포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project March

Transcription

1

2 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project March 20, 2019 대한지질학회 The Geological Society of Korea 포항지진정부조사연구단 Korean Government Commission on the Cause of the Pohang Earthquake Commission Chairperson and PI ( 총괄조사연구단장 ) Kang-Kun Lee ( 이강근 ), Seoul National University Korean Expert Research Team ( 국내조사단 ) In-Wook Yeo ( 여인욱, 국내조사단장 ), Chonnam National University Jin-Yong Lee ( 이진용 ), Kangwon National University Tae-Seob Kang ( 강태섭 ), Pukyong National University Junkee Rhie ( 이준기 ), Seoul National University Dong-Hoon Sheen ( 신동훈 ), Chonnam National University Chandong Chang ( 장찬동 ), Chungnam National University Moon Son ( 손문 ), Pusan National University In-Ky Cho ( 조인기 ), Kangwon National University Seokhoon Oh ( 오석훈 ), Kangwon National University Sukjoon Pyun ( 편석준 ), Inha University Sangwan Kim ( 김상완 ), Sejong University Overseas Research Advisory Committee (ORAC, 해외조사위원회 ) Shemin Ge, University of Colorado, USA William L Ellsworth, Stanford University, USA Domenico Giardini, ETH Zurich, Switzerland John Townend, Victoria University of Wellington, New Zealand Toshihiko Shimamoto, China Earthquake Administration, China

3

4 Organization of Korean Government Commission 포항지진정부조사연구단의구성 The Geological Society of Korea (GSK) Kang-Kun Lee, Seoul Nat l Univ. (Commission Chairperson & President of GSK) Korean Expert Research Team (KERT) (Team Leader: In-Wook Yeo) Overseas Research Advisory Committee (ORAC) (Co-chairs: Shemin Ge, Domenico Giardini) Shemin Ge, Univ. of Colorado William Ellsworth, Stanford Univ. Domenico Giardini, ETH Zurich Toshihiko Shimamoto, China Earthquake Admin. John Townend, Victoria Univ. Seismology Structural Geology & Geomechanics Hydrogeology Geophysical Exploration & Remote Sensing Tae-Seob Kang Pukyong Nat l Univ. Junkee Rhie Seoul Nat l Univ. Dong-Hoon Sheen Chonnam Nat l Univ. Moon Son Pusan Nat l Univ. Chandong Chang Chungnam Nat l Univ. In-Wook Yeo Chonnam Nat l Univ. Jin-Yong Lee Kangwon Nat l Univ. In-Ky Cho Kangwon Nat l Univ. Seokhoon Oh Kangwon Nat l Univ. Sukjoon Pyun Inha Univ. Sang-Wan Kim Sejong Univ. (Representatives from Korean Society of Earth and Exploration Geophysics) * Citizen representatives to commission: Kang-Hun Baek and Manjae Yang i

5

6 서문 포항지진과지열발전의연관성분석연구정부조사연구단은결과발표에앞서먼저지진피해로어려운환경속에서도우리조사연구단의연구과정을기다려주신포항시민들과조사연구과정에서자료제공과협조를해주신분들, 조사연구가원활하게진행되도록지원해주신관계자분들께감사드린다. 포항지진정부조사연구단 ( 대한지질학회주관, 조사연구단장이강근대한지질학회장, 서울대지구환경과학부교수 ) 은지진학, 수리지질학, 지질역학 / 구조지질학, 지구물리탐사등 4개분야의국내전문가로구성된 12명의국내조사단, 각분야의국제적전문가 5명으로구성된해외조사위원회 ( 해외조사연구자문단 ) (ORAC: Overseas Research Advisory Committee), 그리고자문단 2명으로구성되어 2018년 3월부터 1년간조사연구를수행하였다. 국내조사단참여전문가와해외조사연구자문단참여전문가는각자의전문분야에따라서조사연구에공동으로참여하여세부전문분야별조사연구결과를도출하였다. 해외조사위원회는수행된연구결과들을종합하여포항지진과지열발전의연관성에대한해외조사연구자문단차원의논의를독립적으로진행하고그결론을조사연구단장에게제출하도록하였다. 자문단은직접조사연구에참여하지는않고조사연구단의가야할방향에대해의견을제시하고문제를제기하는역할을수행하였다. 조사연구단장은국내조사단의조사연구결과와해외조사연구자문단의논의결과와결론을반영하여조사연구단차원의최종결론을도출하는과정을거쳐이번결과를발표하게되었다. 정부조사연구단은확보된자료를이용하여지진의원인과지열발전의연관성에대해다양한분석을수행하여연구결과를얻었고, 연구결과들은조사연구단의연구진들이종합보고서와학술회의발표, 그리고학술지논문등으로완전히공개해나갈것이다. 또한, 조사연구내용과자료들이총망라된종합보고서는조사연구단의종료시점인 2019 년 4월이후에조사연구단수행결과의평가가완료된이후에절차를거쳐서공개될것이다. 이번결과발표와함께공개되는요약보고서는여러연구성과중에서오늘발표하는결론에이르게된핵심내용을간추린요약보고서의형태이다. 이요약보고서는향후조사연구단의총괄결과보고서가공개되기전까지포항지진의원인에관한조사연구단의설명자료로활용될것이며대한지질학회홈페이지등을통해일반에공개될것이다. 그동안포항지진정부조사연구단이국내학자들과해외학자들로구성되어국제적인조사연구단으로활동할수있도록성원하면서조사연구단의노력을지켜봐주신모든분들께감사를드린다 년 3 월 20 일 포항지진정부조사연구단장대한지질학회장이강근 iii

7 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Preface The Korean Government Commission for the Analysis of the Relationship between the Pohang Earthquake and EGS (Enhanced Geothermal System) project would first like to express its gratitude to the Pohang citizens, who have patiently waited throughout the Commission s research process despite the hardships they faced due to the earthquake, those who facilitated and offered data for the research, and relevant authorities who have aided the study. The Government Commission on the Pohang Earthquake was organized and led by the Geological Society of Korea (Chairperson Kang-Kun Lee, President of the Geological Society of Korea, Professor of School of Earth and Environmental Sciences at Seoul National University) consisted of a National Committee with 12 experts across 4 academic disciplines including seismology, hydrogeology, geomechanics/ structural geology, and geophysical exploration, an Overseas Research Advisory Committee (ORAC) with 5 international experts, and an advisory board with 2 members. The Commission conducted the research and study, beginning in March 2018, for 1 year. The experts of the National Committee and the ORAC administrated their research according to their specialties and acquired findings in each area. The ORAC collected the findings, led independent discussions about the relationship between the Pohang earthquake and the EGS project, and submitted its conclusions to the Commission Chairperson. The Chairperson considered both the findings of the National Committee and the discussion conclusions of the ORAC to derive final conclusions of the research, which led to this publication. The Commission has reached conclusions on the relationship between the causes of the earthquake and the EGS project based on its findings, and the findings and conclusions will be completely published through the final report, conference presentations, and academic papers. In addition, the final report that contains all information on the research conducted and relevant data will be available sometime after April 2019, the official end of the Commission. The abridged report being released with this presentation contains the main and summarized conclusions that constitute the most relevant information to today s conclusions. This abridged report will be utilized as an explanatory resource for the causes of the Pohang earthquake until the final report is iv

8 서문 released and will be made available and downloadable to the public through the Geological Society of Korea website ( We offer our sincerest gratitude to everyone who has supported the work of the Korean Government Commission on the Pohang earthquake consisting of international and national scholars. March 20, 2019 Kang-Kun Lee, Ph.D. Korean Government Commission Chairperson President of the Geological Society of Korea Professor, Seoul National University v

9

10 조사내용및결과요약 1. 주요조사연구항목과내용 정부조사연구단이출범한 2018 년 3 월부터지난 1 년간수행한주요연구내용은다음과같다. 1) 지열발전실증연구팀에자료요청목록제출과자료확보 ( 지진관측자료, 시추자료, 지구물리탐사자료, 지열발전연구초기보고서, 수리자극시험자료등 ) 및분석 2) 현장지구물리탐사추가수행및기존자료해석, 포항지진전후의위성관측자료확보및분석 3) 포항지열발전실증연구시설부지를포함한주변지역의지질구조분석, 이와연계한시추기록및시추암편분석 4) 수리자극이실시되었던지열공 2개에대한시추공영상검층및 PTS 검층조사수행 5) 탄성파탐사자료, 시추공지진계자료, 경상분지속도모델등을이용하여속도구조모델구축 6) 2009 년 1월 1일이후부터포항지진본진발생까지의지진관측자료분석 7) EGS (Enhanced Geothermal System) 실증시설부지인근에서발생한지진들중에서 98개지진의진원결정및규모재산정을통한지진목록작성 8) 수리자극에의한물주입으로심부암반에가해진공극압및지하수변화분석 9) EGS 실증시설부지에서의지중응력상태분석 10) 지열공 (PX-1, PX-2) 의시추시기록된지질검층자료, 시추자료및이수활용자료등에대한분석 2. 주요조사연구결과 조사연구단은다음과같은각부분의조사연구결과들을도출한후, 이를종합하여포항지진의원인에관 한결론을도출하였다. 1) 포항지진원단층에대한지질구조 : 포항지진을발생시킨지진원단층은마이오세분지확장기간동안인장력에의해만들어진북동방향의여러공액상 (conjugate) 정단층들중에서곡강단층의반향단층 (antithetic faults) 중하나가현생응력장하에서우수향수평이동성분을가지는역단층으로재활성된것으로해석된다 ( 관련내용 Fig. 2-10). vii

11 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 2) 지구물리탐사 : 물리탐사자료에서포항지열발전실증연구부지인근에북북동방향의저비저항대가나타나는것으로보이나포항지역지진원이되는단층과의관계를명확하게파악하기어렵다 ( 관련내용 Figs. 3-2, 3-3, 3-4). 탄성파탐사자료를이용하여속도구조와경사구조를파악하였다 ( 관련내용 Fig. 3-10, Table 3-3). 위성관측자료로부터지표변위를분석하였고, 정적변형모델링을통해추정한단층면은지질구조분석과진원분포분석결과와유사하다. 3) 지진자료추출과발생의시간적분포 : EGS 실증부지를중심으로포항지역과그주변에서 2009 년 1월 1일부터 2017 년 11월 15일본진까지 520개지진을식별하였다. EGS 실증시설부지에서진앙거리 5 km 이내, 진원깊이 10 km 미만인지진들만모아서지진과인위적활동의시간적인상관성을파악하였다. EGS 실증부지근접지역에서는 2009 년 1월 1일부터 2015 년 10월까지지진이검출되지않았으나, PX-2 지열정굴착과정에서발생한이수누출 (mud loss) 과함께미소지진이검출되기시작하여그이후진행된 5차례수리자극과시간적으로밀접하게연관된지진들이검출되었다. 즉, EGS 실증연구부지인근에서발생한지진의시간적분포가 EGS 프로젝트의수행과정에서있었던활동과밀접하게연관된다 ( 관련내용 Figs. A-4-1, A-4-2). 4) 진원의위치결정과분포특징 : 진원위치결정을위해속도모델을구축하였다. 천부층의속도는탄성파탐사자료와시추공지진계의 S-P파주시차이를이용하여결정하였다. 심부속도구조는경상분지속도모델을사용하였다 년 1월 1일이후부터본진까지포항지역과그주변에서발생한것으로식별된 520개지진중에서 EGS 실증연구부지에서진앙거리가 5 km 이상인것, 진앙거리 5 km 이내이지만진원깊이가 10 km 보다깊은것을제외한 109개의지진을선별하였다. 이 109개의지진중에서관측자료부족으로상대위치정확도가떨어지는 11개의지진을제외하고 EGS 활동과지진발생의연관성분석을위해필요하다고분류된 98개지진을대상으로정밀지진위치분석을수행하였다. 이렇게결정된지진원의분포중에서 PX-2 지열정을이용한수리자극과연관된지진들의지진원위치들을 3차원공간상에도시하면평면에가까운분포양상을보이며, 이평면 (N214 /43 NW) 은포항지진본진의단층면해주향 / 경사 (N214 /51 NW) 와거의일치한다. 이평면을 PX-2 지열정으로연장시키면약 3,800 m(measured depth) 깊이에서교차하는데이는영상검층, 이수누출등다른관측내용들과일치한다. 이런결과들에기초하여 PX-2 지열정을이용한수리자극과연관되어발생한지진들과포항지진본진은기존에존재하였던동일한단층면상에서유사한단층운동에의해발생한것으로판단할수있다 ( 관련내용 Figs. 5-4, 5-5, 5-7). 이단층면상에서포항지진본진이전에발생한지진들의분포범위내에포항지진본진의진원위치가포함된다. PX-2 지열정을이용한수리자극이 3차례시행되었는데, 각각의수리자극과연관된지진원들의위치는시기가진행됨에따라서해당평면상에서남서쪽방향으로그리고깊이가깊어지는쪽으로이동한다. 이러한이동방향은수리자극에의한공극압모델링에서나타난공극압의확산경향과도유사하다 ( 관련내용 Fig. 5-8). viii

12 조사내용및결과요약 5) 진원분포로부터결정한단층면과타관측자료의연관성 : 진원분포로부터결정한단층면을 PX-2 관정까지연장하면약 3,800 m 심도에서교차하는것으로나타났다. PX-2 시추과정에서확보된시추암편에서 3,790~3,815 m 구간에서단층핵의단층암 ( 비지 ) 을추정할수있는구조가육안관찰과현미경관찰로나타났는데, 다른구간의시추암편에비해원마도가높고풍화를많이받았으며점토물질들의함량이상대적으로매우높았다 ( 관련내용 Figs. 2-16, 2-17, O-3, A-5) 년 8월 PX-1 과 PX-2 지열정에대한시추공 PTS 검층과영상검층을실시하였다. 관측장비가 4,098 m 깊이까지들어갈수있었던 PX-1 관정에비해서 PX-2 관정에서는 3,783 m 심도에서더이상들어갈수없었으며시추공영상이갑자기사라지는현상이나타났다 ( 관련내용 Figs. A-1-1, A-1-3). 이는심도 3,780 m 구간하부에케이싱이파열, 손상, 또는굴절되었거나이와함께파열된관정내로혼탁액이유입되었음을추정할수있다. 포항지진의단층파열이이심도구간을통과하면서 PX-2 관정케이싱을파열시켰음을의미하는것으로, 이를뒷받침하는증거로 PX-2 관정의급격한수위하강이관찰되었다. 포항지진본진발생이전까지지상으로물이배출되고있었던 PX-1( 본진이후에도물배출됨 ) 과 PX-2 지열정은 2018 년 8월관측시에각각지하수위심도가 113 m와 740 m 로나타나서포항지진본진과함께 PX-2 에서최소한 740 m 이상의폭으로수위가급격하게낮아진것으로파악되었다 ( 관련내용 Fig. 6-7). 또한포항지진본진이전과이후의 PX-2 관정의수화학특징의뚜렷한변화가나타났는데, 본진이전에배출된물은주입수의특징을, 본진이후에는빠른속도로심부지열수의특징을반영하고있다. 이는케이싱파열로이구간으로주변심부지하수가유입됨을암시한다 ( 관련내용 Figs. 6-11, 6-12, 6-13). PX-1 과 PX-2 지열정의수위차이는 2019 년 2월 28일현재약 600 m 이상이고, 수위상승속도도약 2배정도차이가나며, 수화학특성도뚜렷한차이를보인다. 두지열정내부에서의수위차이가실제주변지하수의수위차이와는다를수있지만, 현재까지짧은기간동안의수위와수화학특성자료만으로두지열정에서나타나는차이를명확히설명하기어렵다. 두지열정의수위차이에의한비정상적인수리경사는수리환경의급격한변화를야기할수있다. 이런상황에대비하여향후미소지진및안전성에관한장기적인모니터링과분석이필요하다. 6) 수리자극과정의물주입과공극압의확산 : 수리자극과정에서주입된물에의해서주입부 ( 지열정의 open hole section) 로부터거리와시간에따른공극압의변화를모델링하기위해우선수리상수를계산하였다. 수리전도도와저류계수등의수리상수는주입기간에관측된압력과주입률자료를이용하여해석적및수치적인방법으로계산하였다. PX-1 과 PX-2 지열정사이에놓여있는단층대가저투수성단층핵 (fault core) 을갖는경우를 Case A로, 단층핵이없이상대적으로높은투수성의단층손상대 (fault damage zone) 만을갖는경우를 Case B로설정하여수리모델링을수행하여공극압의시공간적인변화를분석하였다. Coulomb 응력변화가 0.01 MPa이상이면단층이임계응력상태일때지진의발생을증가시키거나지진을유발할수있는것으로알려져있는데, 공극압계산결과와지진발생빈도를비교한결과공극압이 0.02~0.06 MPa 증가하는경우지진의발생빈도가커지는것으로분석되었다 ( 관련내용 Fig. 6-6). M W 3.2와 M W 5.5 지 ix

13 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 진이발생한 2017 년 4월 15일과 11월 15일에지열정주변에서의공극압의변화를보면, 2017 년 4월 15일에는 PX-2 주변에 0.1 MPa 이상의공극압의변화가형성되었다 ( 관련내용 Fig. 6-4) 년 11월 15일에는 4 월 15일보다단층면상에넓게공극압이증가되는것으로나타났다. 특히 2017 년 9월 18일 5차수리자극이후약 2개월경과했음에도불구하고, 0.02 MPa이상의공극압의변화가 PX-1, PX-2 및단층에서폭넓게남아있는것으로분석되었다 ( 관련내용 Fig. 6-4) 년 4월 15일 M W 3.2 지진과 2017 년 11월 15일 M W 5.5의지진의진원에서공극압의변화를보면, 특히 PX-2 에서의수리자극에의해공극압이크게증가하는시점에서발생 (2017 년 4월 15일 ) 했거나단층면을따라공극압이넓게확산되어있는시점 (2017 년 11월 15일 ) 에서나타났다 ( 관련내용 Fig. 6-5). 수리자극이후공극압은시간이경과하면서그주변으로압력변화를확산시키며지열정의주입구간에서부터의거리와수리상수에따라서일정시간후에지진이발생할수있음을나타낸다. 7) 이수누출 (mud loss) 로인한지진발생의시간지연관측 : PX-2 지열정의굴착과정에서전체 m 3 의이수 (drilling mud) 가주변지층으로누출되었는데, 심도 3,800 m 지점에서이수누출이집중적으로발생하였다 ( 관련내용 Fig. A-2-1). 3,800 m 해당심도구간굴착시이수의밀도가 1.6 g/cm 3 을상회하여주변지하수압력에비해상대적으로 20 MPa 이상높은압력이만들어졌다. 따라서이수의누출은마치압력 20 MPa 정도로이수누출에해당되는부피의물을지층에주입하는수리자극과같은영향을주변지층에가했다고할수있다. 이수의누출직후에아주작은규모의지진들이발생하였고, 이수의누출이크게발생한시점으로부터약 1개월후에이수누출과연계된가장큰규모의지진 (M W 0.9) 이발생하였다 ( 관련내용 Fig. A-2-2). 이는관정을통한이수누출의발생이주변지층으로공급압을전파하여일정한시간이경과한이후인 11월 30일에 M W 0.9 지진을발생시켰다고판단된다. 이것은인위적인누출이나주입의시점과이와연관된최대규모지진의발생사이에는시간지연이가능함을실제적으로관측한사례라고할수있다. 8) 지중응력상태 : 지진포컬메커니즘응력역산및시추공응력지시자등, 여러기법을기반으로포항지열발전실증연구부지하부의응력상태를분석하였다. 최대수평주응력의방향은깊이에따라변하여얕은깊이 (700~800 m) 에서는 SE-NW, 4 km 하부심도에서는 ENE-WSW 로나타났다. 응력체계또한관측범위에따라변하여지역적규모에서는주향이동단층운동에유리한응력체계를보이고, 보다좁은시추공규모에서는역단층운동에유리한응력체계를보인다. 모든경우에연직응력과최소수평주응력의크기가유사한것으로나타났다 ( 관련내용 Table 4-1). 포항지역의응력장하에서 2017 년 11월 15일포항지진을발생시킨단층의전단성향은 0.57로산정되었으며이는최적단층방향전단성향의 92~95% 수준에이르는높은값이다. 이값은 PX-2 굴착시 4.2 km에서회수된암석코어시료내자연균열의마찰계수 (0.53) 와유사하거나약간큰값이며, 이값들이지진발생깊이에서의단층마찰특성을대표하는경우, 지진발생단층은자연상태에서임계치의응력이거나이에근접한상태에있었음을의미한다 ( 관련내용 Figs. 4-12, 4-13). x

14 조사내용및결과요약 3. 포항지진의원인에관한결론주요연구결과들을종합분석하여포항지진과지열발전실증연구프로젝트의연관성과포항지진의원인에관하여다음과같이판단하였다. 지열발전실증연구수행중지열정굴착과두지열정 (PX-1, PX-2) 을이용한수리자극이시행되었고, 굴착시발생한이수누출과 PX-2 를통해높은압력으로주입한물에의해확산된공극압이포항지진단층면상에남서방향으로깊어지는심도의미소지진들을순차적으로유발시켰다. 시간의경과에따라결과적으로그영향이본진의진원위치에도달되고누적되어거의임계응력상태에있었던단층에서포항지진이촉발되었다. 유발 (induced) 지진과 촉발 (triggered) 지진에관한정의가국제적으로확립되지않은상태이기때문에본보고서에서사용한 유발 과 촉발 의의미를좀더정확히나타내기위해해외조사위원회보고서에영문으로 유발 과 촉발 의의미를정의하였다 ( 해외조사위원회 (ORAC) 보고서페이지 O-10). 이용어의정의는본보고서에서사용한용어를설명하는데국한한다. 정확한의미는영문원문을참조하기바라며, 한글의역은다음과같다. 유발 지진은지구내부에서유체주입의영향으로공극압과응력이변화된암석의공간적범위내에서일어날수있는규모의지진으로, 이때의지진은유체주입과조구조운동으로축적된변형에너지를방출한다. 촉발 지진은인위적인영향이최초의원인이지만그영향으로자극을받은공간적범위를크게벗어나는규모의지진으로, 이때의지진은대부분조구조운동으로축적된변형에너지를방출한다. xi

15

16 조사내용및결과요약 Overseas Research Advisory Committee Report on the Pohang Earthquake Shemin Ge 1, Domenico Giardini 2, William Ellsworth 3, Toshihiko Shimamoto 4, John Townend 5 1 Department of Geological Sciences, University of Colorado Boulder 2 Department of Earth Sciences, ETH Zürich 3 Department of Geophysics, Stanford University 4 State Key Laboratory of Earthquake Dynamics, Institute of Geology, China Earthquake Administration; Shimamoto Earth and Environment Laboratory Ltd. 5 School of Geography, Environment and Earth Sciences, Victoria University of Wellington O-1

17 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Overseas Research Advisory Committee Report on the Pohang Earthquake O-2

18 Table of Contents Page O-7 1. Background O Pohang earthquake of 15 November 2017 O Enquiry terms of reference O Composition and mandate of the Overseas Research Advisory Committee (ORAC) O Pohang EGS project overview O Project timeline O Terminology used in this report O Regional setting O Geological history O Active faulting in southeast Korea O Historical and recent seismicity of Korea O Regional stress field O Site geology and geophysics O Pre-drilling site investigations and local stratigraphy O Petrographic analysis of fault zones identified during drilling O Frictional characteristics of basement rocks O State of stress at the Pohang drill site O Seismicity associated with injection O Temporal patterns of seismicity before and during drilling O Seismicity near the EGS site prior to simulation O Seismicity induced by mud loss during drilling O Spatial patterns of seismicity O Seismicity associated with PX-1 O Seismicity associated with PX-2 O Focal mechanisms O-3

19 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Page O Magnitude-frequency characteristics O Model-based analyses of triggering mechanisms O Effects of Tohoku and Gyeongju earthquakes O Hydrogeologic analysis of fluid pressure perturbations O M W 5.5 Pohang earthquake of 15 November 2017 O Foreshock activity O Location and timing of mainshock O Mainshock focal mechanism O Magnitude of mainshock and previous scaling arguments O Aftershock activity O Findings, conclusion and lessons learned O Findings O Conclusion O Lessons learned O References O-4

20 Overseas Research Advisory Committee Report on the Pohang Earthquake Executive Summary On the afternoon of November 15, 2017, the coastal city of Pohang, Korea, was rocked by a magnitude 5.5 earthquake (M W, USGS). The earthquake injured 135 residents, displaced more than 1,700 people into emergency housing and caused more than $75 M (USD) in direct damage to over 57,000 structures and over $300 M (USD) of total economic impact, as estimated by the Bank of Korea. This was the most damaging earthquake to strike the Korean Peninsula for centuries. Questions soon arose about the possible involvement in the earthquake of the Republic of Korea s first Enhanced Geothermal System (EGS) project, as the epicenter of the quake was located near the project s drill site. Debate within the Korean and international scientific communities did not resolve whether the earthquake as associated with the EGS or of purely tectonic origin. Following this, the Pohang EGS project was suspended and the Korean Government commissioned the Geological Society of Korea to produce an evaluation report. An Overseas Research Advisory Committee (ORAC) was formed, consisting of William Ellsworth (Stanford University, United States), Shemin Ge (University of Colorado Boulder, United States; co-chair), Domenico Giardini (ETH Zürich, Switzerland; co-chair), Toshihiko Shimamoto (China Earthquake Administration) and John Townend (Victoria University of Wellington, New Zealand). ORAC s mandate was to elucidate the origin of the Pohang November 15, 2017 mainshock. The committee s work started in March 2018, and included four meetings in Korea and intense work with Korean colleagues. The work involved performing ORAC s own analyses and taking into account the results and evidences collected by other groups and researchers working on the earthquake sequence, as well as data made available by the NexGeo project operator and by agencies involved in monitoring the seismicity during the EGS development. The Pohang EGS project was intending to create an artificial geothermal reservoir within low-permeability crystalline basement by hydraulically stimulating the rock to form a connected network of fractures between two wells, PX-1 and PX-2. Forensic examination of the tectonic stress conditions, local geology, well drilling data, the five high-pressure well stimulations undertaken to create the EGS reservoir, and the seismicity induced by injection produced definitive evidence that small earthquakes induced by high-pressure injection into the PX-2 well activated the fault that ultimately ruptured in the M W 5.5 earthquake. O-5

21 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Pre-drilling site investigations failed to identify any active faults in the vicinity of the EGS project, but also indicated that the faults in the region capable of generating moderate or large earthquakes were critically stressed, as shown by stress state investigations and by the occurrence in 2016 of the M L 5.8 (M W 5.4, USGS) Gyeongju earthquake nearby. During drilling, a fault zone was crossed by the PX-2 well at almost 4 km depth, where extensive mud loss occurred that triggered seismic events. The fault s position and orientation were delineated by the seismicity induced during subsequent injections into PX-2. The locations of these earthquakes relative to the borehole, their delineation of a planar structure that projects to a fault zone recognizable in borehole logs which subsequent seismological and geodetic observations indicated was likely the mainshock fault and the occurrence of seismicity releasing tectonic strain during each stimulation phase indicate that this pre-existing fault was highly sensitive to perturbations. The events associated with PX-2 injection affected a portion of the fault of approximately 600 m 1,000 m dimensions; foreshocks occurred in the 24 hours preceding the mainshock, which initiated in the lower part of the fault already stimulated by the PX-2 injections. ORAC concludes that the Pohang earthquake was triggered by the EGS stimulation. Seismicity induced by injection activated a previously unmapped fault zone, which in turn triggered the mainshock. Once initiated, the Pohang earthquake grew through the release of tectonic strain. The cuttings extracted during the drilling of PX-2 contained a large amount of fault gouge at depths of about 3,800 m, close to the biggest mud-loss zone, revealing the presence of a fault. The fault s location is close to the intersection with the PX-2 well of the main fault inferred from seismicity. Hydraulic modeling of the injections in PX-1 and PX-2 corroborate the conclusion that the fault responsible for the M W 5.5 earthquake was stimulated by the PX-2 injection. Important lessons of a general nature can be learned from the Pohang experience, and can serve to increase the safety of future EGS projects in Korea and elsewhere. The ORAC report represents the unanimous opinion of the ORAC members. O-6

22 Overseas Research Advisory Committee Report on the Pohang Earthquake 1. Background 1.1. Pohang earthquake of 15 November 2017 On November 15, 2017, a magnitude (M W ) 5.5 earthquake shook the city of Pohang, Korea (Figure O-1). The earthquake injured 135 residents, displaced more than 1700 people into emergency housing and caused more than $75 M (USD) in direct damage to over 57,000 structures and over $300 M (USD) of total economic impact, as estimated by the Bank of Korea. This was the most damaging earthquake to strike the Korean Peninsula for centuries. Questions soon arose about the possible involvement in the earthquake of the Republic of Korea s first Enhanced Geothermal System (EGS) project, as the epicenter of the quake was located near the project s drill site Enquiry terms of reference Shortly after the 2017 Pohang earthquake, a debate arose regarding the cause of the earthquake. The central question in this debate was whether the EGS stimulations had triggered this earthquake. On one side of the debate is the argument that the 2017 Pohang earthquake is a natural event unrelated to EGS activities. Situated on the eastern margin of the Eurasian tectonic plate, the Pohang area and Korea in general exhibit low levels of seismicity in comparison with neighboring Japan and China. However, damaging earthquakes have happened in historical and modern times. Faults active during the Quaternary are recognized [Ree et al., 2003], and the region has experienced some seismicity in recent decades including the M L 4.5 Yongwol event in 1996 and the M L 5.8 (M W 5.4) Gyeongju event in 2016 [Kim et al., 2018b; Kim et al., 2016; Lee et al., 2018 ]. During a period from the 15 th to the 18 th centuries, southeastern Korea experienced elevated levels of seismicity [Lee and Yang, 2006 ]. An alternative view is that the 2017 Pohang earthquake was triggered by the hydraulic stimulations at the Pohang EGS site nearby. The hydraulic stimulations took place over the two years prior to the Pohang earthquake. There are clear spatial and temporal correlations between hydraulic stimulation activity and earthquake occurrences. EGS hydraulic stimulations elsewhere such as in Basel, Switzerland [Deichmann and Giardini, 2009], are known to have caused damaging earthquakes and forced geothermal operations to be shut down. The historical and recent occurrence of tectonic earthquakes nearby does not preclude the possibility that the 2017 Pohang earthquake was triggered by EGS activities. While spatial and temporal correlations are the primary basis for linking hydraulic stimulation to earthquakes, they do not necessarily demonstrate causation and in the case of the Pohang earthquake require specific investigation. O-7

23 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 1.3. Composition and mandate of the Overseas Research Advisory Committee (ORAC) To address the central question of whether EGS activities triggered the Pohang earthquake, the Geological Society of Korea, on behalf of the Korean government, assembled a panel of researchers to form the Overseas Research Advisory Committee (ORAC) with expertise covering seismology (Ellsworth, Giardini), geomechanics (Townend, Giardini), geology (Shimamoto), and hydrogeology (Ge). ORAC s mandate was to answer authoritatively the question Was the Pohang event induced? ORAC worked from March 2018 to March 2019, with the committee members making four visits to Korea and interacting extensively with Korean colleagues. The work involved undertaking new analysis and taking into account the results and evidence collected by other groups and researchers working on the earthquake sequence, as well as data made available by the EGS project team (NexGeo and the Korean Institute of Geoscience and Mineral Resources, KIGAM), the Korea Meteorological Administration (KMA), and university researchers not involved in either the official inquiry or the EGS project. In this report, all times and dates are given in Universal Coordinated Time (UTC), which is nine hours behind Korean Standard Time (KST) Pohang EGS project overview The Pohang Enhanced Geothermal System (EGS) Project was intended to demonstrate the potential of geothermal energy production in a ~4 km-deep granodioritic reservoir overlain by Cretaceous volcanics and sedimentary rocks, Tertiary volcanics and sedimentary rocks, and Quaternary sediments. The Pohang area is one of the highest heat-flow areas in Korea and has been the focus of dedicated geothermal research since 2003 [Lee et al., 2010 ]. Over the course of approximately four years from 2012 to 2016, two exploratory wells named PX-1 and PX-2 were drilled into the bedrock to develop the enhanced geothermal system (Figure O-1). PX-1 had a designed depth of 4,127 m, but the drill pipe became stuck and was broken and not recoverable below a depth of 2,485 m. PX-1 was later side-tracked and extended in the WNW direction to a depth of 4,215 m, measured depth (MD) 4,362 m. PX-2 was drilled to a depth of 4,340 m (MD 4,348 m). Note that all depths were measured from the drill rig floor, which is 9 m above the ground surface. PX-1 and PX-2 are 6 m apart from each other in the north-south direction on the ground surface but they are approximately 600 m apart at the bottom. Both wells are cased along their length except for the bottom 313 m in PX-1 and 140 m in PX-2. These bottom intervals are open for fluid injection and flow back. The inner diameters of the casing and open intervals are 155 mm and 216 mm, respectively. O-8

24 Overseas Research Advisory Committee Report on the Pohang Earthquake Figure O-1. Pohang EGS location and the schematics of the two exploration wells PX-1 and PX Project timeline Five hydraulic stimulations were conducted in PX-1 and PX-2 between January 29, 2016 and September 18, The first, third, and fifth stimulations were conducted in PX-2 and the second and fourth in PX-1. Each hydraulic stimulation involved multiple periods of injection, when water is forced into the formation under a wellhead pressure and repeated periods of shut-in or water flowing back to the surface. The Pohang earthquake occurred during shut-in of PX-1 and flow-back of PX-2 after the fifth stimulation. Injection rates and wellhead pressures for all five stimulations were recorded. The temporal resolutions for these data are seconds for PX-1 and one minute for PX-2. Figure O-2 shows the injection rates and the net injection volume over the entire period of five stimulations. The volumes of water injected into and flowed back from PX-1 are 5,663 m 3 and 3,968 m 3. The volumes of water injected into and flowed back from PX-2 are 7,135 m 3 and 2,989 m 3. Thus, a net volume of 5,841 m 3 of injected water remains in the subsurface. In PX-2, the maximum wellhead pressure and injection rate reached MPa and m 3 /s during the first stimulation. In PX-1, the maximum wellhead pressure and injection rate reached MPa and m 3 /s during the second stimulation. Injection pressures were higher overall for PX-2 than for PX-1 at similar injection rates. Seismicity accompanied each stimulation and continued for up to several months (Figure O-2). O-9

25 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Figure O-2. Timeline of the Pohang EGS stimulations and seismicity leading up to the November 15, 2017 M W 5.5 Pohang earthquake. Earthquakes with measured local magnitudes (M L ) are represented by colored dots (left-hand scale). Daily injection and flow-back volumes and the cumulative net injection volume are illustrated with colored lines (right-hand scales) Terminology used in this report Earthquakes can occur as a consequence of a wide variety of industrial activities, including the impoundment of high dams, underground mining, petroleum production and storage, geothermal energy extraction, CO 2 sequestration and wastewater disposal by injection [Ellsworth, 2013; Grigoli et al., 2017]. The earthquakes caused by these activities are sometimes referred to as induced or triggered to identify them as being of anthropogenic origin. In the scientific literature, induced and triggered are sometimes used to draw a distinction between earthquakes that primarily release strains created by the industrial process (induced) and earthquakes that primarily release natural tectonic strain (triggered; e.g. McGarr et al., 2002). The term induced is also used to refer to all anthropogenic earthquakes, as only human activity can induce earthquakes, while natural earthquakes routinely trigger other earthquakes. Because use of terms induced and triggered can be confusing, we adopt the following definitions for these terms in this report in the specific context of activities connected to the Pohang EGS project: Induced Earthquakes occur within the volume of rock in which pressure or stress changes as a consequence of injection. Their magnitudes are consistent with the spatial dimension of the stimulated volume. They can occur both during injection and after injection ceases. They may release tectonic strains or strains created by injection pressure or volume. Triggered Earthquakes are runaway ruptures, initiated by anthropogenic forcing that grow in size O-10

26 Overseas Research Advisory Committee Report on the Pohang Earthquake beyond the bounds of the stimulated region. They release tectonic strain. As an example, within a volume affected by stimulation with a diameter of 1,000 m, earthquakes as large as approximately magnitude 4 would be classified as induced, as they would largely be contained within the stimulated zone. Earthquakes with magnitudes exceeding 5 would be classified as triggered as their ruptures would extend beyond the stimulated volume. 2. Regional setting 2.1. Geological history The Pohang EGS site is located within the Pohang Basin, one of several sedimentary basins that formed in the early Miocene during back-arc extension and opening of the East Sea [Son et al., 2015]. The basin is bordered to the west and south by the N-striking Western Border Fault and the NE-striking Ocheon Fault System, respectively, which are each composed of strike-slip and normal fault segments that formed during the basin s extensional phase [Cheon et al., 2012; Son et al., 2015]. A change in regional tectonics in the late Miocene resulted in broadly ENE WSW compression across the southeastern Korean Peninsula [Chough et al., 2000; Park et al., 2007]. Surface mapping and borehole investigations undertaken prior to the EGS project revealed a thin layer of Quaternary alluvium overlying a m-thick Tertiary mudstone, ~1000 m-thick Cretaceous sandstone/mudstone sequence interlayered with Eocene volcanic intrusions, ~900 m-thick Cretaceous volcanics, and granodioritic basement below approximately 2.2 km [Kwon et al., 2018; Lee et al., 2015]. The geothermal gradient near Pohang has been recognized since the 1960s as being higher than in most other parts of Korea, in which the average geothermal gradient is 25 /km [Lee et al., 2010]. Exploratory drilling and geophysical surveys conducted by the Korean Institute of Geoscience and Mineral Resources (KIGAM) between 2003 and 2008 revealed a much higher geothermal gradient in Pohang and temperatures at 5 km depth of ~180 [Lee et al., 2015; Lee et al., 2010]. These findings formed the basis of the Pohang EGS project (from 2010) and the drilling of the PX-1 and PX-2 wells Active faulting in southeast Korea Most earthquakes on the Korean Peninsula are too small to leave geological evidence at the surface and neither the 2016 Gyeongju earthquake nor the 2017 Pohang earthquake produced a O-11

27 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 distinct surface rupture although ground damage indicative of faulting was observed at Pohang [Choi et al., 2019; Gihm et al., 2018]. Paleoseismic investigations of active faults have not proven effective in Korea and geologists have instead considered active those faults that dissect Quaternary formations and referred to them as Quaternary faults. Much of the Quaternary faulting recognized in southeastern Korea occurs on subsidiary faults associated with the Yangsan and Ulsan faults [Ree et al., 2003]. Those associated with the Yangsan fault tend to be N- or NNE-striking subvertical dextral strike-slip faults, whereas those associated with the Ulsan fault are typically NNE- to NNW-striking reverse faults [Ree et al., 2003]. The EGS drill site is situated within 5 km of the E-striking Heunghae Fault and the NNEstriking Gokgang Fault, which together bound the northeast corner of the Doumsan structural domain. The blind Gokgang fault has a similar strike to the fault that ruptured in the 2017 Pohang earthquake, but an antithetic (ESE) dip. No Quaternary faulting close to the EGS site was recognized prior to the 2017 earthquake, although Quaternary faults had previously been identified within 15 km of the site at outcrops on the Yangsan fault and Wangsan faults [Ree and Kwon, 2005; Ree et al., 2003] Historical and recent seismicity of Korea The Korean Peninsula exhibits much lower rates of seismicity than surrounding regions, particularly Japan. The Korean Meteorological Administration s online seismicity catalog lists only 10 earthquakes larger than M L 5 in the Korean region since national seismic monitoring began in 1978 ( kma.go.kr/weather/earthquake_volcano/domesticlist.jsp, last accessed 26 February 2019). The historical record of seismicity spans two millennia and reveals that earthquakes have occurred throughout the Korean Peninsula [Lee and Yang, 2006 ]. The attribution of preinstrumental earthquakes to specific faults is difficult [Houng and Hong, 2013] but the historic catalog indicates the occurrence in southeastern Korea of more than 100 felt earthquakes, of which at least 11 produced Modified Mercalli Intensity shaking exceeding VIII [Kim et al., 2018b]. This latter group includes a M~6.7 earthquake in 779 AD and M~6.4 earthquake in 1306 AD. The historical seismicity in southern Korea proves that the major active fault systems identified in the regional geology (such as the Yangsan fault) have been active in historical and recent times [Lee and Yang, 2006 ]. The most recent large event to occur in southeastern Korea prior to the 2017 earthquake was the M L 5.8 (M W 5.4) Gyeongju earthquake of 12 September 2016, which was preceded 48 minutes earlier by a M L 5.1 foreshock. These events occurred approximately 40 km south of the Pohang EGS site. Aftershock relocations and analysis of the foreshock and mainshock focal mechanisms indicated strike-slip motion on a steeply-east-dipping NNE-striking fault plane at mid-crustal depths of approximately 15 km [Hong et al., 2017; Kim et al., 2018b]. O-12

