A HYBRID MACHINE LEARNING APPROACH FOR REAL-TIME FORECASTING OF MINUTE-LEVEL SEISMIC SIGNAL AMPLITUDES

Real-time interpretation of seismic waveform data is essential for monitoring signal dynamics and detecting abnormal activity. However, forecasting seismic amplitudes remains challenging because seismic signals are highly variable, noisy, and non-stationary. This study presents an automated, gap-aware real-time forecasting pipeline for minute-level seismic signal amplitudes built on IRIS FDSN waveform data, together with a rigorous, statistically grounded comparison of seven forecasting models: Naive persistence, ARIMA, Ridge regression, HistGradientBoosting, a standard Gated Recurrent Unit (GRU), an LSTM, and a physics-inspired regularized GRU (PI-GRU) that adds temporal smoothness and amplitude-growth constraints to the training loss. Models were evaluated on 81 independently sampled, non-overlapping forecasting windows drawn from a continuous 91-day observation period across four horizons (5, 15, 30, and 60 minutes), using paired non-parametric significance testing, a quantitative ablation study, and noise- and missing-data robustness testing, and the main findings were additionally verified under an alternative peak-priority signal aggregation. Classical models, particularly Ridge regression and HistGradientBoosting, achieved significantly lower forecasting error than every neural model tested (Friedman and Holm-corrected Wilcoxon tests, p<0.0001) and were robust to synthetic noise and missing data, while the proposed PI-GRU was, contrary to expectation, the weakest-performing model overall and did not produce smoother forecast trajectories than an unregularized baseline GRU. We report this as an honest negative result: as currently configured, the physics-inspired regularization examined here does not demonstrate an accuracy or stability advantage over classical machine learning baselines for this task. The principal contributions of this work are the automated real-time forecasting pipeline itself and a reusable, rigorous evaluation methodology, together with specific, testable hypotheses - concerning hyperparameter tuning, training budget, and calibration of the physics-inspired constraint - for why the neural models underperformed and how this might be addressed in future work.

1. Mojtahedi, F., Yousefpour, N., Chow, S. H., & Cassidy, M. (2025). Deep learning for time series forecasting: Review and applications in geotechnics and geosciences. Archives of Computational Methods in Engineering. Advance online publication. https://doi.org/10.1007/s11831-025-10244-5

2. Wang, L., Jia, B., & Su, G. (2025). Prediction of coal and gas outbursts based on physics-informed neural networks and traditional machine learning models. Scientific Reports, 15, Article 29984. https://doi.org/10.1038/s41598-025-02320-4

3. Jafari, A., Fox, G., Rundle, J. B., Donnellan, A., & Ludwig, L. G. (2024). Time series foundation models and deep learning architectures for earthquake temporal and spatial nowcasting. GeoHazards, 5(4), 1247–1274. https://doi.org/10.3390/geohazards5040059

4. Li, Z., Eaton, D. W., & Davidsen, J. (2023). Physics-informed deep learning for forecasting seismic activity. Scientific Reports, 13, Article 13133. https://doi.org/10.1038/s41598-023-40403-2

5. Bischof, R., & Kraus, M. A. (2025). Multi-objective loss balancing for physics-informed deep learning. Computer Methods in Applied Mechanics and Engineering. Advance online publication. https://doi.org/10.1016/j.cma.2025.117914

6. Hu, H., Qi, L., & Chao, X. (2024). Physics-informed neural networks for computational solid mechanics: Applications and frameworks. Thin-Walled Structures. Advance online publication. https://doi.org/10.1016/j.tws.2024.112495

7. Wang, S., Teng, Y., & Perdikaris, P. (2020). Understanding and mitigating gradient pathologies in physics-informed neural networks. Journal of Computational Physics, 419, Article 109913. https://doi.org/10.48550/arXiv.2001.04536

8. Nasiri, M., Kortelainen, J., & Särkkä, S. (2026). Dual-level models for physics-informed multi-step time series forecasting. arXiv. https://doi.org/10.48550/arXiv.2601.07640

9. Park, K. V., Kim, J., & Seo, J. (2025). PINT: Physics-informed neural time series models with applications to long-term inference on WeatherBench 2m-temperature data. arXiv. https://doi.org/10.48550/arXiv.2502.04018

10. Biswas, S., & Yadav, K. K. (2025). A physics-informed U-Net-LSTM network for nonlinear structural response under seismic excitation. arXiv. https://doi.org/10.48550/arXiv.2511.21276

11. Martinez, Y., Rojas, L., Peña, A., Valenzuela, M., & Garcia, J. (2025). Physics-informed neural networks for the structural analysis and monitoring of railway bridges: A systematic review. Mathematics, 13(10), Article 1571. https://doi.org/10.3390/math13101571

12. Huang, Y., Han, X., & Zhao, L. (2021). Recurrent neural networks for complicated seismic dynamic response prediction of a slope system. Engineering Geology, 289, Article 106198. https://doi.org/10.1016/j.enggeo.2021.106198

13. Huang, Y.-M., Chen, K.-Y., Lin, W.-W., & Chen, D.-Y. (2025). An attention-based framework with multistation information for earthquake early warnings. IEEE Transactions on Geoscience and Remote Sensing. Advance online publication. https://doi.org/10.48550/arXiv.2412.18099

14. Bilal, M. A., Ji, Y., Wang, Y., Akhter, M. P., & Yaqub, M. (2022). An early warning system for earthquake prediction from seismic data using batch normalized graph convolutional neural network with attention mechanism (BNGCNNATT). Sensors, 22(17), Article 6482. https://doi.org/10.3390/s22176482

15. Lyu, D., Nakata, R., Ren, P., Mahoney, M. W., Pitarka, A., Nakata, N., & Erichson, N. B. (2025). Rapid wavefield forecasting for earthquake early warning via deep sequence to sequence learning. Nature Communications. Advance online publication. https://doi.org/10.1038/s41467-025-65435-2

16. Cho, K., van Merriënboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014). Learning phrase representations using RNN encoder-decoder for statistical machine translation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) (pp. 1724-1734). https://doi.org/10.3115/v1/D14-1179

17. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735

18. Hoerl, A. E., & Kennard, R. W. (1970). Ridge regression: Biased estimation for nonorthogonal problems. Technometrics, 12(1), 55-67. https://doi.org/10.1080/00401706.1970.10488634

19. Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics, 29(5), 1189-1232. https://doi.org/10.1214/aos/1013203451

20. Box, G. E. P., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time series analysis: Forecasting and control (5th ed.). John Wiley & Sons.