Abstract
A cross-region prediction model named SeisEML (an acronym for Seismological Ensemble Machine Learning) has been developed in this paper to predict the peak ground acceleration (PGA) at a given site during an earthquake. The SeisEML model consists of hybridized models, kernel-based algorithms, tree regression algorithms, and regression algorithms. The model ablation study is conducted to examine the performance and the selection of meta-machine learning models in the SeisEML. The training and testing dataset consists of 20852 and 6256 accelerograms recorded by the Kyoshin Network, Japan. The mean absolute error (MAE) and root mean square error (RMSE) have been utilized to compare the predicted peak ground acceleration (PGA) for the test data. The SeisEML model yields approximately half the MAE and RMSE values obtained with conventional attenuation relations. The SeisEML model has been used to compute Japan’s iso acceleration contour map of three earthquakes (\(M_{JMA}\) 7.4, 6.6, and 6.1). The qualitative comparison of iso acceleration contours obtained from actual and predicted PGA using SeisEML clearly shows that the model can reliably predict the PGA distribution during an earthquake compared to the regional ground motion prediction equation (GMPE). The cross-region prediction was performed on the datasets of the Iranian earthquakes using SeisEML. The comparison of predicted and observed peak ground acceleration in terms of MAE and RMSE shows that the machine learning model’s performance is superior to the regional attenuation relation. The predictions of PGA from the developed ML model indicate that this trained model has the potential for predicting regional and global scenarios with similar tectonic setups. The ML model developed in this paper can considerably enhance the reliability of PGA prediction for seismic hazard mapping of any region and can serve as a reliable substitute for GMPEs.
Similar content being viewed by others
Data availability
The data used in this research work is taken from National Research Institute for Earth Science and Disaster Resilience (NIED).
Code availability (software application or custom code)
Code used in this paper has been developed in Python.
References
Abrahamson NA, Litehiser JJ (1989) Attenuation of vertical peak acceleration. Bull Seismol Soc Am 79:549–580
Abrahamson NA, Silva WJ, Kamai R (2014) Summary of the ask14 ground motion relation for active crustal regions. Earthq Spectra 30:1025–1055
Adler J, Parmryd I (2010) Quantifying colocalization by correlation: The pearson correlation coefficient is superior to the mander’s overlap coefficient. Cytometry A 77A
Akoğlu H (2018) User’s guide to correlation coefficients. Turk J Emer Med 18:91–93
Alavi AH, Gandomi AH (2011) Prediction of principal ground-motion parameters using a hybrid method coupling artificial neural networks and simulated annealing. Comput Struct 89:2176–2194
Alavi AH, Gandomi AH, Modaresnezhad M et al (2011) New ground-motion prediction equations using multi expression programing. J Earthq Eng 15:511–536
Ambraseys N, Simpson K, Bommer JJ (1996) Prediction of horizontal response spectra in europe. Earthq Eng Struct Dyn 25:371–400
Anbazhagan P, Kumar A, Sitharam TG (2013) Ground motion prediction equation considering combined dataset of recorded and simulated ground motions. Soil Dyn Earthq Eng 53:92–108
Aoi S, Kunugi T, Nakamura H, et al. (2011) Deployment of new strong motion seismographs of k-net and kik-net. In: In: Akkar S, Gülkan P, van Eck T (eds) Earthquake Data in Engineering Seismology. Geotechnical, Geological, and Earthquake Engineering, Springer, Dordrecht
Boore DM (1983) Stochastic simulation of high-frequency ground motions based on seismological models of the radiated spectra. Bull Seismol Soc Am 73:1865–1894
Boore DM, Atkinson GM (1987) Stochastic prediction of ground motion and spectral response parameters at hard-rock sites in eastern north america. Bull Seismol Soc Am 77:440–467