28 Overseas Research Advisory Committee Report on the Pohang Earthquake 2.4. Regional stress field Earthquakes are a manifestation of faults slipping in response to the stresses acting on them, and it is therefore important to understand the state of stress near the Pohang EGS site in order to understand the cause of the Pohang earthquake. The susceptibility of a fault to slip in shear in response to a particular state of stress is governed by the Mohr-Coulomb criterion, τ = μ (S n P f ), where τ is the shear stress, S n is the normal stress, P f is the fluid pressure, and μ is the coefficient of friction. The shear and normal stresses depend on the orientation of the plane of interest, typically represented by its strike and dip, and the orientations and magnitudes of the three orthogonal principal stresses, S 1, S 2 and S 3, where S 1 S 2 S 3 and we adopt the geological convention that compressive stresses are positive numbers. The fluid pressure is the sum of the background fluid pressure, which at shallow depth in the crust generally increases linearly with depth, and any spatial or temporal perturbation. To fully specify the state of stress, we ideally require knowledge of six parameters specifying the magnitudes and orientations of S 1, S 2 and S 3. In many cases, one of the principal stresses is observed to be oriented vertically and is referred to as the vertical stress, S v. The orientations of all three principal stresses can then be described by specifying which of the principal stresses is vertical and the orientation of the azimuth of maximum horizontal compressive stress ( S Hmax ). The terms normal stress state, strike-slip stress state and reverse stress state are used to refer to the case in which S 1, S 2 and S 3, respectively, is the vertical stress. A common situation, especially when estimating stress parameters from seismological data, is to know the orientations of all three principal stresses and a single stress magnitude parameter R = (S 1 S 2 )/(S 1 S 3 ). The state of contemporary tectonic stress in the Korean Peninsula has been studied by several groups in recent years using a variety of borehole and seismological techniques [Kim et al., 2017; Lee et al., 2017b; Soh et al., 2018]. We focus here on those results most pertinent to stress in the vicinity of the Pohang EGS site and at depths comparable to the depth of the 15 November earthquake. Soh et al. [2018] mapped stress parameters throughout the Korean Peninsula using earthquake focal mechanism analysis and documented a strike-slip stress state (S v = S 2 ), R~0.85, and ENE WSW S Hmax orientation in southeastern South Korea. Analysis of focal mechanisms recorded between 1997 and 2016 within 70 km of the Pohang EGS site yields a strike-slip stress state (S v = S 2 ), R~0.88, and S Hmax = 074 (see Chapter 4). This result is similar to that obtained using the focal mechanisms of aftershocks following the 15 November 2017 earthquake (R = 0.87, S Hmax = 086 ). The state of stress at shallow depths within ~10 km of the EGS site was investigated using borehole data by Kim et al. [2017] and Lee et al. [2017a], who derived S Hmax estimates of approximately 130 at depths of ~700 m and inferred the stress state to be strike-slip. Quaternary fault data have also been used to determine the regional stress field representative of O-13

29 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 longer time scales and broader spatial domains. Park et al. [2006] analyzed 24 fault slip measurements from southeastern Korea and inferred a reverse stress state with an azimuth for the subhorizontal S 1 axis of approximately 070. That analysis, however, indicated a stress ratio of R = 0.35, lower than obtained from seismological or borehole measurements. 3. Site geology and geophysics 3.1. Pre-drilling site investigations and local stratigraphy Prior to the drilling of PX-1 and PX-2, an extensive program of geophysical site characterization was undertaken by the Korea Institute of Geoscience and Mineral Resources (KIGAM), as detailed in Chapter 3. Magnetotelluric measurements revealed W-dipping conductive features beneath the EGS site, which were interpreted as fracture zones and potential geothermal targets [Lee et al., 2015]. However, the limited spatial resolution of the models did not enable the presence of a large discrete fault to be determined. During the drilling of PX-1 and PX-2 the drill cuttings were analyzed at regular depth intervals by on-site geologists who created records of lithologic observations referred to as mud logs. These observations were used to identify the geologic units crossed by the drill holes and the depths of formation boundaries. The integrated stratigraphy of the PX-1 and PX-2 wells inferred from mud logs and wireline geophysical measurements undertaken during drilling, and later reanalysis of cuttings, is described in detail in Chapter 2. The stratigraphy consists of Miocene Pohang Basin sediments extending to a depth of ~200 m, overlying Cretaceous sedimentary and volcanic rocks and Paleozoic granodiorite below ~2,350 m. A seismic velocity model based on check-shot data, PX-2 sonic logs and the borehole stratigraphy was constructed during the course of the Korean Government Commission s inquiry (see Chapter 4) and augmented with regional seismological observations to form the composite model used to determine the location of the seismic activity (see Chapter 5) Petrographic analysis of fault zones identified during drilling Identifying faults that cross the drill holes is important for understanding the geologic framework of the Pohang earthquake. Most drill cuttings are fresh and angular as shown in Figure O-3a. However, the cuttings from PX-2 in the depth interval from 3,790~3,816 m contain a large fraction O-14

30 Overseas Research Advisory Committee Report on the Pohang Earthquake of round-shaped mud balls that can be broken easily by hand (Figure O-3b). Figure O-3c illustrates the microstructure of a mud ball showing a typical fault gouge structure in which clasts are scattered Figure O-3. (a, b) Photographs of representative (a) rock fragments and (b) mud balls extracted from cuttings at a depth of 3,798 m in the PX2 borehole (with 1 mm blue grid shown for scale). In this sample, mud balls account for ~65 wt.% of the cuttings. (c, d) Photomicrographs under a stereo-microscope of (c) incohesive fault gouge constituting a mud ball from 3,798 m and (d) a granitic cataclasite from 3,825 m. (e) A summary of fault and mud-loss data for the depths of 3,785~3,840 m in the PX-2 borehole, revealing a large-scale fault. The mud-loss data were quoted from an unpublished compilation of drilling data by Geo-Energie Suisse. O-15

31 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 within a sheared and foliated matrix. Detailed microscopic examination of over 100 mud balls (summarized in Figure O-3e) reveals that they are fragments of fault gouge and breccia. Fragments of cohesive cataclasite such as that shown in Figure O-3d were also observed. Cuttings below 3,791 m contain fragments of granite (e.g., pinkish fragments in Figures O-3a, O-3c and O-3d), in contrast with granodiorite mixed with fine-grained igneous rocks at shallower depths. The data illustrated in Figure O-3e indicate the presence of a fault gouge and breccia zone several meters in thickness. In August 2018, the Korean investigation team ran wireline logs in PX-2 to better understand the borehole environment after the earthquake. The logging tools were unable to descend below 3,783 m (Figure A-1-3) due to obstruction of the well. This depth nearly coincides with the top of the fault gouge zone illustrated in Figure O-3e. It is possible that fault movement during the Pohang earthquake caused damage to the borehole at this depth. The biggest mud loss occurred at depths of 3,830~3,840 m in October November 2015 while PX-2 was being drilled, and required lost circulation material to be pumped into the well to stabilize the hole and prevent further mud loss. The data shown in Figure O-3e suggest that the mud loss occurred in a fault zone within fractured host rocks. The mud loss event is discussed in Section Frictional characteristics of basement rocks The frictional properties of the basement rocks were determined from cuttings from a depth of 3,607 m in PX-2 and analogous granitic lithologies sampled in nearby fault outcrops. Three velocitycycle tests and three normal-stress cycle tests were conducted at a temperature of approximately 200 C, slip rates of 0.17, 1.6 and 17 μm/s, and effective normal stresses to 30 MPa with pore water pressure of 30 MPa. The effective normal stress at which the experiments were conducted is likely to be lower than in situ, but is limited by the particular testing apparatus used. These materials exhibited steady-state friction coefficients in the range typical of crustal rocks ( ): the specific values of the Pohang cuttings were 0.54~0.68 (0.63 on average). These materials exhibited slight velocityweakening, meaning a decrease in steady-state friction with increasing slip speed that can lead to an earthquake State of stress at the Pohang drill site Measurements made during and after drilling enable us to construct a local stress model that shares some features of the regional stress model described above but which reflects measurements specific to the EGS site. Full details of the observations and interpretations underpinning the stress analysis are provided in Chapter 4. Dipole sonic logging of the PX-2 borehole in August 2018 revealed the presence of anisotropy O-16

32 Overseas Research Advisory Committee Report on the Pohang Earthquake features at depths of 3.4~4.3 km that are interpreted to indicate an axis of maximum horizontal compression (S Hmax ) oriented 077±23. This orientation is consistent with the pre-2017 regional orientation computed from focal mechanisms and is our preferred value in the analysis below. The fluid pressure regime surrounding the PX-1 and PX-2 boreholes is presumed here to have been hydrostatic. High mud weights (exceeding 1.6 g/cm 3 ) were used while drilling below ~2.7 km in PX-2, but there are no indications recorded in the drilling reports of suprahydrostatic fluid pressures. Our preferred model of stress corresponds to a critically-stressed reverse stress state evaluated at 4.2 km (S v = S 3 = 106 MPa), with hydrostatic fluid pressure, R = 0.90, and an S Hmax orientation of 077±23. Based on the analysis described in Chapter 4 of the Summary Report, we adopt values for the maximum and minimum horizontal stresses of S Hmax = S 1 = 243 MPa and S hmin = S 2 = 120 MPa, respectively. The S Hmax value is computed assuming that the crust is in a state of frictional equilibrium governed by slip on faults with a coefficient of friction of 0.6. The S hmin value is taken from step-rate tests and fracture propagation analysis of PX-2. A Mohr circle representing the preferred model of stress is illustrated in Figure O-4. This diagram represents the combinations of shear and normal stress acting on planes of different orientations, and the frictional failure constraint for a friction coefficient of μ = 0.6. In Sections 4.3 and 6.3, we also consider an alternative model of stress based on the analysis of regional focal mechanisms recorded prior to the Pohang earthquake ( regional model ; see Chapter 4). The regional model (Table O-1) corresponds to a strike-slip stress state and is based on the estimates of R and S Hmax obtained by Soh et al. [2018], converted to principal stress magnitudes at a depth of 4.2 km assuming that S 2 = S v, and that the state of stress is governed by frictional failure for a friction coefficient of 0.6. Table O-1. Preferred and alternative (regional) stress models. Both models assume hydrostatic fluid pressure. S 1, S 2, S 3 maximum, intermediate, and minimum principal stresses; S v vertical stress; S Hmax maximum horizontal compressive stress; S hmin minimum horizontal compressive stress; R = (S 1-S 2 )/(S 1-S 3 ). Model Stress state S 1 (MPa) S 2 (MPa) S 3 (MPa) S Hmax Azimuth( ) R Preferred Reverse 243 (S Hmax ) 120 (S hmin ) 106 (S v ) Regional Strike-slip 203 (S Hmax ) 106 (S v ) 93 (S hmin ) O-17

33 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Figure O-4. Mohr circle illustrating the preferred model of stress in the vicinity of the EGS site. The unshaded circles represent the state of stress if the fluid pressure were zero, and the purple circles show the corresponding state of stress for a hydrostatic fluid pressure regime. 4. Seismicity associated with injection In this section, we examine the temporal and spatial patterns of seismicity induced by injection into PX-1 and PX-2. Our interpretations are based on a comprehensive reanalysis of the seismicity performed in partnership with our colleagues who are part of the Geological Society of Korea s investigation of the earthquake. Details of the investigations are contained in Chapter 5. Local geological and geophysical data were used to develop a crustal velocity model for locating the earthquakes. A precise calibration of earthquake locations derived from the model was performed using data from a multi-level seismic array installed in PX-2 during the August 2017 stimulation of PX-1. Seismic waveform data were collected from all available seismic stations within 100 km of the site, and earthquakes were identified using a matched-filter technique. Earthquake hypocenters were determined from a combination of phase arrival time readings and waveform cross-correlation measurements O-18

34 Overseas Research Advisory Committee Report on the Pohang Earthquake made by the project team using well-established location procedures. New magnitudes were determined using a calibrated local magnitude scale (M L ). In addition, moment magnitudes (M W ) were computed for many of the events. A total of 519 earthquakes were detected between January 1, 2009 and the time of the Pohang mainshock (Figure O-5). More than half of these events (277) locate further than 10 km from the EGS project drill site. Of the 239 events spatially associated with the drill site, the earliest occurred on November 1, High-precision earthquake hypocenters were determined for 98 of these events. Figure O-5. Epicentral distribution of 519 earthquakes detected between January 1, 2009 and November 15, 2017 in the Pohang region. Earthquakes within 10 km of EGS project drill site (yellow triangle) and shallower than 10 km are shown in green, and the four deeper than 10 km in blue; earthquakes further than 10 km from the drill site are shown in red. Geological lineaments and faults are shown as gray lines. O-19

35 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 4.1. Temporal patterns of seismicity before and during drilling The temporal characteristics of seismicity that occurred before drilling, while PX-1 and PX-2 were being drilled, and after completion when they were stimulated by high pressure injection of water are the key factors for understanding the origin of the November 15, 2017, M W 5.5 Pohang mainshock. The first two phases are discussed here, and the seismicity associated with injection is discussed in Section Seismicity near the EGS site prior to simulation The analysis of the instrumental seismicity recorded by KMA shows that no instrumental seismicity with M L > 2.0 has been detected within 10 km distance of the Pohang EGS site from at least 1978 to October 2015 [Kim et al., 2018c]. Only six events of M L had been detected in the area since In addition, Kim et al. [2018c] used a matched-filter technique to identify uncatalogued earthquakes in the continuous waveform data at station PHA2 of the KMA permanent network. PHA2 is located about 10 km north of the EGS site. The matched-filter analysis revealed no events near the Pohang EGS site for the period from January 2012 to October However, the analysis detected small earthquakes in the month of November 2015 that originated near the EGS project at the time when the PX-2 well was being drilled. Our Korean research partners also used a matched-filter technique to search for events located near the Pohang EGS site (Figure O-5). A total of six earthquakes were detected within a 10 km radius of the site between January 2009 and October The largest, M L 2.2, occurred in March 2013 at a depth of 12 km. None was closer than 7 km to the bottom of PX-2 and they had depths of between 6 and 15 km. This analysis confirms that no earthquakes occurred in the vicinity of the crustal volumes stimulated by injection into PX-1 and PX-2 between January 2009 and November It also establishes that the mid-crust beneath the site was at least weakly seismogenic with tectonic earthquakes. On September 12, 2016, the M L 5.8 (M W 5.4) Gyeongju earthquake occurred approximately 40 km southwest of Pohang within the major right-lateral Yangsan fault system. Grigoli et al. [2018] addressed the possibility that the Gyeongju earthquake might have contributed to triggering the Pohang earthquake, and concluded that the static Coulomb stress perturbation produced by the Gyeongju event on the Pohang fault was negligible, and that a direct triggering effect could be excluded. From these analyses we conclude that no increase of seismicity in the area of the Pohang EGS project is observed prior to November Seismicity induced by mud loss during drilling On October 30 31, 2015, during the drilling of PX-2, a fault zone was encountered near 3,800 O-20

36 Overseas Research Advisory Committee Report on the Pohang Earthquake m depth (Figure O-6; see Section 3.2). A significant loss of heavy drilling mud (density 1.6 g/cm 3 ) occurred at this time, amounting to over 160 m 3 or one well bore volume and transferring an additional pressure of >20 MPa to the formation due to the weight of the mud column. The seismicity detected at station PHA2 started at this time and lasted through the month, with the largest event, M W 0.9, occurring on November 30, 2015 (Figure O-6). Of these events, we have only been able to locate the November 30 event. The seismicity associated with mud loss from PX-2 indicates that the press perturbation was sufficient to induce fault slip and implies that some faults were close to failure prior to stimulation. Further, it suggests that a hydraulically conductive structure was intersected near 3,800 m in PX-2. Previous mud loss of 76 m 3 from PX-1 during the first phase of drilling and mud loss of <40 m 3 from PX-2 in October 2015 had not been associated with discernible seismicity. Figure O-6. Mud loss data and occurrence of seismicity in November 2015 as PX-2 was drilled below 3,800 m Spatial patterns of seismicity Earthquakes large enough to be located precisely occurred during each of the five well stimulations. The earthquakes define two distinct spatial populations that are related to well stimulation activities. Earthquakes that occurred during or shortly after stimulation of PX-1 fall into one population, while those that occurred during or shortly after stimulation of PX-2 fall into the other (Figure O-7). Seismicity continues after individual stimulations ended, sometimes for weeks (Figure O-2). The mud loss event discussed above locates together with the PX-2 events. O-21

37 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Figure O-7. Perspective view of earthquakes associated with activity in PX-1 (blue) and PX-2 (red). Yellow star marks the mainshock hypocenter. Well trajectories are shown with the open hole sections for PX-1 and PX-2 in blue and red, respectively. The range of focal depths of earthquakes associated with the well stimulations is very restricted (Figure O-8). For earthquakes associated with PX-1, depths range between 3.7 and 4.4 km, a similar depth interval to the open-hole section of PX-1 (3.9~4.2 km). Earthquakes associated with PX-2 span the depth range from 3.8 to 4.4 km, compared with the open-hole interval from 4.2 to 4.3 km. Earthquakes in each zone exhibit both upward and downward growth with respect to the open-hole intervals where pressure entered the formation. O-22

38 Overseas Research Advisory Committee Report on the Pohang Earthquake Figure O-8. Histogram of earthquake focal depths. Depth of mainshock at 4.27 km indicated by arrow Seismicity associated with PX-1 Most of the earthquakes associated with PX-1 occurred during or shortly after the initial stimulation of the well in December This stimulation activated an inclined tabular volume with a height of 800 m, horizontal length of 500 m and width of 230 m. Minor seismic activity continued in the zone following injection, with the last located event occurring in mid-january The second stimulation of the well in August 2017 produced only a single earthquake, M W 1.2 that was large enough to locate with the surface seismic networks. A M W 2.0 earthquake occurred in September, 2017, in the PX-1 zone, four weeks after the stimulation ended. Thus, while the majority of activity occurred when the well was pressurized, seismicity lingered for weeks afterwards, as has been observed in many other hydraulic well stimulations [e.g. Yoon et al., 2017]. To better understand the evolution of seismicity in the PX-1 zone, the earthquake locations are projected onto the plane that best fits the distribution (Figure O-9). This plane is only an approximate representation of the structure of the seismicity, as the width and height are almost equal. Consequently, over-interpretation of the plane should be avoided and the projection is for illustrative purposes only. In Figure O-9 the rupture area of each earthquake is approximated by a circular crack with radius appropriate for the earthquake s magnitude. Note the clustering of the larger events cluster together at several locations, with the September 2017 event near the top of the cluster and on the edge of events induced in December As one earthquake occurs, stress is transferred to the periphery of the fault patch that ruptured, increasing the potential for additional activity. O-23

39 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Figure O-9. Earthquakes associated with (left) PX-1 injection and (right) PX-2 injection, projected onto the best-fitting plane in each case. For PX-1, the coordinates are relative to the center of the seismicity; earthquakes during and following the December 2016 stimulation are shown in blue, earthquakes during and following the August 2017 stimulation in magenta. For PX-2, the bottom of the open-hole section of the well is at (0,0), 375 m behind the plane and the intersection of the plane with the well at 3,800 m depth is marked by ; the mud loss event in November 2015 is shown in brown, events during and following first stimulation in February 2016 in green, events during and following second stimulation in April 2017 in blue, events during and following third stimulation in September 2017 in orange, and foreshocks on November 14 and 15 in red. For both images, the faulted area in each earthquake is approximated by the equivalent circular crack for a stress drop of 4 MPa. This value for stress drop is the global average for crustal earthquakes [Allmann and Shearer, 2009]. Song and Lee [2018] estimated the stress drop of the Pohang mainshock in the region near the hypocenter to be in the range from 2 to 4 MPa Seismicity associated with PX-2 The seismicity in the PX-2 cluster forms a tabular body striking 214 and dipping 43 to the NW. The zone has a strike length of 1000 m, dip length of 500 m and a width of 200 m. The best-fitting plane to the zone intersects PX-2 at 3,800 m depth. Most of the earthquakes locate within ±60 m of the plane. The earthquakes are projected onto the plane in Figure O-9 (right) with the approximate area of each earthquake s rupture shown by a circular crack model with a radius appropriate for its magnitude. This plane is a good approximation of the structure of the seismicity. The initial seismicity associated with the PX-2 cluster occurred during the drilling of the PX-2 well in November 2015, discussed above and as a consequence of the major mud loss event at 3,800 m depth. Significant volumes of drilling mud were lost to the formation at this time, accompanied by the first detected occurrence of seismicity at the project site. Of the eighteen earthquakes we detected, O-24

40 Overseas Research Advisory Committee Report on the Pohang Earthquake only the largest, M W 0.9 on November 30, 2015 could be located with confidence (Figure O-6, Figure O-9). It locates near the top of the PX-2 cluster. No further activity was detected after well control was re-established and casing set until the first PX-2 stimulation in February The first PX-2 stimulation produced only a modest seismic response (Figure O-2), with the largest event M W 1.6. More than 6 months after injection ended, a M W 1.1 event occurred in the same cluster. The second PX-2 stimulation in March and April 2017 induced a M W 3.2 earthquake on April 15, at a time when the well was shut in. Declining seismicity continued into mid-may. The third PX-2 stimulation in September 2017 produced only a modest seismic response, similar to the first stimulation, with a maximum magnitude event of M W 2.0. The last earthquake large enough to be located occurred on September 26, Forty-nine days later, on November 15, 2017, activity resumed in the PX-2 cluster with what proved to be the foreshocks of the Pohang earthquake. The foreshocks occurred immediately to the southwest of the area ruptured during the April 2017 stimulation. The largest and last locatable foreshock, M W 2.7, occurred just 7 minutes before the mainshock and expanded the ruptured area down-dip toward the mainshock hypocenter (Figure O-9). It is evident from the distribution of earthquakes in the PX-2 cluster that the Pohang mainshock initiated in an area that was strongly perturbed by not only the foreshocks but also by the entire sequence of earthquakes induced by injection into PX Focal mechanisms Focal mechanisms were obtained during this investigation for 53 earthquakes that occurred during and following the simulations and up until the M W 5.5 earthquake on 15 November Figure O-10 displays the focal mechanisms as a function of time. The geometries of key focal mechanisms and the planes defined by seismicity are listed in Table O-2. The strike/dip/rake parameters listed for each focal mechanism are averages of the suite of solutions compatible with the P-wave first-motion data. The highest-quality focal mechanisms from the three phases of PX-2 stimulation exhibit predominantly oblique strike-slip/reverse faulting. Most of the events, including the largest earthquake during the stimulation (M W 3.2 on April 15, 2017), have similar focal mechanisms to the foreshocks and the mainshock itself. This focal mechanism indicates oblique right-lateral slip on a NW-dipping plane or oblique left-lateral slip on the orthogonal E-dipping plane. The NW-dipping plane has a very similar geometry to the plane defined by PX-2 seismicity and to the fault plane of the M W 5.5 inferred by analysis of regional moment tensor and InSAR analysis [Grigoli et al., 2018]. This plane is well-oriented for slip according to the preferred stress model. O-25

41 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Figure O-10. Summary of the focal mechanisms computed for 53 events that occurred during the five phases of stimulation (red PX2; blue PX-1), the foreshocks of November 2017 and the M W 5.5 Pohang earthquake (black). Bright red and blue colors indicate the highest-quality focal mechanism solutions associated with PX-2 and PX-1, and the paler colors indicate poorer-quality solutions. Table O-2. Planes of interest defined by focal mechanisms, hypocenters, and mainshock observations. FMs first motions; NP nodal plane. No. Feature Basis Strike ( ) Dip ( ) Rake ( ) 1 Mainshock, NP1 First motions Mainshock, NP2 First motions April 2017 M W 3.2, NP1 First motions April 2017 M W 3.2, NP2 First motions PX-2 seismicity plane Fit to hypocenters PX-1 seismicity plane Fit to hypocenters Mainshock fault InSAR modeling [Grigoli et al., 2018] Mainshock fault Moment tensor [Grigoli et al., 2018] Seismicity associated with stimulation of PX-1 shows a broader range of focal mechanisms. Many of the 21 highest-quality events have focal mechanisms similar to that characteristic of PX-2 seismicity, but other events show either purer strike-slip faulting (e.g. 08:04 event on 19 December 2016 and 07:56 event on 20 December 2016) or oblique strike-slip/reverse faulting on N- or S-dipping O-26

42 Overseas Research Advisory Committee Report on the Pohang Earthquake planes (e.g. 10:04 event on 21 December 2016). The orientation of the plane used to project the PX-1 seismicity in Figure O-9 is not represented in individual focal mechanisms. Figure O-11 illustrates the orientations of the planes of interest listed in Table O-2, and the corresponding shear and effective normal stresses calculated using the preferred stress model of Table O-1. The Mohr circle illustrated is the same as shown in Figure O-4, but with the planes of interest added as colored circles. This analysis indicates that most of the fault planes were close to failure for the preferred stress model, and were oriented such that small increases in fluid pressure would cause slip. In particular, the west-dipping nodal plane of the mainshock inferred from local network observations (plane 1) was near-optimally oriented for frictional reshear in the preferred stress model described above. The east-dipping auxiliary plane of the mainshock focal mechanism (plane 2) is less well oriented for shear in the stress field, bolstering our interpretation of the other nodal plane as the mainshock fault plane. Similarly, the west-dipping nodal plane of the largest earthquake that occurred during stimulation (plane 3) was very susceptible to frictional failure whereas the east-dipping plane (plane 4) was not. The plane defined by seismicity associated with PX-2 injection (including the November 2015 mud loss events; plane 5) is also near-optimally oriented for frictional shear, as are the fault planes inferred from InSAR and moment tensor analysis by Grigoli et al. [2018] (planes 7 and 8). In contrast, the plane fit to the PX-1 seismicity (plane 6) is not well oriented for frictional reactivation in the preferred model of stress. Figure O-11. (left) Stereonet showing the orientations of planes of interest and the corresponding normal vectors, colored according to proximity to slip; red denotes planes closest to failure and green denotes planes furthest from failure. Blue dots mark the calculated shear vectors on each plane. (right) Mohr diagram calculated for the preferred stress model and a hydrostatic fluid pressure at a depth of (Table O-1). The black diagonal lines demarcate the stresses required for frictional reshear of a cohesionless plane with a friction coefficient of 0.6. The dots are numbered as in Table O-2 and colored according to the proximity of each plane to frictional failure as in the left-hand image. s(s) and s(n) denote shear and normal stress, respectively. O-27

43 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Similar results are obtained if the regional stress model described in Table O-1 is used, namely that the west-dipping nodal planes of the mainshock and M W 3.2 event s focal mechanisms were close to failure whereas the respective alternate planes were not. Moreover, the plane defined by PX-2 seismicity, which has a very similar geometry to the west-dipping nodal plane of the mainshock, was also close to frictional failure Magnitude-frequency characteristics The number of earthquakes of different magnitudes is universally observed to follow the empirical Gutenberg-Richter magnitude-distribution given by log 10 (N) = a bm, where N is the cumulative number of events greater than or equal to magnitude M. The a-value defines the productivity and the b-value the ratio of small to large events. For a given a-value, the smaller the b-value, the higher the probability of larger magnitude events being present in the sample. In virtually every tectonic region on Earth the b-value is close to 1.0±0.2. The b-value has been hypothesized to relate to stress state, among other variables, with lower values corresponding to higher stresses, according to both field and laboratory data [Scholz, 2015]. During several recent EGS projects, b-values ranged from a high of 1.58 in Basel, Switzerland [Bachmann et al., 2011], between 0.9 and 1.2 in Soultz-sous-Forêts, France [Dorbath et al., 2009], and a low of 0.83 in Cooper Basin, Australia [Baisch et al., 2009]. For the five Pohang stimulations a b-value of 0.73±0.1 is observed (Figure O-12). This relatively low b-value corresponds to a higher Figure O-12. Gutenberg-Richter magnitude frequency diagram. Solid symbols correspond to earthquakes occurring during or following the five well stimulations at the Pohang EGS facility. The dashed line has the formula log 10(N) = M L and was determined using the method of Tinti and Mulargia [1985]. Open symbols include foreshocks, mainshock and stimulation events. O-28

44 Overseas Research Advisory Committee Report on the Pohang Earthquake likelihood of a large magnitude event compared to either the other EGS projects or to global tectonic seismicity generally. It should be noted that the low b-values in the Cooper Basin and the lowest at Soultz-sous-Forêts corresponded to stimulations that activated discrete faults, where the high b-value stimulation at Basel activated a volume. 5. Model-based analyses of triggering mechanisms Models are important scientific tools for relating measurements or other forms of data to observations. In this section, we discuss the ability of certain physical phenomena to explain the occurrence of the Pohang earthquake. Although the individual models are based on very different data and hypotheses, each relates a physical change to its potential to trigger the Pohang earthquake Effects of Tohoku and Gyeongju earthquakes The 2011 M W 9.0 Tohoku earthquake produced small but measurable displacements across the Korean Peninsula [Kim and Bae, 2012]. Sites on the eastern side of the Peninsula were displaced eastward by larger amounts than sites on the western side of the Peninsula, meaning that the induced strains were extensional; that is, the Korean Peninsula was stretched in an east west direction. Hong et al. [2015] considered the changes in stress resulting from these geodetically measured strains and compared them with Coulomb failure stress perturbations. They obtained estimates of the tensional stress changes at mid-crustal depths of 1 7 kpa, which are of similar magnitude to the <3 kpa reductions in Coulomb failure stress they calculated for optimally oriented strike-slip and reverse faults. In other words, the overall effect of the Tohoku earthquake on the Korean Peninsula was to slightly reduce the stresses causing strike-slip or reverse faulting on optimally oriented faults. This effect is referred to as a stress shadow as it reduces the potential for an earthquake to occur [Harris, 1998]. We agree with Hong et al. [2015] that large subduction earthquakes have been observed to influence crustal stresses at regional distances. We also note that a release of seismicity characterized by periods of high activity and long periods of seismic quiescence is observed in several stable or lowseismicity areas of the world, located away from active plate margins, and would not be in itself unusual in Korea. It has been suggested that the effect of the Tohoku earthquake had been to hasten the time of the M L 5.1 (foreshock) and M L 5.8 Gyeongju earthquakes in 2016 and that static stress perturbations caused by those events triggered the M W 5.5 Pohang earthquake in 2017 [Hong et al., 2018]. This O-29

45 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 interpretation is based on the assertion that seismicity rates increased throughout the Korean region after 2011 and that the Gyeongju earthquakes increased Coulomb failure stresses near Pohang by ~200 Pa. This value is substantially smaller than previously observed triggering thresholds [Reasenberg and Simpson, 1992]. In contrast, the Coulomb failure stress analysis by Grigoli et al. [2018] concluded that the Gyeongju earthquake did not play a role in triggering the Pohang earthquake 14 months later. Hong et al. [2018] observed that no seismicity of magnitude 2 or larger was observed within 10 km of the 2017 earthquake s epicenter prior to the 2016 Gyeongju earthquakes, whereas four earthquakes of this size occurred within 3 km of the 2017 earthquake s epicenter after the 2016 earthquakes. They interpreted this to indicate that the Gyeongju earthquakes triggered low-magnitude seismicity near Pohang and ultimately the M W 5.5 Pohang earthquake. The occurrence of seismicity near the Pohang EGS site following the Gyeongju earthquakes and not before does not imply a causative relationship between the Gyeongju and Pohang earthquakes. On the contrary, the locations, timing, and focal mechanisms of the M L 2+ earthquakes observed near Pohang in 2017 show that they were induced by EGS activities, as discussed in Sections 4.1 and Hydrogeologic analysis of fluid pressure perturbations When a well is stimulated by injection of fluid under pressure, it changes the state of stress in the Earth. Because there were no direct measurements of changing fluid pressure or stress made as part of the Pohang EGS project, it is necessary to use physics-based model to develop an understanding of how injection may have affected fault stability. Numerical models of the fluid pressure perturbations associated with the five phases of stimulation were developed during the course of this investigation. Full details of these models are contained in Chapter 6 and omitted here. The hydrogeologic regime surrounding the Pohang EGS site can be treated as the superposition of the pre-drilling state and any perturbations associated with drilling and injection. The models developed to date presume that an undisturbed, hydrostatic fluid pressure regime existed prior to stimulation, and therefore do not address the perturbations associated with the long phase of drilling or the mud loss event in October Two models, referred to as Case A and Case B below, have been developed to illustrate key features of pore pressure diffusing away from the PX-1 and PX-2 injection points. Each model represents a 5 km 5 km 5 km domain and incorporates two faults (Figure O-13). The faults are embedded in bedrock with a homogeneous hydraulic diffusivity of m 2 /s. The existence and geometries of the two faults are based on hydrologic analysis of the stimulation data and the seismological results described in Chapter 5 of the Summary Report. The first fault separates PX-1 and PX-2 and represents the mainshock plane, having an orientation (strike/dip) of 214 /43 and intersecting PX-2 at 3,810 m. It acts to compartmentalize the fluid pressure response. The second O-30

46 Overseas Research Advisory Committee Report on the Pohang Earthquake fault represents a high-permeability feature inferred to be present near PX-1. The hydrologic properties of the faults have been specified on the basis of representative models of fault zone structure [Caine et al., 1996; Choi et al., 2015] and laboratory measurements of the fault gauge and breccia samples from lithologies analogous to the basement rock at Pohang [Kim et al., 2018a]. Figure O-13. Pore pressure model setup, dimension, location of the bottom sections of PX-1 and PX-2 in the model, and the two faults. See Chapter 6 for further details. Case A: The mainshock fault plane is modeled as having a 10 m-thick low-permeability fault core (D = m 2 /s) bounded on both sides by a 85 m-thick high-permeability damage zone (D = 0.1 m 2 /s) [Kim et al., 2018a]. The second fault is a smaller, 130 m-thick, highpermeability feature (D = 1 m 2 /s) near PX-1. Case B: The fault locations and geometries are the same as in Case A but the mainshock fault plane does not have a low-permeability core. The spatial patterns of modeled pore pressure changes are illustrated on vertical cross-sections through the mainshock fault plane in Figure O-14a for two different epochs. The upper row shows the results computed for April 15, 2017 (M W 3.2 earthquake) and the lower row shows the results for November 15, 2017 (M W 5.5 mainshock); the results for Cases A and B are illustrated on the leftand right-hand sides, respectively. Figure O-14b shows the calculated 0.02 MPa isosurfaces on November 15, O-31

47 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 (a) (b) Figure O-14. (a) Modeled pore pressure change in vertical cross section for Case A (left) and Case B (right) for April 15 and November 15, (b) 0.02 MPa isosurface of pressure change. For both cases considered, pressure changes exceeding 0.1 MPa had developed around PX-2 by April 15, The pressure lobe surrounding PX-2 was produced by the third stimulation from March 16 to April 14, 2017, which immediately preceded the M W 3.2 earthquake. By November 2017, the extent of the combined fluid pressure perturbations had expanded, with the two distinct lobes resulting primarily from the fourth stimulation in PX-1 and the fifth in PX-2. The temporal evolution of pore pressure at the hypocenters of the M W 3.2 and M W 5.5 earthquakes is illustrated in Figure O-15. The model results suggest that pore pressure had been O-32

48 Overseas Research Advisory Committee Report on the Pohang Earthquake elevated by MPa at the hypocenter of the M W 3.2 event by April 15, 2017, largely as a consequence of the third stimulation phase in PX-2. By November 15, 2017, the modeling suggests pore pressure had risen by approximately 0.07 MPa at the hypocenter of the M W 5.5 earthquake. Pore pressure changes of more than 0.01 MPa have been shown to reduce fault strength and trigger earthquakes [Reasenberg and Simpson, 1992]. The geomechanical results presented above and in Chapter 4 indicate that the mainshock fault plane was critically stressed prior to the Pohang earthquake, and imply that small increases in fluid pressure would trigger slip. The fluid pressure modeling conducted to date indicates that fluid pressure increases of greater than 0.01 MPa were likely to have occurred at distances of several hundred meters from the injection points and to have persisted for weeks or months after injection ended. Figure O-15. Pore pressure change with time at the M W 3.2 (a) and M W 5.5 (b) hypocenters. 6. M W 5.5 Pohang earthquake of 15 November Foreshock activity In mid-november 2017, seismicity restarted on the fault activated by injection into PX-2 (Figure O-9). The five largest events were recorded over a period of about 10 hours, between 19:55 on November 14 and 05:22 on November 15, with a magnitude progression increasing from M W 1.6 to M W 2.7. These events were immediately followed by the main M W 5.5 shock, occurring at 05:29 on November 15. According to information provided to the ORAC, no further injections or other activities were O-33

49 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 carried out in the boreholes after the third PX-2 injection in September The events of November 14 15, occurring two months after the third injection in PX-2, can therefore be considered as foreshocks of the main Pohang event of November 15. This conclusion is confirmed by the fact that the initial portions of the foreshocks and mainshock seismograms show highly correlated, substantially identical waveforms, indicating that the foreshocks had the same focal mechanisms and were located on the same focal plane of the mainshock Location and timing of mainshock The mainshock of November 15, 2017 occurred 58 days after the last injection activities in PX-2. This delay has been used to argue that the mainshock has no causal connection to the EGS activities in Pohang. A delay of weeks and months between tectonic events occurring on adjacent fault segments is commonly observed, with seismic sequences developing in some cases over years and propagating to adjacent faults. A recent example is the sequence occurring in 2016 in the Central Apennines region of Italy, with four main episodes of seismicity occurring over several months [Chiaraluce et al., 2017]. The causal link in natural seismicity, even with delays of several months, is not disputed. A similar delay has also been observed in well-documented occurrences of induced seismicity, for example in the case of wastewater injection in Oklahoma [Keranen et al., 2014; Schoenball and Ellsworth, 2017]. The first documented case of earthquakes induced by injection occurred in the 1960s near Denver, Colorado, where a deep well was used to dispose of waste by injection at the Rocky Mountain Arsenal [Healy et al., 1968]. Injection into the Precambrian basement took place between March 1962 and February 1966, and the rate of injection was strongly correlated with the earthquake rate. However, the largest earthquake, M W 4.8, struck in April, 1967 more than one year after injection had been terminated. At Basel, Switzerland, activity continued for more than a year after pressure was bled off, with multiple magnitude 3 earthquakes occurring [Deichmann and Giardini, 2009]. On the basis of these observations, of both natural and induced seismicity, the separation in time between stimulation activities in PX-2 ending and the occurrence of the mainshock cannot be considered a reason to exclude a triggering effect of the EGS activities. On the contrary, there are strong elements indicating a causal link between the seismicity induced by the PX-2 stimulations and the foreshocks and mainshock of November Indeed, the foreshocks (November 14, 2017, at 20:04 and 20:59) have the same waveform signature as the events that occurred during the last PX-2 stimulation (September 15, 2017, at 19:33; September 16, 2017, at 08:55), indicating that the PX-2 seismicity and the foreshocks are part of the same sequence of events and occurred on the same focal plane as the mainshock. The same correlation is not found for events associated with PX-1 stimulations. The foreshocks are contiguous with the previously ruptured O-34

