地震研究

采用基于深度学习的实时智能地震处理系统(RISP)检测2022年5月1—14日山东青州地震序列,构建自动检测目录,评估RISP系统在山东的适用性。RISP系统检测的地震数量多于人工编目的地震数量,匹配率达90.14%; 其中系统多检测的地震事件都含有地震信号,漏检测的事件由于不满足系统自动检测的触发条件,没有自动处理结果。此系统的定位精度也与人工处理结果相当,在发震时刻、震中位置、震源深度等方面具有较好的一致性; 但自动检测目录的地震震级的人工目录测量震级略大。根据自动检测的地震序列分布和震源机制反演推测此次地震的发震断层为张店—仁河断裂。检测结果表明:RISP系统检测的微震、小震多,定位精度准确,分析速度快,且不依赖于个人经验,可提供更加完整的地震目录。

The real-time intelligent seismic processing system(RISP)based on deep learning monitored the May 1-14,2022,earthquake sequence in Qingzhou,Shandong Province,and identified more earthquakes than manual identification,boasting a 90.14% accuracy rate.The exceeding earthquakes identified by RISP system do have seismic signals.The undetected events do not meet the trigger conditions of RISP system and have no automatic processing results.The positioning accuracy of RISP system is equal to that of the manual results,and other seismic parameters like original time,epicenter position,focal depth,etc.are in consistent with the manual results.However,the magnitude given by RISP system is slightly larger than that of the manual results.Based on the distribution and the inverted source mechanisms of the Qingzhou earthquake sequence,the Zhangdian-Renhe fault is supposed to be the seismogenic fault of the Qingzhou earthquake sequence.RISP system is proved to be capable of detecting more micro- and small-earthquakes.It has high accuracy for earthquake location,it does rapid analysis of seismic phases,but it does not depend on manual experience.

引言
1 数据和方法
2 结果与讨论
3 结论

图1 山东省地震台站分布图<br/>Fig.1 Distribution of the seismic stations in Shandong

图1 山东省地震台站分布图
Fig.1 Distribution of the seismic stations in Shandong

图2 自动目录(a)与人工目录(b)地震分布图<br/>Fig.2 Distribution of the earthquakes identified by RISP(a)and by manual analysis(b)

图2 自动目录(a)与人工目录(b)地震分布图
Fig.2 Distribution of the earthquakes identified by RISP(a)and by manual analysis(b)

图3 自动拾取震相的走时分布<br/>Fig.3 Distribution of the travel time of the automatically picked phases by RISP

图3 自动拾取震相的走时分布
Fig.3 Distribution of the travel time of the automatically picked phases by RISP

图4 自动目录与人工目录地震参数对比<br/>Fig.4 Comparison between the original time,epicentral location,focal-depth deviation,and magnitude deviation determined by RISP and by manual analysis

图4 自动目录与人工目录地震参数对比
Fig.4 Comparison between the original time,epicentral location,focal-depth deviation,and magnitude deviation determined by RISP and by manual analysis

图5 震中位置偏差较大的事件波形<br/>Fig.5 Waveforms with significant deviation in the earthquake epicenter

图5 震中位置偏差较大的事件波形
Fig.5 Waveforms with significant deviation in the earthquake epicenter

表1 震级偏差大于0.5级的地震事件<br/>Tab.1 The magnitudes(deviation larger than 0.5)of the same earthquakes from the automatic catalogue and the manual catalogue

表1 震级偏差大于0.5级的地震事件
Tab.1 The magnitudes(deviation larger than 0.5)of the same earthquakes from the automatic catalogue and the manual catalogue

图6 单台震级的偏差(匹配成功的振幅)<br/>Fig.6 Magnitude deviation at the same station (amplitudes successfully matched)

图6 单台震级的偏差(匹配成功的振幅)
Fig.6 Magnitude deviation at the same station (amplitudes successfully matched)

图7 自动检测目录与人工目录在不同震级段的地震数量对比<br/>Fig.7 Number of the earthquakes in different magnitude ranges from the automatic catalogue and the manual catalogue

图7 自动检测目录与人工目录在不同震级段的地震数量对比
Fig.7 Number of the earthquakes in different magnitude ranges from the automatic catalogue and the manual catalogue

图8 RISP系统自动识别双震叠加事件波形的震相<br/>Fig.8 Superposition of the waveforms of the double earthquakes

图8 RISP系统自动识别双震叠加事件波形的震相
Fig.8 Superposition of the waveforms of the double earthquakes

图9 拟合残差和震源机制解随震源深度的变化<br/>Fig.9 Fitted errors and focal mechanisms at different focal depths

图9 拟合残差和震源机制解随震源深度的变化
Fig.9 Fitted errors and focal mechanisms at different focal depths

图10 CAP方法反演地震波形拟合<br/>Fig.10 Waveform-fitting inversion with cut-and-paste method

图10 CAP方法反演地震波形拟合
Fig.10 Waveform-fitting inversion with cut-and-paste method

图11 自动目录地震分布及主震震源机制解图<br/>Fig.11 Distribution of the aftershocks given by the automatic catalogue and the focal mechanism of the main shock

图11 自动目录地震分布及主震震源机制解图
Fig.11 Distribution of the aftershocks given by the automatic catalogue and the focal mechanism of the main shock

