Abstract
Predicting high-fidelity ground motions for future earthquakes is crucial for seismic hazard assessment and infrastructure resilience. Conventional empirical simulations suffer from sparse sensor distribution and geographically localized earthquake locations, while physics-based methods are computationally intensive and require accurate representations of Earth structures and earthquake sources. We propose an artificial intelligence (AI) spectrogram generator, Conditional Generative Modeling for Ground Motion (CGM-GM). CGM-GM leverages earthquake magnitudes and geographic coordinates of earthquakes and sensors as inputs, when postprocessed with phase information, capturing spatially continuous Fourier amplitude spectra (FAS) as well as properties such as P and S arrivals, and waveform durations, without explicit physics constraints. This is achieved through a probabilistic autoencoder that extracts latent distributions in the time-frequency domain and variational sequential models for prior and posterior distributions. We evaluate the performance of CGM-GM using small-magnitude earthquake records from the San Francisco Bay Area, a region with high seismic risks. Here, we report that CGM-GM demonstrates potential for complementing physics-based simulations and non-ergodic empirical ground motion models, as well as shows promise in seismology and beyond.
Similar content being viewed by others
Data availability
Waveform data, metadata, or data products for this study were accessed through the Northern California Earthquake Data Center (NCEDC) under DOI: 10.7932/NCEDC (https://doi.org/10.7932/NCEDC). The downloaded datasets were preprocessed and provided from82 under https://doi.org/10.17603/ds2-necm-5q32.
Code availability
All source code to reproduce the results in this study is available on GitHub at https://github.com/paulpuren/cgm-gmand is archived on Zenodo at https://doi.org/10.5281/zenodo.1848027086. We provide the training, generation, evaluation, and visualization details.
References
Naeim, F. & Lew, M. On the use of design spectrum compatible time histories. Earthq. Spectra 11, 111–127 (1995).
Rezaeian, S. & Der Kiureghian, A. Simulation of synthetic ground motions for specified earthquake and site characteristics. Earthq. Eng. Struct. Dyn. 39, 1155–1180 (2010).
Konakli, K. & Der Kiureghian, A. Simulation of spatially varying ground motions including incoherence, wave-passage and differential site-response effects. Earthq. Eng. Struct. Dyn. 41, 495–513 (2012).
Sabetta, F., Pugliese, A., Fiorentino, G., Lanzano, G. & Luzi, L. Simulation of non-stationary stochastic ground motions based on recent Italian earthquakes. Bull. Earthq. Eng. 19, 3287–3315 (2021).
Kelly, K. R., Ward, R. W., Treitel, S. & Alford, R. M. Synthetic seismograms: a finite-difference approach. Geophysics 41, 2–27 (1976).
Virieux, J. P-sv wave propagation in heterogeneous media: velocity-stress finite-difference method. Geophysics 51, 889–901 (1986).
Bao, H. et al. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Comput. Methods Appl. Mech. Eng. 152, 85–102 (1998).
Komatitsch, D. & Vilotte, J.-P. The spectral element method: an efficient tool to simulate the seismic response of 2d and 3d geological structures. Bull. Seismol. Soc. Am. 88, 368–392 (1998).
Komatitsch, D. & Tromp, J. Introduction to the spectral element method for three-dimensional seismic wave propagation. Geophys. J. Int. 139, 806–822 (1999).
Irikura, K. Semi-empirical estimation of strong ground motions during large earthquakes. Bull. Disaster Prev. Res. Inst. 33, 63–104 (1983).
Motazedian, D. & Atkinson, G. M. Stochastic finite-fault modeling based on a dynamic corner frequency. Bull. Seismol. Soc. Am. 95, 995–1010 (2005).
Zeng, Y., Anderson, J. G. & Yu, G. A composite source model for computing realistic synthetic strong ground motions. Geophys. Res. Lett. 21, 725–728 (1994).
Kozdon, J. E. & Dunham, E. M. Rupture to the trench: dynamic rupture simulations of the 11 March 2011 Tohoku earthquake. Bull. Seismol. Soc. Am. 103, 1275–1289 (2013).
