--Manuscript Draft-- Pukyong National University Busan, KOREA, REPUBLIC OF. Tae-Seob Kang, Ph.D.

Transcription

1 Seismological Research Letters Development and application of a real-time warning system based on a MEMS seismic network and response procedure for the day of the national college entrance examination in South Korea --Manuscript Draft-- Manuscript Number: Full Title: Article Type: Corresponding Author: SRL-D R1 Development and application of a real-time warning system based on a MEMS seismic network and response procedure for the day of the national college entrance examination in South Korea EduQuakes Tae-Seob Kang, Ph.D. Pukyong National University Busan, KOREA, REPUBLIC OF Corresponding Author Secondary Information: Corresponding Author's Institution: Pukyong National University Corresponding Author's Secondary Institution: First Author: YoungHee Kim, Ph.D. First Author Secondary Information: Order of Authors: YoungHee Kim, Ph.D. Tae-Seob Kang, Ph.D. Junkee Rhie, Ph.D. Order of Authors Secondary Information: Suggested Reviewers: Dong-Hoon Sheen, Ph.D. Associate Professor, Chonnam National University dhsheen@jnu.ac.kr He is a seismologist in Korea. He knows well on the significance of recent Gyeongju earthquake (M 5.8) in Korea. Yih-Min Wu, Ph.D. Professor, National Taiwan University drymwu@ntu.edu.tw He is a leading expert in earthquake early warning. Moreover he is a developer of P- alert system which is used in this study as well. Maurice Lamontagne, Ph.D. Seismologist, Geological Survey of Canada maurice.lamontagne@canada.ca Dr. Maurice reviewed the earlier version of the manuscript at the first round of review and gave many constructive comments. Thus it would be good for him to check this revised version. Powered by Editorial Manager and ProduXion Manager from Aries Systems Corporation

2 Copyright Form Click here to download Copyright Form srl_copyright_form_rev2016_2nd.pdf Seismological Research Letters COPYRIGHT/COLOR--CHARGES FORM FILL OUT, SCAN, AND SUBMIT THIS FORM ONLINE WHEN SUBMITTING YOUR PAPER Manuscript Number: SRL-D [leave blank for new submissions] Development and application of a real-time warning system based on a MEMS seismic network Title: and response procedure for the day of the national college entrance examination in South Korea Authors: YoungHee Kim, Tae-Seob Kang, and Junkee Rhie COPYRIGHT In accordance with Public Law , copyright to the article listed above is hereby transferred to the Seismological Society of America (for U.S. Government employees, to the extent transferable) effective if and when the article is accepted for publication in Seismological Research Letters. The authors reserve the right to use all or part of the article in future works of their own. In addition, the authors affirm that the article has not been copyrighted and that it is not being submitted for publication elsewhere. To be signed by at least one author (who agrees to inform the others, if any) or, in case of "work made for hire," by the employer. Tae-Seob Kang April 9, 2017 Signature for Copyright Print Name (and title, if not author) Date Authorized PUBLICATION CHARGES Page charges: The Seismological Society of America requests that institutions supporting research share the cost of publicizing results of that research. The Editor has discretion to waive publication charges for authors who do not have institutional support. In addition to regular publication charges there is a nominal fee for publishing electronic supplements. Some article types do not incur page charges. Current rates and information about are available at Color options: Color figures can be published (1) in color both in the online journal and in the printed journal, or (2) in color online and in grayscale in print. Online color is free; authors will be charged for color in print. You must choose one option for all of the color figures within a paper; that is, you cannot choose option (1) for one color figure and option (2) for another color figure. You cannot submit two versions of the same figure, one for color and one for grayscale. You are responsible for ensuring that color figures are understandable when converted to gray scale, and that text references and captions are appropriate for both online and print versions. Color figures must be submitted before the paper is accepted for publication. See additional information at Will publication charges be paid? Check only one: BOTH X PAGE CHARGES AND COLOR CHARGES WILL BE PAID. All color figures for this paper will be color both online and in print. This option requires full payment of publication & color charges. ONLY PAGE CHARGES WILL BE PAID. All figures for this paper will be grayscale in print. Color figures, if any, will be color online. REQUEST A REDUCTION IN PAGE CHARGES. Send a letter of request and explanation to the Editor-in-Chief at SRL@seismosoc.org. Color figures, if any, will be color online but grayscale in print. ONLY COLOR CHARGES WILL BE PAID. The article is exempt from page charges, but you wish to print color figures. This option is available only for select SRL article types that are exempt from page charges and requires full payment of color charges. PUBLICATION CHARGES DO NOT APPLY. The article type doesn t incur page charges and there are no color figures. Send Invoice to (required for ALL SRL submissions) : Tae-Seob Kang, Department of Earth and Environmental Sciences, Pukyong National University, 45 Yongso-ro, Nam-gu, Busan 48513, Republic of Korea tskang@pknu.ac.kr; Phone: If your paper is accepted for publication, SSA will require you to fill out and submit an online billing/offprint-order form. General questions about publication charges should be directed to srl@seismosoc.org. For questions about a specific SSA invoice, please contact the SSA Business Office, accounts@seismosoc.org,

3 Abstract 1 Abstract The largest earthquake in South Korea (ML 5.8) since the modern seismograph network began operating in 1978 occurred in Gyeongju on September 12, Because of the low general level of seismic activity in the country, Korean citizens had not expected or thought about how to respond to such a large earthquake and its related aftershocks. After the event, Koreans were left feeling a certain level of apprehension. A temporary micro-electromechanical system (MEMS)- based seismic network was operated at six sites (at high schools) to monitor seismic activities in Gyeongju from October 18 to November 18, The purpose of this temporary real-time warning system was not only to monitor aftershocks but also to provide an onsite earthquake drill in the event of an earthquake on the day of the College Scholastic Ability Test (which is held in November each year). The national college entrance examination is extremely important in Korea, and the government aims to avoid any interruptions to the exam. An earthquake drill manual was established for the first time for an effective response to possible earthquakes occurring on the day of the exam (November 17) in The MEMS network detected several aftershocks during the operating period, but none on the day of the exam. The real-time MEMS warning system represents a new type of earthquake-monitoring system for South Korea. 18 1