50 Overseas Research Advisory Committee Report on the Pohang Earthquake area along the fault stimulated by injection into PX-2, extending the area approximately 200 m to the SW (Figure O-9). The mainshock hypocenter sits immediately below the foreshocks, where stresses had been increased by the foreshocks and earlier events. From the location of the mainshock hypocenter alone, it is evident that this earthquake is directly related to the preceding activity Mainshock focal mechanism Figure O-16 illustrates the observed focal mechanism representing the initiation of the mainshock and the focal mechanism calculated by resolving different stress models on the best-fitting plane fit to the PX-2 seismicity, assuming that slip occurs in the direction of maximum resolved shear stress. In each case, the calculated focal mechanism is similar to that observed, indicating oblique reverse/ strike-slip motion on the assumed west-dipping fault plane. For the preferred stress model, slip on this plane is calculated to have a rake of 141, while the regional stress model yields a rake of 158. Given uncertainties in the focal mechanism parameters and the stress models, the differences between the observed and predicted focal mechanisms are within acceptable bounds. We conclude from this analysis that the stress models listed in Table O-1 are consistent with the geometry of slip during the mainshock. In other words, the mainshock focal mechanism, and the focal mechanisms of the foreshocks and several events associated with stimulation of PX-2, have a geometry that can be accounted for using a known fault geometry and plausible models of stress. Figure O-16. Observed P-wave mainshock focal mechanism (red beachball; strike/dip/rake = 214 /51 /128 ) and focal mechanisms calculated using the PX-2 seismicity plane and different models of stress (black beachballs). The value of the rake (λ) calculated for each of the stress models is printed above the corresponding beachball. For each focal mechanism, the white dot marks the T axis and the red or black dot the P axis. O-35

51 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 6.4. Magnitude of mainshock and previous scaling arguments It has been argued that the sizes of earthquakes induced by stimulation can be managed by controlling the pressure, rate and location of where fluid enters the rock mass by allowing time for pressure to diffuse when seismicity rates escalated [Hofmann et al., 2018]. The threshold magnitudes for traffic light systems have often been set to avoid earthquakes that pose a shaking nuisance and/or risk of damage. Part of the rationale for selecting the magnitude thresholds comes from an empirical hypothesis that the largest magnitude of induced earthquakes is bounded by a function of the injected volume [Galis et al., 2017; McGarr, 2014]. If correct, this volume hypothesis would enable the hazard to be managed prescriptively by simply maintaining the net injection volume below a certain value. However, an alternative analysis of the same cases found that the observed maximum magnitude was well modeled by independent random sampling of the Gutenberg-Richter distribution log 10 (N) = a bm, where N is the cumulative number of events greater than or equal to M [van der Elst et al., 2016]. In this interpretation, the largest event in an induced seismicity sequence is not related to the Figure O-17. Comparison of injected fluid volume with maximum earthquake magnitude in a global injection data set compiled by Galis et al. [2017]. Scaling lines for maximum arrested rupture from their paper. Scaling line for maximum earthquake according to McGarr [2014] shown in gray. Maximum magnitude at different times during the development of the Pohang EGS project shown by stars. Figure adapted from Figure 4 of Galis et al. [2017]. O-36

52 Overseas Research Advisory Committee Report on the Pohang Earthquake injection volume, but to pre-existing tectonic conditions and the number of earthquakes induced. The greater the number of earthquakes, the higher the odds of one of those earthquake being large. The Pohang earthquake contradicts the volume hypothesis, as the injected volume was less than 1/500 th of the amount expected to produce a M W 5.5 earthquake (Figure O-17). This discrepancy would be larger if the net volume (injection minus extraction) were considered instead of injection alone. Once initiated, the Pohang earthquake grew through the release of tectonic stress rather than being limited by the injected volume. The earthquake was almost two magnitude units larger than the M W 3.7 predicted by one model [McGarr, 2014] and exceeded the maximum arrested earthquake size predicted by the other [Galis et al., 2017] and therefore constituted a runaway earthquake in their terminology Aftershock activity Once initiated, the November 15, 2017 Pohang earthquake grew outward from its hypocenter and beyond the ~1000 m-long segment of the fault that had been activated by the stimulations of PX-2. The aftershock activity that followed the mainshock illuminated this plane further (Figure O-18). Figure O-18. Longitudinal cross section along the Pohang earthquake fault plane showing aftershocks recorded on the day following the M W 5.5 earthquake (November 16, 2017; gray circles). PX-2 well shown by black line. Hypocenters of earthquakes stimulated by injection into PX-2 in red; mainshock yellow star. O-37

53 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 7. Findings, conclusion and lessons learned We first present our findings with respect to (1) regional setting, (2) site geology, (3) fluid pressure perturbations, (4) seismicity associated with injection, and (5) the M W 5.5 mainshock of November 15, Based on these findings, we then formulate our overall conclusion regarding the origin of the Pohang earthquake, and draw some lessons of a general nature Findings ORAC s findings are as follows: 1. Regional setting A. The Korean Peninsula is located on the continental margin of the Eurasian plate, which underwent extension during the opening of the East Sea. The region is now under tectonic compression and previously extensional faults with appropriate orientations can be reactivated with reverse or strike-slip kinematics. B. The present-day regional stress field shows compression oriented ENE WSW and several recognized active fault systems in the region are susceptible to slip in this stress field. C. The stresses acting on regional faults are high, approaching the static stability of the faults, as confirmed by pre-drilling assessment of stress conditions in Pohang. The occurrence of the M W 5.4 Gyeongju event of September 12, 2016, on the Yangsan fault system is consistent with this analysis. D. The historical seismic record shows periods of high activity, including earthquakes exceeding the size of the 2016 Gyeongju and 2017 Pohang earthquakes. E. Regional deformation following the 2011 M W 9.0 Tohoku earthquake may have affected seismic activity in the Korean Peninsula. However, the calculated effects of the regional deformation and the seismicity do not explain the occurrence of the Pohang earthquake. 2. Site geology F. Neither geological investigations in the Pohang area nor geophysical surveys performed during the selection of the EGS site identified the fault activated by the Pohang earthquake. G. Fault gouge observed in drill cuttings from the PX-2 well indicates the presence of a fault at a depth of approximately 3,800 m. 3. Fluid pressure perturbations H. Multiple evidence suggests that the PX-1 and PX-2 wells occupy different hydraulic regimes: Injection tests carried out during hydraulic stimulations indicated the presence of a flow barrier separating the two wells. Two distinct seismicity populations, separated in space and time, were observed during O-38

54 Overseas Research Advisory Committee Report on the Pohang Earthquake successive stimulations of the PX-1 and PX-2 wells. Injection conditions in the two wells were different, requiring a maximum well-head pressure of 24 MPa in PX-1 and almost 90 MPa in PX-2. I. Modeling performed with representative hydrological properties and high-permeability and low-permeability fault cores shows that the pressure perturbations produced by stimulation of PX-2 propagated several hundred meters. The pore pressure increases near the hypocenters of the M W 3.2 and M W 5.5 events exceeded 0.05 MPa. J. Detectable seismicity occurred during drilling of PX-2 over a period of one month, following the mud loss event at about 3,800 m depth, induced by the weight of the mud column entering the formation. 4. Seismicity associated with the PX-1 and PX-2 injections K. Each of the five stimulations induced seismicity. After each stimulation, seismicity continued for up to several months. L. The seismicity induced by the stimulations ranges in depth between 3.7 and 4.4 km, spanning the open sections of the two boreholes. M. Seismicity induced by the three stimulations in PX-2 did not produce a detectable seismic response within 200 m of the well but activated an approximately 1000 m-long, 600 m-high fault zone aligned with the fault traversing PX-2 at 3,800 m and corresponding to the west-dipping plane of the M W 5.5 Pohang mainshock focal mechanism. N. The west-dipping nodal planes of the focal mechanisms of events induced by PX-2 injection agree with the orientation of the stimulated fault zone. Their oblique reverse motion is well explained by the local stress field. O. The magnitude-frequency distribution of the events induced before the mainshock shows a higher proportion of larger magnitudes than normally observed in tectonic sequences, as indicated by the b-value of Mainshock of November 15, 2017 P. The mainshock was preceded by foreshocks over a period of 24 hours, with a sequence of events of increasing size culminating in an event of M W 2.7 seven minutes before the mainshock. These foreshocks extended laterally the fault zone activated by seismicity induced by PX-2 stimulations. They have similar focal mechanisms to the mainshock and the events induced by the PX-2 stimulations. Q. The mainshock initiated within the fault zone activated by the PX-2 stimulations, at 4.3 km depth. R. The delay of almost two months between the last PX-2 stimulated events and the mainshock is consistent with similar delays observed in earlier stimulations in Pohang and commonly observed in natural and induced seismic sequences. A delay of this length does not preclude O-39

55 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 a causal effect. S. The orientation of the fault activated by the mainshock is similar to that of other faults in the region. The geometry of the initial slip in the mainshock is well explained by the combination of the fault geometry and the state of stress surrounding the borehole. T. The size of the mainshock is consistent with a triggered origin according to the published analyses of van der Elst et al. [2016] and Galis et al. [2017], and is inconsistent with the hypotheses of McGarr [2014] or Galis et al. [2017] that relate the maximum magnitude of an induced earthquake to the injected volume Conclusion ORAC concludes that the Pohang earthquake was triggered by the EGS stimulation. Seismicity induced by injection activated a previously unmapped fault zone, which in turn triggered the mainshock. A schematic representation of this conclusion is illustrated in Figure O-19. Figure O-19. Schematic illustration of the sequence of seismicity associated with stimulation of PX-1 and PX-2, and the relationship of the seismicity to the rupture plane of the M W 5.5. Pohang earthquake. The view is towards the northeast. The gray grid has 1 km spacing and extends from the surface to 6 km depth. The mainshock fault plane extends from 2.5 km to 6 km and intersects the PX-2 well at 3.8 km. Open hole section of PX-1 and associated seismicity shown in blue sits above and in the hanging wall of the fault plane. The fault plane cuts the seismicity associated with PX-2, shown in red. O-40

56 Overseas Research Advisory Committee Report on the Pohang Earthquake 7.3. Lessons learned Lessons of a general nature can be learned from the Pohang experience, and can serve to increase the safety of future EGS projects in Korea and elsewhere. 1. The Pohang event had a complex origin: seismicity induced by injection activated a previously unmapped fault, which in turn triggered the mainshock. Current models do not cover adequately this complexity and the possibility that pressure perturbations induced on a fault may trigger run-away events of large magnitudes. Physical and statistical models of induced and triggered seismicity need to be further developed to provide reliable assessments of probabilities and uncertainties for inclusion in risk assessments of future EGS projects. 2. The analyses and investigations referenced in this report were done only after the occurrence of the M W 5.5 Pohang mainshock, but they would have been possible during the sequence of stimulations, lasting almost two years. All the data required to re-evaluate seismic risk were collected and the most important evidence was available in April 2017 after the second stimulation in PX-2. In future EGS projects, the project team and the scientific institutions involved should engage in timely and adequate efforts to monitor, analyze and interpret the evolution of any earthquake sequence, and provide information to the public authorities on the developing seismic risk conditions. 3. Several institutions from Korea and other countries were active in different capacities in the monitoring and analysis of the seismicity in Pohang. This complicated the exchange and analysis of data and samples. Scientific institutions involved in monitoring and evaluation activities with relevance to the assessment and mitigation of seismic risk such as the risk potentially associated with an EGS project in the vicinity of a major city should prioritize an open-access policy for data and samples and clear channels of cooperation to maximize their contribution to the mitigation of seismic risk. 4. The Pohang EGS project was located close to a major city, port and industrial center. This proximity raised clear issues of seismic risk, governance and mitigation. It is crucial that strategies and tools for monitoring, mitigating and communicating the risk of induced seismicity are established together with responsible authorities. Seismic risk scenarios should be developed to evaluate the possible consequences and to identify risk mitigation measures. A risk-based framework for making operational decisions should always be used O-41

57 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 and updated as new knowledge is acquired. 5. Operational decision-making in the EGS project was internal to the project team. An independent oversight committee/authority should be established to provide assurance that all aspects of the project plan, protocols and standards are designed and conducted with appropriate considerations of seismic risk. O-42

58 Overseas Research Advisory Committee Report on the Pohang Earthquake 8. References Allmann, B. P., and P. M. Shearer (2009), Global variations of stress drop for moderate to large earthquakes, Journal of Geophysical Research: Solid Earth, 114 (1), doi: /2008jb Bachmann, C. E., S. Wiemer, J. Woessner, and S. Hainzl (2011), Statistical analysis of the induced Basel 2006 earthquake sequence: Introducing a probability-based monitoring approach for Enhanced Geothermal Systems, Geophysical Journal International, 186 (2), , doi: / j x x. Baisch, S., R. Vörös, R. Weidler, and D. Wyborn (2009), Investigation of fault mechanisms during geothermal reservoir stimulation experiments in the Cooper Basin, Australia, Bulletin of the Seismological Society of America, 99 (1), , doi: / Caine, J. S., J. P. Evans, and C. B. Forster (1996), Fault zone architecture and permeability structure, Geology, 24 (11), , doi: / (1996)024<1025:fzaaps>2.3.co;2. Cheon, Y., M. Son, C. W. Song, J. S. Kim, and Y. K. Sohn (2012), Geometry and kinematics of the Ocheon Fault System along the boundary between the Miocene Pohang and Janggi basins, SE Korea, and its tectonic implications, Geosciences Journal, 16 (3), , doi: / s Chiaraluce, L., et al. (2017), The 2016 central Italy seismic sequence: A first look at the mainshocks, aftershocks, and source models, Seismological Research Letters, 88 (3), , doi: / Choi, J. H., et al. (2019), Surface Deformations and Rupture Processes Associated with the 2017 M W 5.4 Pohang, Korea, EarthquakeSurface Deformations and Rupture Processes Associated with the 2017 M W 5.4 Pohang, Korea, Earthquake, doi: / Choi, J. H., S. J. Yang, S. R. Han, and Y. S. Kim (2015), Fault zone evolution during Cenozoic tectonic inversion in SE Korea, Journal of Asian Earth Sciences, 98, , doi: / j.jseaes Chough, S. K., S. T. Kwon, J. H. Ree, and D. K. Choi (2000), Tectonic and sedimentary evolution of the Korean peninsula: A review and new view, Earth Science Reviews, 52 (1-3), , doi: /s (00) Deichmann, N., and D. Giardini (2009), Earthquakes Induced by the stimulation of an enhanced geothermal system below Basel (Switzerland), Seismological Research Letters, 80 (5), , doi: /gssrl Dorbath, L., N. Cuenot, A. Genter, and M. Frogneux (2009), Seismic response of the fractured and faulted granite of Soultz-sous-Forêts (France) to 5 km deep massive water injections, Geophysical Journal International, 177 (2), , doi: /j x x. O-43

59 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Ellsworth, W. L. (2013), Injection-induced earthquakes, Science, 341 (6142), doi: /science Galis, M., J. P. Ampuero, P. M. Mai, and F. Cappa (2017), Induced seismicity provides insight into why earthquake ruptures stop, Science Advances, 3 (12), doi: /sciadv.aap7528. Gihm, Y. S., et al. (2018), Paleoseismological implications of liquefaction-induced structures caused by the 2017 Pohang Earthquake, Geosciences Journal, 22 (6), , doi: /s y. Grigoli, F., S. Cesca, E. Priolo, A. P. Rinaldi, J. F. Clinton, T. A. Stabile, B. Dost, M. G. Fernandez, S. Wiemer, and T. Dahm (2017), Current challenges in monitoring, discrimination, and management of induced seismicity related to underground industrial activities: A European perspective, Reviews of Geophysics, 55 (2), , doi: /2016rg Grigoli, F., S. Cesca, A. P. Rinaldi, A. Manconi, J. A. López-Comino, J. F. Clinton, R. Westaway, C. Cauzzi, T. Dahm, and S. Wiemer (2018), The November 2017 M w 5.5 Pohang earthquake: A possible case of induced seismicity in South Korea, Science, 360 (6392), , doi: / science.aat2010. Harris, R. A. (1998), Introduction to special section: Stress triggers, stress shadows, and implications for seismic hazard, Journal of Geophysical Research: Solid Earth, 103 (10), Healy, J. H., W. W. Rubey, D. T. Griggs, and C. B. Raleigh (1968), The Denver earthquakes, Science, 161 (3848), , doi: /science Hofmann, H., G. Zimmermann, A. Zang, and K. B. Min (2018), Cyclic soft stimulation (CSS): a new fluid injection protocol and traffic light system to mitigate seismic risks of hydraulic stimulation treatments, Geothermal Energy, 6 (1), doi: /s Hong, T. K., J. Lee, and S. E. Houng (2015), Long-term evolution of intraplate seismicity in stress shadows after a megathrust, Physics of the Earth and Planetary Interiors, 245, 59-70, doi: /j.pepi Hong, T. K., J. Lee, W. Kim, I. K. Hahm, N. C. Woo, and S. Park (2017), The 12 September 2016 M L 5.8 midcrustal earthquake in the Korean Peninsula and its seismic implications, Geophysical Research Letters, 44 (7), , doi: /2017gl Hong, T. K., J. Lee, S. Park, and W. Kim (2018), Time-advanced occurrence of moderate-size earthquakes in a stable intraplate region after a megathrust earthquake and their seismic properties, Scientific Reports, 8 (1), doi: /s Houng, S. E., and T. K. Hong (2013), Probabilistic analysis of the Korean historical earthquake records, Bulletin of the Seismological Society of America, 103 (5), , doi: / Keranen, K. M., M. Weingarten, G. A. Abers, B. A. Bekins, and S. Ge (2014), Sharp increase in central Oklahoma seismicity since 2008 induced by massive wastewater injection, Science, 345 (6195), , doi: /science O-44

60 Overseas Research Advisory Committee Report on the Pohang Earthquake Kim, H., L. Xie, K.-B. Min, S. Bae, and O. Stephansson (2017), Integrated In Situ Stress Estimation by Hydraulic Fracturing, Borehole Observations and Numerical Analysis at the EXP-1 Borehole in Pohang, Korea, Rock Mechanics and Rock Engineering, 50 (12), , doi: /s Kim, J. H., J.-H. Ree, C. Park, C.-M. Kim, R. Han, and T. K. Shimamoto, H.-C. (2018a), Proxies for the 2017 Pohang earthquake fault and modeling of fluid flow (S23B-0519), in AGU Fall Meeting Abstracts, edited, Washington, D.C. Kim, K. H., et al. (2018b), Deep fault plane revealed by high-precision locations of early aftershocks following the 12 September 2016 M L 5.8 Gyeongju, Korea, earthquake, Bulletin of the Seismological Society of America, 108 (1), , doi: / Kim, K. H., J. H. Ree, Y. Kim, S. Kim, S. Y. Kang, and W. Seo (2018c), Assessing whether the 2017 Mw5.4 Pohang earthquake in South Korea was an induced event, Science, 360 (6392), , doi: /science.aat6081. Kim, S. K., and T. S. Bae (2012), Analysis of crustal deformation on the Korea peninsula after the 2011 Tohoku earthquake, Journal of the Korean Society of Surveying Geodesy Photogrammetry and Cartography, 30 (1), 87-96, doi: /ksgpc Kim, Y. H., J. Rhie, T. S. Kang, K. H. Kim, M. Kim, and S. J. Lee (2016), The 12 September 2016 Gyeongju earthquakes: 1. Observation and remaining questions, Geosciences Journal, 20 (6), , doi: /s x. Kwon, S., L. Xie, S. Park, K. I. Kim, K. B. Min, K. Y. Kim, L. Zhuang, J. Choi, H. Kim, and T. J. Lee (2018), Characterization of 4.2-km-Deep Fractured Granodiorite Cores from Pohang Geothermal Reservoir, Korea, Rock Mechanics and Rock Engineering, doi: /s Lee, H., Y. J. Shinn, S. H. Ong, S. W. Woo, K. G. Park, T. J. Lee, and S. W. Moon (2017a), Fault reactivation potential of an offshore CO 2 storage site, Pohang Basin, South Korea, Journal of Petroleum Science and Engineering, 152, , doi: /j.petrol Lee, J., T. K. Hong, and C. Chang (2017b), Crustal stress field perturbations in the continental margin around the Korean Peninsula and Japanese islands, Tectonophysics, 718, , doi: /j.tecto Lee, J., et al. (2018), Seismicity of the 2016 M L 5.8 Gyeongju earthquake and aftershocks in South Korea, Geosciences Journal, 22 (3), , doi: /s z. Lee, K., and W. S. Yang (2006), Historical seismicity of Korea, Bulletin of the Seismological Society of America, 96 (3), , doi: / Lee, T. J., Y. Song, D.-W. Park, J. Jeon, and W. S. Yoon (2015), Three dimensional geological model of Pohang EGS pilot site, Korea, paper presented at Proceedings of the World Geothermal Congress, Melbourne, Australia, April O-45

61 포항지진과지열발전의연관성에관한해외조사위원회요약보고서 Lee, Y., S. Park, J. Kim, H. C. Kim, and M. H. Koo (2010), Geothermal resource assessment in Korea, Renewable and Sustainable Energy Reviews, 14 (8), , doi: /j.rser McGarr, A. (2014), Maximum magnitude earthquakes induced by fluid injection, Journal of Geophysical Research: Solid Earth, 119 (2), , doi: /2013jb Park, J. C., W. Kim, T. W. Chung, C. E. Baag, and J. H. Ree (2007), Focal mechanisms of recent earthquakes in the Southern Korean Peninsula, Geophysical Journal International, 169 (3), , doi: /j x x. Park, Y., J. H. Ree, and S. H. Yoo (2006), Fault slip analysis of Quaternary faults in southeastern Korea, Gondwana Research, 9 (1-2), , doi: /j.gr Reasenberg, P. A., and R. W. Simpson (1992), Response of regional seismicity to the static stress change produced by the Loma Prieta earthquake, Science, 255 (5052), , doi: / science Ree, J. H., and S. T. Kwon (2005), The Wangsan Fault: One of the most 'active' faults in South Korea?, Geosciences Journal, 9 (3), , doi: /bf Ree, J. H., Y. J. Lee, E. J. Rhodes, Y. Park, S. T. Kwon, U. Chwae, J. S. Jeon, and B. Lee (2003), Quaternary reactivation of tertiary faults in the southeastern Korean Peninsula: Age constraint by optically stimulated luminescence dating, Island Arc, 12 (1), 1-12, doi: /j x. Schoenball, M., and W. L. Ellsworth (2017), A Systematic Assessment of the Spatiotemporal Evolution of Fault Activation Through Induced Seismicity in Oklahoma and Southern Kansas, Journal of Geophysical Research: Solid Earth, 122 (12), 10,189-"110,206", doi: /2017jb Scholz, C. H. (2015), On the stress dependence of the earthquake b value, Geophysical Research Letters, 42 (5), , doi: /2014gl Soh, I., C. Chang, J. Lee, T. K. Hong, and E. S. Park (2018), Tectonic stress orientations and magnitudes, and friction of faults, deduced from earthquake focal mechanism inversions over the Korean Peninsula, Geophysical Journal International, 213 (2), , doi: / gji/ggy061. Son, M., C. W. Song, M.-C. Kim, Y. Cheon, H. Cho, and Y. K. Sohn (2015), Miocene tectonic evolution of the basins and fault systems, SE Korea: dextral, simple shear during the East Sea (Sea of Japan) opening, Journal of the Geological Society, 172 (5), , doi: / jgs Song, S. G., and H. Lee (2018), Static Slip Model of the 2017 Mw 5.4 Pohang, South Korea, Earthquake Constrained by the InSAR Data, Seismological Research Letters, 90 (1), , doi: / Tinti, S., and F. Mulargia (1985), Effects of magnitude uncertainties on estimating the parameters in O-46

62 Overseas Research Advisory Committee Report on the Pohang Earthquake the Gutenberg-Richter frequency-magnitude law, Bulletin of the Seismological Society of America, 75 (6), van der Elst, N. J., M. T. Page, D. A. Weiser, T. H. W. Goebel, and S. M. Hosseini (2016), Induced earthquake magnitudes are as large as (statistically) expected, Journal of Geophysical Research: Solid Earth, 121 (6), , doi: /2016jb Yoon, C. E., Y. Huang, W. L. Ellsworth, and G. C. Beroza (2017), Seismicity During the Initial Stages of the Guy-Greenbrier, Arkansas, Earthquake Sequence, Journal of Geophysical Research: Solid Earth, 122 (11), , doi: /2017jb O-47

63

64 Table of Contents Page iii vii 서문 / Preface 조사내용및결과요약 O-1 해외조사위원회요약보고서 Overseas Research Advisory Committee Report on the Pohang Earthquake 1 제 1 장포항지열발전실증연구프로젝트개요 Overview of the Pohang EGS (Enhanced Geothermal System) Demonstration Research Project 3 요약 / Abstract 포항지열발전실증연구시설현황과구축과정 수리자극실시현황 7 제2장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 Analysis of Geological Structures in and around the Pohang EGS Site to Interpret the Relationship between Pohang Earthquake and Hydraulic Stimulations 9 요약 / Abstract 연구배경 연구방법 연구결과

65 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Table of Contents Page 마이오세포항분지일원의광역지질과지질구조특성 포항지열발전실증시설지하의암상분포와단층특성 주요암상경계 단층대 21 1) Master log 분석 22 2) PX-2 커팅시료육안관찰 23 3) PX-2 커팅시료현미경관찰 26 4) PX-2 커팅시료의 X-선회절분석 결론 참고문헌 31 제 3 장지구물리탐사및자료해석 Geophysical Exploration and Data Analysis 33 요약 / Abstract 지구물리탐사배경및필요성 지하구조비저항영상화 MT 탐사자료획득 MT 탐사 2차원자료해석 MT 탐사 3차원자료해석 결과및토의 탄성파탐사자료해석 자료설명 탄성파반사법자료처리결과 탄성파굴절법해석결과

66 Table of Contents Page 속도모델구축 원격탐사기반단층모델링 SAR 위성자료획득 SAR 위성자료를이용한변위관측 SAR 위성관측변위기반단층모델링 본진전후규모 3.0 이상지진에의한지표변위관측 Sentinel-1 PSInSAR 시계열분석을통한미세지표변위관측 결론 지구물리탐사해석결과 참고문헌 59 제 4 장응력상태분석 Stress State Analysis 61 요약 / Abstract 연구배경 포항주변응력장방향 포항지열발전실증사이트현장응력장크기 연직응력 지진포컬메커니즘역산자료로부터의응력크기정보 시추공주입자료로부터의응력크기정보 응력장과포항지진단층의운동학적상관관계 포항지진유발단층의응력상태 결론 참고문헌

67 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Table of Contents Page 77 제 5 장지진분석 Seismological Analysis 79 요약 / Abstract 서론 지진자료 속도모델 지진검출 진원결정 초기위치 상대위치 최종위치 단층면해 지진규모결정 토의 참고문헌 95 제 6 장지중암반공극압확산분석및지하수변화 Analysis of Pore Pressure Perturbation and Groundwater Change 97 요약 / Abstract 서론 ( 연구필요성및목적 ) 지중암반의공극압확산분석 지중암반의수리특성

68 Table of Contents Page 공극압시공간확산분석 지열정및그주변지하수위변화특성 지진발생전지열발전실증시설주변지하수위 지진발생후지하수위변동 지열정의지하수화학특성변화 M W 3.2 지진발생시동위원소특성 포항지진발생후지열수의화학특성 결론 참고문헌 113 Appendix 115 A-1. Acoustic image logging data of PX-1 and PX-2 geothermal wells 118 A-2. Mud loss and microseismicity 120 A-3. Earthquake catalog near the EGS site since 2009 to the 2017 Pohang earthquake 124 A-4. Temporal distribution of earthquakes and EGS project activities 125 A-5. Microstructures of fault gouge and breccia from the 3800 m fault zone in PX-2 borehole

69 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 List of Figures Page 5 Fig Location of Pohang EGS project site and the epicenter of the 15 November 2017 Pohang Earthquake (left) and schematic diagram of PX-1 and PX-2 geothermal wells (right). 6 Fig Injection, flow back and net injection volumes during five hydraulic stimulations conducted at PX-1 and PX-2 geothermal wells. 11 Fig Example from the PX-2 master log. 12 Fig (a) Blocks of cuttings fixed in epoxy and (b) an example of microscopic observation. 13 Fig (a) Tectonic framework of the southern Korean Peninsula, (b) Landsat TM satellite image showing the distribution of the Miocene sedimentary basins, major faults and stratigraphic units in SE Korea (from Son et al., 2015) and (c) Regional structural map of SE Korea showing the Miocene stress regime and strain diagram. The black arrows indicate the mean declination directions of characteristic remanent magnetizations of the basin fills. 14 Fig Geological map of the Pohang Basin with its major bounding and intrabasinal faults (from Song, 2015). 15 Fig (a) Structural map of the Pohang Basin, which is divided into four structural domains named Bomun, Ocheon, Doumsan, and Gojusan (Song, 2015), (b) Fault slip data obtained in and around the Pohabg Basin (Song, 2015). Divergent arrow heads represent minimum horizontal stress direction. R = (σ2 - σ3)/(σ1 -σ2). R = R (σ1 is vertical), 2 -R (σ2 is vertical), or 2 + R (σ3 is vertical) and (c) Kinematic model, wedged-shaped pull-apart basin model, explaining the opening of the Pohang Basin (Son et al., 2015). 16 Fig (a) Simplified borehole logs in the Pohang Basin showing abrupt changes of the basement depths along east-west direction and (b) Contoured depth map of the basin floor, produced using 26 deep drilling boreholes, shows inferred major intrabasinal fault traces. W.B.F.: the western border fault of the Pohang Basin (from Song et al., 2015).

70 List of Figures Page 17 Fig Outcrop photographs showing NE-striking normal faults in (a) the basements (rhyolitic rocks) and (b) the Pohang basin-fill (mudstone) observed along the Gokgang Fault line. 17 Fig NE- or ENE-striking conjugate reverse faults identified on (a) eastern and (b) western trench walls excavated along the Gokgang Fault line, which cut the Quaternary sediment layers (Unit A~H). 18 Fig Examples of ground cracks (a-c), sand blows (c-d) induced by the 2017 Pohang earthquake and (e) Rose diagram showing the orientations of the cracks and sand blows. 19 Fig A schematic diagram showing the distribution of surface deformations across the 2017 Pohang earthquake rupture and proposed mechanism associated with blind oblique-slip including reverse-slip component and their related surface folding (Choi et al., 2019). 19 Fig Regional stress trajectory map showing the distribution of regional stress fields in the central and eastern parts of Eurasian continent. Reddish dashed lines indicate trace of the maximum horizontal stress axes based on the World Stress Map release 2008 (Heidbach et al., 2010; Kim et al., 2016). 20 Fig Geological column showing the boundary between Miocene Pohang and Cretaceous Gyeongsang basins. 21 Fig Geological column showing the boundary between Cretaceous Gyeongsang Basin and Yeongnam Massif. 22 Fig Summary of information indicating the lithological changes and the presence of fault zones from PX-2 master log. Almost is assumed to be the term used to describe the state of being lumped together with clays and rock fragments. 23 Fig Comparison of the cutting fragments at depths of 3,784 and 3,803 m of PX-2. The fragments at 3,784 m are mostly angular and fresh, while the fragments at 3,803 m are mostly rounded and degraded (friable).

71 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 List of Figures Page 24 Fig Stereoscopic photomicrographs of cutting fragments at depths of (a) 3,535 m, (b) 3,544 m, (c) and (d) 3,791 m, (e) 3,804 m and (f) 3,807 m of PX-2. Fragments of (c) to (f) mostly show textures indicating fault gouge or ultracataclasites. (g) and (h) Cutting fragments at a depth of 3,535 m surrounded by drilling mud and no drilling mud, respectively. 25 Fig Polarizing photomicrographs of thin sections of cuttings at depths of 3,791 m (a) under open and (b) crossed polars, 3,804 m (c) under open and (d) crossed polars, and 3,807 m (e) under open and (f) crossed polars. 27 Fig X-ray diffraction patterns of PX-2 cuttings. 36 Fig MT/AMT site map. The solid lines and the green box represent the survey lines for 2D interpretation and 3D inversion area, respectively. 38 Fig Resistivity models from 2D inversion of TM mode data for the survey line (a) V1, (b) V2, (c) V3, (d) V4 and (e) V5, respectively. 39 Fig D resistivity distribution from 3D inversion of MT and AMT data. The black and blue circles indicate the MT sites by KIGAM at 2002 and The red circles represent the AMT sites at Four survey lines (black lines) on the surface are assumed for the 2D interpretation. 40 Fig Resistivity sections along the survey lines shown in Fig. 3-1; (a) A1, (b) A2, (c) A3, and (d) A4, respectively. The low resistivity zone (white circle) at the central part of the section is interpreted as the Heunghae-eup artefact. 42 Fig Locations of the EGS site, boreholes, seismic reflection survey line. 43 Fig (a) Geometry of the borehole geophone array placed at PX-2 and (b) the location map of 6 check shots. 44 Fig (a) Final stack section and (b) its stratigraphic interpretation (the section is displayed only up to 2,500 m in depth axis because the stacked traces at later times are severely contaminated by noise and ground-roll). 45 Fig Shot gather seismograms corresponding to (a) forward and (b) reverse traverse for refraction analysis.

72 List of Figures Page 45 Fig Two-layered dipping interface model obtained by interpreting firstarrival traveltime curves. 47 Fig Modified P-wave velocity model using the sonic log-based model and the two-layered refraction velocity model. 50 Fig Descending InSAR measurements from (a) Sentinel-1A/B, (b) and (c) Cosmo-SkyMed, (d) ALOS-2 PALSAR-2, and (e) ascending InSAR measurement from ALOS-2: LOS measurements (a, b, d, e) and azimuth measurement (c). 51 Fig (a) Horizontal displacement vector field (the vectors indicate the magnitude and directions of the horizontal displacements, and the colored map represents the vertical displacements), (b) displacement vector field of the box A. 52 Fig Histograms of the Okada dislocation model parameters obtained from the InSAR measurements using the Monte-Carlo simulation. 53 Fig Fault plane models estimated from three InSAR and 3D measurements along with the EGS wells, PX-1 and PX-2 (Upper image: COSMO-SkyMed LOS displacement). 54 Fig Sentinel-1 displacement map with the epicenter: (a) 2017/04/ /04/26 pair and (b) 2018/02/ /02/20 pair. 55 Fig (a) Mean velocity map estimated from Sentinel-1 ascending mode data, (b) the enlarged image around the epicenter of three major earthquakes, and (c) close view around the EGS site. 56 Fig Time series histories of LOS displacement at the selected points of P1, P2, P3 and P4 around the EGS site. 62 Fig Maximum horizontal principal stress directions in Korea (data provided by World Stress Map 2016): (a) B-C quality stress data over the country, (b) C-D quality stress data around Pohang. 63 Fig (a) Earthquake focal mechanism solutions for M>2.5 earthquakes that occurred around Pohang from 1997 and 2016, and stress inversion results showing (b) the orientations of the three principal stresses (S 1, S 2, S 3 ) and (c) R value.

73 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 List of Figures Page 64 Fig Earthquake focal mechanism stress inversion result from aftershocks. 65 Fig (a) Dipole sonic shear wave anisotropy in PX-2, indicating the maximum horizontal principal stress azimuth (b, c). 66 Fig The azimuth of the maximum horizontal principal stress (S Hmax ) as a function of depth estimated from various data around Pohang. 67 Fig Rock density measurements using cores extracted nearby borehole (BH-4) and PX Fig First hydraulic stimulation pressure-time and injection rate-time curves in PX-2 and PX Fig Well-head pressure and injected water volume curves in (a) PX-2 and (b) PX Fig Traces of pole orientations of natural fractures that can possibly play as leakage channel when PX-1 borehole pressure is raised by 15 MPa, which are depicted as cyan lines in the stereonet. Three circles indicate poles of faults estimated from the earthquake focal mechanisms that occurred during the initial stage of PX-1 injection (17~18 Dec 2016). 71 Fig Step-rate test results in PX-2 on (a) 02 Feb 2016 and (b) 04 Sep Fig (a) Example of nearly constant pressures (fracture propagation pressures) attained at different constant injection rates and (b) their trend as a function of injection rate. 74 Fig Modelled rake on the Pohang fault plane as a function of S Hmax azimuth and R value for the verification of the constrained stress model using kinematics of the Pohang fault. 75 Fig (a) Slip tendency of the PX-2 (square) injection-related fault and (b) stress condition plotted in Mohr diagram. 81 Fig Location map of seismic stations used in this study. 82 Fig Local 1-D velocity model developed in this study.

74 List of Figures Page 84 Fig Initial locations of 240 earthquakes detected by the template matching method. Yellow triangle represents the PX-2 well. Earthquakes whose epicentral distances are greater than and less than 10 km from the PX-2 well are denoted by red and green circles, respectively. Four earthquakes with focal depth greater than 10 km are plotted as blue circles. Event ID and focal depths of the four events are also represented. Geological lineaments and faults are shown as dashed lines. 86 Fig Final locations of 98 earthquakes. Yellow, gray, and blue circles represent events b, b and a, respectively. Five immediate foreshocks and the mainshock are denoted by red and green circles, respectively. Green and blue curves ended by red represent the PX-1 and PX-2 wells, respectively. Open sections of wells are shown in red curves. 87 Fig Distribution of 53 focal mechanism solution. Colors of beachball diagrams represent faulting types according to the classification of Zoback (1992): Strike-slip (black), Thrust (blue), Strike-slip with thrust component (red), and Strike-slip with normal component (green). 89 Fig Comparison of magnitude estimates. (a) Local magnitudes of the KMA versus and those of this study. (b) Moment magnitudes versus local magnitudes. Red circles represent the events published by the KMA. 90 Fig Illustration of the classification of groups G1 and G2. Magenta and cyan circles represent locations of earthquakes belong to group G1 and G2, respectively. The size of circles scale with the magnitude of earthquakes. Black line in B1-B2 section represents a G2 plane approximated from the principal component analysis. 91 Fig Locations of G2 events projected on the plane approximated by the principal component analysis. Colors of circles represent the occurrence period of earthquakes: G2-0 (yellow), G2-1 (orange), G2-2 (green), G2-3 (blue), and G2-M (purple). Aftershocks of M W 3.2 earthquakes are denoted by open circles. Open square indicates a crossing point of PX-2 borehole and the plane. 99 Fig Wellhead pressure and injection rate measured during five hydraulic stimulations are plotted with hydraulic conductivity estimated by the Jacob straight line method.

