|
|
|
| Rapid magnitude estimation for earthquake early warning based on SVM |
| ZHU Jingbao1,2, SONG Jindong1,2, LI Shanyou1,2 |
1. Institute of Engineering Mechanics, China Earthquake Administration, Harbin 150080, China;
2. Key Laboratory of Earthquake Engineering and Engineering Vibration， China Earthquake Administration, Harbin 150080, China |
|
|
|
|
Abstract In order to more accurately estimate earthquake magnitude, using Japanese K-Net strong earthquake data within 3 seconds after P-wave triggering, 12 P-wave characteristic parameters including 4 kinds of amplitude parameter, period one, energy one and derivative one were selected as input to construct the prediction model of earthquake magnitude based on support vector machine for earthquake magnitude estimation (SVM-M). Results showed that compared with the traditional earthquake magnitude estimation “τc method” and “Pd method”, the prediction error of the established SVM-M model is obviously reduced and not affected by variation of epicentral distance, and the overestimation of small earthquake is obviously improved. The case analysis results of main earthquake of 2016 Kumamoto earthquake (Mj7.3) in Japan and 2008 Wenchuan earthquake (Ms 8.0) in China showed that the 3-second time window can’t match the whole process of focal rupture, and the seismic magnitude is underestimated to a certain extent, but what happening can be determined as a large earthquake event within the short time window after P-wave triggering; the established SVM-M model can be used to rapidly estimate the seismic magnitude for earthquake early warning.
|
|
Received: 16 April 2020
Published: 15 April 2021
|
|
|
|
［1］宋晋东. 高速铁路运行控制用地震动参数及单台地震预警技术研究［D］.哈尔滨: 中国地震局工程力学研究所, 2013.
［2］NAKAMURA Y. On the urgent earthquake detection and alarm system (UrEDAS)［C］∥9th World Conference on Earthquake Engineering. Tokyo： WCEE, 1988.
［3］ALLN R M, KANAMORI H. The potential for earthquake early warning in southern California［J］. Science, 2003, 300(5620): 786-789.
［4］OLSON E L, ALLEN R M. The deterministic nature of earthquake rupture［J］. Nature, 2005, 438(10): 212-215.
［5］KANAMORI H. Real-time seismology and earthquake damage mitigation［J］. Annual Review of Earth and Planetary Science, 2005, 33: 195-214.
［6］WU Y M, KANAMORI H. Experiment on an onsite early warning method for the Taiwan early warning system［J］. Bulletin of the Seismological Society of America, 2005, 95(1): 347-253.
［7］PENG C Y, YANG J S, ZHENG Y, et al. New τc regression relationship derived from all P-wave time windows for rapid magnitude estimation［J］. Geophysical Research Letters, 2017, 44(4): 1724-1731.
［8］ZIV A. New frequency-based real-time magnitude proxy for earthquake early warning［J］. Geophysical Research Letters, 2014, 41(20): 7035-7040.
［9］WU Y M, ZHAO L. Magnitude estimation using the first three seconds P-wave amplitude in earthquake warning［J］. Geophysical Research Letters, 2006, 33(16): 271-284.
［10］ZOLLO A, LANCIERI M, NIELSEN S. Earthquake magnitude estimation from peak amplitudes of very early seismic signals on strong motion records［J］. Geophysical Research Letters, 2006, 33(23): L23312.
［11］FESTA G, ZOLLO A, LANCIERI M. Earthquake magnitude estimation from early radiated energy［J］. Geophysical Research Letters, 2008, 35(22): 113-130.
［12］WANG Z J， ZHAO B M. A new Mw estimation parameter for use in earthquake early warning system［J］. Journal of Seismology, 2017, 22(1)： 1-11.
［13］WANG W T, NI S D, CHEN Y, et al. Magnitude estimation for early warning applications using the initial part of P waves: a case study on the 2008 Wenchuan sequence［J］. Geophysical Research Letters, 2009, 36(16): L16305.
［14］金星, 张红才, 李军，等. 地震预警震级确定方法研究［J］.地震学报, 2012, 34(5): 593-610.
JIN Xing, ZHANG Hongcai, LI Jun, et al. Research on earthquake early warning magnitude estimate［J］. Acta Seismological Sinica, 2012, 34(5): 593-610.
［15］PENG C Y, YANG J S, ZHENG Y, et al. A τc magnitude estimation of the 20 April 2013 Lushan earthquake, Sichuan, China［J］. Science China Earth Sciences, 2014, 57(12): 3118-3124.
［16］PENG C Y, YANG J S, XUE B, et al. Exploring the feasibility of earthquake early warning using records of the 2008 Wenchuan earthquake and its aftershocks［J］. Soil Dynamic and Earthquake Engineering, 2014, 57: 86-93.