McCallen, D. et al. Eqsim-a multidisciplinary framework for fault-to-structure earthquake simulations on exascale computers part I: computational models and workflow. Earthq. Spectra 37, 707–735 (2021).
Graves, R. W. & Pitarka, A. Broadband ground-motion simulation using a hybrid approach. Bull. Seismol. Soc. Am. 100, 2095–2123 (2010).
Mai, P. M., Imperatori, W. & Olsen, K. B. Hybrid broadband ground-motion simulations: combining long-period deterministic synthetics with high-frequency multiple s-to-s backscattering. Bull. Seismol. Soc. Am. 100, 2124–2142 (2010).
Yang, Y., Gao, A. F., Azizzadenesheli, K., Clayton, R. W. & Ross, Z. E. Rapid seismic waveform modeling and inversion with neural operators. IEEE Trans. Geosci. Remote Sens. 61, 1–12 (2023).
Zhu, M., Feng, S., Lin, Y. & Lu, L. Fourier-deeponet: Fourier-enhanced deep operator networks for full waveform inversion with improved accuracy, generalizability, and robustness. Comput. Methods Appl. Mech. Eng. 416, 116300 (2023).
Zou, C., Azizzadenesheli, K., Ross, Z. E. & Clayton, R. W. Deep neural helmholtz operators for 3-d elastic wave propagation and inversion. Geophys. J. Int. 239, 1469–1484 (2024).
Lehmann, F., Gatti, F., Bertin, M. & Clouteau, D. 3d elastic wave propagation with a factorized fourier neural operator (f-fno). Comput. Methods Appl. Mech. Eng. 420, 116718 (2024).
Raissi, M., Perdikaris, P. & Karniadakis, G. E. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys. 378, 686–707 (2019).
Song, C., Alkhalifah, T. & Waheed, U. B. Solving the frequency-domain acoustic vti wave equation using physics-informed neural networks. Geophys. J. Int. 225, 846–859 (2021).
Rasht-Behesht, M., Huber, C., Shukla, K. & Karniadakis, G. E. Physics-informed neural networks (pinns) for wave propagation and full waveform inversions. J. Geophys. Res. Solid Earth 127, e2021JB023120 (2022).
Ren, P. et al. SeismicNet: physics-informed neural networks for seismic wave modeling in semi-infinite domain. Comput. Phys. Commun. 295, 109010 (2024).
Krishnapriyan, A., Gholami, A., Zhe, S., Kirby, R. & Mahoney, M. W. Characterizing possible failure modes in physics-informed neural networks. In Proc. Conference on Neural Information Processing Systems Vol. 34, 26548–26560 (ACM, 2021).
Négiar, G., Mahoney, M. W. & Krishnapriyan, A. S. Learning differentiable solvers for systems with hard constraints. In Proc. International Conference on Learning Representations (OpenReview.net, 2023).
Hansen, D., Maddix, D. C., Alizadeh, S., Gupta, G. & Mahoney, M. W. Learning physical models that can respect conservation laws. In Proc. International Conference on Machine Learning Vol. 202, 12469–12510 (Proceedings of Machine Learning Research, 2023).
Clements, T., Cochran, E. S., Baltay, A., Minson, S. E. & Yoon, C. Grapes: earthquake early warning by passing seismic vectors through the grapevine. Geophys. Res. Lett. 51, e2023GL107389 (2024).
Bloemheuvel, S., van den Hoogen, J., Jozinović, D., Michelini, A. & Atzmueller, M. Graph neural networks for multivariate time series regression with application to seismic data. Int. J. Data Sci. Anal. 16, 317–332 (2023).
Murshed, R. U. et al. Real-time seismic intensity prediction using self-supervised contrastive gnn for earthquake early warning. IEEE Trans. Geosci. Remote Sens. 62, 1–19 (2024).
Sun, L., Han, X., Gao, H., Wang, J.-X. & Liu, L. Unifying predictions of deterministic and stochastic physics in mesh-reduced space with sequential flow generative model. In Proc. Conference on Neural Information Processing Systems Vol. 36 (ACM, 2024).