75 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 List of Figures Page 100 Fig Hydraulic models used in numerical calculation for pore pressure perturbation analysis. The upper left figure shows the hypocenters of earthquakes greater than magnitude of Fig Plot of mud loss versus measured depth of PX-2 (left) and the acoustic images obtained at the depth of 3,783 m in PX-2, below which the signal was completely lost (right). 102 Fig (a) Spatial distribution of pore pressure change (ΔP) along the vertical cross section passing through PX-1 and PX-2 and (b) the isosurface of pore pressure change at 0.02 MPa around PX-1, PX-2, and the fault. 103 Fig Pore pressure change with time on the hypocenters of (a) M W 3.2 and (b) M W Fig Pore pressure change at the hypocenters of earthquakes greater than M 1.0 versus the time of occurrence (left), and the histogram of frequency of earthquakes greater than M 1.0 with increasing pore pressure change (right). 106 Fig The groundwater level measured at PX-1 and PX-2 before and after M W 5.5 earthquake. The zero level indicates the flow back. 107 Fig Water level and water temperature monitored at the PX-1 and PX-2 since August 31, 2018 in the Pohang EGS site. 108 Fig Flow back water from PX-1 (left) and PX-2 (right) after M W 3.2 earthquake. 109 Fig The values of δ 18 O and δd in PX-1 and PX Fig Piper diagram for flow back water from PX-1 and PX Fig Stiff Plots for groundwater from PX-1 and PX-2 before and after M W 5.5 earthquake. 110 Fig The values of 14 C for groundwater from PX-1 and PX-2 before and after the M W 5.5 earthquake.

76 List of Figures Page 115 Fig. A-1-1. PX-1 well structure and acoustic images near the open hole section (from HADES report). The cement shown after the casing section continued from the casing shoe until 4,097 m (measured depth) where the tool stopped. 116 Fig. A-1-2. PX-2 well structure and acoustic image around 1,512 m depth which indicates detection of a hole in casing (from HADES report). The hole matches the casing damage during the 5 th hydraulic stimulation reported by EGS project team. 117 Fig. A-1-3. PX-2 well acoustic image above 3,783 m and complete loss of acoustic signals below 3,783 m (from HADES report). While the PX-1 acoustic signals were obtained below the casing shoe to open hole section, the acoustic signals of PX-2 were not obtainable because the tool stopped at 3,783 m that is 425 m above the casing shoe. 118 Fig. A-2-1. Mud loss depths and mud density of PX-1 (old), PX-1, and PX-2 wells (above) and temporal distribution of accumulated mud loss and seismicity (below). 119 Fig. A-2-2. Temporal distribution of seismicity plotted on mud loss. 124 Fig. A-4-1. Temporal distribution of EGS project activity and seismicity of events with location certainty. 124 Fig. A-4-2. Temporal distribution of EGS project activity and seismicity of events whose magnitude was determined. 126 Fig. A-5-1. Photomicrographs of polished specimens of fault gouge and breccia recovered from the depths of (a, b) 3,791 m, (c) 3,798 m, (d, e) 3,803 m in PX-2 borehole. (b) and (e) are close-ups of rectangular portions in (a) and (d), respectively. Observations were all made under two stereomicroscopes. 127 Fig. A-5-2. Photomicrographs of (a, b) clayey foliated fault gouge nearly free from visible clasts (3,806 m in depth), (c) highly sheared gouge between clast-rich zones (3,813 m), and (d, e) fragments of fault breccia that are both deformed internally (from depths of 3,804 m and 3,812 m, respectively). (b) is a close-up of the rectangular portion in (a); note that the gouge in (b) is finely foliated.

77 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 List of Tables Page 26 Table 2-1. Mineral compositions of the PX-2 cuttings measured by XRD (wt%). Qz: Quartz, Pc: Plagioclase, K-fd: K-feldspar, Am: Amphibole, Ch: Chlorite, Mica: Illite+Muscovite, La: Laumontite, Cc: Calcite, and Gs: Gypsum. 43 Table 3-1. Origin times and coordinates for check shot events. 46 Table 3-2. Simplified lithology model and corresponding P-wave velocity model at PX Table 3-3. Final velocity model (Model I) based on well-logging data and refraction velocity model. 48 Table 3-4. Modified velocity model (Model II) using check shot data. 49 Table 3-5. List of SAR data collected for research. 51 Table 3-6. Input sources for 3D decomposition of surface displacement. 52 Table 3-7. The best-fit model parameters and standard deviation of the 2017 Pohang earthquake estimated by the Okada dislocation model using the descending InSAR and 3D measurements. 73 Table 4-1. Constrained stress model in Pohang geothermal site. 82 Table 5-1. Local 1-D velocity model and its description. 105 Table 6-1. Water levels (DTW, m) at the shallow ( 100 m), intermediate (100~ 300 m), deep (300~1,100 m) and very deep (2,383 m) wells (n=number of wells) before the 2017 Pohang earthquake. 107 Table 6-2. Measured water levels and level logger installed depth.

78 제 1 장 포항지열발전실증연구프로젝트개요 Overview of the Pohang EGS (Enhanced Geothermal System) Demonstration Research Project

79 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project 2

80 제 1 장포항지열발전실증연구프로젝트개요 요약 / Abstract 포항지열발전실증연구시설은포항시흥해읍에위치하며 M W 급지열발전상용화기술개발의정부연구과제로 2010년 12월부터시작되었다. 이연구는지하심부로지열정을굴착하여수리자극을통해지하심부의암반에유체의이동경로를생성시키고이렇게생성된경로에물을순환시켜지하심부의지열을지상으로추출하는과정을포함한다. 포항지열발전실증연구부지에서 2012년 9월부터 2016년 11월까지시추를진행하여두개의지열공인 PX-1과 PX-2를완공하였다. PX-1, PX-2 지열공의수직심도는 4,215 m (measured depth 4,362 m), 4,340 m (measured depth 4,348 m) 이다. 지표에서 PX-1 과 PX-2 두지열정의거리는약 6 m이며, 지열정하단에서의거리는약 600 m이다. 2016년 1월부터 2017년 9월까지 PX-1, PX-2 두지열정을이용하여총 5회의수리자극이실시되었으며, PX-1을이용하여 2회, PX-2를이용하여 3회의수리자극이수행되었다. PX-1에주입된유체의양은 5,663 m 3, 배출량은 3,968 m 3 이며, PX-2의주입량은 7,135 m 3 이며, 배출량은 2,989 m 3 이다. 두지열정의순주입량은 5,841 m 3 이다. PX-2 지열정에서는 1차, 3차, 5차의수리자극이실시되었으며, 각수리자극동안의최대주입압력은 89.2, 88.8, 84.6 MPa에이르렀다. 3차수리자극종료시점인 2017년 4월 15일 M W 3.2의지진이발생하였다. 2017년 9월 18일마지막 5차수리자극이후포항지진이발생한 2017년 11월 15일까지물을주입한지열정 (PX-2) 의밸브를개방하여주입된물을지표로배출시키고있었다. PX-1 지열정에서는 2차, 4차두차례의수리자극이실시되었으며, 2차, 4차수리자극에서최대주입압력은 27.71, MPa에이르렀다. 2017년 11월 15일포항지진이발생하였으며, 그이후실증연구과제의진행이중지되었다. Two boreholes of PX-1 and PX-2 were drilled in granodiorite/granitic gneiss basement to develop the enhanced geothermal system (EGS) in the city of Pohang, Republic of Korea. PX-1 was initially drilled up to the measured depth (MD) of 4,127 m (measured along borehole) from September 2012 to October 2013, but the drill pipe was accidently broken and stuck inside. Therefore, deviated drilling was started at the depth of 2,419 m in the west-northwest (WNW) direction from June 28, 2016, which was completed in November 13, The true vertical depth (TVD) and MD of PX-1 were 4,215 m and 4,362 m, respectively. PX-2 was drilled from April 3, 2015 to December 9, 2015, of which TVD and MD were 4,340 m and 4,348 m, respectively. PX-1 and PX-2 were 6 m apart on ground surface and aligned along the north-south direction. PX-1 and PX-2 were cased, except for 313 m and 140 m long open hole section of PX-1 and PX-2. Five hydraulic stimulations were conducted at PX-1 and PX-2 from January 29, 2016 to September 18, 2017, and no water injection was made after September 19, 2017 until November 15, 2017 when the M W 5.5 Pohang earthquake occurred. Hydraulic stimulation procedures involved injection of highly pressured water, shut-in, and flow back. Flow back and shut-in occurred between hydraulic stimulations. Wellhead pressure and injection rate were simultaneously recorded only during hydraulic stimulation, which were measured every few seconds at PX-1 and every minute at PX-2. Water, which was injected and 3

81 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 flowed back into/from PX-1, amounted to 5,663 and 3,968 m 3, respectively. Total two hydraulic stimulations were conducted at PX-1: the 2nd and the 4th stimulations. Maximum wellhead pressure and injection rate reached MPa and L/s in the 2nd stimulation and MPa and L/s in the 4th stimulation. For PX-2, there were three hydraulic stimulations: the 1st, the 3rd, and the 5th stimulations. Maximum wellhead pressure and injection rate reached 89.2 MPa and L/s in the 1st stimulation, 88.8 MPa and L/s in the 3rd stimulation, and 84.6 MPa and L/s in the 5th stimulation. In PX-2, total injection and flow back volumes of water were 7,135 and 2,989 m 3, respectively. Shut-in for PX-1 and flow back for PX-2 were being done until the November 15 Pohang earthquake after the last hydraulic stimulation 포항지열발전실증연구시설현황과구축과정 포항지열발전실증연구시설은포항시흥해읍에위치하며 M W 급지열발전상용화기술개발의정부연구과제로 2010 년 12월부터시작되었다. 이연구는심부지열원이특별히존재하지않는비화산지대에서청정에너지원을개발하기위한것으로지하심부로지열정을굴착하여수리자극을통해지하심부의암반에유체의이동경로를생성시키고이렇게생성된경로에물을순환시켜지하심부의지열을지상으로추출하는과정을포함한다. 이렇게추출된지열을이용하여발전을하는것을 EGS (Enhanced Geothermal System) 유형의발전이라고부르며, 포항실증연구는최고온도 160 의고온지열수를생산하여 1.2 M W 의전기를생산하는지열발전소를구축하고운영하는것을목표로진행되었다. 연구가최종단계로진행하고있는과정에 2017 년 11 월 15일 M W 5.5의포항지진이발생하였고, 그이후실증연구과제의진행이중지되었다. 포항지열발전실증연구부지에서 2012 년 9월부터 2013 년 10월까지시추를진행하여심도 4,127 m (Measured Depth, MD) 인최초의지열정 (PX-1) 을완성하였다. 시추완료이후공벽정리과정에서드릴파이프가절단되어하부구간의드릴파이프가시추공에잠기게되었다. 이에 PX-1 지열공에서드릴파이프를회수하고지열공을복원하기위한노력이 2015 년 11월부터전개되었고 2,485 m 심도까지파이프회수와시추공을복원할수있었다. 하지만나머지구간이여의치않아 PX-2 굴착이완료된이후, 기존의 PX-1 의심도 (MD) 2,419 m에서경사시추를통해새로운경로로굴착하여현재의심도 4,362 m (MD) 인지열정 PX-1 을 2016 년 11월에완성하게되었다. PX-2 지열정은 2015 년 4월부터시추를시작하여 2015 년 12월에심도 4,348 m (MD) 의깊이로설치를완료하였다. 완성된 PX-1 과 PX-2 지열정의구조는 Fig. 1-1과같다. PX-1 의수직심도는 4,215 m이며, 물이주입또는생산되는 ( 즉물의유출입이일어나는 ) 나공의길이는지열정하단 313 m이며, 나머지상부구간은케이싱으로처리되었다. PX-2 의수직심도는 4,340 m이며, 나공의길이는 140 m이다. 지표에서 PX-1 과 PX-2 두지열정의거리는약 6 m이며, 지열정하단에서의거리는약 600 m이다. 4

82 제 1 장포항지열발전실증연구프로젝트개요 Fig Location of Pohang EGS project site and the epicenter of the 15 November 2017 Pohang Earthquake (left) and schematic diagram of PX-1 and PX-2 geothermal wells (right) 수리자극실시현황 2016 년 1월부터 2017 년 9월까지 PX-1, PX-2 두지열정을이용하여총 5회의수리자극이실시되었으며, PX-1 을이용하여 2회, PX-2 를이용하여 3회의수리자극이실시되었다. PX-1 에주입된유체의양은 5,663 m 3, 배출량은 3,968 m 3 이며, PX-2 의주입량은 7,135 m 3 이며, 배출량은 2,989 m 3 이다. 두지열정의순주입량은 5,841 m 3 이다 (Fig. 1-2). PX-2 지열정에서는 1차, 3차, 5차의수리자극이실시되었으며, 인근저수지의물을끌어와지열정에고압으로주입하였다. 제1차수리자극은 PX-2 관정이완료된이후인 2016 년 1월 29일부터 PX-2 지열정에서실시되었다. 이때최대주입압력은 89.2 MPa에이른다. 3차, 5차역시최대주입압력은 88.8, 84.6 MPa에이르렀다 (Fig. 1-2). 3차수리자극종료시점인 2017 년 4월 15일 M W 3.2의지진이발생하였다. 마지막 5차수리자극은 2017 년 8월 30일부터 9월 18일까지이루어졌으며, M W 5.5의포항지진이발생한 2017 년 11월 15 일까지물을주입한지열정 (PX-2) 의밸브를개방하여주입된물을지표로배출시키고있었다. PX-1 지열정을 2016 년 11월에완공한후 2016 년 12월부터 PX-1 지열정을이용하여제2차수리자극 (PX-1 자체로는첫번째수리자극 ) 이실시되었다. 이후 2017 년 8월 7일부터 8월 14일까지 4차수리자극이실시되었다. 2차, 4차수리자극에서최대주입압력은 27.71, MPa로서 PX-2 에비해상대적으로작다. 5

83 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig Injection, flow back and net injection volumes during five hydraulic stimulations conducted at PX-1 and PX-2 geothermal wells. 6

84 제 1 장포항지열발전실증연구프로젝트개요 제 2 장 포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 Analysis of Geological Structures in and around the Pohang EGS Site to Interpret the Relationship between Pohang Earthquake and Hydraulic Stimulations 7

85 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project 8

86 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 요약 / Abstract 2017년 11월 15일발생한 M W 5.5 포항지진의진앙이위치한마이오세포항분지는북북서방향의우수향주향이동단층과이에수반된북북동내지북동방향의정단층들의운동에의해확장된당겨열림형퇴적분지로시계방향으로회전된지괴와회전되지않은지괴사이에서벌어진쐐기형기하를가진다. 포항분지가서북서-동남동방향의인장응력하에서확장되는동안분지내부에는북동내지북북동방향의퇴적동시성성장정단층들도다수만들어졌으며, 이중규모가가장큰단층인곡강단층이포항지진의진앙과인접하여위치한다. 포항지진진앙지일원의지질구조와지표변형특성, 여진들의 3차원진원분포, 발진기구해결과, InSAR 자료를종합하면, 포항지진은분지확장동안만들어진북북동방향공액상정단층들중서쪽으로약 50~70 경사진곡강단층의반향단층들중하나가우수향수평이동성분을가지는역단층으로재활되어발생한것으로해석된다. 포항지열발전실중시설내시추공의커팅시료를분석하면, 포항분지- 경상분지경계는 PX-1 지점에서심도 206~208 m 사이에그리고 PX-2 지점에서 186~200 m 사이에위치한다. 또한경상분지-영남육괴경계는 PX-1 지점에서심도 2,354~2,356 m 사이에그리고 PX-2 지점에서 2,348~2,350 m 사이에위치한다. 한편, PX-2의 Master log, 커팅시료들의육안관찰과현미경관찰그리고 X-선회절분석을통해서 3,790~3,815 m 심도구간에서단층핵에해당하는단층비지대가존재하는것이이번연구에서확인되었다. 이심도구간커팅시료들은다른구간의시료들과달리육안관찰에서원마도가매우좋고풍화또는변질로손가락으로부스러질정도로약한강도를가지며다량의점토물질로피복된특징을보여준다. 또한실체현미경과편광현미경관찰에서양호한원마도의암편들과입자크기감소, 엽리상단층비지, 파쇄유동조직등의특징을보여주고 X-선회절분석에서단층암에서흔히산출되는이차광물들이존재하고있음이확인되어반복된단층활동으로인해일정한규모로발달하는성숙한단층대가이심도구간에존재하는것으로해석된다. 2018년여름 PX-2에서새로이수행된시추공영상검층에서시추공에삽입된장비가 3,783 m 심도에서막혀더이상아래로내릴수없었다는사실은 3,790~3,815 m 심도구간의기존단층핵을따라 포항지진의원인이되는단층면파열이발생하였을가능성을강하게시사한다. This study aims to determine the structural characteristics in and around the Miocene Pohang Basin, in where the epicenter of the Nov. 15, 2017 M W 5.5 Pohang earthquake is located, and to ascertain vertical variation of lithology and depth of fault zone beneath the Pohang EGS site. The Pohang Basin is a wedged-shaped pull-apart basin opened between clockwise rotated and unrotated blocks under NNW-trending dextral simple shear and associated WNW ESE-directed extension during the East Sea opening. During the basin extension, NE- or NNE-striking intrabasinal conjugate normal faults were produced in places. The largest of them is the Gokgang Fault, which is located adjacent to the epicenter of the Pohang earthquake. A synthetic analysis, using the geological structures, surface deformation, spatial distributions and focal mechanism solutions of the mainshock and aftershocks, and InSAR data, indicates that the Pohang earthquake was generated by the reactivation (as dextral reverse oblique-slip fault) of a 50 to 70 westward-dipping antithetic fault to the Gokgang Fault. Cutting specimens from two deep boreholes at the Pohang EGS site 9

87 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 indicate that the boundaries between Miocene Pohang Basin and Cretaceous Gyeongsang Basin are situated at depths of 207~209 m and 196~200 m of PX-1 and PX-2, respectively. Meanwhile, the boundaries between the Gyeongsang Basin and the Yeongnam Massif are situated at depths of 2,354~2,356 m and 2,348~3,350 m of PX-1 and PX-2, respectively. Based on the master log, visual and microscopic observations, and XRD analyses of the PX-2 cuttings, at depths of 3,790 to 3,815 m, especially, very rounded, easily fragile, weathered and/or altered rock fragments with a large amount of clay materials, typical cataclastic flow textures, and secondary minerals commonly observed in gouge-filled fault cores are significantly identified. It is thus reasonable that there is a mature fault zone made of multiple movements at depths of 3,790 to 3,815 m. This interpretation is strongly supported by an obstruction of the inserted measuring equipment inside the casing of PX-2 well at a depth of 3,783 m during acoustic image logging in summer 2018, which is also indicative of the possibility that the fault rupture for the Pohang earthquake crosses in this depth interval 연구배경 전세계적으로중규모이상지진은대부분기존단층이재활되어발생하므로, 지진발생원인을정확히이해하기위해서는진앙이위치하는구조구역의광역지질구조와진앙인접지역의지표및지하지질구조특성에관한상세정보가필요하다. 또한지하에가해진수압의증가로단층이재활되어발생하는유발지진 (induced earthquake) 은물을주입한시추공과단층이근접하거나시추공이단층면을관통하여단층대내의유효응력 (effective stress) 을감소시킬때발생하므로, 수압에의한지진발생가능성을평가하기위해서는지하단층의위치를정확히파악하고주입시추공과의근접성을분석하는것이필요하다. 이번조사는 2017 년 11월 15일발생한 M W 5.5 포항지진과포항지열발전실증시설의수리자극과의상관성을해석하기위한기초자료인진앙이위치한마이오세포항분지일원의광역지질구조의특성을파악하고실증시설내시추공자료들을분석하여지하암상분포, 단층의존재유무와분포심도구간을규명하고자하였다 연구방법 마이오세동안확장된포항분지일원의광역지질과지질구조관련기존문헌자료와포항지진의진앙이 위치한포항시흥해읍일원의야외조사를통한지표변형과지질구조특성자료를수집하고분석하였다. 또한 10

88 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 포항분지의지질구조, 지표변형특성, 포항지진의본진과주요여진의발진기구해 (focal mechanism), 여진의 3차원진원분포, InSAR 자료등을이용하여포항지진을유발한단층의기하와운동학적특징을해석하였으며, 넥스지오로부터제공된 PX-1과 PX-2 시추공에대한 Master log report (Fig. 2-1) 와 Geological logging report 에기록된 2 m 간격암상기재와함께고단열대 (highly fracture zone), 단층비지 (fault gouge) 의유무와함량, 커팅시료의크기등에대한정보들을이용하여지하암상분포와포항지진과관련된단층암의존재유무와분포구간을분석하였다 (Sources 1과 2). Fig Example from the PX-2 master log. PX-1 과 PX-2 의커팅시료들의육안관찰, 실체및편광현미경관찰, XRD 분석을통해심도별암상변화와단층암의존재와구간을파악하였다. PX-1 커팅시료는모두세척이매우잘되어있어육안암상기재가용이하나단층암 ( 단층각력, 단층비지 ) 의유무를판단하기에힘든반면, PX-2 커팅시료는세척이완전히이루어지지않아세척전에단층대확인을위한미구조와 X-선회절분석을실시하고세척후육안암상기재를실시하였다. PX-2 커팅시료중단층비지로판단되는경우세척하지않고저밀도에폭시로시료를고정한후블록을제작하여실체현미경으로관찰하고표준연마박편을제작하여편광현미경을이용해조직과광물조성을관찰하였으며 (Fig. 2-2), 세립의암편들과점토질입자로주로구성된단층비지의광물동정과함량비분석에적합한 X-선회절을이용한정성및정량분석도실시하였다 (Song et al., 2017). 11

89 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig (a) Blocks of cuttings fixed in epoxy and (b) an example of microscopic observation 연구결과 마이오세포항분지일원의광역지질과지질구조특성 포항지진의진앙은지리적으로경상북도포항시북구흥해읍망천리에위치하며 (Kim et al., 2018), 지질학적으로선캄브리아영남육괴내에서확장한백악기경상분지의남동부에위치한마이오세포항분지에위치한다 (Fig. 2-3a와 b). 한반도의기저를이루고있는영남육괴는편마암과편암복합체그리고화강암질암으로주로구성되며 (GSK, 1998), 이를부정합으로피복하는경상분지충전물인경상누층군은고-태평양판의섭입과관련한배호분지 (back-arc basin) 내육성퇴적물과화산호 (volcanic arc) 에서의화산활동에의한화산쇄설성물질그리고화강암질관입암으로구성된다 (Chang et al., 2003; Chough and Sohn, 2010). 약 17~10 Ma 사이에퇴적된포항분지충전물인연일층군 (Sohn and Son, 2004) 은분지서편경계단층을따라분포하는선상지삼각주 (fan-delta) 역질퇴적암과동쪽으로갈수록세립화되는사암과이암그리고약 15 Ma경에남부에국지적으로관입한알칼리현무암으로구분되며 (Song, 2015), 2017 년포항지진의진앙은포항분지북동부연일층군사암과이암의분포지에위치한다 (Fig. 2-4). 12

90 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 Fig (a) Tectonic framework of the southern Korean Peninsula, (b) Landsat TM satellite image showing the distribution of the Miocene sedimentary basins, major faults and stratigraphic units in SE Korea (from Son et al., 2015) and (c) Regional structural map of SE Korea showing the Miocene stress regime and strain diagram. The black arrows indicate the mean declination directions of characteristic remanent magnetizations of the basin fills. 13

91 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig Geological map of the Pohang Basin with its major bounding and intrabasinal faults (from Song, 2015). 포항분지는약 25 Ma에서 15 Ma 사이에일본열도가한반도에서부터분리되면서동해가확장될당시에한반도동해안육상에가해진북북서방향우수향전단력에의한침강된분지이다 (Son et al., 2015; Figs. 2-3c 과 2-5). 포항분지의서쪽가장자리는양산단층으로부터약 2~5 km 동편에위치한북동내지북북동방향의정단층과북서방향의전이단층 (transfer fault) 이지그재그형태로연결된단층들로경계된다 (Fig. 2-5a; Son et al., 2015). 분지남쪽은북북서방향의우수향주향이동단층인연일구조선으로 (Son et al., 2002) 그리고동 14

92 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 쪽은북동내지북북동방향의단층들로구성된오천단층계로경계되며, 오천단층계는좌수향정이동성사교단층의특징을보이며남서쪽으로갈수록수직변위가줄어드는가위단층 (scissor fault) 의기하를가진다 (Cheon et al., 2012). 분지내외의단층이동자료를이용한고응력장복원자료 (Song, 2015) 는분지가서북서 -동남동방향의인장응력하에서확장되었음을지시한다 (Fig. 2-5b). 또한포항분지일원에대한고자기연구에서분지동편호미곶반도가시계방향으로약 30 회전된반면, 분지서편의지괴는회전되지않은특징을보여주고있어 (Fig. 2-3c; Son et al., 2015), 분지는북북서방향의우수향전단력에의해시계방향으로회전된해안가지괴와보다육지쪽의회전되지않은지괴사이에서확장된쐐기형당겨열림분지 (wedge-shaped pull-apart basin; sphenochasm) 로해석된다 (Fig. 2-5c). Fig (a) Structural map of the Pohang Basin, which is divided into four structural domains named Bomun, Ocheon, Doumsan, and Gojusan (Song, 2015), (b) Fault slip data obtained in and around the Pohabg Basin (Song, 2015). Divergent arrow heads represent minimum horizontal stress direction. R = (σ2 - σ3)/(σ1 - σ2). R = R (σ1 is vertical), 2 - R (σ2 is vertical), or 2 + R (σ3 is vertical) and (c) Kinematic model, wedged-shaped pull-apart basin model, explaining the opening of the Pohang Basin (Son et al., 2015). 15

93 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항분지가서북서 -동남동방향의인장에의해확장될당시에분지내부에는북동내지북북동방향의퇴적동시성성장정단층들 (syn-depositional growth faults) 이다수형성되었으며, 이러한분지내부성장단층들은심부시추코어들에서얻어진분지바닥심도와분지충전물내미고생물의대비를통해확인되었다 (Fig. 2-6a; Yun, 1994; Song et al., 2015). 이중포항지진의진앙과인접한지열발전시설의동편에위치한곡강단층 (Fig. 2-6b) 은북동방향의주향을가지고남동으로경사진정단층의기하를가지는것으로알려져있다 (Yun, 1994). Fig (a) Simplified borehole logs in the Pohang Basin showing abrupt changes of the basement depths along east-west direction and (b) Contoured depth map of the basin floor, produced using 26 deep drilling boreholes, shows inferred major intrabasinal fault traces. W.B.F.: the western border fault of the Pohang Basin (from Song et al., 2015). 16

94 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 곡강단층선주변에대한야외조사결과, 분지기반암인고신생기유문암질화산암내에약 7 m의단층대폭을가지는 N64 E/84 SE 의정단층 (Fig. 2-7a) 과분지충전물을절단하는 N40 E/49 SE 의정단층이발견되었으며 (Fig. 2-7b), 굴착조사를통해북동내지동북동방향으로배열된여러조의공액상 (conjugate) 역단층들이제4기로추정되는지층을절단하고있음이확인되어 (Fig. 2-8) 곡강단층또는인근에서최근까지지표를절단하는단층운동이발생하였을가능성이높다. Fig Outcrop photographs showing NE-striking normal faults in (a) the basements (rhyolitic rocks) and (b) the Pohang basin-fill (mudstone) observed along the Gokgang Fault line. Fig NE- or ENE-striking conjugate reverse faults identified on (a) eastern and (b) western trench walls excavated along the Gokgang Fault line, which cut the Quaternary sediment layers (Unit A~H). 17

95 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 곡강단층선의주로서편에서포항지진에의한지표융기와균열 (surface crack) 그리고액상화에의한모래분출 (sand blow) 이발생하였으며 (Fig. 2-9a~d), 곡강단층선을따라서도지표균열이집중되는경향이야외조사를통해확인되었다 (Ghim et al., 2018; Choi et al., 2019). 또한포항지진에의해만들어진타원형의모래화산체 (sand volcano) 의장축과지표균열들은곡강단층의주향과유사한북동방향으로배열되는특징을보인다 (Choi et al., 2019; Fig. 2-9e). Fig Examples of ground cracks (a-c), sand blows (c-d) induced by the 2017 Pohang earthquake and (e) Rose diagram showing the orientations of the cracks and sand blows. 포항지진을발생시킨지하단층의기하를해석하기위해포항지진의본진과여진들의 3차원진원분포, 주요지진들의발진기구해그리고 InSAR 를이용한지표변형자료들을종합하면, 지하단층의파열면이북동주향을가지며북서방향으로 50~70 의경사를갖는수평이동성분을포함한역단층운동을한것으로해석된다 (Grigoli et al., 2018; Kim et al., 2018). 따라서포항지진을발생시킨단층운동은분지확장기동안인장력에의해만들어진여러북동방향의지하공액상정단층들중곡강단층과인접한반향단층 (antithetic faults) 중하나가현생응력장하에서우수향수평이동성분을포함한역단층으로재활한결과로해석된다 (Fig. 2-10; Choi et al., 2019). 18

96 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 Fig A schematic diagram showing the distribution of surface deformations across the 2017 Pohang earthquake rupture and proposed mechanism associated with blind oblique-slip including reverse-slip component and their related surface folding (Choi et al., 2019). 한편한반도일원의지진발진기구해, 활성단층의운동학적자료로부터복원된응력장그리고천부시추공응력자료들을종합하면 (Kim et al., 2016), 한반도동남부는현재동북동 -서남서또는동서방향의광역압축응력장하에놓여있으며이러한현생응력장은태평양판의서향저각섭입과인도- 유라시아충돌로부터전파된응력장이중첩된결과로해석된다 (Fig. 2-11). Fig Regional stress trajectory map showing the distribution of regional stress fields in the central and eastern parts of Eurasian continent. Reddish dashed lines indicate trace of the maximum horizontal stress axes based on the World Stress Map release 2008 (Heidbach et al., 2010; Kim et al., 2016). 19

97 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지열발전실증시설지하의암상분포와단층특성 주요암상경계시추공 PX-1 의심도 206 m까지마이오세포항분지충전물인연일층군퇴적암만관찰되다가 208 m부터백악기경상분지퇴적암시편이관찰되므로포항분지 -경상분지경계는심도 206에서 208 m 사이로판단된다 (Fig. 2-12). PX-2 의경우에는 186 m까지포항분지충전물만관찰되다가 200 m 이하에서경상분지퇴적암류과산성질화산암류가관찰되므로이시추공에서포항분지바닥심도는 186 m에서 200 m 사이이다. 한편, 이번보고서에서기술된시추공심도는모두시추공을따라측정된심도 (measured depth) 로실제수직심도 (true vertical depth) 는아니다. Fig Geological column showing the boundary between Miocene Pohang and Cretaceous Gyeongsang basins. PX-1 의 208~2,354 m 심도에서경상분지퇴적암류와화산암류시편들이관찰되다가 2,356 m부터약 280 Ma의페름기화강섬록암 (Yoon et al., 2015) 이관찰되므로경상분지 -영남육괴경계는 2,354 m와 2,356 m 사이로판단된다 (Fig. 2-13). PX-2 의경우에는 200~2,348 m에서경상분지퇴적암류와화산암류가관찰되다 2,350 m부터화강섬록암시편이관찰되므로경상분지 -영남육괴의경계는 2,348 m에서 2,350 m 심도사이이 20

98 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 다. 한편, 넥스지오가제공한 PX-1 Master log 에서영남육괴구성암석은화강섬록암과각섬암으로구분되나, PX-2 Master log 에서는염기성암, 섬록암, 화강섬록암, 화강암질편마암으로보다세분되어기술되어있다. 그러나실제육안관찰에서 PX-1 과 PX-2 커팅시료의특별한암상차이는인지되지않았다. Fig Geological column showing the boundary between Cretaceous Gyeongsang Basin and Yeongnam Massif 단층대 1) Master log 분석 PX-1 Master log에는단층의존재와관련한기록이전혀없는반면, PX-2 에는단층과관련한비지 (gouge), 단층핵 (fault core), 손상대 (fault damage zone), 소단층 (small fault), 가지단층 (small branched fault) 등의용어들이기재에사용되고있다 (Fig. 2-14; Source 2). 특히심도 3,115~4,202 m 사이에서단층관련기재가여러곳에서확인되는데, 단층핵은 3,533~3,536 와 3,541~3,546 m 구간그리고 5 m 이상단층비지대는 3,115~3,121, 3,153~3,185, 3,515~3,524, 3,533~3,546, 3,789~3,807, 4,173~4,192 m 구간에서기록되어있다. 21

99 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig Summary of information indicating the lithological changes and the presence of fault zones from PX-2 master log. Almost is assumed to be the term used to describe the state of being lumped together with clays and rock fragments. 2) PX-2 커팅시료육안관찰완전한세척이이루어지지않은 PX-2 의커팅시료는이수점토와단층비지의육안관찰이현재도가능한상태로, 3,790~3,815 m 구간의커팅시료에서다른구간에서관찰되지않는단층핵을구성하는단층암 ( 비지 ) 을지시하는특이한산상들이관찰되었다. 이구간시편들은다른구간에비해원마도가매우좋고풍화또는변질을받아손가락으로살짝누르면부러질정도로경도가매우약한암편들로대부분구성되며단층비지로추정되는점토물질도다량관찰된다 (Fig. 2-15). 특히, PX-2의 3,802~3,815 m 구간에서단층비지로추정되 22

100 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 는점토물질의함량이눈에띠게높았으며, 3,808, 3,811, 3,814 m 시료에서단층비지추정물질과함께원마 도가좋고쉽게부스러지는직경 1 cm 이하의암편들이다량관찰되었다. Fig Comparison of the cutting fragments at depths of 3,784 and 3,803 m of PX-2. The fragments at 3,784 m are mostly angular and fresh, while the fragments at 3,803 m are mostly rounded and degraded (friable). 3) PX-2 커팅시료현미경관찰 Master log에서단층핵으로기재된심도 3,535 m와단층손상대로기재된 3,544 m 시편그리고이번육안관찰에서단층암 ( 비지 ) 으로판단된쉽게부서지고원마도가좋은 3,791, 3,804, 3,807 m의시편들에대한실체현미경관찰을실시하였다 (Fig. 2-16). 먼저 3,535 m에서고철질암편이다수관찰되며일부 2 mm 이하의암편에서단층암조직이관찰된다 (Fig. 2-16a). 고철질암편은대부분신선하고각지며분급이불량하나 (Fig. 2-16h), 단층암조직을보이는것은원마도가고철질암에비해좋으며이수점토가주위를감싸는특징을보인다 (Fig. 2-16g). 3,544 m에는전형적인화강암조직을보이는암편들이다양한크기로관찰되며, 대부분신선하고강도가높으며각진특징을보이며시편주변에이수점토가관찰되지않는다 (Fig. 2-16b). 3,791 m 암편들은크기가다양하며높은원마도와대부분단층암조직을보이고이수점토로둘러싸여진특징을보인다. 단층암조직을보이지않는일부화강암과고철질암편들은 3,535 와 3,544 m 심도암편들에비해상대적으로크기가작고원마도가좋다 (Fig. 2-16c 와 d). 3,804 m의암편대부분은단층암조직을보이며장경 1 cm 이하로분급이불량하고입자주변에이수점토가두껍게피복된다 (Fig. 2-16e). 마지막으로 3,807 m 암편들은 3,791 m 암편들과유사하게대부분단층암조직을보이며원마도가좋고이수점토가시편을감싸는특징을보여주나, 대부분 1 cm 내외의크기를보여준다 (Fig. 2-16f). 23

101 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig Stereoscopic photomicrographs of cutting fragments at depths of (a) 3,535 m, (b) 3,544 m, (c) and (d) 3,791 m, (e) 3,804 m and (f) 3,807 m of PX-2. Fragments of (c) to (f) mostly show textures indicating fault gouge or ultracataclasites. (g) and (h) Cutting fragments at a depth of 3,535 m surrounded by drilling mud and no drilling mud, respectively. 24

102 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 전형적인화강암조직을보여주는 3,544 m 시편을제외한모든심도의시편들은편광현미경하에서취성영역에서생성된점토질단층비지의미구조특징을뚜렷이보여준다 (Fig. 2-17). 현미경상에서시편들은최대 500 μm크기의아각형 -아원형의암편들을포함하며암편들의원마도는전체적으로양호하다. 고배율로관찰된수μm이하의미세암편들은주로석영과장석으로구성되며일부염기성광물및방해석을포함한것도존재한다. 또한단층활동이후생성된침상의 2차광물들도소량관찰된다. 기질부에는개방니콜에서암갈색내지담갈색그리고녹니석으로추정되는녹색의점토광물들과불투명광물들이서로대상 (banded) 으로교호하며뚜렷한엽리가만들어져있다. 엽리의두께는수십μm, 방향은규칙적인부분과불규칙적인부분이모두관찰되며, 상대적으로크기가큰암편주변부에는엽리가암편을감싸듯이발달한다. 엽리사이에는전단에의해암편의깨진입자들이엽리와유사한방향으로흐르는조직인파쇄유동 (cataclastic flow; Chester et al., 1985) 의전형적인모습이관찰된다. 양호한원마도의암편들과입자크기의감소, 엽리상단층비지, 파쇄유동조직등의존재를고려할때, PX-2 의 3,791, 3,804, 3,807 m 심도에서관찰되는시편들은대부분반복된단층활동으로인해단층암이일정한규모로발달하는성숙한 (mature) 단층대로부터유래된것으로판단된다. 또한단층암내방해석및침상의 2차광물들의존재는단층암생성동시혹은이후단층대내부로유체의이동이가능하였음을지시한다 (Fig. 2-17). Fig Polarizing photomicrographs of thin sections of cuttings at depths of 3,791 m (a) under open and (b) crossed polars, 3,804 m (c) under open and (d) crossed polars, and 3,807 m (e) under open and (f) crossed polars. 25

103 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 4) PX-2 커팅시료의 X-선회절분석 X-선회절분석은이수점토를세척하지않은상태에서 315 μm이하의시료를채로걸러건조기에 24시간건조뒤분말화하여실시하였다. 분석결과, 석영 (quartz), 사장석 (plagioclase), K-장석 (K-feldspar), 방해석 (calcite), 각섬석 (amphibole) 과같은일반적인조암광물과녹니석 (chlorite), 운모류 (Mica; 일라이트 (Illite) 와백운모 (muscovite)), 불석광물 ( 로만타이트 ; laumontite) 과같은단층암에서이차로흔히산출되는광물들이함께존재하고있음이확인되었다 (Table 2-1; Fig. 2-18). Table 2-1. Mineral compositions of the PX-2 cuttings measured by XRD (wt%). Qz: Quartz, Pc: Plagioclase, K-fd: K-feldspar, Am: Amphibole, Ch: Chlorite, Mica: Illite+Muscovite, La: Laumontite, Cc: Calcite, and Gs: Gypsum. Depth (m) Qz Pc K-fd Am Ch Mica La Cc Gs 3, , , , , , , , , , 주로화강섬록암으로구성되는 3,544 m 심도커팅시료에서는석영의함량이다른구간에비해매우높으며로만타이트가산출되지않는다. 3,535 m에서 3,791 m 구간에는각섬석이거의산출되지않는반면, 3,792 m 에서 3,814 m 구간에는각섬석이 6.3~12.1% 내외로상당량산출되는것으로보아 (Table 2-1) 심도 3,791 과 3,792 m 사이에서급작스런암상변화가발생하였을가능성을지시한다. 26

104 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 Fig X-ray diffraction patterns of PX-2 cuttings. 27