4 Manuscript Click here to download Manuscript manuscript_version5_2_tk1_englishcheck.docx Development and application of a real-time warning system based on a MEMS seismic network and response procedure for the day of the national college entrance examination in South Korea 4 5 YoungHee Kim 1, Tae-Seob Kang 2*, Junkee Rhie School of Earth and Environmental Sciences, Seoul National University, Seoul 08826, Republic of Korea 2 Division of Earth Environmental System Science, Pukyong National University, Busan 48513, Republic of Korea *Corresponding author: Tae-Seob Kang, Ph.D. Address: Division of Earth Environmental System Science Pukyong National University 45 Yongso-ro, Nam-gu, Busan 48513, Republic of Korea tskang@pknu.ac.kr Phone: Facsimile:

5 21 Abstract The largest earthquake in South Korea (ML 5.8) since the modern seismograph network began operating in 1978 occurred in Gyeongju on September 12, Because of the low general level of seismic activity in the country, Korean citizens had not expected or thought about how to respond to such a large earthquake and its related aftershocks. After the event, Koreans were left feeling a certain level of apprehension. A temporary micro-electromechanical system (MEMS)- based seismic network was operated at six sites (at high schools) to monitor seismic activities in Gyeongju from October 18 to November 18, The purpose of this temporary real-time warning system was not only to monitor aftershocks but also to provide an onsite earthquake drill in the event of an earthquake on the day of the College Scholastic Ability Test (which is held in November each year). The national college entrance examination is extremely important in Korea, and the government aims to avoid any interruptions to the exam. An earthquake drill manual was established for the first time for an effective response to possible earthquakes occurring on the day of the exam (November 17) in The MEMS network detected several aftershocks during the operating period, but none on the day of the exam. The real-time MEMS warning system represents a new type of earthquake-monitoring system for South Korea. 2

6 38 1. Introduction At 10:44:32 UTC (19:44:32 Korea Standard Time; GMT +9 h) and 11:32:55 UTC (20:32:54 Korea Standard Time) on September 12, 2016, earthquakes with local magnitudes (ML) of 5.1 and 5.8 occurred in the historic city of Gyeongju on the southeastern Korean Peninsula (Fig. 1a; Kim et al., 2016a). These two events (fore- and mainshocks) caused only minor to moderate damage in mostly older buildings including at cultural heritage sites. Despite the lack of deaths and only a few injuries, the whole country was shocked by the earthquakes, which changed a widely held misconception that Korea is an earthquake-free zone. Several aftershocks, including an ML 4.5 earthquake on September 19 at 11:33:58 UTC (20:33:58 Korea Standard Time), followed, clustering around the epicenters of the fore- and mainshocks (Fig. 1a; Kim et al., 2016b). Prior to these events, seismic activity had been low since 1905, when the first seismic station was installed in South Korea. However, historical records document high seismic activity in the Gyeongju region in the 15 th to 18 th centuries (Kim et al., 2016a). At least 10 earthquakes of MMI VIII are considered to have occurred in Gyeongju (Lee and Jin, 1991). A major earthquake (and possibly the largestmagnitude event in the region over the past 2,000 years; Lee and Jin, 1991) occurred in 779, claiming the lives of at least 100 people, according to Samguksagi, the oldest surviving chronicle of Korean history. After the September 2016 earthquakes, residents in Gyeongju and nearby regions experienced small to moderate tremors, not knowing when they would cease. The Korea Metrological Administration (KMA) determines the location and size of all earthquakes that occur on the Korean Peninsula and its surrounding regions and disseminates this information to 3

7 national and international agencies, key relevant facilities, and the general public. However, it does not provide aftershock forecasts for significant earthquakes in Korea, whereas the U.S. Geological Survey (USGS) releases aftershock forecasts following large global earthquakes on an ad-hoc basis (Page et al., 2016). For example, the USGS issued a series of aftershock forecasts following the M7.8 Gorhka, Nepal, earthquake on April 25, The regional earthquake catalog provided by the KMA reported that after the origin time of the ML 5.8 earthquake, 76 aftershocks with an ML greater than 1.5 occurred on September 12, The total number of aftershocks had exceeded 500 by early November (Fig. 1b). Accordingly, there was growing concern about aftershocks in Gyeongju on the day of the national college entrance examination (the College Scholastic Ability Test [CSAT]; Suneung in Korean). The CSAT is an important event for high school students, parents, and educators in South Korea because the exam is offered only once a year, in November, and the results are the determining factor in college admission. Some 600,000 students and college applicants took the exam on November 17, On the day of the CSAT, there is increased use of public transportation services such as buses and metro trains, public organizations adjust their work hours (the starting time is delayed by 1 h to 10 a.m.), and the stock market opens 1 h late. Even take-off and landing of aircraft (except for emergencies) are banned for a period of 25 min during the listening comprehension section of the English exam. Due to nationwide concern about earthquakes in Gyeongju, the Korean Ministry of Education (KME) established the Earthquake Preparedness Advisory Committee, charged with developing a response plan for possible earthquakes in Gyeongju on the day of the CSAT, aiming to prevent any interruptions to the exam. The Committee has published an onsite earthquake drill manual (Table 1; also see Supplementary, available in the electronic supplement 4

