu W. Q. , Beroza G. C. , 2019. PhaseNet: a deep-neural-network-based seismic arrival-time picking method. Geophysical Journal International, 216(1): 261—273.