105 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 2.4. 결론 2017 년 11월 15일발생한포항지진의진앙이위치한마이오세포항분지는북북서방향의우수향주향이동단층과이에수반된북북동내지북동방향의정단층들의운동에의해확장된당겨열림형퇴적분지로운동학적으로시계방향으로회전된지괴와회전되지않은지괴사이에서벌어진쐐기형분지이다. 포항분지가서북서- 동남동방향의인장하에서확장되는동안분지내부에는북동내지북북동방향의퇴적동시성성장정단층들도다수만들어졌으며, 이중규모가가장큰단층인곡강단층이포항지진의진앙과인접하게위치한다. 포항지진진앙지일원의지질구조와지표변형특성, 여진들의 3차원진원분포, 발진기구해등의자료를종합하면, 포항지진은서쪽으로약 50~70 경사진분지확장동안만들어진북북동방향공액상정단층들중곡강단층의반향단층하나가우수향수평이동성분을가지는역단층으로재활된결과로해석된다. 포항지열발전실증시설에서수행된시추공의커팅시료를분석하면, 포항분지 -경상분지경계는 PX-1 지점에서심도 206~208 m 사이이며 PX-2 지점에서 186~200 m 사이에위치한다. 또한경상분지 -영남육괴경계는 PX-1 지점에서심도 2,354~2,356 m 사이에그리고 PX-2 지점에서 2,348~2,350 m 사이에위치한다. 한편, PX-2 의 Master log 분석, 커팅시료들의육안관찰과현미경관찰그리고 X-선회절분석을통해서 3,790~ 3,815 m 심도구간에서단층핵에해당하는단층비지대가존재함이이번연구에서확인되었다. 이심도구간커팅시료들은다른구간의시료들과달리육안관찰에서원마도가매우좋고풍화또는변질로손가락으로부스러질정도로약한강도를가지며다량의점토물질로피복된특징을보여준다. 또한실체현미경과편광현미경관찰에서양호한원마도의암편들과입자크기감소, 엽리상단층비지, 파쇄유동조직등의특징을보여주고있고 X-선회절분석에서단층암에서흔히산출되는이차광물들이존재하고있음이확인되어반복된단층활동으로인해일정한규모로발달하는성숙한단층대가이심도구간에존재하는것으로해석된다 년 8월에수행된 PX-2 공의시추공영상검층에서삽입된장비가 3,783 m 심도에서막혀더이상아래로내릴수없었다는사실 ( 부록 A) 은이곳하부에주단층의파열면이존재하여시추공케이싱이심하게변형되었거나파열되었음을암시한다. 또한이번조사에서확인된 3,790~3,815 m 심도구간의단층핵을따라포항지진을발생시킨단층운동이발생하였음을지지한다 참고문헌 Chang, K.H., Suzuki, K., Parka, S.-O., Ishida, K., and Uno, K., 2003, Recent advances in the Cretaceous stratigraphy of Korea. Journal of Asian Earth Sciences, 21, Cheon, Y., Son, M., Song, C.W., Kim, J.-S., and Sohn, Y.K., 2012, Geometry and kinematics of the Ocheon Fault System along the boundary between the Miocene Pohang and Janggi basins, SE Korea, and its tectonic implications. Geosciences Journal, 16, Chester, F.M., Friedman, M., and Logan, J.M., 1985, Foliated cataclasites. Tectonophysics, 111,

106 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 Choi, J.-H., Ko, K., Gihm, Y.S. Cho, C.S. Lee, H., Song, S.G., Bang, E.-S., Lee, H.-J., Bae, H.-K., Kim, S.W. Choi, S.-J., Lee, S.S., and Lee, S.R., 2019, Surface deformations and rupture processes associated with the 2017 M W 5.4 Pohang, Korea, Earthquake. Bulletin of the Seismological Society of America, doi: / Chough, S.K. and Sohn, Y.K., 2010, Tectonic and sedimentary evolution of a Cretaceous continental arc-backarc system in the Korean Peninsula: new view. Earth-Science Reviews, 101, Gihm, Y.S., Kim, S.W., Ko, K., Choi, J.-H., Bae, H., Hong, P.S., Lee, Y., Lee, H., Jin, K., Choi, S.-J., Kim, J.C., Choi, M.S., and Lee, S.R., 2018, Paleoseismological implications of liquefactioninduced structures caused by the 2017 Pohang Earthquake. Geosciences Journal, 22, Grigoli, F., Cesca, S., Rinaldi, A.P., Manconi, A., Lopez-Comino, J.A., Clinton, J.F., Westaway, R., Cauzzi, C., Dahm, T., and Wiemer, S., 2018, The November 15, 2017 Pogang earthquake: A probable induced event of M W 5.5 in South Korea. Science, 360, GSK (The Geological Society of Korea), 1998, Geology of Korea. 802 p (in Korean). Heidbach, O., Tingay, M., Barth, A., Reinecker, J., Kurfeв, D., and Muller, B., 2010, Global crustal stress pattern based on the World Stress Map database release Tectonophysics, 482, Kim, K.-H., Ree, J.-H., Kim, Y., Kim, S., Kang, S.Y., and Seo, W., 2018, Assessing whether the 2017 M W 5.4 Pohang earthquake in South Korea was an induced event. Science, 360, Kim, M.-C., Jung, S., Yoon, S., Jeong, R.-Y., Song, C.W., and Son, M., 2016, Neotectonic crustal deformation and current stress field in the Korean Peninsula and their tectonic implication: a review. Journal of the Petrological Society of Korea, 25(3), (in Korean with English abstract). Sohn, Y.K. and Son, M., 2004, Synrift stratigraphic geometry in a transfer zone coarse-grained delta complex, Miocene Pohang Basin, SE Korea. Sedimentology, 51, Son, M., Chong, H.Y., and Kim, I.-S., 2002, Geology and geological structures in the vicinities of the southern part of the Yonil Tectonic Line, SE Korea. Journal of the Geological Society of Korea, 38, (in Korean with English abstract). Son, M., Song, C.W., Kim, M.-C., Cheon, Y., Cho, H., and Sohn, Y.K., 2015, Miocene tectonic evolution of the basins and fault systems, SE Korea: dextral, simple shear during the East Sea (Sea of Japan) opening. Journal of the Geological Society, 172, Song, C.W., 2015, Study for the evolution of the Miocene Pohang Basin by the analysis of the structural characteristics. Ph. D. thesis, Pusan National University, 144 p (in Korean with English abstract). Song, C.W., Moon, S., Sohn, Y.K., Han, R.H., Shin, Y.J., and Kim, J-C., 2015, A study on potential geologic facility sites for carbon dioxide storage in the Miocene Pohang Basin, SE Korea. Journal of the Geological Society of Korea, 51, (in Korean with English abstract). Song, S.J., Choo, C.O., Chang, C.J., and Jang, Y.D., 2017, A microstructural study of the fault gouge in the granite, Yangbuk, Gyeongju, southeastern Korea, with implications for multiple faulting. 29

107 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Geosciences Journal, 21, Yoon, K.-S., Jeong, J.-S., Hong, H.-K., Kim, H.-G., Hakan, A., Park, J.-H., and Yoon, W.-S., 2015, Deep drilling experience for Pohang Enhanced Geothermal Project in Korea. Proceedings World Geothermal Congress Yun, H.S., 1994, Emended stratigraphy of the miocene formations in the Pohang Basin, Part II: South of the Hyongsan Fault. Journal of the Paleontological Society of Korea, 10, Unpublished sources Source 1: Master log of well PX-1 (Excel file) Source 2: Master log of well PX-2 (Excel file) 30

108 제 2 장포항지열발전실증시설수리자극과포항지진의상관성해석을위한지질구조특성분석 제 3 장 지구물리탐사및자료해석 Geophysical Exploration and Data Analysis 31

109 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project 32

110 제 3 장지구물리탐사및자료해석 요약 / Abstract 포항지진구역의지하구조를영상화하고지표면변위에기반한단층모델링수행을위해지구물리탐사와원격탐사자료해석을수행하였다. 지하구조는 MT 탐사자료를이용한전기비저항분포를파악하여이루어졌으며, 굴절법및반사법탄성파탐사자료를이용하여천부구조와시추공심도까지속도분포를파악하였다. 원격탐사자료해석을위해세종류의위성을사용하였으며, 지표면변위분석, 단층모델링및본진전지표변위의시계열분석을수행하였다. MT 탐사는지표에서수행하는물리탐사방법중가장깊은곳까지조사할수있는방법이며, 본연구에서는한국지질자원연구원에서기존에수행한 70여지점의 MT 탐사자료와금번조사를위해추가적으로수행한 AMT 자료를동시에활용하여 2차원및 3차원역산을통해지하구조를추정하였다. 2차원분석을통해지열발전실증사이트인근에서북동방향의주향을갖는것으로추정되는낮은전기비저항대가나타나는것이확인되었으며이는포항지진을유발한단층대와연관이있는것으로파악된다. 3차원분석결과 EGS 사이트의 ESE 방향에서이어지던높은전기비저항대가 EGS 사이트인근에서저비저항대로전환되는것이관찰되었으며이역시 2차원결과와유사한단층대의존재를의미하는것으로파악된다. 이상의물리탐사결과를토대로포항지진구역의지질구조를파악하여볼때, EGS 사이트인근에낮은전기비저항대로나타나는단층대가존재한것으로추정되었으며, 시추공과일부지점에서교차할가능성을암시하였다. 탄성파탐사자료해석은한국지질자원연구원에서수행한자료를이용하였으며, 반사법탐사자료를통해천부구역의층서구조를파악하였고초동주시발췌를통해서천부의속도분포와경사구조를확인하였다. 초동주시해석결과, 하부약 200 m 심도부터급격한탄성파속도의증가가관측되었으며, 음파검층자료와 check shot, 시추공내암상자료등을이용하여시추공하단까지의속도분포를추정하였다. 탄성파자료로부터추정한속도분포는진원깊이분석을위한연구에공동으로사용되었다. 분석결과, 기존의연구에서사용한속도를전반적으로상회하는속도분포가예측되었다. 원격탐사를이용한지표면변위관측과정밀한단층모델링을위해다양한관측방향에서촬영된 SAR 자료를수집하였다. Sentinel-1A/B, ALOS-2, Cosmo-SkyMed 위성으로부터획득된 SAR 자료의레이더간섭기법을이용하여수 mm~cm 정밀도로지형변위관측을수행하였다. 다수의위성으로부터관측된변위성분으로부터 3차원변위성분을분해하고, 오카다모델에적용하여지표하부에위치한단층의주향과경사및이동량을계산하고분석하였다. 또한, 몬테카를로시뮬레이션을적용하여오카다모델을통해분석된단층의기하학적특성에대한민감도분석도함께실시하였다. Geophysical exploration and remote sensing data were analyzed to visualize the underground structure of the Pohang earthquake zone and to perform fault modeling based on surface displacement. The underground structure was constructed by analyzing the electrical resistivity distribution using MT survey data, and the near surface structure and the velocity distribution up to the borehole depth was obtained by using refraction and reflection seismic survey data. Three kinds of satellites were used to analyze the remote sensing data, and the surface displacement analysis, the fault modeling and the temporal analysis of surface displacement before the main shock were performed. In this study, MT survey data of more than 70 sites obtained by the Korea Institute of Geoscience and Mineral Resources (KIGAM) were analyzed and additional AMT data were collected for integrated 33

111 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 2D and 3D inversion analysis. 2D analysis revealed that a low resistivity zone near EGS site appears to have a tendency toward NE direction, which is related to the fault zone that caused the Pohang Earthquake. As a result of the 3D analysis, it was observed that the high-resistivity zone, which continued from the ESE direction of EGS site, was converted to a low resistivity region near EGS site. Based on the results of the above geophysical surveys, it is assumed that there is a low resistivity fault zone near the EGS site and the possibility of intersection with some boreholes is implied. Seismic survey data obtained by KIGAM were also analyzed, and the stratigraphic structure of the near surface area was identified through reflection survey data. As a result of the analysis of the first-arrival travel-time data, the rapid increase of the seismic velocity was observed from the depth of 200 m, and the velocity distribution to the lower end of the borehole was estimated by using the sonic log, check shots and lithological information. Velocity distributions estimated from seismic data have been used jointly for the identification of focal depth. As a result of the analysis, the velocity distribution which is generally above the speed used in the previous studies was predicted. The SAR data were collected from various observation direction for surface displacement measurement and accurate fault modeling using remote sensing. Observation of surface displacement was performed with a precision of several millimeters to several centimeters using radar interferometry of SAR data obtained from Sentinel-1A/B, ALOS-2, and Cosmo-SkyMed. The 3D displacement component is decomposed from the InSAR measurements observed from multiple satellites and applied to the Okada model to calculate and analyze the direction, slope, and movement of the fault located in the lower part of the surface. In addition, the sensitivity analysis of the geometrical characteristics of the faults estimated by the Okada model was also performed by applying the Monte Carlo simulation 지구물리탐사배경및필요성 포항지진발생지역에대한지하구조정보를구하기위하여지구물리탐사기법이적용되었다. 지구물리탐사는지하의구조를전기비저항, 탄성파속도등주요물성으로해석함으로써지하정보를알아낼수있다. 본조사단에서는지구물리탐사방법중포항현장에서적용가능한다양한지구물리탐사기법을적용하여포항지진발생지역의지하정보를획득하고자하였다. 이를위해기존연구획득자료의재해석과더불어신규지구물리탐사자료를획득하고이를해석하였다. 대상지역에서기수행된지구물리탐사는중자력탐사, MT (magentotelluric) 탐사, 전기비저항탐사, SP탐사, 탄성파반사법및굴절파탐사등이며이에대한기본적인검토가수행되었고필요한경우에대하여재해석이수행되었다. 34

112 제 3 장지구물리탐사및자료해석 지구물리탐사는크게전기비저항을영상화하기위한 MT 및 AMT 탐사, 지하지층의구조와속도구조를규명하기위한탄성파탐사그리고원격탐사기반단층모델링으고구분되어수행되었다. 이중 MT 및 AMT 탐사자료는기존자료의병합과자료처리과정을거쳐지하의비저항구조를 2차원혹은 3차원으로영상화함으로써지열발전실증사이트주변의단층, 파쇄대로해석될수있는낮은전기비저항이상대의발달양상을제시하고자하였다. 또한탄성파탐사는직접적인지층의경계구조와탄성파신호가투과, 반사된개별지층의속도정보를제공할수있으므로, 지층경계를영상화하고대상지역의속도구조를확인하기위한목적으로활용되었다. 한편지진발생위치및단층면의구조를추정하기위한원격탐사기반단층모델링방법이적용되었다. 원격탐사중, 영상레이더 (SAR) 자료를이용한위성영상레이더간섭기법 (SAR interferometry; InSAR) 은지진에의한지표변위를수 mm~cm 의정밀도로관측할수있으며, 이를통해산정된미세지표변위를단층현상에의해발생하는이론적미세지표변위와비교함으로써실제지진발생위치및단층면의구조를추정하는데적용될수있는방법이다. 적용된지구물리탐사는포항지진발생지역의지하정보를간접적으로제공할수있는기법으로지표지질, 직접시추및지진파분석등을통해얻을수있는정보와더불어포항지진의원인과특성을규명하는데기여하였다 지하구조비저항영상화 MT 탐사자료획득 2000 년대초 KIGAM 에서는경상북도포항시흥해읍일대에서지열자원개발을위한심부파쇄대탐지를목적으로자기지전류 (MT) 탐사를수행하였다 ( 이태종외, 2005; Lee et al., 2007). MT 탐사는전기장 2성분과자기장 3성분의텐서측정이가능한캐나다 Phoenix 사의 MTU-5 및 MTU-5A 시스템이사용되었다. 이탐사시스템은측정점의위치및시간정보를제공하는정밀 GPS가장착되어있어, 측정점과원거리기준점간의시간동기화 (synchronize) 가용이하다. 현장조사는총 8대의 MT 탐사시스템을사용하여이루어졌으며, 2002 년도에 33측점, 2003 년도에 37측점에서 MT 자료가획득되었다. 한편포항지진과지열시추공에서수행된수압파쇄의관련여부를규명하기위해서는 EGS 사이트를중심으로하는지하지질구조, 특히단층대의기하학적위치및특성을파악해야한다. 그러나앞서언급한바와같이 2002 년도및 2003 년도에심부지열자원개발을수행된 MT 탐사영역은현재의 EGS 사이트와는공간적으로상당히떨어져있다. 이번조사에서는 EGS 사이트주변에대한추가적인 AMT 탐사를통하여자료를획득하고, 이들자료를기존의 MT 탐사자료와병합하여종합적인해석을수행하였다. 추가조사에서 MT 탐사대신에주파수대역이높은 AMT 탐사를적용한이유는 MT 탐사에사용되는저주파대역의잡음수준이높고, AMT 탐사가 MT 탐사에비하여측정시간이매우짧으면서도효과적으로천부의전기비저항분포에관한정보를제공해줄수있기때문이다. Fig. 3-1은기존 MT 및추가 AMT 측점의위치, 2차원해석을위한 35

113 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 측선및 3차원해석을위한역산영역을나타낸것이다. AMT 탐사에서가장큰문제는조사지역의전자기적잡음수준이매우높다는점이다. 특히포항지역은국내의주요산업단지로잡음수준이매우높은지역에해당하며, 현재는 KIGAM 이자료를획득한 2002~ 2003 년도와비교해도급속한도시화로인하여전자기적잡음수준이매우악화된상태이다. 따라서대부분의 AMT 측점에서심부의정보를제공하는저주파대역의자료가잡음에심하게영향을받아편집과정에서제외하였으며, 상대적으로양호한주파수자료만을선별적으로해석에사용하였다. 추가탐사를통하여얻어진 AMT 자료는자료처리를통하여잡음을제거하는과정을거쳐, 기존의 KIGAM 자료와병합하였다. 이통합된자료에대하여 EGS 사이트주변의지질구조규명을위하여 2차원역산과 3차원역산을수행하였다. 이를통하여 EGS 사이트주변의 2차원및 3차원적물성 ( 전기비저항 ) 분포를영상화하여지질구조, 특히파쇄대의기하학적위치및그특성의해석에필요한기본정보를제공하고자하였다. Fig MT/AMT site map. The solid lines and the green box represent the survey lines for 2D interpretation and 3D inversion area, respectively MT 탐사 2 차원자료해석 2000 년대초 KIGAM 에서획득한자료와 2018 년도에획득된 AMT 탐사자료를병합하여 EGS 사이트를중심으로하는다양한측선에대한 2차원역산을수행하였다. 그러나상당부분의자료, 특히저주파대역자료가해석에사용하기어려울정도도극심한잡음에오염되어자료의선별과정에상당한어려움이있었으며, 연속성이떨어지는자료는편집과정에서제거하였다. 이러한저주파대역자료의결손은역산해석에서심부 36

114 제 3 장지구물리탐사및자료해석 의물성분포추정을어렵게한다. EGS 사이트주변의물성분포영상화를위하여 Fig. 3-1에주어진바와같이조사지역의주된구조선과수직한방향인북서- 남동방향을갖은 5개의측선 (V1~V5) 에대한 2차원역산을수행하였다 년도, 2003 년도에획득된 MT 시계열과 2018 년도에얻어진 AMT 시계열자료의샘플링주파수가각각다르며, 자료편집결과측선마다다른특성을보이기때문에, 측선별로최적의주파수에대한임피던스를추출하였다. 다음 2차원역산을위하여임피던스회전을통하여 Fig. 3-1에나타낸각측선에대하여전기장이수직한 TE 모드및자기장이측선에수직한 TM 모드겉보기비저항및위상을계산하고, 측선별로 2차원역산을수행하였다. MT 2차원모형반응계산은유한요소법 (FEM) 을사용하였으며, 역산은최소제곱역산법 (Uchida, 1993) 을사용하였다. 한편 2차원역산에서 TE mode 자료를포함할경우매우불안정하여 TM mode 자료만을사용하여역산을수행하였다. Fig. 3-2는 5개의북서- 남동방향 5개의측선에대한 2차원역산결과이다. 모든측선에서천부에반고결이암층으로보이는 10 ohm-m 정도의저비저항층이발달하고있으며, 중간에상대적고비저항층이, 그리고심부에는저비저항층이나타나고있다. 천부의저비저항층은위치에따라두께를달리하며전반적으로북쪽에서는얇고, 남쪽으로갈수록두꺼워지는경향을보이고있다. 중간정도의깊이에서는북쪽에서는매우높은전기비저항대가발달하는반면남쪽에서는전기비저항값이상당히감소하는양상을보이고있다. 심부에나타나는낮은전기비저항층은역산에사용된자료, 특히 AMT 탐사자료의가탐심도를벗어나는영역에속하므로실제발달여부를확정하기어렵다. 2차원역산결과에서예상되는낮은전기비저항이상대는 2개정도로추정된다. 우선 EGS 사이트에근접한 V1, V2, V3 측선은모두측선의우측끝 ( 동남방향 ) 에서낮은전기비저항을보이고있으며, 이후고비저항대가나타나고있다. 즉측선 V1의측점 14, 측선 V2의측점 910과측선 V3의측점 715를연결하는전기비저항경계부는북서방향으로급경사를보이며, 단층혹은파쇄대로해석될수있다. 이이상대는 EGS 시추공에서동남방 1,000 m 정도에위치하며상당한깊이까지연장되어있어 EGS 시추공을통과할가능성이높으며, 공간적으로떨어져있어무리가따르지만측선 V4의측점 424에서시작되는북서방향으로경사진저비저항이상대와도연결가능하다. 그러나이낮은전기비저항이상대는측선의가장자리에위치하기때문에역산의가장자리효과혹은동해바닷물에의한영향일수도있다. MT 탐사의분해능을고려할때상당한폭을갖는같은파쇄대에의한이상대일가능성을배제할수없으며, 하나의동일한이상대로해석하는것이보다합리적으로판단된다. 두번째이상대는지표기준으로측선 V1의측점 433, 측선 V2의측점 207, 측선 V3의측점 302에서시작되는저비저항이상대는공간적으로연결성이뛰어나단일파쇄대로해석된다. 비록측선 V4에서는흥해분지내의자료부족으로인하여이상대가나타나지않지만, 측선 V5의측점 407~411 에서시작되는저비저항이상대와연결된것으로해석되며, EGS 시추공에서서북방 2,000 m 정도에위치한다. 37

115 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 (a) (b) (c) (d) (e) Fig Resistivity models from 2D inversion of TM mode data for the survey line (a) V1, (b) V2, (c) V3, (d) V4 and (e) V5, respectively MT 탐사 3 차원자료해석 2 차원역산은지하구조를 2 차원으로가정하는근본적인문제점을가지고있다. 특히본조사지역과같이 지질구조가 3 차원특성을보이는좁은영역의조사에서는 3 차원해석이요구된다. 이조사에서는기존에 38

116 제 3 장지구물리탐사및자료해석 KIGAM 에의해 2002 년과 2003 년에획득된 MT 자료와 2018 년도에추가된 AMT 자료를통합한자료를사용하여 EGS 사이트를포함하는영역에대한 3차원역산을수행하였다. 물론자료획득이불가능한흥해읍과같은도심지역과각종산업시설, 건물, 도로, 철도및고압전력선으로인하여 3차원해석에충분한자료가확보되지는못하였으며, 특히심부의정보를제공하는저주파대역의자료부족으로인하여주된관심사인 4,000 m 이상깊이의물성분포영상화는불가능하였다. 또한자료수가부족할경우가중되는비선형역산의비유일해 (non-unique solution) 문제에서도자유로울수없다. 그러나주어진조건하에서 EGS 사이트주변의 3차원물성분포를영상화하고지질구조에대한정보를추출하고자하였다. Fig. 3-1에나타낸초록색사각형내에위치한 109개의측점에서얻어진 MT 및 AMT 탐사통합자료를사용하여 3차원역산을수행하였다. 모든측점에서 ~390 Hz까지대수적으로등간격인 24개의주파수에대한 TE 및 TM 모드겉보기비저항및위상자료를추출한다음, 이중잡음이심한자료를제외하고 6,754개의자료를사용하여역산을수행하였다. 3차원 MT 모형응답반응은유한차분법 (FDM) 을사용하였으며 (Mackie et al., 1994), 모델링요소의크기는수평방향으로 200 m, 수직방향으로는 50 m에서시작하여깊이가증가할수록그크기를증가시켰다. 역산방법은정적효과를고려한최소제곱법 (Sasaki, 2004) 을적용하였다. Fig. 3-3은 3차원역산결과영상으로측점의위치및역산결과의분석을위하여설정된 4개의측선을함께표시하였다. 이지역의전반적인전기비저항분포는대략 3층구조를보인다. 500 m 이내의천부에는반고결이암층으로해석되는낮은전기비저항층이, 그하부에상대적으로높은전기비저항층이발달하고있으며, 심부에는다시저비저항층이나타나고있다. 물론심부의저비저항층은가탐심도의한계로신뢰도가떨어진다. 천부의저비저항층은남쪽에서북쪽으로갈수록두께가얇아지는특성을보이며, 동서방향으로는위 Fig D resistivity distribution from 3D inversion of MT and AMT data. The black and blue circles indicate the MT sites by KIGAM at 2002 and The red circles represent the AMT sites at Four survey lines (black lines) on the surface are assumed for the 2D interpretation. 39

117 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 치에따라두께를달리한다. 또한역산결과에서는흥해읍을중심으로천부저비저항대가 1,000 m 이상의깊이까지나타나고있으나, 흥해읍주변의측점부족으로인하여그진위여부를판정하기는어렵다. 한편 EGS 시추공의북쪽천부에나타나는고비저항이상대는관입화강암에의한것으로해석된다. Fig. 3-4는 EGS 시추공부근에서파쇄대로해석가능한저비저항대의분포양상을분석하기위하여 Fig. 3-3에표시된측선 A1부터측선 A4 하부의전기비저항분포단면을나타낸것이다. 모든측선에서북서쪽으로경사진저비저항이상대가북동- 남서방향으로일관되게나타나고있으며, 이는 2차원해석에서시추공동남부에나타난저비저항이상대와잘일치하고있다. 이저비저항이상대는공간적연결성이뛰어나며, 2차원역산결과와도잘부합하므로파쇄대일가능성이높다. 특히 Fig. 3-4(d) 에나타낸측선 A4는 2차원역산영상인 Fig. 3-2(d) 에나타낸측선 V4와거의일치한다. 이두역산결과를비교하면전반적으로유사한전기비저항분포를보인다. 특히 2차원역산결과에서측선 V4 측점 424에서시작되는북서방향으로경사진저비저항이상대는 3차원역산결과에서도나타나고있다. 그러나 3차원역산결과에서는전기비저항대비가크지 (a) (b) (c) (d) Fig Resistivity sections along the survey lines shown in Fig. 3-1; (a) A1, (b) A2, (c) A3, and (d) A4, respectively. The low resistivity zone (white circle) at the central part of the section is interpreted as the Heunghae-eup artefact. 40

118 제 3 장지구물리탐사및자료해석 않아 2 차원역산영상에비하여이상대가뚜렷하지못하며, 모든단면중앙에광범위하게나타나는낮은전기 비저항대는흥해읍일대의자료부족으로나타난역산잡음이출현하고있다 결과및토의 KIGAM 에서는경상북도포항시흥해읍일대에서심부파쇄대탐지를목적으로 2002 년부터 2003 년에걸쳐 MT탐사를수행하였다. 그러나 KIGAM 에의해수행된 MT 탐사조사에서는현재의지열발전실증사이트가주된탐사대상이아니었기때문에지열발전실증사이트인근의측점빈도가매우부족하다. 이조사에서는이러한문제점을극복하기위하여지열발전실증사이트를중심으로하는영역에대하여추가적인 AMT 탐사를실시하고, 기존자료와추가탐사자료를병합한자료에대한 2차원및 3차원역산을통하여지열발전실증사이트주변의전기비저항분포양상을영상화하였다. MT 및 AMT 통합자료에대한 2차원역산을통하여 2조의낮은전기비저항이상대를확인하였다. 이들이상대는모두북동- 남서방향의주향과북서방향의경사를보이며, 지열발전실증사이트시추공에서동남방 500~1,000 m 및서북방 2,000 m 정도에각각위치한다. 그러나이들이상대는측선의가장자리에위치하거나흥해분지와같은자료결손지역을통과하기때문에신뢰도에문제가있다. 한편통합 MT 자료에대한 3 차원역산결과에서는지열발전실증사이트시추공에서동남방으로대략 500~1,000 m 거리에낮은전기비저항이상대가확인되었으며, 주향및경사방향도 2차원역산결과와잘부합되는것으로나타났다. 그러나 2차원역산에서지열발전실증사이트북서방 2,000 m 지점에나타난낮은전기비저항이상대는 3차원역산결과에서확인할수없었다. 이상에서지열발전실증사이트주변에발달한파쇄대로해석가능한저비저항이상대는시추공에서동남방으로대략 1,000 m 정도의거리에위치하며, 북동- 남서방향의주향과북서방향의경사를가진것으로해석된다. 또한이이상대는북서방향의경사를보이며, 지열발전실증사이트시추공 PX-2 와대략 3,000 m 이상의깊이에서교차하는것으로해석된다. 한편포항지역은전자기적잡음이극심한지역으로 2018 년도에획득한 AMT 탐사자료의질이낮다는문제점을가지고있다. 즉이번조사지역은 KIGAM 이자료를획득한 2002 년및 2003 년과비교해도각종건물, 도로, 철도및고압전력선등에의한전자기적잡음이크게악화되어양질의자료획득이어려웠으며, 방법론적으로가탐심도가낮은 AMT 법을적용하였으므로가탐심도가제한된다는문제점을가지고있다. 또한급속한도시화로인하여공간적으로자료획득이불가능한영역이많아역산결과의신뢰도및해석에상당한문제점을나타내었다. 비록자료처리과정에서최대한잡음을억제하여양질의자료를확보하고자하였으나 2018 년도에획득된 AMT 탐사자료의질은좋다고볼수없다. 따라서 MT 탐사자료의역산을통하여제시된전기비저항영상의해석은지질조사, 시추조사, 지진자료분석및다른계측자료들과종합하여신중하게해석되어야할것으로판단된다. 41

119 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 3.3. 탄성파탐사자료해석 자료설명 지표탄성파탐사자료및물리검층자료 포항지열발전실증사이트부근탄성파자료는약 1.2 L-km 규모의탄성파반사법탐사자료와지열시추공 PX-2 에서측정한음파검층자료이다. 추가적으로 PX-2 로부터약 2.5 km 떨어진곳에위치한시추공 BH-3, BH-4 에서측정된음파검층자료를연구에활용하였다. Fig. 3-5는지열발전실증시설위치와탄성파반사법탐사측선, 시추공 (PX-2, BH-3, BH-4) 의위치를표시한지도이다. Fig Locations of the EGS site, boreholes, seismic reflection survey line. 시추공배열식지진계자료포항지열발전미소진동모니터링자료중가장품질이좋은시추공배열식지진계자료 (3성분지오폰 17 개 ) 에서지표인공송신원에의한기록 (check shot data) 을속도모델검증및교정 (calibration) 에활용하였다 (Fig. 3-6). Table 3-1은인공송신원에대한위치좌표와발파시간을나타낸다. 위치좌표는시추공 PX-2 를기준으로한상대좌표로미터단위이고시간은세계협정시 (UTC) 이다. 42

120 제 3 장지구물리탐사및자료해석 (a) (b) Fig (a) Geometry of the borehole geophone array placed at PX-2 and (b) the location map of 6 check shots. Table 3-1. Origin times and coordinates for check shot events. Shot # Time (UTC) X (North, m) Y (East, m) Altitude (m) 1 00:31: , :17: , :17: , :12: , :57: ,137-1, :33: ,594-2, 탄성파반사법자료처리결과 지표탄성파반사법자료는 2012 년 2월지열개발사업을통해수행한자료로진동형송신원 (Vibroseis) 과폭발형송신원 (dynamite) 2가지를사용하여취득되었다. 측선길이는약 1.2 L-km이고기록시간은송신원에따라 2초또는 3초로되어있어깊은심도의지층구조를규명하기에는한계가있다. 또한신호대잡음비가좋지못하여의미있는속도정보를얻기는매우어려운것으로판단된다. 자료처리는 Schlumberger 사의 VISTA 소프트웨어를이용하여트레이스편집 (trace editing), 정적보정 (static correction), 표면파제거 (ground-roll suppression), 겹쌓기속도분석 (stacking velocity analysis), 곱풀기 (deconvolution), 수직시간차보정 (NMO correction), 겹쌓기 (stacking) 등기초적인탄성파반사법자료처리를수행하였다. 43

121 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 반사법자료처리결과기본적인반사법자료처리단계를거쳐겹쌓기 (stack) 단면을얻고깊이변환을하였다 (Fig. 3-7). 단면구조를볼때약 200 m 깊이에반사면이존재하는것으로보이는데, 이반사면은천부의미고결이암 (unconsolidated mudstone) 층과그하부의결정질응회암 (crystal tuff) 층의경계인것으로추정된다. 공통중간점 (CMP) 300번과 500번사이구간에는 600~1,000 m 깊이에서서쪽으로경사진반사면이보이는데, 이반사면은 3기층과중생대백악기퇴적층 ( 사암층과이암층이혼재된지층 ) 을구분하는것으로추정된다. 1,800 m 이상심부에서동쪽으로경사진이벤트또한매우뚜렷하게나타나지만지층경계면으로단정하기에는정보가매우부족하다. Fig (a) Final stack section and (b) its stratigraphic interpretation (the section is displayed only up to 2,500 m in depth axis because the stacked traces at later times are severely contaminated by noise and ground-roll). 44

122 제 3 장지구물리탐사및자료해석 탄성파굴절법해석결과 반사법자료처리에서속도분석이거의불가능하였기때문에천부지층의속도분포를얻기위해굴절법해석을수행하였다. 측선왼쪽 ( 서쪽 ) 에서오른쪽 ( 동쪽 ) 으로진행하는선두파 (head wave) 는빠른속도를보이고반대방향의선두파는느린속도를보이므로서쪽으로갈수록깊어지는경사층구조를가정할수있다 (Fig. 3-8). 지표탄성파탐사자료로부터초동주시를발췌하여 2층경사모델을구축하였다. (a) (b) Fig Shot gather seismograms corresponding to (a) forward and (b) reverse traverse for refraction analysis. Fig. 3-9 는해석결과로얻은속도모델을보여준다. 1 층의속도는 1,669±5 m/s ( 신뢰수준 95%) 로미고 결이암층의속도와비슷한것을확인할수있다. 2 층의속도는 4,009±8 m/s ( 신뢰수준 95%) 로결정질응회 암이라는추정을뒷받침해주는것으로보인다. Fig Two-layered dipping interface model obtained by interpreting first-arrival traveltime curves. 45

123 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 속도모델구축 속도모델을구축하기위해지표탄성파탐사자료와시추공물리검층자료, 암상분석자료등을종합적으로사용하였다. 속도모델구축과정은다음과같다. 1 암상분석자료를바탕으로층서모델구축 2 음파검층자료를이용하여 P파속도결정 3 굴절법해석결과를이용한천부 (1, 2층 ) 속도수정 4 물리검층자료를이용한심부지층 S파속도결정 5 Check shot 자료를이용한중간지층 (3, 4층 ) P파속도보정 암상분석자료와음파검층자료에의한 P파속도모델구축시추공 BH-4 와 PX-2 에서의지질주상도가유사하다는가정하에암상과 P파속도의관계를이용하여초기속도모델을구축하였다 (Table 3-2). 5개의층으로구분하여음파검층으로부터구한속도의대푯값을할당하였다. Table 3-2. Simplified lithology model and corresponding P-wave velocity model at PX-2. Layer # Depth (m) Lithology v p (m/s) 1 0~203 Mudstone ~670 Crystal Tuff & Tuff Breccia 3, ~1,185 Mudstone, Sandstone & Andesitic Tuff 4, ,185~2,450 Andesite & Tuff Breccia 4, ,450~ Granodiorite, Diorite & Granite - PX-2 음파검층자료와굴절법해석자료의병합 Table 3-2에서구축한속도모델에 PX-2 의음파검층자료를추가하고, 굴절법해석으로부터얻은속도모델을병합한결과는 Fig 과같다. BH-4 에서측정한음파검층자료는 1,680 m까지만존재하므로 1,680 m 이상의심도에는 PX-2 에서측정한음파검층자료를적용하였다. 46

124 제 3 장지구물리탐사및자료해석 Fig Modified P-wave velocity model using the sonic log-based model and the two-layered refraction velocity model. S파속도모델구축 S파속도에대한정보는거의찾을수없고 PX-2 에서의물리검층에서유일하게 S파속도를측정하였다. 3,400~4,000 m 구간에서측정한음파검층자료로부터 S파속도의중앙값을구하면 3,310 m이고이때포아송비는약 이다. 이값을포함하여구축한속도모델은 Table 3-3과같다. 속도는반올림하여유효숫자 3자리로나타내었다. Table 3-3. Final velocity model (Model I) based on well-logging data and refraction velocity model. Layer # Depth (m) Lithology v p (m/s) v s (m/s) Poisson s ratio Remarks 1 0~203 Semi-consolidated Mudstone 1, Refraction 2 203~670 Crystal Tuff & Tuff Breccia 4, Refraction 3 670~1,185 Mudstone, Sandstone & Andesitic Tuff 4, BH-4 4 1,185~1,680 Andesite & Tuff Breccia 4, BH-4 5 1,680~2,450 Andesite & Tuff Breccia ,450~3,400 Granodiorite, Diorite & Granite 5, PX-2 7 3,400~4,344 Granodiorite, Diorite & Granite 5,860 3, PX-2 47

125 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Check shot을이용한속도모델보정미소진동모니터링에사용된시추공배열식지진계자료중지표인공송신원에의한신호 (check shot data) 를이용하여속도모델을보정하였다. Check shot자료를분석하여 P파에대한시간차 (moveout) 를이용한정보추출이가능할것으로판단하였다. 속도모델의 1, 2층은굴절법해석으로결정한값이므로고정하였고, 시추공센서가가장많은영향을받는 3, 4층속도를결정하고자하였다. 시추공센서가 1360~1520 m 구간에설치되었기때문에지표에서발생한이벤트는 5층이상의속도에는영향을받지않는다고가정하였다. Check shot 자료의 P파시간차정보로속도를결정하는문제는매우편향된부족결정 (under-determined) 문제이므로 3층과 4층의속도가동일하다는가정하에전역탐색을수행하였다. 역산결과최적속도는 5,050±3 m/s 로결정되었다. 기존속도모델의 3층과 4층속도가각각 4,450, 4,780 m/s인데비해약간크게결정되었다. 이결과를반영하여보정한최종속도모델은 Table 3-4와같다. Table 3-4. Modified velocity model (Model II) using check shot data. Layer # Depth (m) Lithology v p (m/s) v s (m/s) Poisson s ratio Remarks 1 0~203 Semi-consolidated Mudstone 1, Refraction 2 203~670 Crystal Tuff & Tuff Breccia 4, Refraction 3 670~1,185 Mudstone, Sandstone & Andesitic Tuff 5, Check shot 4 1,185~1,680 Andesite & Tuff Breccia 5, Check shot 5 1,680~2,450 Andesite & Tuff Breccia ,450~3,400 Granodiorite, Diorite & Granite 5, PX-2 7 3,400~4,344 Granodiorite, Diorite & Granite 5,860 3, PX 원격탐사기반단층모델링 SAR 위성자료획득 2017 년포항지진에의한지표변위관측을위해여러파장대역의위성자료로부터 SAR 자료를수집하였다. 지진에의한지표변위의응력 (stress) 방향에대한정보를제공하는단층이동모델링의정확도를높이기위해서는다수의 SAR 자료로부터관측된변위를이용하는것이효과적이다 (Jo et al., 2017). 따라서지진으로인한지형변위분석을위해레이더간섭기법을적용할수있는국내외인공위성원격탐사자료를수집하였다 (Table 3-5). 관측위성의종류는일본의 ALOS-2 PALSAR-2 (L-band), 유럽의 Sentinel-1A/B TOPSAR 48