Boore DM, Atkinson GM (2008) Ground-motion prediction equations for the average horizontal component of pga, pgv, and 5%-damped psa at spectral periods between 0.01 s and 10.0 s. Earthq Spectra 24:138–99
Boore DM, Thompson EM, Cadet H (2011) Regional correlations of vs30 and velocities averaged over depths less than and greater than 30 meters. Bull Seismol Soc Am 101:3046–3059
Boore DM, Stewart JP, Seyhan E et al (2014) Nga-west2 equations for predicting pga, pgv, and 5% damped psa for shallow crustal earthquakes. Earthq Spectra 30:1057–1085
Borcherdt R (2012) Vs30 - a site-characterization parameter for use in building codes, simplified earthquake resistant design, gmpes, and shakemaps. In: 15th World Conf. on Earthquake Engineering
Breiman L (1996) Bagging predictors. Mach Learn 24(2):123–140
Breiman L (2004) Random forests. Mach Learn 45:5–32
Breiman L, Friedman JH, Olshen RA, et al (1984) Classification and regression trees. In: Belmont, Calif.: Wadsworth
BSSC (2004) Nehrp recommended provisions for seismic regulations for new buildings and other structures (fema 450). 2003 edition. part 1: Provisions, buildin seismic safety council, national institute of building sciences, washington, d.c
Campbell KW (1997) Empirical near-source attenuation relationships for horizontal and vertical components of peak ground acceleration, peak ground velocity, and pseudo-absolute acceleration response spectra. Seismol Res Lett 68:154–179
Campbell KW, Bozorgnia Y (2014) Nga-west2 ground motion model for the average horizontal components of pga, pgv, and 5% damped linear acceleration response spectra. Earthq Spectra 30:1087–1115
Chan YH (2003) Biostatistics 104: correlational analysis. Singap Med J 44:614–619
Chen T, Guestrin C (2016) Xgboost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
Dancey CP, Reidy J (1999) Statistics Without Maths For Psychology. Pearson Education Limited, https://api.semanticscholar.org/CorpusID:117870366
Derras B, Bard PY, Cotton F et al (2012) Adapting the neural network approach to pga prediction: an example based on the kik-net data. Bull Seismol Soc Am 102:1446–1461
Derras B, Bard PY, Cotton F (2013) Towards fully data driven ground-motion prediction models for europe. Bull Earthq Eng 12:495–516
Derras B, Bard PY, Cotton F (2016) Site-condition proxies, ground motion variability, and data-driven gmpes: insights from the nga-west2 and resorce data sets. Earthq Spectra 32:2027–2056
Douglas J (2003) Earthquake ground motion estimation using strong-motion records: a review of equations for the estimation of peak ground acceleration and response spectral ordinates. Earth Sci Rev 61(1):43–104. https://doi.org/10.1016/S0012-8252(02)00112-5
Douglas J (2003) Earthquake ground motion estimation using strong-motion records: a review of equations for the estimation of peak ground acceleration and response spectral ordinates. Earth Sci Rev 61:43–104
Douglas J, Edwards B (2016) Recent and future developments in earthquake ground motion estimation. Earth Sci Rev 160:203–219
Efron B, Tibshirani R (1994) An Introduction to the Bootstrap. Chapman and Hall/CRC
Frazier P (2018) A tutorial on bayesian optimization. ArXiv abs/1807.02811
Freund Y, Schapire RE (1995) A decision-theoretic generalization of on-line learning and an application to boosting. In: EuroCOLT
Freund Y, Schapire RE (1996) Experiments with a new boosting algorithm. In: ICML
Gandomi AH, Alavi AH, Mousavi M et al (2011) A hybrid computational approach to derive new ground-motion prediction equations. Eng Appl Artif Intell 24:717–732
Geurts P, Ernst D, Wehenkel L (2006) Extremely randomized trees. Mach Learn 63:3–42
Hadley DM, Helmberger DV (1980) Simulation of strong ground motions. Bull Seismol Soc Am 70:617–630
Hartzell S (1982) Simulation of ground accelerations for the May 1980 Mammoth Lakes, California, earthquakes. Bull Seismol Soc Am 72(6A):2381–2387. https://doi.org/10.1785/BSSA07206A2381
Hartzell SH (1978) Earthquake aftershocks as green’s functions. Geophys Res Lett 5:1–4