8 to this article), which guides examination invigilators and exam candidates in how to immediately respond to sudden ground motions and earthquake-related damage such as cracks in ceilings and walls, and doors and windows that will not open or close. Efforts to develop the manual were featured in the media on the day of the most recent CSAT. Psychologists were stationed at each exam center (high school) in Gyeongju so that, in the event of an earthquake, they could counsel any candidates experiencing anxiety. There were also designated surveyors who inspected the conditions of the buildings at each site. Here, we describe the onsite manual, which outlines three steps for invigilators and exam candidates to follow in the event of an earthquake on the day of the CSAT, as well as the temporary micro-electromechanical system (MEMS) network developed by seismologists to monitor seismic activities from October 18 to November 18, 2016 at the CSAT exam centers Onsite earthquake drill manual The Earthquake Early Warning (EEW) system of the KMA, which was established in 2005, provides the magnitude, origin time, and epicenter of any earthquake exceeding ML 5.0 within ~30 s after detecting a P-wave from 6 or more of the ~150 seismometers in the country. The accuracy of the epicenter location is within ±10 km, and that of the ML is ±0.5. On September 12, 2016, the EEW provided such information for the ML 5.8 earthquake 26 s after its occurrence. However, this would not be particularly useful on the day of the CSAT if the epicenter is near the exam centers. The distance between the exam centers in Gyeongju and the ML 5.8 earthquake of September 12 was only km (Fig. 1a). Furthermore, depending on the epicenter s vicinity and other factors, such as the conditions of exam centers and the seismic resistance of 5

9 buildings, people at different centers might respond differently to sudden ground shaking. There were 2,600 exam candidates in Gyeongju on November 17, Following the September 2016 earthquakes, the advisory committee mentioned above, consisting of seismologists and earthquake engineers, produced the aforementioned earthquake response manual. It has a concise format to minimize any confusion that may arise during an earthquake (Table 1; also see Supplementary). Invigilators are trained to immediately follow the guidelines, whether or not they receive notification from the KMA. The KME distributed the manual to all 85 exam districts, as well as to the general public on November 8, On the day of the CSAT, officials from the KME were placed at the National Earthquake and Volcano Center (NEVC) at the KMA to provide information on any earthquake and the levels of response (Table 1; also see Supplementary) to any such event. In addition, all exam candidates were notified about the manual and informed that they would follow it in the event of an earthquake MEMS-based real-time warning system at six high schools in Gyeongju In addition to the manual, a real-time warning system is needed. We employed a real-time warning system with MEMS for the first time in Gyeongju, encouraged by recent success using an MEMS sensor network for monitoring earthquakes and issuing early warnings in Taiwan (Palert, P-wave Seismic Alarm System; Wu et al., 2013 and references therein). Each P-alert device is equipped with a three-component accelerometer with 16 bit resolution and a dynamic range of gal. The signal from each field station is processed to detect P-wave arrivals and is continuously double-integrated into a displacement signal to calculate the vertical peak amplitude of displacement from the P waves (Wu and Kanamori, 6

10 ). Once the captured signal exceeds the predefined threshold of 25 gal, the P-alert device sends an alert with a warning signal. The real-time warning system was established at six exam centers (all were high schools where power and internet connections were available; Fig. 1a) on October 18, 2016 to monitor seismic activity. These schools were all north of the 2016 September events (only ~10 km from the mainshock location) (Fig. 1a). Each school was equipped with an MEMS sensor on each floor and one monitoring system. On November 7, 2016, the P-alert group from Taiwan visited Gyeongju to install sensors at one of the sites. Data were transferred continuously to a central processing station located in the KME building in Gyeongju. Several aftershocks were recorded by the P-alert system. Figure 2 shows strong motion data from one of the P-alert stations, located approximately 11 km from the epicenter of the Gyeongju earthquakes (Fig. 1a). On November 17, 2016 (the day of the CSAT), two university students (from Pukyong National University and Seoul National University) were assigned to each high school exam site to monitor the real-time seismic signals recorded by the MEMS sensors. Three seismologists were present at the central processing station to monitor all signals from the six sites. No earthquake activity was recorded by the MEMS on this day. The monitoring officially started at 8:40 a.m. and ended at either 5:40 p.m. or 10:00 p.m. (the latter time being for exam candidates with disabilities). Overall, the system was considered a success by both local residents in Gyeongju and the KME. Although there is still room to improve seismic activity monitoring in Gyeongju using MEMS, it can be applied to other intraplate areas where there is a record of historical earthquakes (despite low seismic activity during the modern monitoring period) or areas where aftershocks are likely on important days, such as the day of the CSAT in South Korea. 7

11 Concluding remarks The 2016 Gyeongju earthquake events (including the ML 5.8 event, the largest earthquake in Korea since 1978) have changed perceptions of earthquakes among the Korean public. These events have served as forceful reminders that earthquakes have occurred in the past (based on Korean historical records) and can strike the region again at any time. The onsite earthquake drill manual was established to provide a rapid and effective response to possible earthquakes on the day of the CSAT. In addition, a temporary MEMS network was operated at six high-school exam centers in Gyeongju to monitor seismic activity from October 18 to November 18, 2016, and to effectively inform invigilators and exam candidates how to respond to an earthquake on the day of the CSAT (November 17, 2016). The MEMS network detected several aftershocks during the 1-month operating period but no seismic events on exam day. Our system may be a model for future earthquake monitoring systems in South Korea. It can be used in other intraplate regions and regions where aftershocks are likely Data and Resources The data recorded by the MEMS sensors used in this study are available from the authors upon request. The regional earthquake catalog was provided by the KMA. Plots were generated using Generic Mapping Tools, version (gmt.soest.hawaii.edu; Wessel et al., 2013) Acknowledgements 8