126 제 3 장지구물리탐사및자료해석 (C-band), 이탈리아의 Cosmo-SkyMed (X-band) 및한국의 KOMPSAT-5 (X-band) 이며수집된자료의 획득일은 2015 년부터 2018 년까지이다. Table 3-5. List of SAR data collected for research. Satellite Date Direction Mode Path/ Frame Incidence Angle Polarization Resolution/ Swath Wavelength Descending Stripmap (SM1) 24/ HH 3x3 m/ 50 km ALOS-2 PALSAR Ascending Stripmap (SM3) ScanSAR (WD1) 131/ / HH/HV 4x6 m/ 50 km 4x20 m/ 350 km L-band (24 cm) Ascending VV/ VH Sentinel-1A/B TOPSAR (IW) 5x20 m/ 250 km C-band (5.6 cm) Descending VV/ VH Cosmo- SkyMed Descending HImage HH 3 m/ 40 km X-band (3.1 cm)

127 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 SAR 위성자료를이용한변위관측 수집된 SAR 자료를이용하여지표변위를관측하기위하여레이더간섭기법을적용하였다. 정밀지표변위관측을위해다중개구간섭기법 (Multiple Aperture Interferometry, MAI) 을이용한이온효과보정등의기술을적용하였다 (Jung et al., 2011). Fig 은 Sentinel-1A/B, Cosmo-SkyMed, ALOS-2 PALSAR-2 를이용하여관측한지표변위도이다. Fig Descending InSAR measurements from (a) Sentinel-1A/B, (b) and (c) Cosmo-SkyMed, (d) ALOS-2 PALSAR-2, and (e) ascending InSAR measurement from ALOS-2: LOS measurements (a, b, d, e) and azimuth measurement (c). 50

128 제 3 장지구물리탐사및자료해석 Fig. 3-11c를제외한나머지지표변위도는위성센서가바라보는방향 (Line-of-sight, LOS) 의변위만을나타내고있다. 3차원지표변위를효과적으로추출하기위해서는위성의비행방향 (azimuth) 에대한변위도가필요하므로, 이탈리아항공우주국에요청을통하여 Cosmo-SkyMed 의원시자료를구매수집하였다. MAI 자료처리과정 (Bechor and Zebker, 2006) 을거쳐위성비행방향으로의변위도를추가적으로생성하였다 (Fig. 3-11c). 또한 ascending 궤도에서변위관측수행을위해일본항공우주국에영상을요청하여 ALOS-2 PALSAR-2 자료를추가로수집하고변위관측을수행하였다 (Fig. 3-11e: Lindsey et al., 2015). 3차원지표변위분석을위해 Table 3-6의위성관측값을입력자료로선택하였다. 비행방향에대한지표변위도와관측방향을고려하여 Cosmo-SkyMed 와 ALOS-2 위성의지표변위도가성분분해에사용되었다. Fig 는 3차원지표변위성분분해를통해얻어진결과로, 수평변위성분의크기및방향은벡터로표현되어있으며, 컬러맵은수직변위성분의크기를나타낸다. Table 3-6. Input sources for 3D decomposition of surface displacement. No. Satellite Component Acquisition mode Heading angle( ) Incidence angle( ) 1 Cosmo-SkyMed LOS Descending ALOS-2 LOS Ascending Cosmo-SkyMed MAI Descending Fig (a) Horizontal displacement vector field (the vectors indicate the magnitude and directions of the horizontal displacements, and the colored map represents the vertical displacements), (b) displacement vector field of the box A. 51

129 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 SAR 위성관측변위기반단층모델링 단층에의한지표변위를수리적으로나타내기위해서많은경우에 Okada dislocation model 을활용하였다 (Okada, 1985). Okada 모델은단층의기하정보와 slip양으로부터균일 / 등방성 / 반무한탄성매질의지표에서의변위를계산해주는모델이다. 단층면의 strike/dip 은지진파분석자료로부터도출된 214 /43 으로고정하고, 단층의기하정보 ( 길이, 폭, 깊이, 위치 (X,Y)) 와 slip 양 (strike-slip, dip-slip) 으로구성된 7개변수를추정하는역산을수행하였다. 단층면모델링수행본연구에서는우선적으로하향궤도에서촬영된 ALOS-2, Sentinel-1, CSK 자료의 InSAR 처리결과를바탕으로단층기하모델링 (inverse modeling) 을수행하였다. 한편 InSAR 관측치는 LOS 방향의변위만을파악할수있기때문에단층의실제변화를판단하기에는한계가존재한다. 따라서 3차원지표변위관측치를활용하여추가적인단층기하모델링을수행하였다. 최종모델파라미터는 1,000 회의 Monte-Carlo 시뮬레이션을통하여얻어진각파라미터의히스토그램에서나타나는중앙값으로결정하였다. Fig 은 Monte-Carlo 시뮬레이션을통하여획득된모델파라미터의히스토그램분포를나타낸다. Table 3-7은하향궤도에서관측된 3개의 InSAR 관측치와 3차원관측치를활용한최적모델파라미터를나타낸다. Fig Histograms of the Okada dislocation model parameters obtained from the InSAR measurements using the Monte-Carlo simulation. Table 3-7. The best-fit model parameters and standard deviation of the 2017 Pohang earthquake estimated by the Okada dislocation model using the descending InSAR and 3D measurements. Model parameters Input Length (Km) Width (Km) Depth (Km) Strike (Deg.) Dip (Deg.) X (Km) Y (Km) Strike-Slip (mm) Dip-Slip (mm) InSAR (LOS) 6.76± ± ± ± ± ± ± D 4.79± ± ± ± ± ± ±

130 제 3 장지구물리탐사및자료해석 단층면모델 3차원가시화및해석 Sentinel-1, CSK, ALOS-2의 descending InSAR 관측치를활용하여구한단층면 (Table 3-7) 을 COSMO-SkyMed 영상으로부터구한변위도와시추공의위치벡터자료 PX-1, PX-2 를이용하여 3차원도시하였다 (Fig. 3-14). 추정된단층면은 PX-2 의최하부지점아래에위치하고있다. 단층면과 PX-2 끝점간의수직거리는 m, m 이다. Fig Fault plane models estimated from three InSAR and 3D measurements along with the EGS wells, PX-1 and PX-2 (Upper image: COSMO-SkyMed LOS displacement). 단층면모델링결과를이용하여 seismic moment (Mo) 값을구하고, 다시 Mo 를지진규모 (M W ) 로변환하 기위해 USGS 에서사용하는다음식을사용하였다 ( magnitude-types.php). Mo = (Area) (Net slip) (shear modulus) ( 단위 : Newton-meter) M W = 2/3 * (log10(mo) ) Table 3-7 에제시되어있는모델값을이용하여계산된 M W 는 5.29 (LOS) 와 5.23 (3D) 으로지진파관측 을통해산정된값과유사한값을보인다. 따라서모델링결과로제시된단층면의크기와이동변위가적절하 게추정되었음을확인할수있다. 53

131 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 본진전후규모 3.0 이상지진에의한지표변위관측 2017 년 4월 15일발생한규모 3.2의지진과 2018 년 2월 10일발생한규모 4.6에의한지표변위발생여부를확인하기위해 Sentinel-1 과 COSMO-SkyMED SAR 영상을수집분석하였다 년 4월 15일전후에획득된세개의 CSK 영상을이용한분석결과, 대부분긴밀도가낮아분석에한계가있지만대기영향에의한오차가매우작은 2017/04/ /05/20 간섭쌍으로부터진앙주위에서뚜렷한지표변위가발생하지않은것으로판단된다. Sentinel-1 위성은 C-band 이며간섭쌍의짧은시간간격으로인해 CSK 영상을이용한결과보다정확한변위도생성이가능하였다 년 4월 15일전후로획득된 Sentinel-1 위성자료중가장긴밀도가좋은 2017/04/ /04/26 간섭쌍으로부터구한지표변위도는 Fig 와같다. 연구지역전체에서약 ±0.5 cm에해당하는신호가관측되지만, 이는연구지역주변부신호들로볼때대기에의한성분으로판단되며, 진앙주변부에서의미있는변위는발생하지하지않은것으로보인다 년 2월 10일규모 4.6의지진변위관측을위해 2018 년 2월 8일과 2018 년 2월 20일획득된 Sentinel-1 영상을이용하여변위도를생성하였다 (Fig. 3-15b). 대기오차로판단되는 ±0.5 cm 정도의신호가전반적으로관찰되며, 한동대학교북동쪽지역좁은구간에서 1.5 cm의위성에서멀어지는방향변위가관측된다. 이지점은진앙과 7 km 이상떨어져있으며, 기존의토양액상화현상과관련된지역으로국지적인지표변위로해석된다. Fig Sentinel-1 displacement map with the epicenter: (a) 2017/04/ /04/26 pair and (b) 2018/02/ /02/20 pair. 54

132 제 3 장지구물리탐사및자료해석 Sentinel-1 PSInSAR 시계열분석을통한미세지표변위관측 2017 년 11월 15일지진발생전까지연구지역에서발생하는미세지표변위를정밀관측하기위해 Sentinel- 1A/1B 위성으로부터 2016 년 1월이후로촬영된자료를모두수집하였다. Ascending 및 descending 궤도로부터 2016 년 1월 20일이후 34개영상, 2016 년 10월 4일이후 31개의영상이수집되었다. 시계열분석을위해고정산란체를이용한 PSInSAR 기법을적용하였으며, 지진발생지역과포항시를포함한지역의자료처리를수행하였다. SAR 영상위에색상을가지고있는관측점은최종분석후시계열긴밀도가 0.7 이상의높은신뢰도를갖는지점만을선택하여출력한것으로, ascending 자료에서 550,501 개 descending 자료에서 710,408 개의 PS (Persistent Scatterer) 가추출되었다. Fig. 3-16a 는 ascending 자료분석결과로, 청록색은변위속도가 0 이고녹색과적색은위성으로부터멀어지는방향으로지반침하로해석될수있으며각각 1, 2.5 cm/ yr를나타낸다. 청색에서보라색은위성으로가까워지는방향의변위로지표상승에해당한다. 포항시일부지역에서약 2~3 cm/yr 정도의지반침하에해당하는지표변위가국지적으로관측되지만, 이변위는지열발전실증부지와멀리떨어져있어다른원인에의한지반침하현상으로해석된다. Fig (a) Mean velocity map estimated from Sentinel-1 ascending mode data, (b) the enlarged image around the epicenter of three major earthquakes, and (c) close view around the EGS site. 55

133 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 주요지진발생지역을중심으로확대한결과는 Fig. 3-16b, 그리고지열발전실증사이트주변부를확대한결과는 Fig. 3-16c 에제시되어있다. 지열발전실증사이트주변관측지점에서평균지표변위속도는대부분 -2~1 mm/yr 이내의값으로, 뚜렷한변위를보이는지표면은없는것으로해석된다. 각관측지점에서잔여오차는 1~5 mm 정도의표준편차를보인다. 지열발전실증사이트주변에서선정된 4개의관측지점 (Fig. 3-16c 의 P1~P4) 에서의시계열변위분석결과는 Fig 과같다. P1, P2, P3, P4 지점에서의변위속도는 0.1, -1.7, -2.6, 1.2 mm/yr 로거의 0의값을보이고있으며, 선형변위를기준으로관측변위의표준편차는 1.8~3.0 mm에불과해, 매우높은관측정밀도를제공함을확인할수있다. Fig Time series histories of LOS displacement at the selected points of P1, P2, P3 and P4 around the EGS site 결론 Sentinel-1 (C-band, 5.6 cm) 위성및 ALOS-2 (L-band, 23 cm) 위성의 ascending/descending 궤도와 COSMO-SkyMed (X-band, 3.1 cm) 위성 descending 궤도로부터포항지진전 후의 SAR 위성자료를수집하고, 지표변위를관측하였다. 오카다모델과몬테카를로시뮬레이션을통해단층면과이동벡터모델링을수행하였다. 추정된단층면은 PX-2 시추공의최하부지점과수직거리약 150~250 m의범위에위치하고있으며, 모델링결과로부터구한 moment 는 M W 5.4로지진파관측을통해산정된 moment 와유사한값을보인다. Sentinel-1 과 COSMO-SkyMED SAR 영상을이용하여 2017 년 4월 15일발생한규모 3.2 지진과

134 제 3 장지구물리탐사및자료해석 년 2월 10일발생한규모 4.6 지진에의한지표변위발생여부를확인한결과, 진앙주변지역에서뚜렷한변위는발생하지하지않은것으로해석된다. 또한, 2017 년 11월 15일지진발생전까지지열발전실증사이트주변지역에서발생하는미세지표변위를정밀관측하기위해 Sentinel-1A/1B 위성으로부터 2016 년 1월이후로촬영된자료 65개영상을수집하고, 고정산란체를이용한 PSInSAR 기법을적용하여미세지표변위를정밀관측하였다. 지열발전실증사이트주변관측지점에서평균지표변위속도는대부분 2~1 mm/yr 이내의값으로, 뚜렷한변위를보이는지표면은없는것으로해석된다 지구물리탐사해석결과 본장에서는포항지진구역의지하구조영상화를위해 MT 및 AMT 자료를활용한전기비저항분포의파악, 탄성파탐사자료분석을통한지질구조및속도구조분석, 원격탐사자료를활용한단층모델링등을수행하였다. MT 및 AMT 통합자료분석의경우확인된낮은비저항대가측선의가장자리에위치하거나자료결손지역을통과하여해석의정확도에많은영향을미치는것으로나타났다. 또한 MT자료의해상도및가탐심도의문제로제시된전기비저항영상으로부터정확한단층대의위치를파악하는것은무리가있는것으로판단된다. 이러한한계에도불구하고지열발전실증사이트동남쪽의이상대는 2차원및 3차원역산에서도확인되며주향및경사도잘부합된다. 단층대또는파쇄대로해석이가능한이전기비저항이상대는 PX-2 시추공과 3 km 이상의깊이에서교차하는것으로추정된다. 기존의단순중합단면만존재하는반사법탐사자료를다양한방법으로재처리한결과천부의구조는일부확인할수있었으나잡음으로인해심도 1 km 이하의구조는파악할수없었다. 반사법자료의한계를극복하기위해굴절법자료및검층자료를종합적으로분석하여속도구조를예측하였으며이결과는독립적으로추정한지진연구팀의결과와잘부합하였으며기존의연구결과보다는다소높은값으로추정되었다. 원격탐사자료는 Sentinel-1 위성및 ALOS-2 위성의 ascending/descending 궤도와 COSMO-SkyMed 위성 descending 궤도로부터포항지진전 후의 SAR 위성자료를수집하고, 지표변위를관측하였다. 오카다모델과몬테카를로시뮬레이션을통해단층면과이동벡터모델링을수행하였다. LOS 자료를이용하여추정된단층면은 PX-2 시추공과약지하 3980 m 부근에서만나며, 모델링결과로부터구한 moment 는 M W 5.29/5.23 로지진파관측을통해산정된 moment 와유사한값을보인다. Sentinel-1 과 COSMO-SkyMED SAR 영상을이용하여 2017 년 4월 15일발생한규모 3.2 지진과 2018 년 2월 10일발생한규모 4.6 지진에의한지표변위발생여부를확인한결과, 진앙주변지역에서뚜렷한변위는발생하지하지않은것으로해석된다. 57

135 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 또한, 2017 년 11 월 15 일지진발생전까지지열발전실증사이트주변지역에서발생하는미세지표변위를 정밀관측한결과, 지열발전실증사이트주변관측지점에서평균지표변위속도는대부분 -2~1 mm/yr 이내 의값으로, 뚜렷한변위를보이는지표면은없는것으로해석된다 참고문헌 송윤호, 김형찬, 심병완, 이창범, 박덕원, 이성곤, 이종철, 이병태, 박인화, 이태종, 이철우, 문상호, 김연기, 이병대, 임현철, 2004, 지열자원부존특성규명및활용기반기술연구. 한국지질자원연구원연구보고서, KR-04( 연차 )-08, 국무총리, 123. 이태종, 송윤호, Uchida, T., 2005, 심부지열자원개발을위한원거리기준점 MT 탐사자료의 2차원역산해석. 한국지구물리 물리탐사학회, 8(2), Bechor, N.B.D. and Zebker, H.A., 2006, Measuring two-dimensional movements using a single InSAR pair. Geophysical Research Letter, 33(16): L Jo, M.J., Jung, H.S., and Yun, S.H., 2017, Retrieving precise three-dimensional deformation on the 2014 M6.0 South Napa Earthquake by joint inversion of multi-sensor SAR. Scientific Reports, 7(1), Jung, H.S., Lu, Z., Won, J.S., Poland, M.P., Miklius, A, 2011, Mapping three-dimensional surface deformation by combining multiple aperture interferometry and conventional interferometry: Application to the June 2007 eruption of Kilauea volcano, Hawaii. IEEE Geoscience and Remote Sensing Letters, 8(1), Kim, K.H., Ree, J.H., Kim, Y., Kim, S., Kang, S.Y., and Seo, W., 2018, Assessing whether the 2017 M W 5.4 Pohang earthquake in South Korea was an induced event. Science, 360(6392), Lee T.J., Song, Y., and Uchida, T., 2007, Three-dimensional magnetotelluric surveys for geothermal development in Pohang, Korea. Exploration Geophysics, 38, Lindsey, E.O., Natsuaki, R., Xu, X., Shimada, M., Hashimoto, M., Melgar, D., and Sandwell, D.T., 2015, Line-of-sight displacement from ALOS-2 interferometry: M W 7.8 Gorkha Earthquake and M W 7.3 aftershock. Geophysical Research Letters, 42(16), Mackie, R.L., Smith, J.T., and Madden, T.R., 1994, Three-dimensional electromagnetic modeling using finite difference equations: The magnetotelluric example. Radio Science, 29, Okada, Y., 1985, Surface deformation due to shear and tensile faults in a half-space. Bulletin of the Seismological Society of America, 75, Sasaki, Y., 2004, Three-dimensional inversion of static-shifted magnetotelluric data. Earth, Planets and Space, 56, Uchida, T., 1993, Smooth 2-D inversion for magnetotelluric data based on statistical criterion ABIC. Journal of Geomagnetism and Geoelectricity, 45,

136 제 3 장지구물리탐사및자료해석 제 4 장 응력상태분석 Stress State Analysis 59

137 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project 60

138 제 4 장응력상태분석 요약 / Abstract 본연구에서는지진포컬메커니즘응력역산및시추공응력지시자를포함한다양한기법을기반으로포항지열발전실증사이트현장의응력상태를규명하였다. 최대수평주응력의방향은깊이에따라변화하여얕은깊이 (700~800 m) 에서는 SE-NW, 4 km 하부심도에서는 ENE-WSW 으로나타났다. 응력체계또한관측규모에따라변화하여지역적규모에서는주향이동단층운동에유리한응력체계를보이고, 보다좁은시추공규모에서는역단층운동에유리한응력체계를보인다. 모든자료의분석결과연직응력과최소수평주응력의크기가유사한것으로나타났다. 최대수평주응력의크기는지각내에존재하는다양한방향의단층들중미끄러지기가장쉬운방향성을갖는단층들의전단성향이 0.6일경우를가정하여산정하였다. 이러한응력장하에서 2017년 11월 15일포항지진을발생시킨단층의전단성향은 0.55~0.57로산정되었으며이값은최적방향을갖는단층의전단성향과비교할때 92~95% 수준의높은값이다. 포항지진단층의전단성향은 PX-2의 4.2 km 심도에서회수된암석코어시편내자연균열의마찰계수 (0.53), 3.6 km 심도에서회수된시추암편마찰계수 ( ) 의하한치와유사하거나약간큰값으로, 이값이지진발생깊이에서의단층마찰특성을대표하는경우, 포항지진단층은자연상태에서임계치의응력상태이거나이에아주근접한상태에있었음을시사한다. We estimate the stress state at the Pohang EGS site based on various techniques including earthquake focal mechanism inversions and borehole stress indicators. The orientation of the maximum horizontal principal stress (S Hmax) varies with depth: SE-NW at shallow depths (700~800 m) and ENE-WSW at greater depths (below 4 km). The stress regime also varies depending on the scale of observation: strike-slip faulting regime at regional scale observation, and reverse faulting stress regime at a narrower borehole scale. All analyses including focal mechanism stress inversion and borehole tests show that the magnitudes of S v and S hmin are similar. Assuming that S Hmax magnitude is constrained by friction of the most optimally oriented fault for slip, we estimate the slip tendency (shear stress normalized by effective normal stress on the fault plane) of the fault that caused the 15 Nov 2017 Pohang earthquake. The slip tendency of the fault is determined to be 0.55~0.57, which is 92~95% of that (0.6) of the most optimally oriented fault. This value is analogous to, and slightly greater than the laboratory measured friction (0.53) of the fracture in rock cores recovered from 4.2 km and the friction lower bound value of basement rock cutting ( ) from 3.6 km in PX-2. If these values represent characteristics of the fault friction at the depth of the earthquake emergence, the seismogenic fault may be critically or near-critically stressed in its natural state 연구배경 단층주변의지각응력에대한정보는지진발생원인인단층운동가능성을파악하기위한필수적인요소 이다. 포항지진을발생시킨단층에대한응력장 ( 방향과크기 ) 규명을통해포항지열발전실증연구에서의물주 61

139 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 입과 2017 년 11월 15일 M W 5.5 지진발생의상관성을이해하기위해본연구가수행되었다. 지진은단층의전단운동에의해발생하며그원리는쿨롱마찰법칙 (Coulomb friction law), cr P p 로설명할수있다. 여기서 은단층의임계전단강도, σ는단층면에작용하는수직응력, P p 는공극압, μ는단층의마찰계수이다. 실제단층면에작용하는전단응력 τ가 이되면단층은전단운동을하고지진을발생시킨다. 본연구는포항지열발전실증사이트하부지각응력장의방향과크기를다양한자료로부터규명하여단층에실제작용하는 τ와 σ를규명함으로써포항지진을발생시킨단층의전단운동가능성을분석하는데목적이있다. 지열발전실증부지인근포항지역의응력장은다양한기법으로규명되었으며이들자료를수집또는분석하여심도에따른응력장의방향과크기를분석하였다. 다양한자료로부터도출되는응력정보는응력이가해지는지하암체의스케일과그응력값이대표하는지각심도가다르므로결과를종합하여포항지진발생지역의응력해석에이용하였다 포항주변응력장방향 기존의많은한반도현생응력장연구로부터한반도는 ENE-WSW 의최대압축응력을받고있는것으로알려져있으며이는다양한기법으로측정된응력자료로부터알수있다 (Jun, 1991; Chang et al., 2010; Hong and Choi, 2012; Soh et al., 2018). Fig. 4-1의응력방향자료들은단일지진포컬메커니즘의 P, T축을이용하거나시추공응력측정시험자료를이용하여구한자료로대부분세계응력지도 (World Stress Map) 기준 C 또는 D등급에해당하는자료로서응력의방향에 ±20~40 의불확실성을갖고있다 (Heidbach et al., 2016). Fig Maximum horizontal principal stress directions in Korea (data provided by World Stress Map 2016): (a) B-C quality stress data over the country, (b) C-D quality stress data around Pohang. 62

140 제 4 장응력상태분석 포항지열발전실증부지주변의광역응력장에대한좀더신뢰할수있는응력방향을구하기위해포항지진진앙 (36.12, ) 을중심으로반경 70 km 내의여러지진포컬메커니즘자료를역산하여응력을도출하였다. 응력역산을위해수집된포컬메커니즘자료는 1997 년부터 2016 년사이에발생한규모 2.5이상의 21개지진자료이며평균심도 12.8±5.1 km의이벤트들이다 (Fig. 4-2a). 응력역산결과포항주변의광역응력장은최대주응력 (S 1 ) 방향이 N74 E 의수평방향으로나타났으며중간주응력 (S 2 ) 이연직에서 22 기울어진방향, 그리고최소주응력 (S 3 ) 이 N345 E 방위각에서수평으로부터 21 기울어진방향으로나타났다 (Fig. 4-2b). 응력방향에대한 95% 신뢰구간에서보듯이 S 1 의방향은상당히잘제한되어도출된반면 S 2 와 S 3 방향분포가 S 1 에직교하는면상으로길쭉하게분포하여 S 2 와 S 3 간의명확한구분이안되는양상을보인다. 이는 S 2 와 S 3 의응력크기가유사하기때문인것으로판단된다. 실제응력역산과정에서구해지는 R(=(S 1 -S 2 )/(S 1 -S 3 )) 값은평균 0.88의높은수치를보여 R의정의대로 S 2 와 S 3 의값이크게차이가나지않음을시사한다 (Fig. 4-2c). 이결과는응력역산에이용된지진자료심도에해당하는 7.6~17.9 km 심도에서의최대압축응력방향 (N74 E) 이한반도의전체적인압축응력방향인 ENE-WSW 방향과일치함을보여준다. Fig (a) Earthquake focal mechanism solutions for M>2.5 earthquakes that occurred around Pohang from 1997 and 2016, and stress inversion results showing (b) the orientations of the three principal stresses (S 1, S 2, S 3) and (c) R value. 포항지열발전실증연구시설주변천부심도 (<1 km) 의응력장은실증부지남서쪽약 4 km에위치한한국지질자원연구원포항분원에위치한심도 1 km의시추공 (EXP-1) 에서수압파쇄응력측정과시추공벽응력지시자 (borehole breakouts, drilling-induced tensile fractures) 분석을통해측정되었으며그결과는 Kim et al. (2017) 에보고되었다. 약 650~810 m 심도에걸쳐최대수평주응력 (S Hmax ) 방향은 NW-SE (N E) 로규명되었다. 응력의크기는 S Hmax 가연직응력 (S v ) 의 1.1~1.4 배로가장크고최소수평주응력 (S hmin ) 이연직응력의 0.8배로가장작은크기의주응력으로나타나응력체계가 S Hmax > S v > S hmin 인주향이동단층운동에유리한체계를보인다. 63

141 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진의진원심도에해당하는 3~7 km 심도에서의응력은포항지진이후의여진자료에대한포컬메커니즘응력역산을통해구할수있다. 포항지진이후발생한총 202개의여진자료로부터각각의포컬메커니즘을규명하였으며그결과 41%(83 개 ) 는역단층성, 24%(48 개 ) 는주향이동단층성, 4%(9 개 ) 는정단층성, 그리고나머지 31%(62 개 ) 는그이외의유형으로분류되었다 (Fig. 4-3). 이들 202개여진의포컬메커니즘자료를모두이용한응력역산결과최대압축응력방향은 N86 E 를보이며 S Hmax > S v > S hmin 순의상대적응력크기를보인다. 응력역산과정에서유도된응력매개변수 R값은평균 0.87의값을보이며이는 R값의정의에의하여 S hmin 의크기가 S v 의크기에접근되어있음을시사한다. 그러나이응력결과는포항지진발생이후나타날수있는지진에의한응력교란의효과를포함할수있어지진발생이전의응력장을지시할수있는지에대해서는신중해야할필요가있다. Fig Earthquake focal mechanism stress inversion result from aftershocks. 포항지열발전실증사이트주입심도부근에서의보다직접적인응력방향은시추공벽에서관찰되는응력지시자 (borehole breakouts, drilling-induced tensile fractures) 로부터파악할수있으나이를파악하고자시도한시추공나공부분에대한영상검층으로부터자료를획득할수없었기때문에직접적인응력지시자의관찰이불가하였다. 대신포항지진발생직전 PX-2 지열정시추과정에서수행된시추공쌍극자음파속도비등방성 (borehole dipole sonic anisotropy) 자료로부터응력방향에대한유추를시도하였다. 시추공쌍극자음파속도자료는시추공벽을따라 S파전파속도를측정한자료인데 S파의특성상진동방향이시추공벽에수직한방향과평행한방향으로진동하는두가지 S파성분의속도차이를측정한자료이다. 시추공에서시추공벽압축파쇄대등의현상이발생하면시추공벽암반에많은균열들이생성되는데이러한균열들은대체로시추공벽에평행한방향성을갖는다. 이때문에시추공벽에직교한방향으로진동하면서전파되는 S파성분이시추공벽에평행한방향으로진동하면서전파되는 S파성분보다속도가느려지는경향이있으며이를 S파비등방성 (shear wave anisotropy) 이라한다. 이러한 S파비등방성을시추공에서검층한자료가시추공쌍극자음파속도자료이며석유업계에서공내영상검층자료가없을경우응력방향을유추하는데흔히사용되고있다 (Brie et al., 1998). 64

142 제 4 장응력상태분석 Fig. 4-4는 PX-2 시추공 3,400~4,350 m 구간에서마지막케이싱설치직전검층한시추공쌍극자음파속도비등방성자료로서공벽방향에따른 S파속도의비등방성을보여준다. 비등방성은방향에따라 0~5% 의범위로변화하는데비등방성이높은부분이시추공벽에서 180 벌어진두방향으로관찰되었다. 또한비등방성이최대가되는중심위치가 3,400~4,100 m 심도에서상당히일관된방향을유지하는것으로관찰되어시추공응력지시자인시추공벽압축파쇄대 (borehole breakouts) 발생양상과매우유사하다. 시추공벽압축파쇄대는비등방성이높은위치, 즉 S hmin 방향에서발생하며, 따라서비등방성이낮은중심위치가 S Hmax 방향을지시하게된다. 비등방성이낮은방향을통계내어보면 3,400~4,100 m 구간에서평균 N75 E 로나타났으며 3,400~4,350 m 전구간에서의평균방향은 N77 E 로나타났다 (Fig. 4-4b, c). 이방향은지진포컬메커니즘응력역산을통해구한포항지역광역최대주응력방향 (N74 E) 과상당히유사한방향으로포항지진발생심도주변에서의응력의방향도광역응력장의방향과유사하다는점을보인다. Fig (a) Dipole sonic shear wave anisotropy in PX-2, indicating the maximum horizontal principal stress azimuth (b, c). 이상의결과들을정리하면 4 km 이하심도에서의최대주응력의방향은 N74-86 E 의방향을보이며한 반도에서우세한 ENE-WSW 의최대압축응력의방향과일관된결과를보인다 (Fig. 4-5). 반면 1 km 이내의 천부심도에서만최대주응력의방향이 NW-SE 로회전되어있는경향을보인다. 65

143 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig The azimuth of the maximum horizontal principal stress (S Hmax) as a function of depth estimated from various data around Pohang 포항지열발전실증사이트현장응력장크기 연직응력 지하심부의지각응력을표현하는세주응력 (S v, S hmin, S Hmax ) 중연직응력 (S v ) 은상재하중 (overburden) 에 의해가해진다는가정하에해당심도상부의암석중량또는밀도를이용하여독립적으로산정할수있으며 S v z z gd 로구한다. 여기서 z 는심도, ρ 는암석밀도, g 는중력가속도이다. 가능한정밀한 S v 산정을위해포항지역심 도에따른암석밀도자료를수집하였다. Fig. 4-6 에보인암석밀도자료는지열사이트부근다른시추공 (BH-4, Yoon et al., 2011) 과지열 PX-2 공의코어시편 (Kwon et al., 2018) 을이용해직접측정한자료이다. 66

144 제 4 장응력상태분석 2,000 m 상부의암석밀도는여러암종의교호에의해분산이심하지만전반적으로심도에따라증가하는추세를보인다. PX-2 의 4,219 m에서회수한코어시편을이용해측정한밀도는 11회의시험을통해측정한가중치가높은결과이다. S v 계산을위해 1,500 m 상부는주어진밀도자료에대한선형회귀를통해, 1,500 m 하부는 4,219 m 화강섬록암의밀도로대표된다는가정하에연직응력을산정하였다. PX-2 지열정나공구간최상부인심도 4.2 km에서의연직응력은 106 MPa로계산되었다. Fig Rock density measurements using cores extracted nearby borehole (BH-4) and PX 지진포컬메커니즘역산자료로부터의응력크기정보 응력의크기는지하로갈수록증가하는경향이있어이를효과적으로표현하기위해각응력성분을연직응력 (S v ) 으로정규화한형태로표현한다. 시추공에서수행되는수압파쇄시험이나 borehole breakout, drillinginduced tensile fractures 등은응력의절대적크기에대한정보를제공하지만, 지진포컬메커니즘응력역산방법은응력의상대적크기에대한정보만을제공할수있다. 포항지열발전실증사이트인근의 EXP-1 시추공에서측정된응력의절대적크기에대한정보는 Kim et al. (2017) 이보고하였으며 659~810 m의심도에서 S hmin 의경우 S v 의 0.8배, S Hmax 의경우 S v 의 1.1~1.4 배로규명되었다. 67

145 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 지진포컬메커니즘을통한응력역산을통해응력의방향과세주응력의상대적크기에대한정보를얻을수있으며특히역산과정에서도출되는 R값은응력의절대크기는제공하지는못하지만응력크기의특성에대한정보를제공한다. R값은 R = (S 1 S 2 ) / (S 1 S 3 ) 이며이는최대차응력 (maximum differential stress, S 1 -S 3 ) 구간에서중간주응력 (S 2 ) 이어디에위치하는지를나타내는요소이다. S 2 가 S 1 에접근할수록 R값은 0에가까워지고, S 2 가 S 3 에접근할수록 R값은 1에가까워진다. 앞에서계산된포컬메커니즘응력역산결과 8~18 km 심도의광역응력장, 3~7 km 심도의여진자료를이용한응력장의경우에 R값은각각 0.88과 0.87로얻어져 S 2 와 S 3 의크기가상당히유사하다는점을알수있다. 두역산경우모두포항지역 3 km 이하심도에서주향이동단층성응력체계를나타내어 S Hmax > S v > S hmin 의크기순을보이며, 높은 R값은 S 2 인 S v 와 S 3 인 S hmin 이큰차이를보이지않는다는점을의미한다 시추공주입자료로부터의응력크기정보 포항지열발전실증부지주입심도에서의응력크기에대한보다직접적인규명을위해 PX-2 와 PX-1 지열정에서 2016 년이후수행된수리자극자료를분석하였다. PX-2 에서의첫번째수리자극은 2016 년 1월 29 일부터 2월 20일까지수행되었으며 PX-1 공에서는 2016 년 12월 15일부터 12월 27일까지수행되었다 (Fig. 4-7). 각시추공에서첫주입시기에는수리자극에의해시추공상태가크게교란되기전상태이므로압력 (pressure) 과주입률 (injection rate) 자료가응력크기에대한정보를제공할가능성이높다. Fig First hydraulic stimulation pressure-time and injection rate-time curves in PX-2 and PX-1. 68

146 제 4 장응력상태분석 첫번째유형의자료는주입초기물주입에따른시추공압력증가양상을보여주는자료이다. PX-2 의경우첫 3일동안 (2016 년 1월 29~31 일 ) 매일수리자극후 bleed off 하여 well-head pressure 를제거하였으며그로인해매일압력곡선을갱신하여관찰할수있다. 첫 3일동안물주입량에따른압력증가곡선은초기에거의동일한형태로선형증가하였다 (Fig. 4-8a). 압력을최대한증가시키지않은 Day 1을제외하고 Day 2와 Day 3에서공히약 64~66 MPa의압력에도달했을때선형에서벗어나는양상을보였다. 물주입초기압력의동일한선형증가양상은케이싱과나공전구간에걸쳐시추공이잘밀폐되어있으며탄성변형하고있다는사실을보여준다. 주입량에따른압력증가곡선이주어진압력에서선형성에서벗어나는이유는 a) 나공구간에서수압파쇄에의한새로운균열의형성되어누수되었을가능성과 b) 기존자연균열의개방에의한누수때문으로해석할수있다. 140 m에달하는 PX-2 나공구간에는다양한방향의자연균열들이존재할것으로판단되지만가장먼저개방될가능성이있는균열은가장개방저항력이낮은최소압축방향에수직한균열일가능성이가장클것이고따라서이압력은최소압축응력과유사한크기의압력일것으로추정된다. Well-head 에서의 64~66 MPa 압력이가해지면나공구간 (4,208~4,348 m) 에는 105~109 MPa의압력이가해지게되며, 이는독립적으로산정된 PX-2 나공심도에서의연직응력 (S v ) 106 MPa과매우유사하다. 이결과는 PX-2 나공구간심도에서최소주응력 (S 3 ) 이연직응력 (S v ) 일가능성이높음을시사한다. Fig Well-head pressure and injected water volume curves in (a) PX-2 and (b) PX-1. 같은방법으로분석하기위해 PX-1 에서최초로주입을실시한 2016 년 12월 15일 3회의주입자료는 Fig. 4-8b 와같다. 주입량에따른압력증가곡선은초기에선형을보이다가 well-head pressure 가약 15 MPa에도달했을때공히선형에서벗어난다. 이때 PX-1 나공부분 (MD: 4,049~4,362 m, TVD: 3,915~4,218 m) 의압력은 54~57 MPa에해당한다. 이압력은이심도에서의연직응력 (106 MPa) 보다현저히낮은압력이므로최소수평주응력 (S hmin ) 값으로해석할수있으나, 그럴경우문제는 PX-2에서관찰된상당히높은수준의압력증가 (105~109 MPa) 를절대설명할수없다는점에있다. 만일 4.2 km 심도에서의 S hmin 값이 54~57 MPa 이 69

147 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 라면 PX-2 에서도이정도의압력에도달했을때수압파쇄에의한누수가발생해야하고압력곡선에이현상이반영되어야하기때문에 PX-2 에서 100 MPa 이상선형으로압력이올라간현상을절대설명할방법이없다. 반대로 PX-2 에서분석한대로 4.2 km 심도에서최소주응력 S 3 가 105 MPa이라면 PX-1 에서의낮은압력조건에서누수현상이발생한것은설명가능하다. PX-2 에서분석한응력조건하에서새로운수압파쇄균열 (hydraulic fracturing) 의개방이나기존자연균열의인장개방 (hydraulic jacking) 을통한누수는 PX-1에서도 S 3 를초과하는압력이작용하지않는한가능하지않다. 대신낮은압력에서도누수가가능한유일한기작은시추공에존재하는자연균열의전단운동에의한균열면확장 (shear-induced dilation) 과그에의한균열투수율 (fracture permeability) 증가이다 (Barton et al., 2009). 주어진응력장하에서이러한누수시나리오가가능한지확인하기위해서 15 MPa의시추공압력이증가되었을때균열전단에기인한누수가가능한균열들의방향을분석해보았다. Fig. 4-9는응력상태 S v ~S hmin =106 MPa, S Hmax =243 MPa 을가정하고 PX-1 지열정에서물주입에의해 15 MPa 의압력증가가생겼을때전단운동이가능한균열의방향들을스테레오넷에나타낸결과이다. Fig. 4-9의예에보인선상에극점이놓이는균열이 PX-1 나공구간에존재한다면 15 MPa 의압력증가로인해이균열에전단이발생하여이를통해누수가생길수있다. 이정도의체계적인누수를야기시킬수있는불연속면이라면그크기도클것으로보인다. 참고로 PX-1 주입초기인 2016 년 12월 17일과 18 일에발생한세건의미소지진자료 ( 부록 A-3) 를이용하여유추한단층들의극점을표시하면우연히도이방향들과유사하다. 이들세단층들은 NNW-SSE 의주향을갖는공액단층조 (conjugate set) 를이루고있다. PX-1 의나공구간에이들단층들과유사한방향성을갖는균열이존재하면이균열들은누수통로의역할을한다. 이상의 PX-2 와 PX-1 에서이루어진수리자극시험결과분석을통해 PX-2 의압력분석결과가응력의크기를지시하고있음을알수있으며, PX-1 에서는수압증가에의한균열이나단층의전단운동에의한누수로인해응력의크기를지시하지못한다는점을알수있다. 따라서이후의추가적인응력크기분석은 PX-2 의자료를이용하여수행되었다. Fig Traces of pole orientations of natural fractures that can possibly play as leakage channel when PX-1 borehole pressure is raised by 15 MPa, which are depicted as cyan lines in the stereonet. Three circles indicate poles of faults estimated from the earthquake focal mechanisms that occurred during the initial stage of PX-1 injection (17~18 Dec 2016). 70

