Enhancing Seismic Event Association: Leveraging Signal Similarity and Correlation Detection with Machine Learning

Louisa Barama, Ana C. Aguiar, & Moira Pyle

Submitted September 7, 2025, SCEC Contribution #14925, 2025 SCEC Annual Meeting Poster #TBD

Association remains a challenging step in seismic monitoring due to the presence of signals from multiple seismic sources and high rates of false detections caused by natural and anthropogenic noise. The complexity increases further in regions with clustered signals and aftershock sequences. In this study, we investigate how machine learning can leverage signal similarity and correlation detection to improve the association process, particularly for small, discrete events that are not widely recorded across stations. Our benchmark dataset is based on Dodge et al. (2025), which focused on five years of continuous waveform data from the Hawaiian Volcano Observatory, focusing on a densely clustered region on Hawaii’s Big Island. This study uses their correlated detections generated from the Detection Framework Testbed and Toolkit (DFTT), a database and suite of Java programs designed to dynamically build correlation detectors by identifying novel signals in continuous data streams, eliminating the need for user-specified templates. Using DFTT vetted detections from five stations, we evaluate a Deep Neural Network (DNN) seismic associator that combines phase timing, event and path pattern recognition, and signal similarity analysis. The results from this work will shed light on the extent to which signal similarity and correlated detections can aid machine learning associators in distinguishing dense temporal and spatial clusters of events, location, and in reducing incorrect waveform associations.

This Low Yield Nuclear Monitoring (LYNM) research was funded by the National Nuclear Security Administration, Defense Nuclear Nonproliferation Research and Development (NNSA DNN R&D). The authors acknowledge important interdisciplinary collaboration with scientists and engineers from LANL, LLNL, NNSS, PNNL, and SNL.

This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under contract DE-AC52-07NA27344. Lawrence Livermore National Security, LLC.

Key Words
Association, detection, machine learning

Citation
Barama, L., Aguiar, A. C., & Pyle, M. (2025, 09). Enhancing Seismic Event Association: Leveraging Signal Similarity and Correlation Detection with Machine Learning. Poster Presentation at 2025 SCEC Annual Meeting.

Related Projects & Working Groups
Seismology