Gao, H. et al. Bayesian conditional diffusion models for versatile spatiotemporal turbulence generation. Comput. Methods Appl. Mech. Eng. 427, 117023 (2024).
Eckmann, P. et al. LIMO: latent inceptionism for targeted molecule generation. In Proc. International Conference on Machine Learning, Vol. 162, 5777–5792 (Proceedings of Machine Learning Research, 2022).
Xu, M., Powers, A. S., Dror, R. O., Ermon, S. & Leskovec, J. Geometric latent diffusion models for 3D molecule generation. In Proc. International Conference on Machine Learning, 38592–38610 (Proceedings of Machine Learning Research, 2023).
Florez, M. A. et al. Data-driven synthesis of broadband earthquake ground motions using artificial intelligence. Bull. Seismol. Soc. Am. 112, 1979–1996 (2022).
Esfahani, R. D., Cotton, F., Ohrnberger, M. & Scherbaum, F. Tfcgan: nonstationary ground-motion simulation in the time–frequency domain using conditional generative adversarial network (cgan) and phase retrieval methods. Bull. Seismol. Soc. Am. 113, 453–467 (2023).
Matinfar, M., Khaji, N. & Ahmadi, G. Deep convolutional generative adversarial networks for the generation of numerous artificial spectrum-compatible earthquake accelerograms using a limited number of ground motion records. Comput. Aided Civ. Infrastruct. Eng. 38, 225–240 (2023).
Shi, Y., Lavrentiadis, G., Asimaki, D., Ross, Z. E. & Azizzadenesheli, K. Broadband ground-motion synthesis via generative adversarial neural operators: development and validation. Bull. Seismol. Soc. Am. 114, 2151–2171 (2024).
Aquib, T. A. & Mai, P. M. Broadband ground-motion simulations with machine-learning-based high-frequency waves from Fourier neural operators. Bull. Seismol. Soc. Am. 114, 2846–2868 (2024).
Arjovsky, M., Chintala, S. & Bottou, L. Wasserstein generative adversarial networks. In Proc. International Conference on Machine Learning, 214–223 (Proceedings of Machine Learning Research, 2017).
Naiman, I., Erichson, N. B., Ren, P., Mahoney, M. W. & Azencot, O. Generative modeling of regular and irregular time series data via Koopman Vaes. In Proc. International Conference on Learning Representations (OpenReview.net, 2024).
Lavrentiadis, G. & Abrahamson, N. A non-ergodic spectral acceleration ground motion model for California developed with random vibration theory. Bull. Earthq. Eng. 21, 5265–5291 (2023).
Chiou, B., Youngs, R., Abrahamson, N. & Addo, K. Ground-motion attenuation model for small-to-moderate shallow crustal earthquakes in California and its implications on regionalization of ground-motion prediction models. Earthq. Spectra 26, 907–926 (2010).
Allen, J. B. & Rabiner, L. R. A unified approach to short-time fourier analysis and synthesis. Proc. IEEE 65, 1558–1564 (1977).
Boashash, B.Time-frequency Signal Analysis and Processing: A Comprehensive Reference (Academic Press, 2015).
Goyal, A., Sordoni, A., Côté, M., Ke, N. R. & Bengio, Y. Z-forcing: training stochastic recurrent networks. In Proc. Conference on Neural Information Processing Systems 6713–6723 (ACM, 2017).
Lavrentiadis, G. et al. Overview and introduction to development of non-ergodic earthquake ground-motion models. Bull. Earthq. Eng. 21, 5121–5150 (2023).
Lacour, M. Efficient non-ergodic ground-motion prediction for large datasets. Bull. Earthq. Eng. 21, 5209–5232 (2023).
Lacour, M., Abrahamson, N., Nakata, R., Nakata, N. & Pinilla-Ramos, C. I. A nonergodic ground-motion model for the San Francisco Bay Area for small-magnitude earthquakes. Bull. Seismol. Soc. Am. https://doi.org/10.1785/0120250094 (2026).
