Skip to content

Detecting and Classifying Volcano Seismicity using a Generalized Deep Learning Model

1David Fee , 1Darren Tan, 2John Lyons, 3Mariangela Sciotto, 3,4Andrea Cannata, 5Alicia Hotovec-Ellis, 1Társilo Girona, 2Aaron Wech, 6Diana Roman, 2Matthew M. Haney, 7Silvio De Angelis

  • Affiliations: 1Alaska Volcano Observatory, Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK, USA; 2U.S. Geological Survey, Volcano Science Center, Alaska Volcano Observatory, Anchorage, AK, USA.; 3Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Etneo -- Sezione di Catania, Catania, Italy; 4Dipartimento di Scienze Biologiche, Geologiche e Ambientali - Sezione di Scienze della Terra, Università degli Studi di Catania, Catania, Italy; 5California Volcano Observatory, U.S. Geological Survey, Moffett Field, CA, USA; 6Earth and Planets Laboratory, Carnegie Institution for Science, Washington DC, USA; 7School of Environmental Sciences, University of Liverpool, Liverpool, England.

  • Presentation type: Talk

  • Presentation time: Friday 11:00 - 11:15, Room S160

  • Programme No: 2.1.6

  • Theme 2 > Session 1

Abstract

Volcano seismicity is often detected and classified based on its spectral properties. However, the wide variety of volcano seismic signals and increasing amounts of data make accurate, consistent, and efficient detection and classification challenging. Progress has been made applying machine learning to volcano seismicity, but efforts have typically been focused on a single volcano with limited training data. We build on the method of Tan et al. [2024] to generalize a spectrogram-based convolutional neural network termed the VOlcano Infrasound and Seismic Spectrogram Neural Network (VOISS-Net) to detect and classify volcano seismicity at any volcano. We use a diverse training dataset of over 270,000 spectrograms from nine volcanoes worldwide that present a range of volcano seismic signals, source-receiver distances, and eruption styles. We apply VOISS-Net to continuous data from several volcanoes and eruptions within and outside the training set including Mt. Etna, Italy; Shishaldin and Semisopochnoi, Alaska; and Kilauea, Hawaii. VOISS-Net successfully detects and classifies multiple types of tremor, explosions, earthquakes, long-period events, and noise, but occasionally confuses earthquakes and explosions and misclassifies seismicity not included in the training dataset. We envision the generalized VOISS-Net model to be applicable in both research and operational volcano monitoring settings, as demonstrated by its recent application at the Alaska Volcano Observatory. Reference: Tan, D., Fee, D., Witsil, A., Girona, T., Haney, M., Wech, A., Waythomas, C., & Lopez, T. (2024). Detection and Characterization of Seismic and Acoustic Signals at Pavlof Volcano, Alaska, Using Deep Learning. Journal of Geophysical Research: Solid Earth, 129(6), e2024JB029194. https://doi.org/10.1029/2024JB029194