12 The work in this study was supported by the Korea Meteorological Administration Research and Development Program under Grant KMIPA The authors acknowledge the support and cooperation of this work by the Ministry of Education, Gyeongsangbuk-do Office of Education, Gyeongju Office of Education, Geunhwa Girl s High School, Gyeongju High School, Gyeongju Girl s High School, Gyerim High School, Moonhwa High School, and Seondeok Girl s High School, Republic of Korea. The authors deeply appreciate Taiwan P-alert group who helped installing MEMS sensors and supplementary toolkits in one high-school site in Gyeongju. Also, the authors appreciate students from Pukyong National University and Seoul National University for participating in the project. The participated students are ByeongSeok Ahn, Gyeong-Don Chai, Dabeen Heo, Minook Kim, ChangHwan Kong, Heekyoung Lee, Euna Park, and Hyejin Park from Pukyong National University (in alphabetical orders) and Hyunsun Kang, Sungwon Cho, Hyoihn Jang, Juhwan Kim, Sang-Jun Lee, Hobin Lim, Jung-Hun Song, and Jeong-Ung Woo from Seoul National University (in alphabetical orders). The Korea Metrological Administration (KMA) is thanked for the earthquake catalog. The authors would like to thank the reviewer Maurice Lamontagne and the Associate Editor Alan Kafka for their comments that have improved the manuscript References Chang, C., J. B. Lee, and T.-S. Kang (2010). Interaction between regional stress state and faults: Complementary analysis of borehole in situ stress and earthquake focal mechanism in southeastern Korea. Tectonophysics 485(1-4), , doi: /j.tecto

13 Chung, T. W., and W. H. Kim (2000). Fault plane solutions for the June 26, 1997 Kyong-ju Earthquake. Journal of the Korean Geophysical Society 3(4), [in Korean with English abstract] Jo, N. D., and C.-E. Baag (2003). Estimation of spectrum decay parameter k and stochastic prediction of strong ground motions in southeastern Korea. Journal of the Earthquake Engineering Society of Korea 7(6), [in Korean with English abstract] Kim, Y., J. Rhie, T.-S. Kang, K.-H. Kim, M. Kim, and S.-J. Lee (2016a). The 12 September 2016 Gyeongju earthquakes: 1. Observation and remaining questions. Geosciences Journal 20(6), , doi: /s x. Kim, K.-H., T.-S. Kang, J. Rhie, Y. Kim, Y. Park, S. Y. Kang, M. Han, J. Kim, J. Park, M. Kim, C. Kong, D. Heo, H. Lee, E. Park, H. Park, S.-J. Lee, S. Cho, J.-U. Woo, S.-H. Lee, and J. Kim (2016b). The 12 September 2016 Gyeongju earthquakes: 2. Temporary seismic network for monitoring aftershocks. Geosciences Journal 20(6), , doi: /s Lee, K. and Y. G. Jin (1991). Segmentation of the Yangsan Fault System: Geophysical studies on major faults in the Kyeongsang Basin. Journal of the Geological Society of Korea 27, Page, M. T., N. van der Elst, J. Hardebeck, K. Felzer, and A. J. Michael (2016). Three ingredients for improved global aftershock forecasts: Tectonic region, time dependent catalog incompleteness, and intersequence variability. Bulletin of the Seismological Society of America 106(5), , doi: /

14 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. Wessel, P., W. H. F. Smith, R. Scharroo, J. F. Luis, and F. Wobbe, Generic Mapping Tools: Improved version released. EOS, Transactions, American Geophysical Union 94(45), , doi: /2013eo Wu, Y. M., and H. Kanamori (2005). Rapid assessment of damaging potential of earthquakes in Taiwan from the beginning of P-waves. Bulletin of the Seismological Society of America 95, , doi: / Wu, Y. M., D. Y. Chen, T. L. Lin, C. Y. Hsieh, T. L. Chin, W. Y. Chang, W. S. Li, and S. H. Ker (2013). A high-density seismic network for earthquake early warning in Taiwan based on low cost sensors. Seismological Research Letters 84, , doi: /

15 Table 1. Earthquake response guidelines in the event of an earthquake on the day of the College Scholastic Ability Test in South Korea Response Level Level I Response Guidelines In the event of a minor tremor, invigilators do not stop the exam and continue with normal procedures. In certain cases, invigilators can stop the exam temporarily and guide the exam candidates to take cover under their desks. In cases where the tremor is felt but the exam candidates safety is not compromised, the candidates may temporarily take shelter under their desks then restart the exam once the tremor has stopped. Level II In cases where broken glass, fallen materials from the ceiling, damaged lights, cracks on walls, and micro-cracks on columns are witnessed, the room may be evacuated depending on the circumstances of the exam candidates and/or facility. In cases where tremors are strong and the exam candidates safety may be compromised due to structural damage, exam candidates are directed Level III to evacuate the building once the tremors have ended. If damage to exam site facilities are minor and exam candidates are unaffected, the exam can be resumed

16 Figure 1. Map showing the 2016 Gyeongju earthquakes, MEMS network, and histogram of the number of detected earthquakes since the ML 5.8 earthquake. a. Map showing the location of exam centers in Gyeongju, the 2016 Gyeongju earthquakes, and previous seismic events. Thin black lines denote lineaments and faults. Red focal mechanisms with a date show the source mechanisms of the 2016 Gyeongju events ( a for ML 5.1 earthquake, b for ML 5.8 earthquake, and for ML 4.5 earthquake) (Kim et al., 2016a). Gray focal mechanisms indicate past earthquakes in the region (Chung and Kim, 2000; Chang et al., 2010; Jo and Baag, 2003; Park et al., 2007). Circles mark the locations of aftershock events. b. Number of earthquakes since the ML 5.8 event. The catalog is provided by the Korea Metrological Administration and includes aftershock events with a local magnitude

