[S-VC29] Volcanic and igneous activities, and these long-term forecasting

[SVC29-06] Development of tephra identification system using a reference geochemical database

*Tatsu Kuwatani¹, Kenta Ueki¹, Morihisa Hamada¹, Kenta Yoshida¹, Satoru Haraguchi^1,2, Iona McIntosh¹, Takashi Miyazaki¹, Takeshi Hanyu¹ (1.Japan Agency for Marine-Earth Science and Technology, 2.Earthquake Research Institute, University of Tokyo)

Keywords:tephra, chemical composition, statistical analysis, data-driven

Tephra layers within sediments provide information on past explosive eruptions. They also can be used as an isochronous geologic marker horizon for correlating and dating past geological events. In this study, we constructed a system for identifying tephra based on chemical composition data including major and trace elements. By inputting the chemical composition of an unknown sample, the system suggests tephra samples which have similar compositions from the reference geochemical database based on the statistical distance in the multidimensional composition space. To evaluate the effectiveness of the system, we investigated how appropriately the system suggests the correct tephras by using well-studied tephra samples including the tephra layers in the Lake Suigetsu samples. As a reference geochemical database, we used the geochemical database (Haraguchi et al., 2021, JpGU) which mainly consists of tephras from the volcanoes in the South Kyushu, Japan, such as K-Ah, AT, and K-Tz, and also includes representative Quaternary widespread tephras in Japan. As a result, the correct tephras were suggested in the top ranks with a high probability, as long as the corresponding references are included in the reference database. We have also confirmed that the system tends to work successfully for the cases where only major-element composition is considered. We will continue to enrich the reference database and add various useful functions, such as probabilistic discrimination and introducing geological information, to the system.