Tsunami Early Warning from Global Navigation Satellite System Data using Convolutional Neural Networks

Abstract

We investigate the potential of using Global Navigation Satellite System (GNSS) observations to directly forecast full tsunami waveforms in real time. We train convolutional neural networks (CNNs) to use less than 9 minutes of GNSS data to forecast the full tsunami waveforms over 6 hours at select locations, and obtain accurate forecasts on a test dataset. Our training and test data consists of synthetic earthquakes and associated GNSS data generated for the Cascadia Subduction Zone (CSZ) using the MudPy software, and corresponding tsunami waveforms in Puget Sound computed using GeoClaw. We use the same suite of synthetic earthquakes and waveforms as in earlier work where tsunami waveforms were used for forecasting, and provide a comparison. We also explore varying the number of GNSS stations, their locations, and their observation durations.

Recent advances in earthquake seismology using machine learning

“…Tsunamis are another phenomenon caused by earthquakes besides ground shaking. Several studies have applied ML to tsunami prediction (Fauzi and Mizutani 2020;Makinoshima et al 2021;Liu et al 2021;Kamiya et al 2022;Rim et al 2022).…”