Hoerl AE, Kennard RW (2000) Ridge regression: biased estimation for nonorthogonal problems. Technometrics 42:80–86
Hoshiba M, Kamigaichi O, Saito M et al (2008) Earthquake early warning starts nationwide in Japan. Eos, Trans Am Geophys Union 89:73–74
Hu P, Pan JS, Chu SC (2020) Improved binary grey wolf optimizer and its application for feature selection. Knowl Based Syst 195(105):746
Hutchings L (1985) Modelling earthquakes with empirical green’s functions (abs). Earthq Notes 56:14
Irikura K (1983) Semi-empirical estimation of strong ground motions during large earthquakes. Bull Disaster Prev Res Inst 33:63–104
Irikura K, Muramatu I (1982) Synthesis of strong ground motions from large earthquakes using observed seismograms of small events. In: Proc 3rd Int Microzonation Conprence, Seattle, pp 447-458
Irikura KK (1986) Prediction of strong acceleration motions using empirical green’s function. In: Proc 7th Japan Earthq Eng Symp, pp 151–156
Joshi AP, Midorikawa S (2005) Attenuation characteristics of ground motion intensity from earthquakes with intermediate depth. J Seismol 9:23–37
Kamae K, Irikura K (1992) Prediction of site specific strong ground motion using semi empirical methods. In: Earthq Eng 10th World Conf, pp 801-806
Kanamori H (1979) A semi-empirical approach to prediction of long-period ground motions from great earthquakes. Bull Seismol Soc Am
Ke G, Meng Q, Finley T et al. (2017) Lightgbm: a highly efficient gradient boosting decision tree. In: NIPS
Kinoshita S (1998) Kyoshin net (k-net). Seismol Res Lett 69:309–332
Kubo H, Kunugi T, Suzuki W, et al (2020) Hybrid predictor for ground-motion intensity with machine learning and conventional ground motion prediction equation. Sci Rep 10
Maclin R, Opitz DW (1999) Popular ensemble methods: an empirical study. J Artif Intell Res 11:169–198
Mandal P, Mandal P (2022) Peak ground acceleration prediction using supervised machine learning algorithm for earthquakes of mw5.6-7.9 occurring in India and Nepal. https://doi.org/10.21203/rs.3.rs-1806354/v1, preprint from Research Square
Midorikawa S (1993) Semi-empirical estimation of peak ground acceleration from large earthquakes. Tectonophysics 218(1):287–295. https://doi.org/10.1016/0040-1951(93)90275-O
Mikumo T, Irikura K, Imagawa K (1981) Near field strong motion synthesis front foreshock and aftershock records and rupture process of the main shock fault. IASPEI 21st General Assembly,London
Mirjalili SM, Mirjalili SM, Lewis A (2014) Grey wolf optimizer. Adv Eng Softw 69:46–61
Morikawa N, Fujiwara H (2013) A new ground motion prediction equation for japan applicable up to m9 mega-earthquake. J Disaster Res 8:878–888
Motazedian D, Atkinson GM (2005) Stochastic finite-fault modeling based on a dynamic corner frequency. Bull Seismol Soc Am 95:995–1010
Munguía L, Brune JN (1984) Simulations of strong ground motion for earthquakes in the mexicali-imperial valley region. Geophys J Int 79:747–771
Nath SK, Raj A, Thingbaijam KKS et al (2009) Ground motion synthesis and seismic scenario in Guwahati city-a stochastic approach. Seismol Res Lett 80:233–242
Okada Y (2004) Recent progress of seismic observation networks in Japan. J Phys: Conf Seri 433(012):039
Ostroumova L, Gusev G, Vorobev A, et al (2018) Catboost: unbiased boosting with categorical features. In: NeurIPS
Panahi H (2017) Sarpol zahab earthquake. https://doi.org/10.13140/RG.2.2.31717.50408
Sheng Paul Lai S (1982) Statistical characterization of strong ground motions using power spectral density function. Bull Seismol Soc Am 72:259–274
Prasanth S, Singh U, Kumar A et al (2020) Forecasting spread of covid-19 using google trends: a hybrid gwo-deep learning approach. Chaos, Solitons, and Fractals 142:110–336
Rasmussen CE, Williams CKI (2009) Gaussian processes for machine learning. In: Adaptive computation and machine learning
Saidi R, Waad B, Essoussi N (2019) Hybrid feature selection method based on the genetic algorithm and pearson correlation coefficient. In: Machine Learning Paradigms