Paolucci, R. et al. Broadband ground motions from 3d physics-based numerical simulations using artificial neural networks. Bull. Seismol. Soc. Am. 108, 1272–1286 (2018).
Rodgers, A. J. et al. Broadband (0–5 hz) fully deterministic 3d ground-motion simulations of a magnitude 7.0 hayward fault earthquake: comparison with empirical ground-motion models and 3d path and site effects from source normalized intensities. Seismol. Res. Lett. 90, 1268–1284 (2019).
Liou, I. & Abrahamson, N. Framework for aleatory variability and epistemic uncertainty for the ground-motion characterization based on the level of simplification. Bull. Seismol. Soc. Am. 115, 296–314 (2024).
Zhu, W. & Beroza, G. C. Phasenet: a deep-neural-network-based seismic arrival-time picking method. Geophys. J. Int. 216, 261–273 (2019).
Trifunac, M. D. & Brady, A. G. A study on the duration of strong earthquake ground motion. Bull. Seismol. Soc. Am. 65, 581–626 (1975).
Lavrentiadis, G., Abrahamson, N. A. & Kuehn, N. M. A non-ergodic effective amplitude ground-motion model for California. Bull. Earthq. Eng. 21, 5233–5264 (2023).
Sun, Y. et al. Test-time training with self-supervision for generalization under distribution shifts. In Proc. International Conference on Machine Learning 9229–9248 (ACM, 2020).
Nakata, R. et al. Simulating seismic wavefields using generative artificial intelligence. LeadEdge 44, 123–132 (2025).
Bi, Z. et al. Advancing data-driven broadband seismic wavefield simulation with multiconditional diffusion model. IEEE Trans. Geosci. Remote Sens. 63, 1–9 (2025).
Lin, G. & Wu, B. Seismic velocity structure and characteristics of induced seismicity at the Geysers Geothermal Field, eastern California. Geothermics 71, 225–233 (2018).
Gatti, F. & Clouteau, D. Towards blending physics-based numerical simulations and seismic databases using generative adversarial network. Comput. Methods Appl. Mech. Eng. 372, 113421 (2020).
Genton, M. G. Classes of kernels for machine learning: a statistics perspective. J. Mach. Learn. Res. 2, 299–312 (2001).
Minasny, B. & McBratney, A. B. The matérn function as a general model for soil variograms. Geoderma 128, 192–207 (2005).
Takida, Y. et al. Preventing oversmoothing in vae via generalized variance parameterization. Neurocomputing 509, 137–156 (2022).
Liang, F. T., Hodgkinson, L. & Mahoney, M. W. A heavy-tailed algebra for probabilistic programming. In Proc. Conference on Neural Information Processing Systems Vol. 36 (ACM, 2024).
Allen, J. Short term spectral analysis, synthesis, and modification by discrete Fourier transform. IEEE Trans. Acoust., Speech, Signal Process. 25, 235–238 (1977).
Allen, J. Applications of the short time fourier transform to speech processing and spectral analysis. In Proc. ICASSP’82. IEEE International Conference on Acoustics, Speech, and Signal Processing Vol. 7, 1012–1015 (IEEE, 1982).
Castagna, J. P., Sun, S. & Siegfried, R. W. Instantaneous spectral analysis: detection of low-frequency shadows associated with hydrocarbons. Lead. Edge 22, 120–127 (2003).
Huang, Z. -l, Zhang, J., Zhao, T. -h & Sun, Y. Synchrosqueezing s-transform and its application in seismic spectral decomposition. IEEE Trans. Geosci. Remote Sens. 54, 817–825 (2015).
Wu, G. & Zhou, Y. Seismic data analysis using synchrosqueezing short time fourier transform. J. Geophys. Eng. 15, 1663–1672 (2018).
Cohen, L. Time-frequency Analysis Vol. 778 (Prentice hall New Jersey, 1995).
Kumar, K. et al. Melgan: generative adversarial networks for conditional waveform synthesis. In Proc. Conference on Neural Information Processing Systems Vol. 32 (ACM, 2019).
