A Prediction Method of Ground Motion for Regions without Available Observation Data (LGB-FS) and Its Application to both Yangbi and Maduo Earthquakes in 2021

Abstract

Currently available earthquake attenuation equations are locally applicable, and methods based on observation data are not applicable in areas without available observation data. To solve the above problems and further improve the prediction accuracy of ground motion parameters, we present a prediction model referred to as a light gradient boosting machine with feature selection (LGB-FS). It is based on a light gradient boosting machine (LightGBM) constructed using historical strong motion data from the NGA-west2 database and can quickly simulate the distribution of strong motion near the epicenter after an earthquake. Cases study shows that compared with GMPE methods and those based on real-time observation data, the model has a better prediction effect in areas without available observation data and can be applied to Yangbi Earthquake and Maduo Earthquake. The feature importance evaluation based on both information gains and partial dependence plots (PDPs) reveals the complex relationships between multiple factors and ground motion parameters, allowing us to better understand their mechanisms and connections.

Keywords

LightGBM / ground motion parameters / NGA-west2 / feature selection / Yangbi / Maduo / earthquakes

References

[1] Abrahamson N A, Silva W J, Kamai R. Summary of the ASK14 Ground Motion Relation for Active Crustal Regions. Earthquake Spectra, 2014, 30(3): 1025-1055.

[2] Abrahamson N, Silva W. Summary of the Abrahamson & Silva NGA Ground-Motion Relations. Earthquake Spectra, 2008, 24(1): 67-97.

[3] Akkar S, Sandıkkaya M A, Bommer J J. Empirical Ground-Motion Models for Point- and Extended-Source Crustal Earthquake Scenarios in Europe and the Middle East. Bulletin of Earthquake Engineering, 2014, 12(1): 359-387.

[4] Alavi A H, Gandomi A H. Prediction of Principal Ground-Motion Parameters Using a Hybrid Method Coupling Artificial Neural Networks and Simulated Annealing. Computers & Structures, 2011, 89(23): 2176-2194. 24

[5] Alavi A H, Gandomi A H, Modaresnezhad M, . New Ground-Motion Prediction Equations Using Multi Expression Programing. Journal of Earthquake Engineering, 2011, 15(4): 511-536.

[6] Ambraseys N N, Douglas J. Near-Field Horizontal and Vertical Earthquake Ground Motions. Soil Dynamics and Earthquake Engineering, 2003, 23(1): 1-18.

[7] Ambraseys N N, Simpson K A, Bommer J J. Prediction of Horizontal Response Spectra in Europe. Earthquake Engineering & Structural Dynamics, 1996, 25(4): 371-400.

[8] Aptikaev, F., Kopnichev, J., 1980. Correlation between Seismic Vibration Parameters and Type of Faulting. In: Proceedings of Seventh World Conference on Earthquake Engineering. September 8–13, 1980, Istanbul

[9] Boore D M, Stewart J P, Seyhan E, . NGA-West2 Equations for Predicting Response Spectral Accelerations for Shallow Crustal Earthquakes, 2013, Berkeley: Pacific Earthquake Engineering Research Center, University of California

[10] Bozorgnia Y, Abrahamson N A, Atik L A, . NGA-West2 Research Project. Earthquake Spectra, 2014, 131(3): 409-444

[11] Breiman L, Friedman J H, Olshen R A, . Classification and Regression Trees (CART). Biometrics, 1984, 40(3): 874

[12] Campbell K W. Strong Motion Attenuation Relations: A Ten-Year Perspective. Earthquake Spectra, 1985, 1 4 759-804.

[13] Campbell K W, Bozorgnia Y. NGA Ground Motion Model for the Geometric Mean Horizontal Component of PGA, PGV, PGD and 5% Damped Linear Elastic Response Spectra for Periods Ranging from 0.01 to 10 S. Earthquake Spectra, 2008, 24(1): 139-171.

[14] Campbell K W, Bozorgnia Y. NGA-West2 Ground Motion Model for the Average Horizontal Components of PGA, PGV, and 5% Damped Linear Acceleration Response Spectra. Earthquake Spectra, 2014, 30(3): 1087-1115.

[15] Cheng H L, Zhou J M, Chen Z Y, . A Comparative Study of the Seismic Performances and Failure Mechanisms of Slopes Using Dynamic Centrifuge Modeling. Journal of Earth Science, 2021, 32(5): 1166-1173.

[16] Chiou B S-J, Youngs R R. Update of the Chiou and Youngs NGA Model for the Average Horizontal Component of Peak Ground Motion and Response Spectra. Earthquake Spectra, 2014, 30(3): 1117-1153.

[17] Derakhshani A, Foruzan A H. Predicting the Principal Strong Ground Motion Parameters: A Deep Learning Approach. Applied Soft Computing, 2019, 80: 192-201.

[18] Derras B, Bard P Y, Cotton F. Towards Fully Data Driven Ground-Motion Prediction Models for Europe. Bulletin of Earthquake Engineering, 2014, 12(1): 495-516.

[19] Derras B, Bard P Y, Cotton F, . Adapting the Neural Network Approach to PGA Prediction: An Example Based on the KIK-NET Data. Bulletin of the Seismological Society of America, 2012, 102(4): 1446-1461.