Saikia CK (1993) Ground motion studies in great los angles due to \(m_w\)=7.0 earthquake on the elysian thrust fault. Bull Seismol Soc Am
Saikia CK, Herrmann RB (1985) Application of waveform modelling to determine focal mechanisms of four 1982 miramichi aftershocks. Bull Seismol Soc Am
Schapire RE (1990) The strength of weak learnability. Mach Learn 5(2):197–227. https://doi.org/10.1023/A:1022648800760
Scordilis E (2005) Globally valid relations converting ms, mb and mjma to mw. NATO advanced research workshop on earthquake monitoring and seismic hazard mitigation in Balkan Countries, 11–17 September 2005, the Rila Mountains-Resort Village Borovetz. Bulgaria, Abstracts book, pp 158–161
Shoushtari AV, bin Adnan A, Zare M (2018) Ground motion prediction equations for distant subduction interface earthquakes based on empirical data in the malay peninsula and Japan. Soil Dyn Earthq Eng
Si H, Midorikawa S (1999) New attenuation relationships for peak ground acceleration and velocity considering effects of fault type and site condition. J Struct Constr (transactions of Aij) 64:63–70
Somala SN, Chanda S, Karthikeyan K, et al (2021) Explainable machine learning on new zealand strong motion for pgv and pga. Structures
Thiagarajan B, Srinivasan L, Sharma A, et al. (2017) A machine learning approach for prediction of on-time performance of flights. 2017 IEEE/AIAA 36th Digital Avionics Systems Conference (DASC), pp 1–6
Thomas S, Pillai GN, Pal K (2017) Prediction of peak ground acceleration using \(\epsilon\)-svr, \(\nu\)-svr and ls-svr algorithm. Geomat Nat Hazards Risk 8:177–193
Trugman DT, Shearer PM (2018) Strong correlation between stress drop and peak ground acceleration for recent m 1–4 earthquakes in the san francisco bay area. Bull Seismol Soc Am 108:929–945
Victoria AH, Maragatham G (2021) Automatic tuning of hyperparameters using bayesian optimization. Evol Syst 12:217–223
Wang M, Chen H, Li H et al (2017) Grey wolf optimization evolving kernel extreme learning machine: application to bankruptcy prediction. Eng Appl Artif Intell 63:54–68
Yu GY (1994) Some aspects of earthquake seismology: slip partitioning along major convergent plate boundaries; composite source model for estimation of strong motion; and nonlinear soil response modeling. In: NIPS
Yu GY, Khattri KN, Anderson JG, et al. (1995) Strong ground motion from the uttarkashi, himalaya, india, earthquake: Comparison of observations with synthetics using the composite source model. Bull Seismol Soc Am
Zeng Y, Anderson JG, Yu GY (1994) A composite source model for computing realistic synthetic strong ground motions. Geophys Res Lett 21:725–728
Zhou T, Jiao H (2022) Exploration of the stacking ensemble machine learning algorithm for cheating detection in large-scale assessment. Educ Psychol Meas
Acknowledgements
We would like to thank Prof. John Clague and reviewers for manuscript suggestions. We thank Prof. Nagendra Kumar and Ms. Swati Singh for proof reading the manuscript and correcting grammatical errors. We would also like to thank National Research Institute for Earth Science and Disaster Resilience (NIED) for providing earthquake data.
Funding
This project is supported from grant for PhD research by Prime Minister Research Fellowship (Grant code: PM-31-22-626-414).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
We declare that there are no competing interests related to any financial concern with regard to the publication of a study. It is being declared that there are no personal relationships with people or organization that may influence or may be perceived to influence the research work described in this paper.
Ethical approval
The work presented in this paper does not involve research using humans and/or animals.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Joshi, A., Raman, B., Mohan, C.K. et al. Application of a new machine learning model to improve earthquake ground motion predictions. Nat Hazards 120, 729–753 (2024). https://doi.org/10.1007/s11069-023-06230-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11069-023-06230-4