Comparison of Machine Learning Approaches for Tsunami Forecasting from Sparse Observations

We have explored various different machine learning (ML) approaches for forecasting tsunami amplitudes at a set of forecast points, based on hypothetical short-time observations at one or more observation points. As a case study, we chose an observation point near the entrance of the Strait of Juan de Fuca, and two forecast points in the Salish Sea, one in Discovery Bay and the other in Admiralty Inlet, the waterway leading to southern Puget Sound. One ML approach considered is to train a support vector machine to predict the maximum amplitude at the forecast points. We also explored the use of two deep convolutional neural networks, a denoising autoencoder and a variational autoencoder to predict the full time series at the forecast points. These latter approaches also provide an estimate of the uncertainty in the predictions. As training data we use a subset of the 1300 synthetic CSZ earthquakes generated in the work of Melgar et al. (J Geophys Res Solid Earth 121:6658-6674, 2016b), reserving some as test data. As additional tests, the trained ML models have also been applied to other hypothetical CSZ earthquakes produced by very different approaches, such as the “L1 event” from the work of Witter et al. (Geosphere 9(6):1783-1803, 2013) that is used in the generation of tsunami inundation maps in Washington State. The ML models are capable of providing very good predictions from short duration observations, even when truncated before the first wave peak has reached the observation point.