[S-VC30] Volcanic and igneous activities, and these long-term forecasting

[SVC30-07] Latent thermodynamic parameter in the chemical compositional variation of basalts of the Fuji volcano, central Japan, identified by unsupervised machine learning

*Yusuke Katsuki^1,2, Shuhei Sakata³, Tatsuji Nishizawa⁴, Masaki Takahashi⁵, Shun-ichi Nakai³, Hitomi Nakamura², Satoru Haraguchi³, Hikaru Iwamori³ (1.Department of Earth and Planetary Science, The University of Tokyo, 2.Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology, 3.Earthquake Research Institute, The University of Tokyo, 4.Volcanic Disaster Research Center, Mount Fuji Research Institute, Yamanashi Prefectural Government, 5.Institute of Natural Sciences, College of Humanities and Sciences, Nihon University)

Keywords:subduction related igneous activity, chemical composition, unsupervised machine learning, Fuji volcano

Mt. Fuji has erupted voluminous basaltic to basaltic-andesitic lava flows for more than one hundred thousand years. The limited range in composition makes it difficult to elucidate magma processes beneath the Fuji volcano. To resolve the characteristics of such limited compositional variation and to extract statistically independent features hidden in the dataset, we applied an unsupervised machine learning technique, Independent Component Analysis, to the major element compositional dataset of whole rock samples. We identify three independent components (IC1-3) as the compositional vectors. Crystal fractionation simulation by MELTS with a variable H₂O content, pressure, and oxygen fugacity of crystallization was also performed to deduce the geological processes/sources corresponding to the individual ICs. As a result, IC1 represents a vector with increasing Si and decreasing Ti, Fe, which broadly reflects chemical compositional changes with temperature descendant in any pressure, water content, and oxygen fugacity condition. IC2 represents a vector with increasing Al and Ca and decreasing Ti, Fe, K, and P, which broadly reflects the compositional variability associated with different H₂O contents at a similar temperature and oxygen fugacity condition along the crystallization paths. IC3 represents a vector with increasing Si, Ti, Al, Na, K, and P and decreasing Fe and Mg, which reflects the compositional variability associated with different oxygen fugacity at high water content conditions with more than ca. 1wt%. Three independent components identified in this study may correspond to temperature, water content, and oxygen fugacity in the magma reservoirs of Mt. Fuji, respectively.