Kong, J., Kim, J. & Bae, J. Hifi-gan: generative adversarial networks for efficient and high fidelity speech synthesis. In Proc. Conference on Neural Information Processing Systems Vol. 33, 17022–17033 (ACM, 2020).
Griffin, D. & Lim, J. Signal estimation from modified short-time fourier transform. IEEE Trans. Acoust. Speech, Signal Process. 32, 236–243 (1984).
Vial, P.-H., Magron, P., Oberlin, T. & Févotte, C. Phase retrieval with Bregman divergences and application to audio signal recovery. IEEE J. Sel. Top. Signal Process. 15, 51–64 (2021).
Hsu, W.-N., Zhang, Y. & Glass, J. Unsupervised learning of disentangled and interpretable representations from sequential data. In Proc. Conference on Neural Information Processing Systems Vol. 30 (2017).
Girin, L. et al. Dynamical variational autoencoders: a comprehensive review. Found. Trends Mach. Learn. 15, 1–175 (2022).
Hochreiter, S. & Schmidhuber, J. Long short-term memory. Neural Comput. 9, 1735–1780 (1997).
Cho, K. et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation. In Empirical Methods in Natural Language Processing, 1724–1734 (ACL, 2014).
Chung, J. et al. A recurrent latent variable model for sequential data. In Proc. Conference on Neural Information Processing Systems, Vol. 28 (2015).
Mildenhall, B. et al. Nerf: representing scenes as neural radiance fields for view synthesis. Commun. ACM 65, 99–106 (2021).
Konno, K. & Ohmachi, T. Ground-motion characteristics estimated from spectral ratio between horizontal and vertical components of microtremor. Bull. Seismol. Soc. Am. 88, 228–241 (1998).
Lacour, M., Nakata, R., Nakata, N. & Abrahamson, N. Earthquake dataset for the study of small magnitude earthquakes in the san francisco region. https://www.designsafe-ci.org/data/browser/public/designsafe.storage.published/PRJ-4573. DesignSafe-CI, https://doi.org/10.17603/ds2-necm-5q32 (2024).
Welch, P. The use of fast fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Trans. Audio Electroacoust. 15, 70–73 (1967).
Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization. In Proc. International Conference on Learning Representations (Scientific Research, 2015).
Bayless, J. & Abrahamson, N. A. Summary of the ba18 ground-motion model for fourier amplitude spectra for crustal earthquakes in California. Bull. Seismol. Soc. Am. 109, 2088–2105 (2019).
Ren, P. & Song, J. Learning earthquake ground motions via conditional generative modeling. Zenodo, https://doi.org/10.5281/zenodo.18480271 (2026).
Acknowledgements
This work was supported by the Laboratory Directed Research and Development Program of Lawrence Berkeley National Laboratory under U.S. Department of Energy Contract No. DE-AC02-05CH11231. M.W.M. and R.N. acknowledge the support by the Statewide California Earthquake Center (Awards Nos. 24123 and 25303). SCEC is funded by NSF Cooperative Agreement EAR-2225216 and USGS Cooperative Agreement G24AC00072-00. Additional support was provided by DOE Grant DE-SC0016520. M.W.M. would also like to acknowledge the DOE Competitive Portfolios grant and the DOE SciGPT grant. P.R. would like to thank Dr. Rasmus Malik Hoeegh Lindrup for his valuable discussions on generative modeling.
Author information
Authors and Affiliations
Contributions
P.R. developed the algorithm and performed the tests and analysis. P.R., R.N., and M.W.M. co-designed the study. P.R., I.N., and J.S. conducted the experiments. M.L. built the SFBA dataset and developed the ergodic and non-ergodic ground motion models. M.L., Z.B., and N.N. performed the geophysical evaluations. O.A.M., D.M., O.A., and N.B.E. contributed to the algorithm design and provided support for model development. R.N. and M.W.M. advised the research. All authors contributed to the research discussions, writing, and editing of the paper.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Filippo Gatti and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Ren, P., Nakata, R., Lacour, M. et al. Learning earthquake ground motions via conditional generative modeling. Nat Commun (2026). https://doi.org/10.1038/s41467-026-70719-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-026-70719-2