17 Figure 2. Seismic waveform recorded at one high school site by P-alert on November 3,

18 253 Captions for table and figures List of Table Table 1. Earthquake response guidelines in the event of an earthquake on the day of the College Scholastic Ability Test in South Korea List of Figures Fig. 1. Map showing the 2016 Gyeongju earthquakes, MEMS network, and histogram of the number of detected earthquakes since the ML 5.8 earthquake. a. Map showing the location of exam centers in Gyeongju, the 2016 Gyeongju earthquakes, and previous seismic events. Thin black lines denote lineaments and faults. Red focal mechanisms with a date show the source mechanisms of the 2016 Gyeongju events ( a for ML 5.1 earthquake, b for ML 5.8 earthquake, and for ML 4.5 earthquake) (Kim et al., 2016a). Gray focal mechanisms indicate past earthquakes in the region (Chung and Kim, 2000; Chang et al., 2010; Jo and Baag, 2003; Park et al., 2007). Circles mark the locations of aftershock events. b. Number of earthquakes since the ML 5.8 event. The catalog is provided by the Korea Metrological Administration and includes aftershock events with a local magnitude Fig. 2. Seismic waveform recorded at one high school site by P-alert on November 3,

19 Annotated Manuscript Click here to download Annotated Manuscript manuscript_version5_2_tk1_englishcheck_annotated.docx Development and application of a real-time warning system based on a MEMS seismic network and response procedure for the day of the national college entrance examination in South KoreaFirst earthquake response monitoring based on MEMS network on the day of College Scholastic Ability Test in South Korea 5 6 YoungHee Kim 1, Tae-Seob Kang 2*, Junkee Rhie School of Earth and Environmental Sciences, Seoul National University, Seoul 08826, Republic of Korea 2 Division of Earth Environmental System Science, Pukyong National University, Busan 48513, Republic of Korea *Corresponding author: Tae-Seob Kang, Ph.D. Address: Division of Earth Environmental System Science Pukyong National University 45 Yongso-ro, Nam-gu, Busan 48513, Republic of Korea tskang@pknu.ac.kr Phone: Facsimile:

20 23 Abstract The largest earthquake in South Korea (ML 5.8) since the modern seismograph network began operating in 1978 occurred in Gyeongju on September 12, 2016Since the modern seismograph network had been operated in 1978, the largest earthquake (ML 5.8) occurred in Gyeongju, South Korea on 12 September Because of the low general level of seismic activity has been low in the country, Korean citizens had generally neithernot expected nor or thought about how to respond to such a major large earthquake and its related aftershocks. After the event, Koreans were left feeling a certain level of apprehension. A Temporary temporary micro-electromechanical systems (MEMS) )-based seismic network was operated in at six sites (at high schools) to monitor seismic activities in Gyeongju during from October 18 to November 18, The purpose of this temporary real-time warning system in schools is was not only to monitor aftershocks but also to exercise provide an onsite earthquake drill manual in the case event of anpossible earthquake outbreak on the day of the College Scholastic Ability Test (which is offered everyheld in November each year). The national college entrance examination is extremely important in Korea, and the government aims to avoid any interruptions to the exam. The An earthquake drill manual was established for the first time established tofor an effectively respond response to possible earthquakes occurring on the day of the test exam (17 November 17) in 2016). The MEMS network detected a fewseveral aftershocks during the operation period, and no seismic events in the testing sitesbut none on the day of the testexam. The real-time MEMS warning system by MEMS opened represents the possibility of successfully operating a new type of earthquake earthquake-monitoring system for post-earthquake rapid assessment in South Korea. 2

21

22 48 1. Introduction At 10:44:32 UTC (19:44:32 Korea Standard Time; GMT + 9 hoursh) and 11:32:55 UTC (20:32:54 Korea Standard Time) on 12 September 12, 2016, earthquakes with local magnitudes (ML) of 5.1 and 5.8 ML 5.1 and 5.8 earthquakes occurred approximately one hour apart in the historic city of Gyeongju in on the southeastern part of the Korean Peninsula (Fig. 1a; Kim et al., 2016a). These two events (fore- and mainshocks) only caused only small minor to moderate damages in mostly old buildings structures including at cultural heritage sites. Despite the lack of deaths and only a few injuriesregardless of no fatalities (only a few casualties), the whole country was shocked by the earthquakes, these two large events andwhich changed a widely held misconception that Korea is an earthquake earthquake-free zone. Several Aftershocksaftershocks, including an ML 4.5 earthquake on 19 September 19 at 11:33:58 UTC (20:33:58 Korea Standard Time), continued to occurfollowed, clustering around the epicenters of the two eventsfore- and mainshocks (Fig. 1a; Kim et al., 2016b). Prior to these events, Although seismic activities activity had been low in modern monitoring period since 1905, when the first seismic station was installed in South Korea. (the year that the first seismic station was installed in S. Korea), However, historical records document high seismic activity in the Gyeongju region in the 15 th to 18 th centuries (Kim et al., 2016a). At least 10 earthquakes of MMI VIII are considered to have occurred in Gyeongju (Lee and Jin, 1991). A major earthquake (and possibly the largestmagnitude event in the region over the past 2,000 years; Lee and Jin, 1991) occurred in 779, claiming the lives of at least 100 people, according to Samguksagi, the oldest surviving chronicle of Korean history. there are written evidences noting for high seismic activities in the region near/at Gyeongju in 15 th 18 th centuries from the historical literature record (Kim et al., 2016). 4

