2025 CSDMS meeting-106
From CSDMS
(if you haven't already)
Log in (or create account for non-CSDMS members)
Forgot username? Search or email:CSDMSweb@colorado.edu
Browse abstracts
Tsunami Early Warning from GNSS Data using Convolutional Neural Networks
Randall LeVeque,
University of Washington Seattle Washington, United States. rjleveque@gmail.com
Donsub Rim, Washington University St Louis Missouri, United States. rim@wustl.edu
Robert Baraldi, , United States.
Christopher Liu, , United States.
Kenjiro Terada, Tohoku University Sendai , Japan. tei@irides.tohoku.ac.jp
Accurate tsunami forecasts often rely on direct measurements of the waves, which are only available at sparse locations, and only after the tsunami has passed the sensors. By contrast, we are investigating the use of a convolutional neural network (CNN) that can forecast tsunamis based only on Global Navigation Satellite System (GNSS) data, which is available within minutes at many existing stations in earthquake-prone regions. Training the model requires a large set of hypothetical earthquakes that are generated using a Karhunen-Loeve expansion, as implemented in the MudPy software, which provides synthetic GNSS data. It also provides seafloor motions that are used for GeoClaw simulations of the tsunami resulting from each event. We consider forecasting both multi-hour time series at select locations and inundation maps for target communities. Successes and limitations of this approach will be discussed.