[20] Douglas J. Earthquake Ground Motion Estimation Using Strong-Motion Records: A Review of Equations for the Estimation of Peak Ground Acceleration and Response Spectral Ordinates. Earth-Science Reviews, 2003, 61(1): 43-104. 2

[21] Friedman J H. Greedy Function Approximation: A Gradient Boosting Machine. The Annals of Statistics, 2001, 29(5): 1189-1232.

[22] Gandomi A H, Alavi A H, Mousavi M, . A Hybrid Computational Approach to Derive New Ground-Motion Prediction Equations. Engineering Applications of Artificial Intelligence, 2011, 24(4): 717-732.

[23] Hao H Z, Gu Q, Hu X M. Research Advances and Prospective in Mineral Intelligent Identification Based on Machine Learning. Earth Science, 2021, 46(9): 3091-3106

[24] Heath D C, Wald D J, Worden C B, . A Global Hybrid V _S30 Map with a Topographic Slope-Based Default and Regional Map Insets. Earthquake Spectra, 2020, 36(3): 1570-1584.

[25] Idriss I M. NGA-West2 Model for Estimating Average Horizontal Values of Pseudo-Absolute Spectral Accelerations Generated by Crustal Earthquakes, 2013, Berkeley: Pacific Earthquake Engineering Research Center, University of California

[26] Jafariavval Y, Derakhshani A. New Formulae for Capacity Energy-Based Assessment of Liquefaction Triggering. Marine Georesources & Geotechnology, 2020, 38(2): 214-222.

[27] Kafaei Mohammadnejad A, Mousavi S M, Torabi M, . Robust Attenuation Relations for Peak Time-Domain Parameters of Strong Ground Motions. Environmental Earth Sciences, 2012, 67(1): 53-70.

[28] Kayabali K, Beyaz T. Strong Motion Attenuation Relationship for Turkey—A Different Perspective. Bulletin of Engineering Geology and the Environment, 2011, 70(3): 467-481.

[29] Ke, G. L., Meng, Q., Finley, T., et al., 2017. LightGBM: A Highly Efficient Gradient Boosting Decision Tree. In: NIPS’ 17: Proceeding of the 31st International Conference on Neural Information Processing Systems, December 2017, New York

[30] Molnar, C., 2019. Interpretable Machine Learning: A Guide for Making Black Box Models Explainable. Leanpub

[31] Nagelkerke N J D. A Note on a General Definition of the Coefficient of Determination. Biometrika, 1991, 78(3): 691-692.

[32] Şen Z. Supervised Fuzzy Logic Modeling for Building Earthquake Hazard Assessment. Expert Systems with Applications, 2011, 38(12): 14564-14573.

[33] Shiuly A, Roy N, Sahu R B. Prediction of Peak Ground Acceleration for Himalayan Region Using Artificial Neural Network and Genetic Algorithm. Arabian Journal of Geosciences, 2020, 13(5): 1-10.

[34] Singer D A. How Deep Learning Networks could be Designed to Locate Mineral Deposits. Journal of Earth Science, 2021, 32(2): 288-292.

[35] Thomas S, Pillai G N, Pal K. Prediction of Peak Ground Acceleration Using ϵ-SVR, N-SVR and Ls-SVR Algorithm. Geomatics, Natural Hazards and Risk, 2017, 8(2): 177-193.

[36] Tobler W R. A Computer Movie Simulating Urban Growth in the Detroit Region. Economic Geography, 1970, 46: 234-240. Sup1

[37] Tuv E, Borisov A, Runger G, . Feature Selection with Ensembles, Artificial Variables, and Redundancy Elimination. Journal of Machine Learning Research, 2009, 10(3): 1341-1366

[38] Wen R Z, Xu P B, Ren Y F, . Development of the Strong-Motion Flatfile. Earthquake Engineering and Engineering Dynamics, 2017, 37(3): 38-47. (in Chinese with English Abstract)

[39] Worden C B, Wald D J, Allen T I, . A Revised Ground-Motion and Intensity Interpolation Scheme for ShakeMap. Bulletin of the Seismological Society of America, 2010, 100(6): 3083-3096.

[40] Yenier, E., Erdoğan, Ö., Akkar, S., 2008. Empirical Relationships for Magnitude and Source-to-Site Distance Conversions Using Recently Compiled Turkish Strong-Ground Motion Database. In: The 14th World Conference on Earthquake Engineering. October 12–17, 2008, Beijing

[41] Youngs R R, Chiou B S-J, Silva W J, . Strong Ground Motion Attenuation Relationships for Subduction Zone Earthquakes. Seismological Research Letters, 1997, 68(1): 58-73.

[42] Youngs R R, Day S M, Stevens J L. Thun J L V. Near Field Ground Motions on Rock for Large Subduction Earthquakes. Earthquake Engineering and Soil Dynamics II: Recent Advances in Ground-Motion Evaluation, 1988, Reston: American Society of Civil Engineers, 445-462

[43] Zuo R G, Peng Y, Li T, . Challenges of Geological Prospecting Big Data Mining and Integration Using Deep Learning Algorithms. Earth Science, 2021, 46(1): 350-358