23 After the September 2016 earthquakes, residents in Gyeongju and nearby regions had experienced small to moderate tremors, not knowing when they would cease without knowing when they would cease. The Korea Metrological Administration (KMA) determines the location and size of all earthquakes that occur on the Korean Peninsula and its surrounding regions and disseminates this information to national and international agencies, key relevant facilities, and the general public. However, it does not provide aftershock forecasts for significant earthquakes in Korea, whereas the U.S. Geological Survey (USGS) releases aftershock forecasts following large global earthquakes on an ad-hoc basis (Page et al., 2016). For example, the USGS issued a series of aftershock forecasts following the M7.8 Gorhka, Nepal, earthquake on April 25, The regional earthquake catalog provided by the Korea Metrological Administration (KMA) reports reported that since after the origin time of the ML 5.8 earthquakes, the 76 aftershocks with an MLlocal magnitude greater than 1.5 were reportedoccurred on 12 September 12, The cumulative total number of the aftershocks had exceeded 500 on by early November (Fig. 1b). Accordingly, there was a growing concern about aftershocks in Gyeongju on the day for of the national college- entrance examination (the College Scholastic Ability Test [CSAT]; Suneung (in Korean); the College Scholastic Ability Test (CSAT)). The college entrance examinationcsat is considered to be an upmost agendaimportant event for high- school students, parents, and educators in S. outh Korea because this the examination is only offered only once a year, in November, and this the test results is are the determining factor in the college admission. On 17 November 2016, there were ssome 600,000 students and college applicants who took this testthe exam on November 17, in South Korea. On the day of the CSAT day, there is increased use of public transportation services, such as buses and subwaysmetro trains, are expanded for CSAT, along withpublic organizations 5

24 adjustments in their work hours (the starting time is delayed by 1 hour to 10 o clocka.m.) among public organizations, and the stock market opens 1 h late.. Even aircraft take off/ and landing of aircraft (except for emergency aircrafties) is strictlyare banned for a period of 25 min by law during a shortthe listening comprehension session section of the college-entranceenglish exam (for 25 minutes). Due to Considering such nationwide concern for about earthquakes in Gyeongju, the Korean Ministry of Education of Korea(KME) called seismologists and earthquake engineers to be inestablished the Earthquake Preparedness Advisory Committee, charged with developing a response plan for possible earthquakes in Gyeongju on the day of the CSAT day, aiming to prevent any interruptions to the exam. The committee Committee newly has developed published the an onsite earthquake drill manual (Table 1; also see Supplementary, available in the electronic supplement to this article), which guides examination invigilators and test takersexam candidates in how to immediately respond to sudden ground motions and earthquake-related damage such as unusual or new cracks in ceilings and walls, and doors and windows that would will not open or shutclose, etc. Efforts for to developing such the manual for earthquakes were featured in news the media and also newspaper articles on the day of the most recent CSAT day. Special ppsychologists were placed stationed in at each testing siteexam center (high school) in Gyeongju so that, in the event of an earthquake, they could counsel any candidates experiencing anxiety.to cope with test takers who may feel anxieties in the case of earthquake outbreak, and There were also designated person surveyors who inspects inspected the conditions of the buildings facility inat each site. In this articlehere, we describe (1) the onsite manual, which concisely describesoutlines three steps that for invigilators and test takersexam candidates can to follow in the case event of the an earthquake on the day of the CSAT, and also (2)as well as the 6

25 temporary micro-electro-mechanical systems (MEMS) network that developed by seismologists put together to monitor seismic activities during from October 18 to November 17, 2016 in at the CSAT testing sitesexam centers Onsite earthquake drill manual The Earthquake Early warning Warning (EEW) system in of the Korea Metrological Administration (KMA), which was established in 2005, provides the earthquake magnitude, origin time, and epicenter of any earthquake within ~30 seconds after the earthquake origin time when an earthquake with the local magnitude exceeding ML 5.0 within ~30 s after detecting a P- wave from 6 or more of the ~150 seismometers in the countryoccurs in S. Korea. The accuracy of the epicenter location is within ±10 km, and that of the ML is ±0.5. On September 12, 2016, the EEW provided such information for the ML 5.8 earthquake 26 s after its occurrence. However, This this wouldseismological information may not be critically partticularly useful on the day of the CSAT if the epicenter is too close tonear the testing sitesexam centers. The distance between the testing sitesexam centers in Gyeongju and the ML 5.8 earthquake of September 12 is was only ~10-13 km (Fig. 1a). Furthermore, depending on the epicenter s vicinity and various other factors (, such as the conditions of exam centers and the seismic resistance of buildings, people at different centers might respond distance between the epicenter and the testing site, site condition, seismic performance of the building, etc.), test takers and invigilators in different sites in Gyeongju would differently respond to the sudden ground shaking. On 17 November 2016, tthere were 2,600 test takersexam candidates in Gyeongju on November 17,

26 Following the September 2016 earthquakes, the advisory committee mentioned above, consisting of seismologists and earthquake engineers, produced the aforementioned earthquake response manual. It has a concise format to minimize any confusion that may arise during an earthquake (Table 1; also see Supplementary). Invigilators are trained to immediately follow the guidelines, whether or not they receive notification from the KMA. The KME distributed the manual to all 85 exam districts, as well as to the general public on November 8, On the day of the CSAT, officials from the KME were placed at the National Earthquake and Volcano Center (NEVC) at the KMA to provide information on any earthquake and the levels of response (Table 1; also see Supplementary) to any such event. In addition, all exam candidates were notified about the manual and informed that they would follow it in the event of an earthquake. After the September 2016 earthquakes, the advisory committee consisting of seismologists and earthquake engineers was formed to establish a new onsite earthquake drill manual in the case of earthquakes on the CSAT day. This manual guides invigilators and test takers to immediately respond to sudden strong ground motions and earthquake-related damages. It is prepared concisely to minimize confusion that may arise during the earthquake. With or without the earthquake notification from KMA, the invigilators are trained to immediately follow the guideline shown in Table 1, during the earthquake MEMS-based real-time warning system in six high schools in Gyeongju 162 8

