Magmatic evolution and timescales of the Naruko caldera system (NE Honshu): Insight from orthopyroxene and quartz mineral zoning

Abstract

Silicic magmas from large caldera-forming eruptions often represent crystal-poor melts separated from highly crystalline mush zones under the influence of heat and volatiles. The cascade of magmatic processes from the formation of partially molten crustal zones to the extraction of magma from mush and its interaction with hotter intrusions creates intricate chemical zoning and textures in phenocrysts, reflecting diverse magmatic processes. Understanding the mechanisms and timescales of eruptible magma extraction and storage remains a fundamental challenge in the volcanology. To address this, we present diffusion chronometry from FeMg in orthopyroxenes and Ti in quartz from two Pleistocene caldera-forming eruptions at Naruko Caldera, Japan with the Nizaka (73 ka, ∼2.5 km³) and Yanagisawa (45 ka, ∼10 km³) Tuffs. Our findings suggest that both eruptions tapped a broadly similar crustal storage zone, but differences in extraction and localization prior to the eruption influenced the mineral assemblages (e.g., amphibole presence in Yanagisawa but not in Nizaka). Olivine and high-Mg pyroxenes, along with widespread reverse zoning, indicate significant mafic recharge during magma extraction and immediately before eruption. We propose that the Nizaka magma was physically separated from the mush, stored under cooler and less oxidized conditions, and experienced less intrusion by hot magmas, whereas the Yanagisawa magma remained in situ, forming a hotter reservoir that preserved amphibole stability. Despite these differences, the FeMg interdiffusion in orthopyroxenes suggests that both magmas were generated within ∼300 years, with peak accumulation rates occurring within ∼30 years of eruption. Ti diffusion in quartz has a shorter timescale (<30 years), capturing only the final stages of crystallization. Our results support the notion that eruptible silicic magmas might be assembled over short timescales prior to eruption regardless of their volume, which is consistent with the results of previous studies.