Modelling of historical tsunami in eastern Indonesia

Abstract

In order to reliably assess tsunami hazard in eastern Indonesia, we need to understand how historical events were generated. Here we consider two such events: the 1674 Ambon and the 1992 Flores tsunamis. Firstly, Ambon Island suffered a devastating earthquake that generated a tsunami with 100 m run-up height on the north coast of the island in 1674. However, there is no known active fault around the island capable of generating such a gigantic wave. Rumphius' report describes that the initial wave was coming from three villages that collapsed immediately after the earthquake with width as far as a musket shot. Moreover, a very high tsunami was only observed locally. We suspect that a submarine landslide was the main cause of the gigantic tsunami on the north side of Ambon Island. Unfortunately, there is no data available to confirm if landslide have occurred in this region. Secondly, several tsunami source models for the 1992 Flores event have been suggested. However, the fault strike is quite different compare to the existing Flores back-arc thrust and has not been well validated against a tide gauge waveform at Palopo, Sulawesi. We considered a tsunami model based on Griffin, et al., 2015, extended with high resolution bathymetry laround Palopo, in order to validate the latest tsunami source model available. In general, the model produces a good agreement with tsunami waveforms, but arrives 10 minutes late compared to observed data. In addition, the source overestimates the tsunami inundation west of Maumere, and does not account for the presumed landslide tsunami on the east side of Flores Island.