分布式光纤高密度地震观测揭示门源M_(S)6.9地震余震活动与隐伏断层  

Detecting aftershocks of the M_(S)6.9 Menyuan earthquake and mapping blind faults by a distributed acoustic sensing array with an existing fiber-optic cable

作　　者：包丰 曾祥方 林融冰 迟本鑫 吕昊 沙成宁[2] Feng Bao;Xiangfang Zeng;Rongbin Lin;Benxin Chi;Hao Lü;Chengning Sha(State Key Laboratory of Geodesy and Earth’s Dynamics,Innovation Academy for Precision Measurement Science and Technology,Chinese Academy of Sciences,Wuhan 430077,China;Qinghai Earthquake Agency,Xining 810001,China)

机构地区：[1]中国科学院精密测量科学与技术创新研究院,大地测量与地球动力学国家重点实验室,武汉430077 [2]青海省地震局,西宁810001

出　　处：《科学通报》2022年第27期3340-3347,共8页Chinese Science Bulletin

基　　金：国家重点基础研究发展计划(2021YFA0716800);国家自然科学基金(41974067,42142008)资助。

摘　　要：高密度的地震观测台阵对地震监测和结构成像都具有重要意义,利用分布式光纤地震传感技术结合现有通讯光缆可以快速构建高密度地震监测网.2022年门源M_(S)6.9地震发生后,利用震中区20.8 km通讯光缆构建了间距10 m的高密度地震台阵,记录到了丰富的余震信号.发展了地震自动检测算法,并在3天观测中检测到了166个余震,与固定台网目录形成互补.在高密度地震波场记录中识别出多个断层相关信号,推测其为断层散射波,由此发现了门源盆地内部的多个断层.研究表明,分布式光纤地震传感技术结合地震自动检测算法和波场分析技术可以为发震趋势判断及隐伏断层探测提供可靠参考的数据,在高原重要基础设施沿线的地震灾害研究中具有良好的应用前景.Dense spatial sampling of the seismic wave field is required for high-resolution structure imaging and comprehensive earthquake monitoring.However,deployment of a dense seismic array in many situations faces challenges,such as short timing and harsh environments.Distributed acoustic sensing(DAS)with existing telecom fiber-optic cable provides an economical solution.On January 8,2022,an M_(S)6.9 earthquake struck Menyuan County,Qinghai,China,causing extensive destruction.Several important infrastructures were destroyed,especially the bridge and tunnel of the Lanzhou-Xinjiang high-speed railway.As the main shock occurred in a mountainous area on the northeastern margin of the Tibetan Plateau,the aftershock monitoring ability of permanent seismograph network was inadequate.In order to enhance monitoring capability,especially for early aftershocks,researchers deployed a DAS interrogator that turned a 20.8 km long fiber-optic cable into a seismic array with a spacing of 10 m one day after the main shock.The observation system consisted of more than 2000 sensors,and abundant aftershock signals were recorded.Because early aftershock activity is of great significance for characterizing the morphology of the seismogenic fault,our initial effort focused on earthquake detection.In this paper,an algorithm(ADE-Mini)based on machine learning and convolutional neural network was developed for automatic earthquake detection.It converted the DAS array observations into images for automatically extracting their features with the AI algorithm.Only a small number of positive samples were required for training the ADE-Mini network.After detection,a total of 166 aftershocks were detected from January 10 to 13,36%of which were missed in the China Earthquake Network Center(CENC)catalog.However,there are a number of missed detections due to the fast attenuation of DAS monitoring ability with distance and the small coverage area of the fiber-optic cable in the study area.The fiber-optic cable used in this study was distributed almost line

关 键 词：门源地震 分布式光纤地震传感 高密度地震台阵 余震监测 断层探测 

分 类 号：P315.7[天文地球—地震学]