［17］马强. 地震预警技术研究及应用［D］. 哈尔滨: 中国地震局工程力学研究所， 2008.
［18］REDDY R,NAIR R R. The efficacy of support vector machines(SVM) in robust determination of earthquake early warning magnitudes in central Japan［J］. Journal of Earth System Science, 2013, 122(5): 1423-1434.
［19］宋晋东, 教聪聪, 李山有, 等. 一种基于地震早期辐射P波能量的高速铁路Ⅰ级地震警报预测方法［J］. 振动与冲击, 2018, 37(19): 22-30.
SONG Jindong, JIAO Congcong, LI Shanyou, et al. A predicting method for magnitude 1 earthquake alarm of high-speed railways based on seismic early radiated P-wave energy［J］. Journal of Vibration and Shock, 2018, 37(19): 22-30.
［20］马强, 金星, 李山有, 等. 用于地震预警的P波震相到时自动拾取［J］. 地球物理学报, 2013, 56(7): 2313-2321.
MA Qiang, JIN Xin, LI Shanyou, et al. Automatic P-arrival detection for earthquake early warning［J］. Chinese Journal of Geophysics, 2013, 56(7): 2313-2321.
［21］TEZCAN J, CHENG Q.Support vector regression for estimating earthquake response spectra［J］. Bulletin of Earthquake Engineering, 2012, 10(4): 1205-1219.
［22］彭朝勇, 杨建思, 薛兵, 等. 基于汶川主震及余震的预警参数与震级相关性研究［J］. 地球物理报, 2013, 56(10): 3404-3415.
PENG Chaoyong, YANG Jiansi, XUE Bing, et a1. Research on correlation between early-warning parameters and magnitude for the Wenchuan Earthquake and its aftershocks［J］. Chinese Journal of Geophysics, 2013, 56(10): 3404-3415.
［23］YAMADA M, HEATON T, BECK J. Real-time estimation of fault rupture extent using near-source versus far-source classification［J］. Bulletin of the Seismological Society of America, 2007, 97(6): 1890-1910.
［24］HUANG P L, LIN T L, WU Y M. Application of τ*cPd in earthquake early warning［J］. Geophysical Research Letters, 2015, 42(5): 1403-1410.
［25］REED J W, KASSAWARA R P. A criterion for determining exceedance of the operating basis earthquake［J］. Nuclear Engineering & Design, 1988, 123(2/3): 387-396.
［26］NAKAMURA Y. A new concept for the earthquake vulnerability estimation and its application to the early warning system ［M］. Berlin Heidelberg: Springer Berlin Heidelberg, 2003.
［27］SAUNDERS C, STITSON M O, WESTON J, et al. Support vector machine［J］. Computer Science, 2002, 1(4): 1-28.
［28］李娇. 支持向量机参数优化研究［D］. 武汉: 华中师范大学, 2011.
［29］VAPNIK V N. The nature of statistical learning theory［M］. New York: Springer, 1995.
［30］STEINWART I, HUSH D, SCOVEL C. Learning from dependent observations［J］. Journal of Multivariate Analysis, 2009, 100(1): 175-194.
［31］CHERKASSKY V, MA Y. Selection of meta-parameters for support vector regression［J］. Artificial Neural Networks, 2002, 2415: 687-693.
［32］CHAPELLE O, VAPNIK V, BOUSQUET O, et al. Choosing multiple parameters for support vector machines［J］. Machine Learning, 2002, 46(1/2/3): 131-159.
［33］TSUBOI, CHUJI. Determination of the gutenberg-richter’s magnitude of earthquakes occurring in and near Japan［J］. Zisin II, 1954, 7(3): 185-193.
［34］BORMANN P, LIU R F, REN X, et al. Chinese national network magnitudes, their relation to NEIC magnitudes, and recommendations for new IASPEI magnitude standards［J］. Bulletin of the Seismological Society of America, 2007, 97(1): 114-127.
［35］王卫民, 赵连峰, 李娟, 等. 四川汶川8.0级地震震源过程［J］. 地球物理学报, 2008, 51(5): 1403-3415.
WANG Weimin, ZHAO Lianfeng, LI Juan, et al. Rupture process of the Ms8.0 Wenchuan earthquake of Sichuan, China［J］. Chinese Journal of Geophysics, 2008, 51(5): 1403-3415.
［36］FESTA G, ZOLLO A, LANCIERI M. Earthquake magnitude estimation from early radiated energy［J］. Geophysical Research Letters, 2008, 35(22): L22307.
|
|
|
|