148 제 4 장응력상태분석 PX-2 에서 2016 년 2월 2일과 2017 년 9월 4일에 Step-rate test (SRT) 가수행되었다. 이시험은주입률을단계적으로증가시켜가면서 well-head pressure 가어떻게증가되는지를측정하여시추공에서어떤일이발생하는지를추적하는시험이다. 결과적으로말하면이시험을통해최소수평주응력 (S hmin ) 의크기를유추할수있다 (Krietsch et al., 2018). Fig 에보인두 SRT 자료는일정한주입률을단계적으로증가시키면서압력변화를모니터링한결과를보여준다. 두 SRT 결과에서공히주입률이증가함에따라도달한최대압력의변화양상이이중선형의형태를보인다. 즉, 주입초기와후기에압력증가양상이현저하게차이가나는데, 그이유는어느압력부터시추공나공에서균열의추가개방에의한물의누수가발생하여압력증가양상이주입초기보다더뎌지기때문이다. 이 well-head 압력이 2016 년 2월에는 81 MPa, 2017 년 9월에는 79 MPa로규명되어두시험간에상당한시차가있음에도유사한결과를얻었다. 이압력에도달했을때시추공나공구간과그주변으로개방되기용이한방향의균열이개방되면서누수가발생한것인데이러한균열은 S v 다음으로큰응력, 즉, S hmin 에수직한방향의균열이며, 이때의압력은 S hmin 의크기를지시한다. 두 SRT로부터규명한 S hmin 의크기는 4.2 km심도에서 120~122 MPa로계산되며동일심도의연직응력 (106 MPa) 보다크다. 이결과로부터 Fig. 4-8a 에서규명한 106 MPa은 S v 를측정한압력일가능성이높고 SRT에서는추가로 S hmin 의크기를측정했을가능성이높음을알수있다. Fig Step-rate test results in PX-2 on (a) 02 Feb 2016 and (b) 04 Sep

149 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 마지막으로응력의크기를추정하기위해분석한자료는균열전파압력 (fracture propagation pressure, FPP) 이다. PX-2 의여러번의주입에서일정한주입률로지속적으로물을주입하여측정된압력자료는시추공을벗어난주변에서의최소주응력의크기에대한정보제공할수있는데, 그예가 Fig. 4-11a 에제시되어있다. 이들결과는일정한주입률로지속적으로물을주입할경우압력이거의일정한수준을유지하는경향을보인다. 이압력에서시추공주변으로균열이지속적으로확장되면서주입된물이확장된균열로누수되어압력이일정하게유지된다. 이때균열의벌어지는개방방향은암체에작용하는압축응력이가장낮은최소주응력방향으로발생하게되며, 이에따라균열의확장방향은최소주응력에수직한방향이된다. 즉, FPP는균열에수직한응력을초과하면서균열을확장시키므로최소주응력의크기에대한정보를제공한다. PX-2 에서수행된세차례의수리자극에서균열전파압력과주입률을측정하여모두도시하면 Fig. 4-11b 와같다. 이결과는주입률이높아지면서 FPP 또한증가하는양상을보여준다. 세차례의수리자극결과로부터각각선형회귀방법으로추적한 zero injection rate의압력은 74~79 MPa의범위로얻어졌으며이를 4.2 km 심도에적용하면약 115~120 MPa의범위로구해진다. 이값은연직응력보다는명확히높으므로그보다는높은 S hmin 의크기를지시하는것으로보인다. (a) (b) Fig (a) Example of nearly constant pressures (fracture propagation pressures) attained at different constant injection rates and (b) their trend as a function of injection rate. 72

150 제 4 장응력상태분석 이상의 PX-2 에서얻은모든결과를종합하면 S v 는 106 MPa, S hmin 은 115~121 MPa로구해지며포항지열사이트하부 4.2 km에는 S v 와 S hmin 이비슷한가운데, 엄밀하게 S v < S hmin < S Hmax 의순을갖는역단층성운동에유리한응력체계임을보여준다. 마지막남은응력성분인 S Hmax 의크기는직접크기를지시하는자료가부재하므로주어진응력장에서가장전단운동하기용이한단층의마찰에의해 S Hmax 의크기가제한된다는일반적인가정하에설정하였다 (Townend and Zoback, 2000). 응력의크기를제한하는가장전단운동하기용이한단층의마찰계수 (μ) 는보통보수적으로가정하는값인 0.6으로가정하였다 (Townend and Zoback, 2000). 이러한가정하에최대수평주응력은 (Zoback, 2010) S H max S v P p P p 으로계산할수있는데이를통해산정한 S Hmax 값은 4.2 km 심도에서 243 MPa 이다. 지금까지의분석을통해 규명된주입심도에서의응력모델은 Table 4-1 에요약하였다. Table 4-1. Constrained stress model in Pohang geothermal site. Parameter Value Brief explanation Target depth 4.2 km PX-2 injection depth S v 106 MPa based on density measurements in cores extracted from a nearby hole (BH-4) and PX-2 P p 41.3 MPa assumed hydrostatic S hmin 115~122 MPa from step-rate tests and fracture propagation pressure analysis in PX-2 S Hmax ~243 MPa constrained by Coulomb friction limit (μ=0.6) along the optimally oriented faults for slip S Hmax azimuth N77 E from borehole dipole sonic data Stress regime Reverse faulting based on borehole-scale stress estimations R 0.88~0.93 R = (S 1-S 2)/(S 1-S 3) = (S Hmax-S hmim)/(s Hmax-S v) 응력장과포항지진단층의운동학적상관관계 규명된응력장이포항지진을유발한단층운동의운동학적 (kinematics) 특성과일치하는지에대한분석을통해규명된응력장의신뢰성을검토하는시도를하였다. 포항지진을발생시킨단층은지진자료분석을통해자세와면선각 (rake) 으로표현되는미끌림방향이규명되었다. PX-2 지열정주입과정에서발생한여러지진들의진원분포를통해단층면의주향과경사는 214 /43 로규명되었으며포항지진의 main-shock focal mechanism solution 중이러한방향을갖는단층면해의면선각은 122~136 범위에있는것으로분석되었다. 주어진단층면상에서의미끌림방향은응력의방향과크기에의해좌우되기때문에, 본연구에서규명된응력장이이러한단층운동을발생시킬수있는지를확인하기위한간단한모델분석을시도하였다. Fig 에 73

151 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 보인모델결과는최대수평주응력방향과 R값으로표현된응력비 (stress ratio) 에따라 214 /43 의자세를갖는단층면에서어떤방향의미끌림이발생하는지를면선각으로표현한결과이다. 지진자료를통해규명한면선각 122~136 가가능한 S Hmax 방위각과 R값의조합을점선으로표현하였다. 본연구를통해규명된응력조건에서는주어진포항지진단층면상에서 140 의면선각을보이며단층운동이발생하는것으로나타나지진자료를통해규명된결과와 4 의차이를보이며거의일치하는결과를보인다. 즉, 본연구를통해규명한응력상태는지진자료를통해규명한실제단층운동의운동학적거동과거의일치한다. Fig Modelled rake on the Pohang fault plane as a function of S Hmax azimuth and R value for the verification of the constrained stress model using kinematics of the Pohang fault 포항지진유발단층의응력상태 4.2 km 심도의주어진응력장하에서단층의방향에따른전단성향 (slip tendency) 을구하면 Fig 의스테레오넷과같다. 이결과는단층의방향에따라전단성향 ( 단층면에작용하는전단응력 / 유효수직응력 ) 이 0~0.6 의범위를보이며변화함을보여준다. 포항지진을유발한단층면의경우 (214 /43 ) 전단성향은 0.55~ 0.57로얻어졌다. 즉, 포항지진을발생시킨단층은지진발생전, 가장전단이유리한자세를갖는가상단층의전단성향 (0.6) 의 92~95% 수준의높은전단성향을띠고있음을알수있으며상당히임계치의전단응력을받고있음을시사한다. 74

152 제 4 장응력상태분석 Fig (a) Slip tendency of the PX-2 (square) injection-related fault and (b) stress condition plotted in Mohr diagram 결론 본연구를통해규명한응력장하에서포항지진을발생시킨단층의전단성향은정수압의공극압상태에서 0.55~0.57 로유추되었으며, 이는주어진응력장하에서가장취약한방향성을갖는단층에비해 92~95% 수준의높은임계치의응력상태에있음을시사한다. 현재까지확보된포항지열발전실증사이트심부불연속물질의전단물성으로는 PX-2 지열정 4.2 km 깊이에서회수된코어시편에존재하는자연균열의마찰계수 (0.53, Kwon et al., 2018) 와 3.6 km 깊이로부터의시추암편을이용해측정한마찰계수 (0.54~0.68) 가측정되었다. 이결과는포항지진을일으킨단층이임계응력상태이거나이에아주근접한상태에있었음을시사한다 참고문헌 Barton, C., Moos, D., and Tezuka, K., 2009, Geomechanical wellbore imaging: Implications for reservoir fracture permeability. AAPG Bulletin, 93,

153 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Brie, A., Endo, T., Hoyle, D., Codazzi, D., Esmersoy, C., Hsu, K., Denoo, S., Mueller, M.C., Plona, T., Shenoy, R., and Sinha, B., 1998, New directions in sonic logging. Oilfield Review, 10, Chang, C., Lee, J.B., and Kang, T.-S., 2010, Interaction between regional stress state and faults: Complementary analysis of borehole in situ stress and earthquake focal mechanism in southeastern Korea. Tectonophysics, 485, Heidbach, O., Rajabi, M., Reiter, K., Ziegler, M., and WSM Team, 2016, World Stress Map Database Release GFZ Data Services, Hong, T.-K. and Choi, H., 2012, Seismological constraints on the collision belt between the North and South China blocks in the Yellow Sea. Tectonophysics, 570, Jun, M.-S., 1991, Body-wave analysis for shallow intraplate earthquakes in the Korean Peninsula and Yellow Sea. Tectonophysics, 192, Kim, H., Xie, L., Min, K.-B., Bae, S., and Stephansson, O., 2017, Integrated in situ stress estimation by hydraulic fracturing, borehole observations and numerical analysis at the EXP-1 borehole in Pohang, Korea. Rock Mechanics and Rock Engineering, 50, Krietsch, H., Gischig, V., Evans, K., Doetsch, J., Dutler, N.O., Valley, B., and Amann, F., 2018, Stress measurements for an in situ stimulation experiment in crystalline rock: integration of induced seismicity, stress relief and hydraulic methods. Rock Mechanics and Rock Engineering, 52, Kwon, S., Xie, L., Park, S., Kim, K.I., Min, K.B., Kim, K.Y., Zhuang, L., Choi, J., Kim, H., and Lee, T.J., 2018, Characterization of 4.2-km-deep fractured granodiorite cores from Pohang Geothermal Reservoir, Korea. Rock Mechanics and Rock Engineering, 52(3), Soh, I., Chang, C., Lee, J., Hong, T.K., and Park, E.S., 2018, Tectonic stress orientations and magnitudes, and friction of faults, deduced from earthquake focal mechanism inversions over the Korean Peninsula. Geophysical Journal International, 213, Townend, J. and Zoback, M.D., 2000, How faulting keeps the crust strong. Geology, 28, Yoon, W. et al., 2011, Annual Report for Technology Development of Geothermal Demonstration Plant for M W Class. KETEP. Zoback, M.D., 2010, Reservoir geomechanics. Cambridge University Press. 76

154 제 5 장 지진분석 Seismological Analysis

155 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project 78

156 제 5 장지진분석 요약 / Abstract 2017년 11월 15일포항지진본진과이전지진활동에대한지진원분석을수행하였다. 진앙주변에서수행된다양한지구물리탐사와시추공조사결과및광역속도모델을토대로 1차원지진파속도모델을결정하여지진원분석에사용하였다. 실증연구시설을중심으로 2009년 1월 1일부터 2017년 11월 15일본진까지총 520개지진을식별하고진원결정이가능한지진의규모를측정하였다. EGS 활동과지진발생사이의연관성분석을위한 98개지진을분류하고정밀지진위치분석을수행하였다. 진앙은 PX-2 지열정을기준으로주로북서쪽에위치하고있으며, 진원깊이는 3.5~4.5 km 범위에있다. 포항본진의진원은위도, 경도, 깊이가각각 , , 4.27 km이며, 규모는 M W 5.5, 단층면의주향, 경사, 미끌림각을각각 214, 51, 128 로결정하였다. 정밀지진원분석결과를주입시기에따라 PX-1 지열정주입시기 (G1) 와 PX-2 지열정주입시기 (G2) 로분류하였다. G1 진원은북서-남동방향, G2 진원은북동-남서방향을따라분포한다. G2 진원분포는 N214 E/43 NW 인평면으로근사되며, G2 근사평면은포항본진의단층면해와유사하다. G2 근사평면상에서지진은순차적으로 PX-2 지열정굴착시기에발생한이수누출지진의진원으로부터포항본진의진원방향으로이동하였다. G2 근사평면을 PX-2 지열정방향으로연장하면, 약 3,800 m 깊이에서 PX-2 지열정과교차한다. 이깊이는 PX-2 지열정파손으로추정되는깊이 (3,783 m 하부 ) 에상응한다. Seismic analysis for the November 15, 2017 Pohang earthquake and its preceding events was carried out. Based on various geophysical explorations around the epicenter, well-log results, and crustal velocity models of South Korea, a 1-D seismic velocity model was established and used for the analysis. A total of 520 earthquakes were identified during the period from January 1, 2009 to the occurrence of the mainshock in November 15, and magnitudes of locatable events were determined. For the investigation, we selected 98 earthquakes and performed a precise location analysis. The epicenters of the earthquakes are located in the northwest from the PX-2 well, and the depths are in the range of 3.5 to 4.5 km. The strike, dip, and rake of the fault plane of the mainshock were determined as 214, 51 and 128, respectively, and the mainshock was located at N, E, and a depth of 4.27 km. Moment magnitude (M W) of the mainshock was determined to be 5.5. The results enable us to classify the earthquakes into PX-1 and PX-2 groups (G1 and G2), respectively. The epicenters of G1 group are distributed along the northwest-southeast direction and those of G2 along the northeast- southwest direction. The hypocenters of G2 group can be approximated by a plane of N-214 E-/-43 NW. It is shown that the G2 plane is similar to the fault plane solution of the mainshock and also found that the hypocenters of the G2 migrate sequentially from the earthquakes occurred during the drilling of the PX-2 well to the direction of the mainshock on the G2 plane. If the G2 plane is extended to the PX-2 well, it meets the PX-2 well at a depth of about 3,800 m. This depth corresponds to a depth (below 3,783 m) where the failure of the PX-2 well was inferred. 79

157 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 5.1. 서론 2017 년 11월 15일 14시 29분 ( 지역시간 ; UT+9h) 국지규모 (M L ) 5.4의지진이한반도남동부에위치한경상북도포항시흥해읍에서발생하였다. 이지진은기상청이지진관측망을운영하기시작한 1978 년이후한반도에서발생한두번째로큰규모이며, 이기간동안에가장큰인명과재산피해를일으킨지진이다. 이지진의진앙인근에서는지열발전실증연구프로젝트가진행되었으며, 약 4 km 깊이에이르는두개의시추공을굴착하였다. 다섯차례의수리자극시험이수행되었으며, 최종수리자극시험을실시한 2017 년 9월이후약 2개월이경과한시점에서포항지진의본진이발생하였다. 이후크고작은여진활동이수개월이상지속되었다. 이연구에서는 2017 년 11월 15일포항지진본진과지열정수리자극시험의상관성여부를파악하기위하여본진과실증연구시설인근에서본진이전에발생했던지진활동을조사하고정밀분석하였다 지진자료 이연구는본진과그이전에실증연구시설인근에서발생한지진들을대상으로한다. 지진활동조사와분석을위하여사용한지진파형자료는해당기간에운영된세종류의지진관측망에기록된것이다. 첫번째관측망은기상청, 한국지질자원연구원, 한국수력원자력에서운영중인상시관측망으로광대역속도계, 단주기속도계, 가속도계로구성되어있으며본진을기준으로진앙거리약 50 km 이내의지진계에기록된자료들을주로사용하였다. 또한본진의규모비교및일부지진의단층면해결정에는더먼거리에있는지진계의자료를추가적으로사용하였다. 두번째관측망은실증연구와관련하여설치된지진계들로구성되어있으며, 11개의지표면속도계, 100~150 m 깊이에설치된 9개의시추공지진계, 2017 년 7월부터한달간운영된 PX-2 지열정지하 1.36 km 하부부터 10 m 간격으로 1.52 km 깊이까지 17개의센서로이루어진시추공배열식지진계 (borehole geophone array), 1~3차수리자극시험기간중 PX-2 지열정하부 1.35~1.55 km에서운영된 ( ~ ; ~ ; ) 수직탄성파탐사지오폰배열 (Vertical Seismic Profile), 그리고 PX-2 지열정에서남서쪽으로약 2 km 떨어진지점의지표지진계 (POH01) 와약 2.3 km 깊이에설치된시추공지진계 (BH4) 로구성되어있다. 마지막세번째관측망은포항지진본진발생직전에설치된임시관측망 (Kim et al., 2018) 으로, 본진을기준으로진앙거리 3 km 이내지역에서운영된 8개의단주기속도계로구성되어있다 (Fig. 5-1). 80

158 제 5 장지진분석 POH01,BH4 Fig Location map of seismic stations used in this study 속도모델 수리자극시험과지진의연관성을분석하기위해서는진원을정확하게결정해야한다. 진원은 P파와 S파의주시자료와속도모델의결합으로결정되며, 지진파속도모델에따라수 km의진원차이가발생할수있다. 따라서한반도전체에통용되는일반적인지진파속도모델을사용할경우진원의불확실성이커진다. 이를줄이기위해서는실증연구시설인근을대표하는국지적인지진파속도모델개발이필수적이다. 지진파속도모델은지진파속도의변화를기준으로층경계가결정되며, 암상이다르면물리적인특성도달라지기때문에속도가변화한다는가정으로층경계를설정하였다. 하지만암상이바뀌어도속도의변화가없다면, 지진파속도모델에서는같은물리적특징을갖는층으로정의할수있다. 실증연구시설주변을포괄하는기존지질단면도와속도자료를토대로 6개의천부층과 Kim et al. (2011) 에서제시한경상분지광역속도모델을이용하여 3개의심부층으로속도모델의층경계를설정하였다. 제 1층과제 2층의 P파속도는이지역에서수행된탄성파탐사자료의굴절법해석을통하여결정하였고, S파속도는지표관측소와 110~150 m 사이에설치된시추공관측소의 S-P파의주시차이를이용하여결 81

159 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 정하였다. 제 3층의속도는 PX-2 지열정의 1,360 m와 1,520 m 깊이에설치된시추공지진계에서 2017 년 8월 13일 M W 1.2 지진 ( a; 지진 ID는부록 A-3 참조 ) 을기록한자료의주시차이를이용하여결정하였다. 제 4층과제 5층의 P파속도는이들층의속도결정에필요한정보가부족하기때문에, 제 3층과제 6층의속도범위에서 0.01 km/s 간격으로속도를변화시키며진원을반복적으로결정하는격자탐색방법을사용하였다. 제 4층과제 5층의 S파속도는제 6층에서계산된 Vp/Vs 비를이용하여결정하였다. 제 6층의속도는시추공검층자료를이용하여결정하였으며, P파와 S파속도는각각 5.85, 3.31 km/s 이다. 심부층에해당하는제 7층, 제 8층, 제 9층의속도는실증연구시설을포함하는경상분지속도모델로제시된 Kim et al. (2011) 의결과를사용하였다. 이들심부층광역속도모델을제 6층까지의천부층속도모델과결합하여 Fig. 5-2와 Table 5-1과같은 1차원지진파속도모델을결정하였다. Fig Local 1-D velocity model developed in this study. Table 5-1. Local 1-D velocity model and its description. Top of layer (km) P-wave velocity S-wave velocity (km/s) (km/s) Vp/Vs ratio Remark Refraction analysis and comparison of phase arrival times Refraction analysis and comparison of phase arrival times Comparison of phase arrival times and check shots Measurement of the least arrival time error Well logging data Regional model Regional model Regional model 82

160 제 5 장지진분석 5.4. 지진검출 개별지진의지진원특성분석에앞서 template matching method (Shelly et al., 2007; Zhang and Wen, 2015; Kato et al., 2016) 을사용하여실증연구시설주변지역에서 2017 년 11월 15일본진이전에발생한지진들을검출하였다. Template matching method 는관측된지진파형을 template 으로사용하여동일한관측소에기록된연속파형자료에서 template 파형과유사한파형을상호상관 (cross-correlation) 을이용하여찾아내는방법으로신호대잡음비가낮은경우에도지진을검출할수있다는장점이있다. 본진직전에발생한 5 개의전진들을포함하여, 본진이전에발생한지진들중명확히지진파형으로판단되는 39개지진의파형을 template 으로사용하였다. 규모가큰지진들은신호대잡음비가커서쉽게검출이가능하기때문에 template matching 을통해서는식별이어려운작은규모의지진들을검출하는것을목표로하였으며, 이에따라비교적잘관측된작은규모지진들의파형들로 template 을구성하였다. 수리자극기간뿐아니라사업이전에발생한지진이있는지확인하기위하여, 장기간지진관측이이루어진실증연구시설인근의상시관측소중가장가까운곳에위치한 PHA2( 혹은 PHA; 관측소의공식명칭이변경되었으나편의를위해앞으로 PHA2 로만부른다 ) 에기록된 2009 년 1월 1일부터본진발생사이의연속파형을대상으로지진을검출하였다. 기상청에서운영하는 PHA2 관측소는 PX-2 지열정을기준으로북쪽으로약 10 km 떨어진곳에위치하고있으며, 이관측소의단주기속도자료중 sampling rate가 100 Hz인것을사용하였다. 한국지질자원연구원의 HAK 관측소가상시관측망중에서는 PX-2 지열정으로부터두번째로가까우나, 실증연구시설로부터거리가약 23 km로 PHA2 에비해두배이상멀기때문에일관성있는분석을위해 PHA2의자료만을사용하였다. Template 파형은 S파도달시간의 1초전에서 3초후까지 4초간의시간창에포함된자료를사용하였으며, 상호상관을수행하기전에 5-20 Hz의대역통과필터를 template 파형과연속파형에동일하게적용하였다. 각채널에서계산된 correlogram 을중합 (stack) 하여신호대잡음비를향상시켰으며, 하루단위로중합된 correlogram 으로부터중앙값절대편차를계산하고그값의 14배를기준으로설정하여 3,547 개의후보지진들을검출하였다. 검출된후보지진들의파형을육안으로검토하여검출오류와중복검출된지진들을제거하였으며, 이과정에서는 PHA2 이외의관측소에기록된파형들도보조적으로사용하였다. 최종적으로 template matching 을이용하여검출한지진은 519개이다. Template 으로사용한지진들과규모차로인한주파수특성의차이로 template matching 을통해검출되지않은본진을포함하여, 총 520개의지진을분석하였다 진원결정 지진의진원을정확하게결정하기위해세단계를거쳤다. 첫번째와두번째단계에서사용한방법은전 통적으로진원결정에많이사용되는 single difference 와 double difference 방법이며, 편의를위해앞으로각 각의방법으로결정된진원들을초기위치와상대위치로부른다. 마지막단계에서는 2017 년 8 월 13 일 21 시 42 분 83

161 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 37초 ( 세계표준시 ) 에발생한지진 ( a) 의상대위치와동일지진에대하여독립적인방법으로결정된절대위치의차이만큼모든지진의상대위치를이동시켰으며, 이결과를최종위치로정의한다 초기위치 Single difference 방법의기본적인원리는 P파와 S파의관측시간과이론적도달시간의잔차를최소화하는위치를선형역산을통해결정하는것으로, 실제계산에는 Hypoellipse (Lahr, 1999) 를사용하였다. 앞서본연구의분석대상으로삼은 520개지진의발생시각을기준으로당시에운영되던지진계들에기록된지진파형에서 P파와 S파의도달시간을육안으로결정하여, 4개이상의도달시간이측정된경우만초기위치를역산하였다. 속도모델에서가정한첫번째층의깊이보다깊은곳에위치한시추공지진계 (VSP stations, PX-2 borehole chains, and BH4 station) 의기록은사용한지진원결정프로그램 (Hypoellipse) 의한계로인하여고도보정만으로는도달시간을정확하게보정할수없으므로, 위치결정에서제외하였다. Template matching 방법을거쳐식별된총 520개의지진들중초기위치가결정된총 253개의진앙을 Fig. 5-3에도시하였다. 이지진들을 PX-2 지열정으로부터진앙거리와진원깊이를기준으로세가지유형으로구분하였다. 첫번째유형은진앙거리 5 km 이상에서발생한 140개의지진들로최대진앙거리는 62 km이다. 두번째유형은진앙거리 5 km 이내에서발생하였지만진원깊이가 10 km 보다깊은네개의지진들로 2013, 2015, 2016, 2017 년에각각발생하였다. 세번째유형은진앙거리 5 km 이내이면서진원깊이가 10 km 보다얕은 109개의지진이다. 첫번째와두번째유형의지진들은먼진앙거리또는깊은진원깊이로미루어수리자극시험과의직접적인연관성은낮은것으로판단된다. 본연구의목적은지열발전실증연구활동과지진발생사이의연관성을분석하는것이기때문에, 세번째유형의지진들만이본연구의관심대상이다. 따라서세번째유형에속한 109개지진들만다음단계의정밀위치분석을수행하였다. Fig Initial locations of 240 earthquakes detected by the template matching method. Yellow triangle represents the PX-2 well. Earthquakes whose epicentral distances are greater than and less than 10 km from the PX-2 well are denoted by red and green circles, respectively. Four earthquakes with focal depth greater than 10 km are plotted as blue circles. Event ID and focal depths of the four events are also represented. Geological lineaments and faults are shown as dashed lines. 84

162 제 5 장지진분석 상대위치 초기위치를기준으로추가분석대상으로선정된지진들의상대적인위치분포를보다정확하게결정하기위하여동일관측소에기록된여러지진들의관측과이론도달시간의상대적인차이를최소화하는 double difference 방법을적용하여총 98개지진들의상대위치를결정하였다. 이때역산에는 HypoDD (Waldhauser and Ellsworth, 2000) 을사용하였다. 초기위치가결정된 109개의지진들중측정된도달시각이 8개미만이거나상대도달시간자료가 9개미만인 11개의지진들은자료부족으로인하여상대위치결정의정확도가떨어질가능성이높으므로분석에서제외하였으며, 초기위치결정과정과마찬가지로사용한속도모델의첫번째층보다깊은곳에설치된지진계의기록은사용하지않았다. Double difference 방법의장점을극대화하기위해서는동일관측소에서측정된서로다른지진들의상대도달시간을정확하게측정하는것이매우중요하다. 상대도달시간을정확하게측정하기위하여, 진원위치와지진발생메커니즘이유사하여지진파형이비슷한자료를이용하였다. 지진파도달시간차이는 P파 ( 혹은 S파 ) 의파형상호상관 (waveform cross-correlation) 을이용하여측정하였다. 이때수동으로측정된 P파와 S파의도달시간을기준으로 -0.5에서 0.5초사이 1초간의파형자료를선별하고, 3차스플라인보간법 (cubic spline interpolation) 을이용하여 sampling rate를 1,000 Hz로증가시켰으며, 2~10 Hz의대역통과필터를적용한후에상호상관을수행하였다. 파형상호상관을이용한상대도달시간측정이어려운지진들간에는초기위치결정에사용되었던도달시간의차이로부터상대도달시간을측정하였다. 역산에는대표적인반복선형역산법 (iterative linear solver) 인 damped least-squares QR (LSQR) 알고리즘이사용되었으며, 결정된상대위치의오차를아래두가지방법으로추정하여교차검증하였다. 첫째, LSQR 결과에서계산된상대주시잔차를무작위로샘플링하여합성상대주시자료를만들어서다시역산을수행하였다 (Waldhauser and Ellsworth, 2000). 이과정을 200회반복하여결정된상대위치정보를통계적으로분석하여오차를추정하였다. 둘째, 상대위치결정시마지막반복과정에서사용된상대주시잔차자료를사용하여구성한선형방정식으로부터 singular value decomposition (SVD) 를통해최소제곱오차를추정하였다. 첫번째와두번째방법을통해추정된상대위치들의 x-, y-, z-방향의평균오차는 2σ를기준으로각각 20, 13, 25 m와 15, 10, 19 m로서로큰차이를보이지않는다 최종위치 수리자극시험기간에발생한지진들중 2017년 8월 13일발생한 M W 1.2 지진 ( a) 은지표지진계이외에 PX-2 지열정내부에 2017 년 7월 26일에서 2017 년 8월 23일까지심부 1,360 m에서 1,520 m까지 10 m 간격으로설치되어운영된시추공배열식지진계에도기록되었다. 이지진계에기록된 P파와 S파의도달시각차이와 P-wave 입자운동분석결과를이용하여 a 지진의진원을독립적인방법으로결정하고, 이를 tube wave를사용하여평가하였다 (Grigoli, 2018, personal communication). 이분석에서는가장가까운거리에서기록된지진자료를활용하였으며, 전체지진가운데유일하게서로독립적인방법 ( 주시와입자운동, tube wave) 을이용하여위치를결정하였기때문에분석결과의신뢰도가가장높은것으로판단하였다. 따라서본연구에서는이결과를최종위치결정에활용하였다. Grigoli (2018, personal communication) 가결정한 a 지진의진원 ( 위도 , 경도 , 깊이 4.21 km) 과본 85

163 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 연구에서결정한진원과의상대위치차이는서쪽, 남쪽, 깊이 (downward positive) 방향으로각각 3 m, 430 m, -161 m이며, 이값을전체지진의상대위치에서보정하여최종위치를결정하였다. 98개의지진들의진앙은 PX-2 지열정을기준으로주로북서쪽에위치하고있으며, 북서- 남동및북동- 남서방향의분포를갖는다 (Fig. 5-4). 본진의최종위치는위도, 경도, 깊이가각각 , , 4.27 km이며, 다른지진들의진원깊이는 3.5에서 4.5 km 사이로결정되었다. Fig Final locations of 98 earthquakes. Yellow, gray, and blue circles represent events b, b and a, respectively. Five immediate foreshocks and the mainshock are denoted by red and green circles, respectively. Green and blue curves ended by red represent the PX-1 and PX-2 wells, respectively. Open sections of wells are shown in red curves 단층면해 최종위치분포 (Fig. 5-4) 로부터대략적인단층면의자세를파악할수는있으나, 각각의지진이실제어떤 단층운동에의해발생했는지를파악하기위해서는단층면해가필요하다. 단층면해를결정하는방법에는여러 86

164 제 5 장지진분석 가지가있으나, 이연구에서는 P파의초동극성정보를이용하였다 (Lay and Wallace, 1995). 수직성분의지진기록으로부터 P파의초동극성을측정하였으며, 신호대잡음비가낮아구분이어려운일부지진들의경우파형이유사하면서상대적으로규모가큰지진기록을참고하여극성을결정하였다. 일부시추공지진계의경우수직성분파형의극성이뒤바뀐것으로의심되는경우가발견되었다. 따라서 PX-2 지열정으로부터남서쪽으로 37 km 떨어진곳에서발생한단층면해가잘결정된 a 지진의이론적초동극성을각관측소에기록된초동극성과비교하여수직성분파형의극성을확인하였으며, 필요할경우극성을보정하였다. 단층면해결정에는위치결정에서사용하지못한깊은곳에설치된시추공지진계들도사용하였으며, 초동정보와최종위치를바탕으로 HASH (Hardebeck and Shearer, 2002) 를사용하여단층면해를결정하였다. 각지진에대해서 1개까지의극성오차의예외를허용하는단층면해의후보들을계산하고, 후보단층면들의평균값을각지진의단층면해로결정하였다. Hardebeck and Shearer (2002) 에서제시한기준을따라결정된단층면해의품질을평가하였으며, Qualiy A와 Quality B에해당하는지진들은각각 28개와 25개이다. 포항본진의단층면해는주향, 경사, 미끌림각이각각 214, 51, 128 로결정되었다 (Fig. 5-5). Zoback (1992) 에제시된방법을따라단층면해로부터단층운동형태를분류하였고, 단층면해가결정된 53개의지진들중주향이동이 14개, 역단층이 22개, 주향이동과역단층이결합된형태가 15개로평가되었다. 정단층지진은발견되지않았으며, 정단층과주향이동이결합된형태는 1개가발생하였다 (Fig. 5-5). 분류된단층형태로미루어연구지역에서는주로주향이동과역단층형태의지진들이발생하는것으로판단된다. Fig Distribution of 53 focal mechanism solution. Colors of beachball diagrams represent faulting types according to the classification of Zoback (1992): Strike-slip (black), Thrust (blue), Strike-slip with thrust component (red), and Strike-slip with normal component (green). 87

165 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 5.7. 지진규모결정 식별된총 520개의지진중에서 137개지진의규모를결정할수있었다. 전통적으로지진규모는소숫점첫째자리까지표현하지만, 규모결정과정과결과를설명하기위해여기서는소숫점둘째자리까지표현하였다. 이연구에서는 Sheen et al. (2018) 에서제시한수직성분지진규모식을사용하여수리자극시험동안발생한지진의국지규모 (M L ) 를결정하였다. 그런데상시지진관측망에속한지진관측소조차도부지특성에기인한관측소규모보정값이수평성분자료의경우 0.54 정도, 수직성분자료는 0.29 정도의차이를보이는만큼 (Sheen et al., 2018), 안정적인 M L 규모결정을위해서는사용된지진관측소의관측소보정값을도출할필요가있다. 이를위해본진이후 2017 년 11월 16일부터 12월 31일까지발생한다수의여진자료를사용하여관측소보정값을결정하였다. 관측소보정항을사용하여 2015 년 11월 30일부터 2017 년 11월 15일본진까지실증부지주변에서발생한지진들중에서 40개지진의 M L 규모를결정하였는데, 규모 0.15에서 5.33의범위를보이며평균적으로 0.14 정도의표준편차를가지는것으로확인되었다. 그리고 2013 년 3월 1일실증연구시설부근의하부약 12 km 지점에서발생한지진의규모는 M L 2.16으로결정되었다. 부지주변약 50 km 이내에있는상시지진관측망에속한 11개의광대역속도계, 단주기속도계와가속도계자료를이용한본진의규모는 M L 5.33±0.14 로, 한반도전역에설치된 77개의광대역속도계를이용한규모는 M L 5.34±0.18 로결정되어서로거의동일한값을얻었다. M L 규모는 Richter (1935) 의정의에따라 Wood-Anderson 변위계에서지진을관측한것처럼지진자료를모사해야하며, 이과정에서저주파수신호가증폭된다. 이로인해모사된 Wood-Anderson 변위자료에서는미세한진동의최대진폭을측정하기어려워, M L 규모를결정하지못한지진들도많이포함되어있다. 이러한 72개의지진은이연구에서도출된속도또는가속도최대진폭과 M L 규모의상관관계를이용해지진규모를결정하였다. 또한 template matching method 를이용하여지진파가식별된지진들중에는진원결정에필요한관측자료의수가부족하여정확한진원을결정하지못한지진들도많이포함되어있다. 이러한작은크기의지진들중에서 PHA2 관측소에서최대진폭식별이가능한지진중 24개지진의규모를추가로결정하였는데, 규모 에서규모 0.60 정도의분포를가진다. 이지진들은진원오차를고려할때, 진원오차에따른추가결정된지진규모의오차는약 ±0.2 정도일것으로예상된다. 지진파의최대진폭을이용해규모를결정한 137개지진중에서 48개지진의모멘트규모 (M W ) 를시간영역또는주파수영역에서결정하였다. 이를위해지진원근처에서의암석밀도를 2.8 g/cm 3, P파의속도를 6 km/s로가정하였다. 시간영역에서의 M W 규모결정은 Tsuboi et al. (1995) 와 Prejean and Ellsworth (2001) 의방법을따랐으며, PHA2 관측소에서관측된 P파변위의초기파형의면적으로부터지진모멘트 (Aki and Richards, 1980) 를측정하여 46개지진의 M W 규모를결정하였다 년 4월 15일 M L 3.27 지진 ( b) 은 P파의파형의복잡성으로인해상대적인규모측정방법인주파수영역의스펙트럼비 (spectral ratio) 를이용해 M W 규모를결정하였다. 이를위해 2017 년 4월 15일 08시 16분 47초 ( 세계표준시 ) 에발생한 M L 2.06(M W 2.15) 지진 ( a) 의자료를경험적그린함수 (empirical Green s function) 로사용하였다. 두지진을모두관측한상시지진관측소 5개소의 P파스펙트 88

166 제 5 장지진분석 럼비를이용해두지진의지진모멘트 (seismic moment) 를비교한결과, b 지진이약 14배정도더큰모멘트를가지고있는것으로측정되었다. 이로부터 b 지진의 M W 는 3.29인것으로판단된다 년 11월 15일포항본진은 Rhee and Sheen (2016) 에서사용한방법을따라 P파의변위스펙트럼에서 M W 규모를결정하였다. P파와 S파가충분히구별되는진앙거리 150 km 이상의 58개소광대역관측소의속도자료로부터 초의시간창길이를가지는 P파신호를이용하였고, 지진파전파과정에서발생하는비탄성적감쇠를보정하기위해 Kim et al. (2016) 의 P파감쇠모델을사용하였다. Brune (1970) 의지진원모델에따라본진의지진원요소를결정하였으며, 본진의 M W 규모는 5.56±0.18 로결정되었고, P파의모서리주파수는 0.60 Hz, 응력강하량은 5.6 MPa, 단층반경은 3.44 km로결정되었다. 부록 A-3에이연구에서도출된 98개지진의지진원목록을제시하였다. Fig. 5-6은이연구에서결정한 M L 규모와 M W 규모의상관관계를비교한것이다. 규모 2.0의이상의지진규모가기상청에서발표한결과와서로잘일치하는것을확인할수있으며, M W 규모와최대진폭을이용한규모도서로좋은상관관계를보이는것을알수있다. 기존연구들 (Grigoli et al., 2018; Kim et al., 2018) 에서발표한본진의 M W 규모가 5.5 또는 5.4 인것을감안할때, 이연구에서결정한 M W 규모는최소 0.02 에서최대 0.2 정도크게결정된것으로판단된다. 다른연구결과들과의비교를위해주요지진의 M W 규모는소수점둘째자리를절삭하여표현하기로한다. Fig Comparison of magnitude estimates. (a) Local magnitudes of the KMA versus and those of this study. (b) Moment magnitudes versus local magnitudes. Red circles represent the events published by the KMA. 89