晁洪太,王琦,李家灵,等.1997.山东省地震构造图、山东省新构造图及其说明书[M].济南:山东省地图出版社.
Chao H T,Wang Q,Li J L,et al.1997.Earthquake tectonic map of Shandong Province,new tectonic map of Shandong Province and manual[M].Jinan:Shandong Province Curtographic Publishing House.(in Chinese)
戴宗辉,李冬梅,王鹏,等.2022.2020年2月18日长清MS4.1地震发震构造研究[J].地震,42(1):111-121.
Dai Z H,Li D M,Wang P,et al.2022.Study on the seismotectonics of the MS4.1 Changqing earthquake on 18 February,2022[J].Earthquake,42(1):111-121.(in Chinese)
洪德全,王行舟,韩立波,等.2011.用CAP方法研究安庆MS4.8级地震震源机制[J].中国地震,27(2):207-214.
Hong D Q,Wang X Z,Han L B,et al.2011.Source mechanism of the 2011 MS4.8 Anqing earthquake determined from CAP method[J].Earthquake Research in China,27(2):207-214.(in Chinese)
李姣,姜金钟,杨晶琼.2020.2017年漾濞MS4.8和MS5.1地震序列的微震检测及重定位[J].地震学报,42(5):527-542.
Li J,Jiang J Z,Yang J Q.2020.Microseismic detection and relocation of the 2017 Yangbi MS4.8 and MS5.1 earthquake sequences,Yunnan[J].Acta Seismologica Sinica,42(5):527-542.(in Chinese)
廖诗荣,张红才,范莉苹,等.2021.实时智能地震处理系统研发及其在2021年云南漾濞地震MS6.4地震中的应用[J].地球物理学报,64(10):3632-3645.
Liao S R,Zhang H C,Fan L P,et al.2021.Development of real-time intelligent seismic processing system and its application in the 2021 Yunnan Yangbi MS6.4 earthquake[J].Chinese Journal of Geophysics,64(10):3632-3645(in Chinese).
刘媛,颜利君,廖诗荣,等.2022.实时智能地震处理系统在2020年四川石渠M5.6地震序列中的应用[J].地质论评,68(4):1375-1381.
Liu Y,Yan L J,Liao S R,et al.2022.Application of real-time intelligent seismic processing system in the 2020 M5.6 earthquake in Shiqu,Sichuan[J].Geological Review,68(4):1375-1381.(in Chinese)
王纪强,王冬雷,鹿子林,等.2020.双山—李家庄断裂地表破裂特征与最新活动性研究[J].地震,40(4):115-128.
Wang J Q,Wang D L,Lu Z L,et al.2020.The surface rupture characteristics and latest activities of the Shuangshan-Lijiazhuang Fault[J].Earthquake,40(4):115-128.(in Chinese)
王祖东,杨晓鹏,尹欣欣,等.2022.2022年1月8日青海门源MS6.9地震余震序列自动处理结果探讨[J].地震工程学报,44(2):408-414.
Wang Z D,Yang X P,Yin X X,et al.2022.Discussion on the automatic processing results of aftershock sequence of Menyuan,Qinghai MS6.9 earthquake on 8 January,2022[J].China Earthquake Engineering Journal,44(2):408-414.(in Chinese)
颜利君,刘媛,廖诗荣,等.2022.2022年6月10日四川马尔康地震序列实时智能检测结果分析与研究[J].地震工程学报,44(6):1450-1458.
Yan L J,Liu Y,Liao S R,et al.2022.Real time automatic detection results of the Maerkang,Sichuan earthquake sequence on June 10,2022[J].China Earthquake Engineering Journal,44(6):1450-1458.(in Chinese)
张演,姜金钟,王光明,等.2024.云南盐津地区两种地震序列的地震活动性分析[J].地震研究,47(2):178-190.
Zhang Y,Jiang J Z,Wang G M,et al.202.Analysis of the seismicity of two types of earthquake sequences in Yanjin,Yunnan[J].Journal of Seismological Research,47(2):178-190.(in Chinese)
Lomax A,Satriano C,Vassallo M.2012.Automatic picker developments and optimization:filter Picker-a robust,broadband picker for realtime seismic monitoring and earthquake early warning[J].Seismological Research Letters,83(3):531-540.
Ross Z E,Meier M A,Hauksson E,et al.2018.Generalized seismic phase detection with deep learning[J].Bulletin of the Seismological Society of America,108(5A):2894-2901.
Scafidi D,Spallarossa D,Ferretti G,et al.2019.A complete automatic procedure to compile reliable seismic catalogs and travel-time and strong-motion parameters datasets[J].Seismological Research Letters,90(3):1308-1317.
Spallarossa D,Cattaneo M,Scafidi D,et al.2020.An automatically generated high-resolution earthquake catalogue for the 2016-2017 Central Italy seismic sequence,including P and S phase arrival times[J].Geophysical Journal International,225:555-571.
Tamaribuchi K.2018.Evaluation of automatic hypocenter determination in thr JMA unified catalog[J].Earth Planets and Space,70(1):1-13.
Yeck W L,Patton J M,Johnson C E,et al.2019.GLASS3:A standalone multiscale seismic detection associator[J].Bulletin of the Seismological Society of America,109(4):1-10.
Zhu W Q,Beroza G C.2019.PhaseNet:a deep neural network based seismic arrival time picking method[J].Geophysical Journal International,216(1):261-273.

备注

引言

1 数据和方法

2 结果与讨论

3 结论

期刊信息