High Resolution Seismic Waveform Generation Using Denoising Diffusion

Abstract

Accurate prediction and synthesis of seismic waveforms are crucial for seismic‐hazard assessment and earthquake‐resistant infrastructure design. Existing prediction methods, such as ground‐motion models and physics‐based wavefield simulations, often fail to capture the full complexity of seismic wavefields, particularly at higher frequencies. This study introduces HighFEM , a novel, computationally efficient, and scalable generative model for broadband stochastic seismic‐waveform generation. Our approach leverages a spectrogram representation of the seismic‐waveform data, which is reduced to a lower‐dimensional manifold via an autoencoder. A state‐of‐the‐art diffusion model is trained to generate this latent representation conditioned on key input parameters: earthquake magnitude, recording distance, site conditions, hypocenter depth, and azimuthal gap. The model generates waveforms with frequency content up to 50 Hz. Any scalar ground‐motion statistic, such as peak ground‐motion amplitudes and spectral accelerations, can be readily derived from the synthesized waveforms. We validate our model using commonly employed seismological metrics and performance metrics from image‐generation studies. Our results demonstrate that the openly available model can generate realistic high‐frequency seismic waveforms across a wide range of input parameters, even in data‐sparse regions. For the scalar ground‐motion statistics commonly used in seismic‐hazard and earthquake‐engineering studies, we show that our model accurately reproduces both the median trends of the real data and their variability. To evaluate and compare the growing number of these and similar “Generative Waveform Models” (GWMs), we argue that they should be openly available and included in community ground‐motion‐model evaluation efforts. Predicting how the ground shakes during an earthquake is crucial for understanding earthquake hazards and for designing earthquake‐resistant buildings. In this study, we use a recently developed artificial intelligence (AI) method to generate realistic synthetic earthquake seismograms. After transforming the training seismograms from a time‐domain into a time‐frequency representation, we use a special type of AI model called a diffusion model—originally successful in generating images—to create synthetic seismograms. Our model takes five input parameters (earthquake magnitude, recording distance, site condition, hypocenter depth, and azimuthal gap) and can produce any number of realistic synthetic seismograms for these parameter choices, with high‐frequency details up to 50 Hz. Our study shows that the open‐source model we present can create realistic seismograms in a wide variety of settings, even in areas with limited training data. For a suite of performance metrics commonly used in assessing earthquake risk and for designing safe buildings, the model closely matches the average trends and variations shown in real earthquake records. To help evaluate and compare the increasing number of such Generative Waveform Models, we argue that such models should generally be made publicly available and included in community efforts to assess ground motion prediction models. A novel generative latent denoising diffusion model generates realistic synthetic seismic waveforms with frequency content up to 50 Hz The model predicts peak amplitudes at least as accurately as local ground motion models, and with the same variability as in real data We introduce tqdne , an open‐source Python library for using the pre‐trained model, and to train new generative models