Estimating the Topography Before Volcanic Sector Collapses Using Tsunami Survey Data and Numerical Simulations

Large sector collapses and landslides have the potential to cause significant disasters. Estimating the topography and conditions, such as volume, before the collapse is thus important for analyzing the behavior of moving collapsed material and hazard risks. This study considers three historical volcanic sector collapses in Japan that caused tsunamis: the collapses of the Komagatake Volcano in 1640, Oshima–Oshima Island in 1741, and Unzen–Mayuyama Volcano in 1792. Numerical simulations of the tsunamis generated by each event were first carried out based on assumed collapse scenarios. The primary objective of this study is to present conditions related to the topography before the events based on inverse models of the topography from those results and tsunami survey data. The Oshima–Oshima Tsunami, which is the subject of many previous studies, was first simulated to validate the model accuracy and evaluate how run-up heights changed during the simulation as the topographic conditions changed. The run-up height was especially sensitive to the collapsed volume and frictional acceleration affecting the collapsed material; however, the observed run-up heights could be reproduced with high accuracy using proper conditions of frictional acceleration for the scenarios, even if they were not exact. A minimum requirement for the collapsed volume to generate the observed run-up height was introduced and quantitatively evaluated using the results of numerical tsunami simulations. The minimum volumes of the collapses of the Komagatake and Unzen–Mayuyama volcanoes were estimated to be approximately 1.2 and 0.3 km3, respectively.