167 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 5.8. 토의 지진발생과두지열정에서의주입간의관계를살펴보기위하여지진발생시기에따라최종위치가결정된 98개의지진들을두개의그룹으로나누었다. 첫번째그룹 (G1) 은 PX-1 지열정에서주입이이루어진시기에발생한지진들이고, 두번째그룹 (G2) 는 PX-2 지열정에서이루어진시기에발생한지진들이다. 수리자극시험전 2015 년 11월 1일부터 12월 1일사이에발생한지진들과포항지진본진및직전에발생한전진들은최종위치가 G2에속하는지진들과유사하기때문에 G2에추가적으로포함시켰다. 최종위치를도시한 Fig. 5-7의 B1-B2 수직단면도에서지진들이 PX-1 지열정을기준으로 G1과 G2 로명확하게나뉘는것을확인할수있다. G1은평면도에서주로북서- 남동방향을따라분포하는지진들에해당하고 G2는북동- 남서방향을따라보이는지진들에해당한다. G2의지진분포로부터주성분분석 (principal component analysis) 을사용하여근사평면을계산하면주향 / 경사가 N214 /43 NW 이다. G1의경우북서쪽에위치한지진들의진원깊이가남동쪽지진들에비해깊어지는양상을보이나전체적인진원분포는평면보다는타원체에가깝다. G2 근사평면의주향 / 경사는본진단층면해의주향 / 경사인 N214 /51 NW 와유사하다. G2의근사평면을 PX-2 쪽으로연장하면약 3,800 m 깊이에서 PX-2 와만나게되며 (Fig. 5-7 참조 ), 이깊이는본진발생 Fig Illustration of the classification of groups G1 and G2. Magenta and cyan circles represent locations of earthquakes belong to group G1 and G2, respectively. The size of circles scale with the magnitude of earthquakes. Black line in B1-B2 section represents a G2 plane approximated from the principal component analysis. 90

168 제 5 장지진분석 이후시추공의상태를점검하기위하여실시된영상검층에서장비가더이상아래로내려가지않아시추공이파손된것으로추정한깊이 3,783 m와거의일치한다 ( 부록 A-3). 이상의결과를종합하면 PX2 주입과연관되어발생한지진들과본진은기존에존재하던동일한단층면상에서유사한단층운동에의해발생한것으로보이며, 본진으로인한단층운동으로인하여 PX2의케이싱이약 3,783 m 깊이에서파손된것으로추정하는것이합리적이다. G1에속하는지진들이다양한단층면해를가지는데비해 G2에속하는지진들은본진을포함하여대부분유사한단층면해를보여주고, 이는지진분포로부터추정한 G2 근사평면의자세와도일치한다. 앞서나눈 G2를시기별로다시세분하여지진의발생특성을분석하였다. PX-2 지열정에서이루어진세번의수리자극시기에따라 G2-1, G2-2, G2-3 으로세분하였고, 추가로주입이전에발생한지진들과포항지진본진및바로전에발생한전진들은각각 G2-0 과 G2-M 으로세분하였다. 각시기별최종위치를 G2 근사평면에투영해보면각세부시기별로지진의발생위치가변화하는것을확인할수있다. G2-0 에속하는지진들중 1개의지진만최종위치가결정되었다. 이지진은 PX-2 와 PX-1 에서수리자극시험이시작되기전에발생하였기때문에수리자극과의연관성은없으나, PX-2 지열정의굴착과정에서상당량의이수누출 (mud loss) 이발생 ( 부록 A-2) 한시기에지진이발생했기때문에밀도가높은고압의이수를주입하는과정에서지진들이유발된것으로보인다. G2-1, G2-2, G2-3, G2-M 으로시기가달라짐에따라순차적으로지진의위치가 G2-0 에서부터남-서쪽그리고깊어지는쪽으로이동하는것을확인할수있다 (Fig. 5-8). 이때 Fig Locations of G2 events projected on the plane approximated by the principal component analysis. Colors of circles represent the occurrence period of earthquakes: G2-0 (yellow), G2-1 (orange), G2-2 (green), G2-3 (blue), and G2-M (purple). Aftershocks of M W 3.2 earthquakes are denoted by open circles. Open square indicates a crossing point of PX-2 borehole and the plane. 91

169 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 G2-2 시기에발생한지진들은상당히넓은범위에분포하고있다. 다만이시기에는본진이전에발생한지진들중가장큰규모인 M W 3.2의지진 ( b) 이발생하였으며, M W 3.2 지진이후에발생한지진들을 M W 3.2의여진으로가정하여위치분포에서제거한후남은지진들의분포만살펴보면시기에따라달라지는지진의위치변화패턴이더명확하게드러난다 참고문헌 Aki, K. and Richards, P.G., 1980, Quantitative Seismology: Theory and Methods. W. H. Freeman and Co., San Francisco, 913 p. Brune, J.N., 1970, Tectonic stress and the spectra of seismic shear waves from earthquakes. Journal of Geophysical Research, 75, Grigoli, F., Cesca, S., Rinaldi, A.P., Manconi, A., Lopez-Comino, J.A., Clinton, J.F., Westaway, R., Cauzzi, C., Dahm, T., and Wiemer, S., 2018, The November 15, 2017 Pogang earthquake: A probable induced event of M W 5.5 in South Korea. Science, 360, Hardebeck, J.L. and Shearer, P.M., 2002, A new method for determining first-motion focal mechanisms. Bulletin of the Seismological Society of America, 92, Kato, A., Fukuda, J.I., Nakagawa, S., and Obara, K., 2016, Foreshock migration preceding the 2016 M W 7.0 Kumamoto earthquake, Japan. Geophysical Research Letters, 43, Kim, K.-H., Ree, J.-H., Kim, Y., Kim, S., Kang, S.Y., and Seo, W., 2018, Assessing whether the 2017 M W 5.4 Pohang earthquake in South Korea was an induced event. Science, 360, Kim, S., Rhie, J., and Kim, G., 2011, Forward waveform modelling procedure for 1-D crustal velocity structure and its application to the southern Korean Peninsula. Geophysical Journal International, 185, Lahr, J.C., 1999, revised 2012, HYPOELLIPSE: a computer program for determining local earthquake hypocentral parameters, magnitude, and first-motion pattern: U.S. Geological Survey Open-File Report Lay, T. and Wallace, T.C., 1995, Modern Global Seismology (Vol. 58). Elsevier. Prejean, S.G. and Ellsworth, W.L., 2001, Observations of earthquake source parameters at 2 km depth in the Long Valley Caldera, Eastern California. Bulletin of the Seismological Society of America, 91, Richter, C.F., 1935, An instrumental earthquake magnitude scale. Bulletin of the Seismological Society of America, 25, Rhee, H.-H. and Sheen, D.-H., 2016, Lateral variation in the source parameters of earthquakes in the Korean Peninsula. Bulletin of the Seismological Society of America, 106,

170 제 5 장지진분석 Shelly, D.R., Beroza, G.C., and Ide, S., 2007, Non-volcanic tremor and low-frequency earthquake swarms. Nature, 446, 305. Sheen, D.-H, Kang, T.-S., and Rhie, J., 2018, A local magnitude scale for South Korea. Bulletin of the Seismological Society of America, 108, Tsuboi, K., Abe, K., Takano, K., and Yamanaka, Y., 1995, Rapid determination of M W from broadband P waveforms. Bulletin of the Seismological Society of America, 85, Waldhauser, F. and 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, Zhang, M. and Wen, L., 2015, An effective method for small event detection: match and locate (M&L). Geophysical Journal International, 200, Zoback, M.L., 1992, First-and second-order patterns of stress in the lithosphere: The World Stress Map Project. Journal of Geophysical Research: Solid Earth, 97,

171

172 제 6 장 지중암반공극압확산분석및지하수변화 Analysis of Pore Pressure Perturbation and Groundwater Change

173 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project 96

174 제 6 장지중암반공극압확산분석및지하수변화 요약 / Abstract 포항지열발전실증시설에서수리자극동안측정된주입압력과주입률자료를이용하여지중암반의수리특성을파악하였다. Jacob 직선법을이용하여추정된수리전도도는대부분 ~ m/s 범위를보였다. 수리모델링을통해추정된수리전도도는 PX-1의경우 m/s, PX-2는 m/s로산출되었으며, 수리확산계수는 PX-1, PX-2 모두 m 2 /s로산출되었다. 지진분석, 영상검층, 커팅시료분석결과에따라두개의단층을 PX-1과 PX-2 주변에놓이도록수리모델을구성하였다. 포항남북지역동시대의암반에발달하는단층핵과단층손상대의샘플을채취하여실내수리시험을수행하여추정된수리특성과수리모델링을이용하여보정한결과를근거로 PX-1과 PX-2 사이의단층 (M W 5.5 지진이발생한단층 ) 은단층손상대의경우수리확산계수를 0.1 m 2 /s, 단층핵은 m 2 /s, PX-1 서쪽의단층은 1 m 2 /s로결정하였다. 수리모델링결과, M W 3.2의지진이발생한시점인 2017년 4월 15일 (3차수리자극이종료된직후 ) PX-2 주변과 PX-1과 PX-2 사이단층에 0.1 MPa 이상의공극압이증가하였다. 포항지진이발생한 2017년 11월 15일에공극압의절대크기는 4월 15일보다감소했지만, M W 5.5 지진이발생한단층을포함한넓은지역에공극압이증가된것으로나타났다. 특히 2017년 9월 18일 5차수리자극이후약 2개월이경과됐음에도불구하고, 0.02 MPa 이상의공극압변화가 PX-1, PX-2 및단층대에발생한것으로분석되었다. 공극압이 0.02~0.06 MPa 만큼증가하는경우지진의발생빈도가커지는것으로분석되었다. 단층이임계응력상태일때 0.01 MPa 이상의 Coulomb 응력변화는지진의발생을증가시키거나많은경우에지진을촉발할수있는것으로알려져있다. PX-2에서의 3차수리자극이후 0.01 MPa 의 Coulomb 응력변화를일으킬정도의공극압의변화가지중암반과단층에서발생한것으로분석되었다. 2018년 8월 6일 PX-1과 PX-2에서측정된지하수위심도는각각 113 m, 740 m였다. PX-1과 PX-2 에서 2018년 8월 31일부터 2019년 2월 28일까지지하수위를자동모니터링한결과, PX-1의경우지하수위는 13.7 m 상승하였고, PX-2는 35.9 m 상승하였다. 지열정개발시의초기지하수위자료는없지만그럼에도불구하고 PX-2의지하수위는지나치게낮다. 또한 4.0 km 하부 PX-1과 PX-2 나공사이의거리 600 m를고려할때두지점의지하수위의차이는비정상적으로크다. 2018년 8월 31일채취한지하수의지화학성분과동위원소 14 C 성분모두 PX-1과 PX-2 사이에서확연히다르게나타났다. 이러한지화학특성과지하수위의차이는 PX-1과 PX-2가단층에의해구분되는서로다른수리환경에놓여있다는가능성을시사한다. 두지열정의수위차이에의한비정상적인수리경사는수리환경의급격한변화를야기할수있다. 이런상황에대비하여향후미소지진및안전성에관한장기적인모니터링과분석이필요하다. This study is to analyze spatiotemporal change of pore pressure caused by water injection under high pressure into the bottom holes of two geothermal wells, PX-1 and PX-2, to develop the enhanced geothermal system in the city of Pohang, Korea. The hydraulic conductivity (K) was estimated with the Jacob straight-line time-drawdown method, which indicated that K ranged from to m/s, reaching the order of m/s only under the higher wellhead pressure. Hydraulic modeling calibration suggested that hydraulic diffusivity of rock formation be m 2 /s. This value served as hydraulic diffusivity of basement rock in the hydraulic models. Based on the results of seismic analysis, image logging, and cutting sample analysis, hydraulic models were constructed with two faults placed near PX-1 and PX-2: the fault between PX-1 and PX-2 and the 97

175 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 fault in the west of PX-1. Hydraulic diffusivities of fault core and damage zone were estimated as m 2 /s and 0.1~1 m 2 /s by laboratory hydraulic tests on fault samples from basement rock in the Pohang area. Hydraulic modeling showed that on April 15, 2017, immediately after the end of the third hydraulic stimulation, pore pressure increased by over 0.1 MPa around the open hole section of PX-2 and in the fault between PX-1 and PX-2. On November 15, 2017, when the M W 5.5 Pohang earthquake occurred, the absolute magnitude of pore pressure became smaller than that of April 15, 2017, but pore pressure as high as 0.02 MPa built up over a larger area of the fault plane between PX-1 and PX-2. It has been shown that Coulomb stress increases of 0.01 MPa are associated with seismicity rate increase and in many cases triggering earthquakes. Water injection under high pressure in particular through PX-2 caused pore pressure, enough to incur the Coulomb failure stress change of 0.01 MPa, to build up in the fault between PX-1 and PX-2. Groundwater depth measured at PX-1 and PX-2 on August 6, 2018 was 113 m and 740 m, respectively. Groundwater level monitored in PX-1 and PX-2 from August 31, 2018 to February 28, 2019 showed that groundwater level increased by 13.7 m in PX-1 and by 35.9 m in PX-2. Even if there is no data on water level in the early stages of the development of geothermal wells, groundwater level in PX-2 is too low. Given the distance of approximately 600 m between the open hole sections of PX-1 and PX-2, the difference in groundwater level at the two points is abnormally large. Geochemical and isotopic 14 C components of groundwater collected from PX-1 and PX-2 on August 31, 2018 were significantly different between them, which suggests, with abnormal difference in groundwater levels, that the hydraulic condition of PX-1 and PX-2 be separated by the fault between them. Abnormal hydraulic gradient between the two geothermal wells is likely to be unstable, which requires long-term monitoring and analysis of groundwater level and micro-seismicity 서론 ( 연구필요성및목적 ) 실증시설의지중에인공저류층을형성하기위해 2016 년 1월 29일부터 2017 년 9월 18일까지지열정 (PX-1, PX-2) 을통해물을주입하였다. 단층이임계응력상태인경우작은공극압의변화에도지진이촉발될수있기때문에수리자극시험을실시한기간과포항지진이발생한시점까지지열정주변지중의공극압변화를시공간적으로분석할필요가있다. 특히 2017 년 9월 18일 5차수리자극종료이후포항지진이발생한 11월 15일까지물을배출하고있는상황이어서압력이해소되고있는중이라는주장이제기되고있다. 따라서이기간동안지중공극압의변화를분석하여이러한주장에대한검증이필요하다. 또한지열정주변의지하수시스템을이해하고자 2017 년 11월 15일포항지진발생전후두지열정과포항지진진앙주변의지하수위및지하수질을분석하였다. 98

176 제 6 장지중암반공극압확산분석및지하수변화 6.2. 지중암반의공극압확산분석 지중암반의수리특성 수리전도도 (K ) 와비저류계수 (Ss) 는공극압의확산정도와속도를결정하기때문에공극압분석에매우 중요한수리특성이다. 해석적인방법과수치적인방법을이용하여포항지열발전실증시설지중암반의수리특 성을산출하였다. 먼저해석적인방법은아래의 Jacob 직선법을이용하였다. 여기서 는시간 에서의수위강하이며, 는 PX-1, PX-2 나공 (open hole section) 의길이이다. 수리전도도는 관계식을이용하여계산하였다. 지중암반의수리전도도는대부분 ~ m/s 범위를보인다 (Fig. 6-1) m/s 이상의수리전도도는수리자극에의해일시적으로상승된값으로서지중암반의일반적인수리전도도로보기어렵다. 물을주입한동일한관정에서압력의변화를측정하는단공시험의경우, Jacob 직선법으로산출된비저류계수는그신뢰도가떨어진다. 따라서 Jacob 직선법대신다음의관계식을이용하여비저류계수를계산하였다. Fig Wellhead pressure and injection rate measured during five hydraulic stimulations are plotted with hydraulic conductivity estimated by the Jacob straight line method. 99

177 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 여기서 ρ는물의밀도, g는중력가속도, α는대수층의압축률, n은공극률, β는물의압축률이다. α는 PX-2 의 4 km 심부지열정굴착시회수된암편시료를이용하여측정된포아송비 (0.21) 와영률 (33.5 GPa) 을이용하여계산한값을, 공극률은동일한시료를이용하여측정된 0.48% 를사용하였다. 위식을이용하여계산된비저류계수는 m -1 이다. 해석적방법외에, 수리모델링을통해수리특성을도출하였다. 수리자극동안측정된주입압력을경계조건으로하고, 수리모델링을통해 PX-1, PX-2에서의주입률 (injection rate) 과주입량 (injection volume) 을계산하였다. 주입률이측정되지않은구간이일부존재하기때문에, 수리모델링으로추정된순주입량은측정주입량에비해크게나타났다. 따라서누적순주입량보다는전체적인주입률의변화를비교하여지중암반의수리특성을추정하였다. 수리모델링을통해추정된수리전도도는 PX-1 의경우 m/s, PX-2 의경우 m/s로산출되었으며, 모두해석적방법으로추정된범위에속한다. 수리확산계수는 PX-1, PX-2 모두 m 2 /s로산출되었다 공극압시공간확산분석 다섯차례의수리자극에의한지중암반에서의공극압변화를분석하기위해수리모델링을수행하였다. 수리모델링은 COMSOL Multiphysics 을이용하였다 (COMSOL, 2018). 수리모델은지중암반을기반으로하여, 두개의단층을 PX-1 과 PX-2 주변에놓이도록구성하였다 (Fig. 6-2). 지진분석결과 PX-1 서쪽의단층은투수성이좋은파쇄대일수있지만, 편의상이후기술에서는단층이라는용어를사용하였다. 두개단층은지진분석결과, 영상검층, 포항지열발전실증연구팀 ( 주관기관넥스지오 ) 이제공한일일시추보고서및지질검층보고서를종합적으로검토하여결정하였다. 단단층의크기는수리모델링목적으로결정한것으로서실제크기와다를수있다. 전체수리모델의크기는 5 km 5 km 5 km로구성하였다. Fig Hydraulic models used in numerical calculation for pore pressure perturbation analysis. The upper left figure shows the hypocenters of earthquakes greater than magnitude of

178 제 6 장지중암반공극압확산분석및지하수변화 PX-1 과 PX-2 사이에위치한단층은 2017 년 11월 15일 M W 5.5의지진이발생한단층으로서주향과경사는 214/43 이다. 이단층을연장하면 PX-2 지열정의약심도 3,810 m를통과한다. 이심도는시추시이수의손실이가장크게발생한지점이다 (Fig. 6-3의왼쪽그림 ). 또한 PX-2 커팅시료분석결과, 3,790~3,815 m 심도에서단층핵 (fault core) 이관찰된다 ( 본요약보고서 2장참조 ). 또한 3,783 m에서장애물로막혀하부심도에대한영상검층을수행할수없었다 (Fig. 6-3의오른쪽그림 ). PX-1 과 PX-2 사이에존재하는단층대가단층핵을갖는경우를 Case A로, 단층핵없이단층손상대 (fault damage zone) 만을갖는경우를 Case B로설정하여수리모델링을수행하였다. Fig Plot of mud loss versus measured depth of PX-2 (left) and the acoustic images obtained at the depth of 3,783 m in PX-2, below which the signal was completely lost (right). PX-1 의서쪽에위치한단층은단층핵없이단층손상대만을갖는것으로수리모델에반영하였다. Kim et al. (2018) 은포항지열발전실증시설사이트의지중암반과동시대의암반에발달하는단층핵과단층손상대의시료를포항남북지역에서채취하여실내수리시험을수행하여단층핵과단층손상대의고유투수계수 (permeability) 를각각 ~10-19 m 2, ~10-11 m 2, 단층핵의비저류계수를 10-9 Pa -1 로제시하였다. 이결과에근거하여, 단층핵의수리확산계수를 m 2 /s, PX-1 과 PX-2 사이의단층손상대는 0.1 m 2 /s로, PX-1 의서쪽단층손상대는 1 m 2 /s로결정하였다 ( 수리자극동안발생한규모 1 이상의지진분석결과를이용하여수리모델을보정한결과 ). 101

179 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 (a) (b) Fig (a) Spatial distribution of pore pressure change (ΔP) along the vertical cross section passing through PX-1 and PX-2 and (b) the isosurface of pore pressure change at 0.02 MPa around PX-1, PX-2, and the fault. 102

180 제 6 장지중암반공극압확산분석및지하수변화 M W 3.2, M W 5.5 지진이발생한 2017 년 4월 15일, 11월 15일에공극압의변화를 Fig. 6-4에도시하였다 년 4월 15일은 PX-2 에서 2017 년 3월 16일부터 4월 14일까지실시된 3차수리자극이종료된직후이다. PX-2 주변과 PX-1 과 PX-2 사이단층에 0.1 MPa 이상의공극압이증가하였다. Case A의경우, 단층핵이불투수경계로작용하기때문에공극압의변화가단층하반에만크게나타난다. 반면단층핵이없는 Case B의경우, 단층상반까지공극압의변화가일어났다 년 11월 15일에는 4월 15일보다공극압의절대크기는감소했지만, PX-1 과 PX-2 사이의단층대를포함하여보다넓은지역에걸쳐공극압이증가되는것으로나타났다. 특히 2017 년 9월 18일 5차수리자극이후약 2개월이경과됐음에도불구하고, 0.02 MPa 이상의공극압변화가 PX-1, PX-2 및단층대에발생한것으로분석되었다 년 4월 15일 M W 3.2, 2017 년 11월 15일 M W 5.5 지진의진원에서의공극압변화를도시하였다 (Fig. 6-5). Case A, B 모두에서수리자극이후공극압이증가하며, 특히 PX-2 에서의 3차물주입이후공극압의변화가가장크게나타났다. 4차, 5차물주입후두진원에서공극압의절대크기는 3차보다감소했지만, 1차물주입이후공극압은추세적으로증가하는것으로나타났다. Fig Pore pressure change with time on the hypocenters of (a) M W 3.2 and (b) M W 차례의수리자극동안발생한규모 1 이상의지진이발생한위치와시점에서의공극압변화를 Fig. 6-6 ( 왼쪽열 ) 에도시하였다. 수리자극동안또는직후에대부분의지진이발생했으며, 그때공극압도대부분 0.02 MPa 이상으로나타났다. 1차수리자극이후 2016 년 3월 12일발생한지진의경우, Case A에서공극압의변화가나타나지않았다. 이는이지진의진원이단층대 (PX-1 과 PX-2 사이 ) 의상부에위치하기때문에 PX-2 에의한공극압이단층핵으로인해전달되지못하기때문이다. 따라서 Case B의경우처럼단층핵이없는경우, 공극압은크게증가한다. 실제단층대에서단층핵이일부사라지면서단층대를통해서공극압이전달될수있기때문에실제수리자극에의한공극압변화범위는 Case A와 B의사이일것으로판단된다. 103

181 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig Pore pressure change at the hypocenters of earthquakes greater than M 1.0 versus the time of occurrence (left), and the histogram of frequency of earthquakes greater than M 1.0 with increasing pore pressure change (right). 공극압이 0.02~0.06 MPa 만큼증가하는경우지진의발생빈도가커지는것으로분석되었다 (Fig. 6-6). 단층이임계응력상태일때 0.01 MPa 이상의 Coulomb 응력변화는지진의발생을증가시키거나지진을촉발할수있는것으로알려져있다 (Reasenberg and Simpson, 1992; Stein, 1999) MPa의 Coulomb 응력변화를일으킬정도의공극압의변화가 PX-2 에서의 3차, 5차수리자극에의해지중암반과단층에서발생한것으로분석되었다. 104

182 제 6 장지중암반공극압확산분석및지하수변화 6.3. 지열정및그주변지하수위변화특성 지진발생전지열발전실증시설주변지하수위 지진 (2017 년 11월 15일 ) 이발생하기전포항지열발전실증시설주변관정의수위는한국지질자원연구원에서지열에너지개발사업을위해포항에서수행한조사보고서 ( 한국지질자원연구원, 2005, 2007, 2008) 와 2012, 2013, 2014, 2015, 2017 년지열발전실증시설반경 5 km 이내의관정에서측정한자료 ( 넥스지오제공 ) 에서취득하였다. 지열발전실증시설주변관정의지하수위 (depth to water (DTW), m) 는 Table 6-1과같다. 100 m 이하의천부관정의수위는 0.3~23.8 m로매우높다. 그러나중간심도 (100~300 m) 관정의경우평균적으로 31 m 내외의수위를보였다 ( 자분정제외 ). 한편심도 300 m 이상의깊은관정에서지하수위의최대심도는 m 이고, 대체로 63 m 정도의수위를보였다 (Table 6-1). 한국지질자원연구원에서조사한관정중가장깊은것은 2012, 2013, 2014 년세번에걸쳐조사된심도 2,383 m의관정으로지열발전실증시설에설치된두지열정의심도 ( 약 4,300 m) 와가장유사하다. 이관정의지하수위는 102.6~130.0 m 범위이며평균수위는 m로나타났다. 심도를모르는관정의경우수위가 0.6~140.9 m의범위를보였다. Table 6-1. Water levels (DTW, m) at the shallow ( 100 m), intermediate (100~300 m), deep (300~1,100 m) and very deep (2,383 m) wells (n=number of wells) before the 2017 Pohang earthquake. Year n Shallow Intermediate Deep Very deep Unknown Max Min Max Min Max Min Max Min Max Min

183 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 한편지열발전실증시설의지열정 (PX-1, PX-2) 의경우지진발생전의수위를확보할수없었다. 지열정의설치과정에서는이수 (drilling fluid) 를지속적으로순환시킴으로써수위의확인이불가능하였으며관정의완성단계에서이수를깨끗한물로치환하였을당시상당기간동안수위가지표부근에서내려가지않은것으로알려진다. 이후수리자극시에는주입과배출에따라수위가관정최상부에머물렀다 (Fig. 6-7). Fig The groundwater level measured at PX-1 and PX-2 before and after M W 5.5 earthquake. The zero level indicates the flow back 지진발생후지하수위변동 지진발생후실증시설에위치한지열정 (PX-1, PX-2) 의수위변화를모니터링하기위해 2018 년 8월 31일자동센서 (vanessen 사의 TD-level logger) 를지열정에각각한개씩설치하였다. 수위와수온을 10분간격으로측정하였으며측정된수위에대해서는기압보정을실시하였다. Table 6-2는현장에서수동수위계및케이블을이용해측정한 PX-1 과 PX-2 의수위이다 년 8월 6일의지하수위는지열정영상검층시에측정한것이다. 수동측정된수위자료를보면지난 207일 ( ~ ) 동안 PX-1(27.9 m 상승 ) 과 PX-2(40.5 m 상승 ) 모두점차적으로수위가상승하였다 (Table 6-2). 한편실증시설주변심부관정의지하수위 (Table 6-1 참조 ) 를고려하였을때 PX-1 의수위는정상적인수위로생각할수있으나 PX-2 는지나치게수위가낮아비정상적인상태로판단된다. 106

184 제 6 장지중암반공극압확산분석및지하수변화 Table 6-2. Measured water levels and level logger installed depth. Date Well 6 Aug Aug Oct Jan Feb 2019 WL (m, DTW) Level logger installed depth (m, bgs) PX PX PX PX 한편자동모니터링에서도두지열정모두지속적인수위상승및수온하강을보였다 (Fig. 6-8). 모니터링기간 ( ~ ) 동안 PX-1 의경우수위가 13.7 m 상승하였고수온은 0.53 하강하였다. 또 PX-2 는수위가 35.9 m 상승하였고수온은 0.21 하강하였다. 수위의상승속도를비교해보면 PX-2 ( m/day, r 2 =0.9981) 가 PX-1 ( m/day, r 2 =0.9971) 보다약 2.6배빠르다. 한편수위가상승함에따라수온이점진적으로하강하는양상을보인다. Fig Water level and water temperature monitored at the PX-1 and PX-2 since August 31, 2018 in the Pohang EGS site. 107

185 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 6.4. 지열정의지하수화학특성변화 M W 3.2 지진발생시동위원소특성 수리자극기간에발생한가장큰규모의지진은 M W 3.2의지진이다. 이지진은 PX-2 에서실시된 3번째수리자극의종료시점인 shut-in 기간에발생하였다. 지진직후 PX-1 과 PX-2 두지열정으로부터물을배출하였으며, PX-1 과 PX-2 에서배출되는물의색깔이확연히차이가났다 (Fig. 6-9). Fig Flow back water from PX-1 (left) and PX-2 (right) after M W 3.2 earthquake. PX-1 과 PX-2 지열정에서배출되는물의수질과안정동위원소의양상이크게다르게나타났다 (Fig. 6-10). PX-2 의경우주입수로사용되었던주변저수지의물과거의비슷한산소수소동위원소성분을보여주는반면, PX-1 에서배출된물의산소수소동위원소는지열수의특징을보여주고있다 (Fig. 6-10). 지열수의산소수소동위원소는지하심부의고온, 고압하에서물암석반응의증가로인해강우기원의물과다른양상을보여준다 (Stefansson et al., 2017; Piña et al., 2018). PX-2 에서배출된물의산소수소동위원소값은지역순환수선 (LMWL) 에좀더가까운값을보인다. 이는 PX-2 수리자극에사용한강우기원의저수지물과지열수와혼합되기때문이다. 108

186 제 6 장지중암반공극압확산분석및지하수변화 Fig The values of δ 18 O and δd in PX-1 and PX-2. PX-1 과 PX-2 지열정에서배출된물과저수지물의지화학분석결과를 Piper diagram 에도시하였다 (Fig. 6-11). PX-2 에수리자극시인근저수지의물을사용하였기때문에 PX-2 에서배출된물과저수지의 지화학성분은비슷하지만 PX-1 에서배출된물은저수지성분과는다른지열수의특성을보여주었다. Fig Piper diagram for flow back water from PX-1 and PX

187 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진발생후지열수의화학특성 2018 년 8 월 PX-1 과 PX-2 지열정에서채취한물은확연히서로다른지화학특성을보여주고있다 (Fig. 6-12). 이러한지화학특성과지하수위의차이 (Fig. 6-7) 는 PX-1 과 PX-2 가서로다른수리환경에있음을시사한다. Fig Stiff Plots for groundwater from PX-1 and PX-2 before and after M W 5.5 earthquake. 환경추적자인탄소동위원소 ( 14 C) 를활용하여지하수유동을평가하기위하여 M W 3.2 지진과 M W 5.5 지진후 PX-1 과 PX-2 지열정의탄소동위원소값을비교하였다 (Fig. 6-13). M W 3.2 지진후인 2017 년 4월에채취한물의탄소동위원소값은 PX-1 과 PX-2 지열정이큰차이가없었으나, M W 5.5 지진후물의탄소동위원소값은 PX-1 과 PX-2 지열정이큰차이를보인다. PX-1 지열정의 14 C 농도는크게증가하여젊은지하수와의혼합을나타내고있으며 PX-2 지열정의 14 C 농도는약간감소하여좀더연령이증가한것으로보인다. 또한 δ 13 C-DIC (Dissolved Inorganic Carbon) 동위원소값도 M W 5.5 지진후 δ 13 C-DIC 동위원소값도 PX-1 과 PX-2 지열정사이에서차이를보였다. Fig The values of 14 C for groundwater from PX-1 and PX-2 before and after the M W 5.5 earthquake. 110

188 제 6 장지중암반공극압확산분석및지하수변화 6.5. 결론 포항지열발전실증연구팀에서제공한주입압력 (wellhead pressure) 과주입률을이용하여심부지중암반의수리확산계수 m 2 /s를해석적인방법과수치적인방법으로계산하였다. 포항남북지역동시대의암반에발달하는단층핵과단층손상대의시료를이용하여수행된실내시험과수리모델링을통해 M W 5.5의지진이발생한단층 (PX-1 과 PX-2 사이에존재 ) 의단층손상대수리확산계수를 0.1 m 2 /s, 단층핵은 m2 /s, PX-1 서쪽의단층은 1 m 2 /s로결정하였다. 지진분석, 영상검층, 커팅시료분석결과에따라두개의단층을 PX-1 과 PX-2 주변에놓이도록수리모델을구성하였다. 수리모델링결과, M W 3.2의지진이발생한시점인 2017 년 4월 15일 PX-2 주변과 PX-1 과 PX-2 사이단층에 0.1 MPa 이상의공극압이증가하였다. 포항지진이발생한 2017 년 11월 15일에공극압의절대크기는 4월 15일보다감소했지만, M W 5.5의지진이발생한단층을포함한넓은지역에공극압이증가된것으로나타났다. 단층이임계응력상태일때 0.01 MPa 이상의 Coulomb 응력변화는지진의발생을증가시키거나많은경우에지진을촉발할수있는것으로알려져있다. PX-2에서의 3차, 5차수리자극에의해 0.01 MPa의 Coulomb 응력변화를일으킬정도의공극압의변화가지중암반과단층에서발생한것으로분석되었다 년 8월 6일 PX-1 과 PX-2 에서측정된지하수위심도는각각 113 m, 740 m였다. 지열정개발시초기지하수위자료가없다고하더라도, PX-2 의지하수위는지나치게낮다. 4.0 km 하부 PX-1 과 PX-2 나공사이의거리 600 m를고려할때두지점의지하수위의차이는비정상적으로크다 년 8월 31일채취한지하수의지화학성분과동위원소 14 C 성분모두 PX-1 과 PX-2 사이에서확연히다르게나타났다. 지화학특성과지하수위의차이는 PX-1 과 PX-2 가단층에의해구분되는서로다른수리환경에놓여있다는가능성을시사한다. 두지열정의수위차이에의한비정상적인수리경사는수리환경의급격한변화를야기할수있다. 이런상황에대비하여향후미소지진및안전성에관한장기적인모니터링과분석이필요하다 참고문헌 한국지질자원연구원, 2005, 심부지열에너지개발사업, 15. 한국지질자원연구원, 2007, 지열수자원실용화기술개발, 한국지질자원연구원, 2008, 지열수자원실용화기술개발, COMSOL Inc., 2018, COMSOL Multiphysics Manual, version 5.4, Stockholm. Kim, J.H., Ree, J.-H., Park, C., Kim, C.-M., Han, R., Shimamoto, T., and Kang, H.-C., 2018, Proxies for the 2017 Pohang earthquake fault and modeling of fluid flow. AGU Fall Meeting Abstract. Piña, A., Donado, L.D., Blake, S., and Cramer, T., 2018, Compositional multivariate statistical analysis of the hydrogeochemical processes in a fractured massif: La Línea tunnel project, Colombia. Applied Geochemistry, 95,

189 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Reasenberg, A.A. and Simpson, R.W., 1992, Response of regional seismicity to the static stress change produced by the Loma Prieta earthquake. Science, 255, Stefansson, A., Hilton, D., Sveinbjornsdottir, A., Torssander, P., Heinemeier, J., Barnes, J., Ono, S., Halldorsson, S., Fiebig, J., and Arnorsson, S., 2017, Isotope systematics of Iceland thermal fluids. Journal of Volcanology and Geothermal Research, 337, Stein, R.S., 1999, The role of stress transfer in earthquake occurrence. Nature, 402,

190 Appendix

191 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Summary Report of the Korean Government Commission on Relations between the 2017 Pohang Earthquake and EGS Project 114

192 Appendix A-1. Acoustic image logging data of PX-1 and PX-2 geothermal wells Fig. A-1-1. PX-1 well structure and acoustic images near the open hole section (from HADES report). The cement shown after the casing section continued from the casing shoe until 4,097 m (measured depth) where the tool stopped. 115

193 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Fig. A-1-2. PX-2 well structure and acoustic image around 1,512 m depth which indicates detection of a hole in casing (from HADES report). The hole matches the casing damage during the 5 th hydraulic stimulation reported by EGS project team. 116

194 Appendix Fig. A-1-3. PX-2 well acoustic image above 3,783 m and complete loss of acoustic signals below 3,783 m (from HADES report). While the PX-1 acoustic signals were obtained below the casing shoe to open hole section, the acoustic signals of PX-2 were not obtainable because the tool stopped at 3,783 m that is 425 m above the casing shoe. 117

195 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 A-2. Mud loss and microseismicity Fig. A-2-1. Mud loss depths and mud density of PX-1 (old), PX-1, and PX-2 wells (above) and temporal distribution of accumulated mud loss and seismicity (below). 118

196 Appendix Fig. A-2-2. Temporal distribution of seismicity plotted on mud loss. 119

197 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 A-3. Earthquake catalog near the EGS site since 2009 to the 2017 Pohang earthquake Event ID Origin time (UTC) (mm/dd/yy hh:mm:ss.sss) Latitude ( N) Longitude ( E) Depth (km) Magnitude M L M w KMA b 11/30/ :52: a 02/02/ :16: a 02/02/ :48: a 02/03/ :08: a 02/04/ :55: a 02/04/ :09: a 02/06/ :11: a 02/06/ :01: a 02/07/ :04: b 02/07/ :04: a 02/07/ :05: a 02/16/ :32: a 02/17/ :43: a 02/18/ :08: a 02/18/ :18: a 03/12/ :25: a 03/28/ :25: a 08/22/ :48: a 12/17/ :42: a 12/17/ :59: a 12/17/ :28: a 12/18/ :06: a 12/18/ :43: a 12/19/ :20: a 12/19/ :18: a 12/19/ :04: a 12/19/ :24:

198 Appendix Event ID Origin time (UTC) (mm/dd/yy hh:mm:ss.sss) Latitude ( N) Longitude ( E) Depth (km) Magnitude M L M w KMA a 12/19/ :02: a 12/20/ :56: a 12/20/ :10: a 12/20/ :22: a 12/20/ :49: a 12/21/ :40: a 12/21/ :07: a 12/21/ :45: a 12/21/ :09: a 12/22/ :53: a 12/22/ :31: a 12/23/ :22: a 12/24/ :51: a 12/24/ :45: a 12/25/ :59: a 12/25/ :13: a 12/25/ :30: a 12/28/ :12: a 12/28/ :46: a 12/29/ :32: a 12/29/ :34: a 12/29/ :35: a 12/29/ :50: a 12/29/ :40: a 01/15/ :13: a 04/08/ :13: a 04/13/ :01: a 04/15/ :17:

199 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 Event ID Origin time (UTC) (mm/dd/yy hh:mm:ss.sss) Latitude ( N) Longitude ( E) Depth (km) Magnitude M L M w KMA a 04/15/ :13: a 04/15/ :31: b 04/15/ :31: c 04/15/ :31: a 04/15/ :32: a 04/15/ :35: a 04/15/ :36: a 04/15/ :11: a 04/15/ :27: a 04/15/ :31: a 04/15/ :12: a 04/15/ :28: a 04/15/ :16: a 04/15/ :47: a 04/15/ :02: a 04/16/ :23: a 04/16/ :44: a 04/16/ :50: a 04/16/ :41: a 04/17/ :23: a 04/20/ :01: a 04/20/ :14: a 04/21/ :51: a 04/21/ :58: a 04/28/ :00: a 04/30/ :37: a 05/06/ :31: a 05/18/ :04:

200 Appendix Event ID Origin time (UTC) (mm/dd/yy hh:mm:ss.sss) Latitude ( N) Longitude ( E) Depth (km) Magnitude M L M w KMA a 08/13/ :42: a 09/11/ :19: a 09/15/ :33: a 09/16/ :55: a 09/16/ :32: a 09/22/ :27: b 09/22/ :27: a 09/22/ :09: a 09/26/ :46: a 11/14/ :55: a 11/14/ :04: a 11/14/ :59: a 11/15/ :22: c 11/15/ :22: b 11/15/ :29:

201 포항지진과지열발전의연관성에관한정부조사연구단요약보고서 A-4. Temporal distribution of earthquakes and EGS project activities Fig. A-4-1. Temporal distribution of EGS project activity and seismicity of events with location certainty. Fig. A-4-2. Temporal distribution of EGS project activity and seismicity of events whose magnitude was determined. 124