Inverse Algorithm for Tsunami Forecasts

Abstract

This paper describes a methodology to assess the severity of a tsunami in progress based on real-time water-level data near the source. The inverse method, which uses tsunami signals in water-level data to infer seismic source parameters, is extended to predict the tsunami waveforms away from the source. This study focuses on the Alaska-Aleutian source region and its potential threat to Hawaii. The algorithm divides the source region into 41 subfaults based on previous analyses of major tsunamigenic earthquakes from 1938 to 1986. For unit slip of the subfaults, a linear long-wave model generates a database of synthetic mareograms at 13 water-level stations near the source and at six strategic locations in the Pacific. Regression of recorded tsunami signals using the mareograms provides the slip distribution at the source and the expected waveforms near Hawaii. A jackknife resampling scheme provides the confidence interval bounds of the predictions. The algorithm along with the database is tested and verified using numerically simulated and actual water-level data of past tsunami events.