27 In addition to the onsite earthquake drill manual, a real-time warning system is needed to effectively notify invigilators and test takers what to do in the case of earthquakes on the day of CSAT. We employed the a real-time warning system with MEMS for the first time in Gyeongju, encouraged by recent success of utilizingusing an MEMS sensor network for monitoring earthquakes and issuing early warning in Taiwan (P-alert;, P-wave Seismic Alarm System) ; for monitoring earthquakes and issuing early warning in Taiwan (Wu et al., 2013 and references therein). Each P-alert device is equipped with a three-component accelerometer with 16-bit resolution and a dynamic range of gal. According to the software algorithm embedded in the P-alert, tthe signal from each field station is processed for theto detection of P-wave arrivals and is continuously double double-integrated into the a displacement signal for to calculating calculate the vertical peak amplitude of displacement from the P waves (Wu and Kanamori, 2005). Once the captured signal exceeds the predefined thresholdsthreshold of 25 gal, the P-alert device sends an alert with a warning soundsignal. In 18 October 2016, tthe real-time warning system with MEMS was established in at six testing sitesexam centers (all were high schools where power and internet connections are providedwere available; Fig. 1a), ) on October 18, 2016 to monitor seismic activitiesactivity. These schools are all situated north of the 2016 September events (only ~10- km distance from the mainshock location) (Fig. 1a). Each school is was equipped with an MEMS sensor in on each floor and one monitoring system. On 7 November 7, 2016, the low-cost earthquake early warning (EEW) system (P-alert) group from Taiwan visited Gyeongju and assisted withto installing sensors in at one of the sites. Data from these P-alert systems are were transferred continuously to a central processing station located, which is set in the KME building of 9

28 Education of Ministry in Gyeongju. A few aftershock events were recorded in the P-alert system, and Fig. ure 2 shows strong- motion data from one of the P-alert station, locatedwhich is about approximated 11 km away from the epicenter of the Gyeongju earthquakes (Fig 1a). On 17 November 17, 2016 (the day of the CSAT), two university students (from Pukyong National University and Seoul National University) were assigned in to each high- school exam site to monitor the real-time seismic signals recorded in by the MEMS sensors. Three seismologists were present in at the site for central processing station to monitor all signals from the six testing sites. There were nno earthquake activities activity was recorded by the MEMS on this day. The monitoring officially started at 8:40 a.m. and ended at two times (either 5:40 p.m. and or 10:00 p.m.;. (the latter time being for test takersexam candidates with disabilities). Overall, the system was considered a success by both local residents in Gyeongju and the KME. Although there is still room to improve seismic activity monitoring in Gyeongju using MEMS, it can be applied to other intraplate areas where there is a record of historical earthquakes (despite low seismic activity during the modern monitoring period) or areas where aftershocks are likely on important days, such as the day of the CSAT in South Korea Concluding remarks The 2016 Gyeongju earthquake events (including the ML 5.8 event, the largest earthquake in Korea since 1978) have changed perceptions of earthquake that mostamong the Korean people hadpublic. These events now becomehave served as forceful reminders that earthquakes have occurred in the past (based on Korean historical literature datarecords) and can hit strike the region again at any time. The Onsite onsite earthquake drill manual was newly established to 10

29 provide a rapid immediately and effectively respond response to possible earthquakes on the day of the College Scholastic Ability Test (CSAT). In addition, a temporary MEMS network was operated at six high-school testing sitesexam centers in Gyeongju to monitor seismic activities activity during from October 18- to November 18, 2016, and to effectively notify inform invigilators and test takersexam candidates what how to do respond in the case ofto an earthquakes on the day of the CSAT (17 November 17, 2016). The MEMS network detected a fewseveral aftershocks during the operation period of one1 month operating period, and but no seismic events on exam dayin the testing sites on 17 November Our system may be a model for future earthquake monitoring systems in South Korea. It can be used in other intraplate regions and regions where aftershocks are likely.the first earthquake response monitoring based on MEMS on the day of CSAT opened a possibility of operating a new earthquake monitoring system for real-time early warning in South Korea Data and Resources The Data data recorded by the MEMS sensors used in this study can be obtainedare available from the authors upon request. The regional earthquake catalog is was provided by Korea Meteorological Administration (the KMA). Plots were made generated using the Generic Mapping Tools, version (gmt.soest.hawaii.edu ; Wessel et al., 2013) Acknowledgements

30 The work in this study was supported by the Korea Meteorological Administration Research and Development Program under Grant KMIPA The authors deeply appreciate Taiwan P-alert group who helped installing MEMS sensors and supplementary toolkits in one highschool site in Gyeongju. Also, the authors appreciate students from Pukyong National University and Seoul National University for participating in the project. The participated students are ByeongSeok Ahn, Gyeong-Don Chai, Dabeen Heo, Minook Kim, ChangHwan Kong, Heekyoung Lee, Euna Park, and Hyejin Park from Pukyong National University (in alphabetical orders) and Hyunsun Kang, Sungwon Cho, Hyoihn Jang, Juhwan Kim, Sang-Jun Lee, Hobin Lim, Jung-Hun Song, and Jeong-Ung Woo from Seoul National University (in alphabetical orders). The Korea Metrological Administration (KMA) is thanked for the earthquake catalog. The authors would like to thank the reviewer Maurice Lamontagne and the Associate Editor Alan Kafka for their comments that have improved the manuscript.finally, the authors thank Korea Metrological Administration (KMA) for the earthquake catalog References Chang, C., J. B. Lee, and T.-S. Kang (2010). Interaction between regional stress state and faults: Complementary analysis of borehole in situ stress and earthquake focal mechanism in southeastern Korea. Tectonophysics 485(1-4), , doi: /j.tecto Chung, T. W., and W. H. Kim (2000). Fault plane solutions for the June 26, 1997 Kyong-ju Earthquake. Journal of the Korean Geophysical Society 3(4), [in Korean with English abstract] 12

