SEDIMENTOLOGICAL SIGNATURES OF LARGE SCALE INUNDATION EVENTS: RECORDS OF THE 1707 AD HOEI TSUNAMI FROM THE BUNGO CHANNEL, JAPAN

Along the Nankai Trough, the subduction of the Filipino plate beneath the Eurasian plate generates earthquakes (M_W>8), which present a serious tsunami risk for regions of southwestern Japan. A tsunami generated by the Hoei earthquake of 1707 AD remains the event of record for the region. However, using historical accounts of the Hoei event to assess the tsunami risk associated with Nankai Trough earthquakes is difficult, as these accounts are qualitative, scarce, and inconsistent with flood elevations derived from tsunami models of the event in many areas. These model-observation comparisons are particularly inconsistent within the Bungo Channel, where there are two documented accounts of a 5-meter tsunami, and model-derived maximum flood elevations are only between 1 and 4 meters. This channel is a critical area for accurately assessing flood risk, as it contains Shikoku’s sole nuclear power plant.

To provide further constraints on the 1707 AD tsunami, we present results from the sedimentary records of three coastal lakes located within the Bungo Channel: Lake Ryuuo (33.372°N, 132.360°E), Lake Kamega (33.465°N, 132.286°E), and Lake Amida (33.369°N, 132.107°E). The lakes are located on the Pacific coast of southwestern Japan and are well situated to preserve marine sediments washed onshore by Nankai Trough-generated tsunamis. Based on a multi-proxy approach for identifying anomalously coarse-grained, clastic event deposits enriched in marine-affiliated material, we have identified a prominent event deposit at each of the three sites that dates to the timing of the 1707 AD Hoei tsunami. Both the regional coherency of the deposit, and its anomalous thickness and grain size composition point to the 1707 tsunami being of significant flood intensity within the Bungo Channel.

Finally, we will simulate the Hoei tsunami’s inundation of Lake Ryuuoo using GeoClaw, a 2D depth-averaged shallow water equations model for flow over varying topography. We will use the simulated flow patterns along with the observed tsunami deposit to assess the application of inverse models that employ different sediment transport mechanisms for calculating the flow speed and depth of an inundating tsunami current based on a deposit’s sedimentary signature.