10:45 AM - 12:15 PM
[MIS16-P01] Detection of sand particles by elemental mapping analysis for inland tracking of paleo-tsunami traces
Keywords:Tsunami deposits, Deep neural network, SEM-EPMA
The distribution of tsunami deposits is important for understanding the extent of past inundation. However, it is known that the distribution of sandy and muddy sediments was limited to 60–90% of the inundation extent in the 2011 Tohoku-oki earthquake tsunami (Abe et al., 2012, Sediment. Geol.). The sand and mud layers with a thickness of a few millimeters are not likely to be preserved due to post-depositional disturbance and weathering, and the distribution of the sand layers is considered to deviate from the actual tsunami inundation extent. In this study, we attempted to detect tsunami traces that cannot be observed with the naked eye inland from the observation limit of the sand layer, targeting mineral particles that have high preservation potential.
The survey area was Hyakuninhama, Erimo, on the Pacific coast of Hokkaido. The previous study has reported the distribution of sandy tsunami deposits in the 17th (S1 layer) and 12th century (S2 layer: Nakanishi et al., 2022, G-cubed). This study targeted the S2 layer, which is easily identifiable and well preserved.
In the field survey, geoslicers were used to sample at intervals of 5–10 m from 1000 m to 1150 m from the coast. The peat layers between the S1 and B-Tm tephra layers, which is corresponded to the S2 layer level, were resin-filled, polished and subjected to SEM-EPMA analysis for elemental mapping and point analysis of sand particles. As a result, it was confirmed that there were mineral particles in the S2 layer level, which were not included in other peat layers. The chemical compositions of plagioclase and K-feldspar in the S2 layer level were similar to that of the visible S2 layer and beach sand, and were distinct from river sand and volcanic ash layers.
The grain size distributions were estimated using ImageJ on the sand particle images of the mapping data. The particles that have short diameters of 0.08–0.12 mm were observed up to 40 m from the site where the sand layer was observed, but almost no sand particles were detected at around 80 m. These very fine sand corresponded the finest component of the beach sand and S2 sand layers.
To investigate whether these sand particles were transported by tsunami, we estimated the inundation extent using FITTNUSS-DNN (Mitra et al., 2020, JGR-ES). The estimated inundation extent (30–50 m from the site observed the sand layer) covered the distribution of sand particles in correlate to the S2 layer. The results of this study indicate that it is possible to track tsunami traces further inland than the sand layer distribution observed with the naked eye by analyzing on the order of tens of micrometers.
The survey area was Hyakuninhama, Erimo, on the Pacific coast of Hokkaido. The previous study has reported the distribution of sandy tsunami deposits in the 17th (S1 layer) and 12th century (S2 layer: Nakanishi et al., 2022, G-cubed). This study targeted the S2 layer, which is easily identifiable and well preserved.
In the field survey, geoslicers were used to sample at intervals of 5–10 m from 1000 m to 1150 m from the coast. The peat layers between the S1 and B-Tm tephra layers, which is corresponded to the S2 layer level, were resin-filled, polished and subjected to SEM-EPMA analysis for elemental mapping and point analysis of sand particles. As a result, it was confirmed that there were mineral particles in the S2 layer level, which were not included in other peat layers. The chemical compositions of plagioclase and K-feldspar in the S2 layer level were similar to that of the visible S2 layer and beach sand, and were distinct from river sand and volcanic ash layers.
The grain size distributions were estimated using ImageJ on the sand particle images of the mapping data. The particles that have short diameters of 0.08–0.12 mm were observed up to 40 m from the site where the sand layer was observed, but almost no sand particles were detected at around 80 m. These very fine sand corresponded the finest component of the beach sand and S2 sand layers.
To investigate whether these sand particles were transported by tsunami, we estimated the inundation extent using FITTNUSS-DNN (Mitra et al., 2020, JGR-ES). The estimated inundation extent (30–50 m from the site observed the sand layer) covered the distribution of sand particles in correlate to the S2 layer. The results of this study indicate that it is possible to track tsunami traces further inland than the sand layer distribution observed with the naked eye by analyzing on the order of tens of micrometers.