31 Jo, N. D., and C.-E. Baag (2003). Estimation of spectrum decay parameter k and stochastic prediction of strong ground motions in southeastern Korea. Journal of the Earthquake Engineering Society of Korea 7(6), [in Korean with English abstract] Kim, Y., J. Rhie, T.-S. Kang, K.-H. Kim, M. Kim, and S.-J. Lee (2016a). The 12 September 2016 Gyeongju earthquakes: 1. Observation and remaining questions. Geosciences Journal 20(6), , doi: /s x. Kim, K.-H., T.-S. Kang, J. Rhie, Y. Kim, Y. Park, S. Y. Kang, M. Han, J. Kim, J. Park, M. Kim, C. Kong, D. Heo, H. Lee, E. Park, H. Park, S.-J. Lee, S. Cho, J.-U. Woo, S.-H. Lee, and J. Kim (2016b). The 12 September 2016 Gyeongju earthquakes: 2. Temporary seismic network for monitoring aftershocks. Geosciences Journal 20(6), , doi: /s Lee, K. and Y. G. Jin (1991). Segmentation of the Yangsan Fault System: Geophysical studies on major faults in the Kyeongsang Basin. Journal of the Geological Society of Korea 27, Page, M. T., N. van der Elst, J. Hardebeck, K. Felzer, and A. J. Michael (2016). Three ingredients for improved global aftershock forecasts: Tectonic region, time dependent catalog incompleteness, and intersequence variability. Bulletin of the Seismological Society of America 106(5), , doi: / 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. 13

32 Wessel, P., W. H. F. Smith, R. Scharroo, J. F. Luis, and F. Wobbe, Generic Mapping Tools: Improved version released. EOS, Transactions, American Geophysical Union 94(45), , doi: /2013eo Wu, Y. M., and H. Kanamori (2005). Rapid assessment of damaging potential of earthquakes in Taiwan from the beginning of P-waves, Bulletin of the Seismological Society of America 95, , doi: / Wu, Y. M., D. Y. Chen, T. L. Lin, C. Y. Hsieh, T. L. Chin, W. Y. Chang, W. S. Li, and S. H. Ker (2013). A high-density seismic network for earthquake early warning in Taiwan based on low cost sensors. Seismological Research Letters 84, , doi: /

33 Table 1. Earthquake response guidelines in the event of an earthquake on the day of the College Scholastic Ability Test in South KoreaGuideline in the case of earthquakes on the day of College Scholastic Ability Test in South Korea. 287 Response Response GuidelinesDescription Level Level 1I In the event of a minor tremor, invigilators do not stop the exam and continue with normal procedures. In certain cases, invigilators can stop the exam temporarily and guide the exam candidates to take cover under their desks.invigilators do not stop the exam regardless of experiencing minor tremor. In certain case, invigilators can stop the exam temporarily and let test takers hide under the desk. Level 2II In cases where the tremor is felt but the exam candidates safety is not compromised, the candidates may temporarily take shelter under their desks then restart the exam once the tremor has stopped. In cases where broken glass, fallen materials from the ceiling, damaged lights, cracks on walls, and micro-cracks on columns are witnessed, the room may be evacuated depending on the circumstances of the exam candidates and/or facility.invigilators stop the exam temporarily and let test takers hide under the desk while experiencing tremors. If the earthquake stops, the invigilators can restart the exam. Test takers can go outside the classroom if there are structural damages (such as glass shattering, ceiling fallout, ceiling light damage, wall crack, and minor cracks on the pillars) are identified. Level 3III In cases where tremors are strong and the exam candidates safety may be compromised due to structural damage, exam candidates are directed to 15

34 evacuate the building once the tremors have ended. If damage to exam site facilities are minor and exam candidates are unaffected, the exam can be resumed.invigilators direct test takers to go outside the building and stay in the school field due to severe shaking and expected large structural damages. Test takers can return back the classroom and take the exam if no major structural damages are identified

35 Fig. 1. Map showing the 2016 Gyeongju earthquakes, MEMS network, and histogram of the number of detected earthquakes since the ML 5.8 earthquake. a. Map showing the location of testing sitesexam centers in Gyeongju, the 2016 Gyeongju earthquakes, and previous seismicityseismic events. Thin black lines denote the lineaments and faults. Red beach ballsfocal mechanisms with the a date labeled show the source mechanisms of the 2016 Gyeongju events ( a for ML 5.1 earthquake, b for ML 5.8 earthquake, and for ML 4.5 earthquake)(kim et al., 2016a). Gray-colored beach balls focal mechanisms show indicate past earthquakes in the region (Chung and Kim, 2000; Chang et al., 2010; Jo and Baag, 2003; Park et al., 2007). Open ccircles are mark the locations of aftershock events. b. The number of earthquakes since the ML 5.8 event. The catalog is provided by Korea Metrological Administration (KMA). The catalog and includes aftershock events with a local magnitude 1.5. The CSAT day is on 67 th day since 12 September