2:30 PM - 2:45 PM
[SVC30-04] Timescales of magmatic processes prior two catastrophic eruptions of Naruko caldera: insight from diffusion chronometry in quartz and orthopyroxene.
Keywords:Rejuvenation, Caldera eruption, Diffusion timescales, Orthopyroxene, Ti diffusion in quartz
The timescales over which magmas in large silicic systems are reactivated, assembled, and stored remains a fundamental question in volcanology. To address this question, we study timescales from Fe–Mg interdiffusion in orthopyroxenes and Ti diffusion in quartz from two catastrophic eruptions of Naruko caldera, NE Honshu. Naruko volcano is an active volcano that caused two caldera-forming eruptions at 72 ka (Nizaka episode) and 45 ka (Yanagisawa episode). Field constraints and the similarities of crystal textures and compositions and glass chemistries of both eruption deposits demonstrate that they came from one overall magmatic system with a common crystal mush source. However, compositions of pumice and mineral assemblage from both eruptions show variations consistent with there being multiple discrete melt-dominant magma bodies tapped during the caldera-forming eruptions.
Both caldera-forming deposits consist of plagioclase, quartz, orthopyroxene and Fe-Ti oxides. Low-Al Mg-Hornblende occurs as a phenocryst in Yanagisawa deposit but is absent in Nizaka. Naruko rhyolite magmas has been stored at temperatures of 777–832ºC and 790–838ºC for the Nizaka and Yanagisawa, respectively. Both magmas stored at a depth of 4.5–7.7 km under the pressure of 116–197 MPa.
Phenocrysts of plagioclase and orthopyroxene from deposits of both eruptions are mostly zoned, with a prevalence reverse-zoned crystal. Reverse zoning is caused by high-Ca or high-Mg mantle zones or rims in plagioclase and orthopyroxene, respectively. High-Mg mantle zones in orthopyroxene enclosed in the crystal at a distance from Mg-depleted rim and core. Thus, we consider that the dark mantle zones correspond to the interaction of the hot magma through underplating at the base of the mush column, essentially primed a volume of material for eruption through thermal rejuvenation and/or volatile exchange, rather than being the eruption trigger itself. The ubiquitous presence of such zoning patterns in Opx suggests that magmatic rejuvenation might have a system-wide influence. Based on Fe-Mg diffusion, this process occurred within 15-930 years (centuries are predominant). At the same time, Opx often have narrow high-Mg rims at the contact with groundmass, which is also caused by interaction with a less evolved magma. However, the position strictly at the rim implies that such interaction may be the trigger of an eruption. In the case of both eruptions outermost high-Mg rims yield 2-54 years, with the ages peak at ∼2-6 years before eruption.
CL imaging of quartz reveals oscillatory and complexly zoned grains with the greyscale intensity inferred to reflect Ti composition. Crystals from both eruptions have a multiple zoning pattern and abundant melt embayments. Crystals with dark (low-Ti) cores predominate in both eruptions, which agrees with the observed patterns in Opx (low-Mg cores and high-Mg rims or mantles). Naruko quartz grains exhibit signs of multiple resorption episodes, indicating that quartz undergoes cycles of dissolution and growth over relatively short geological periods. We deduce that the timeframes represented by most quartz occurrences likely capture only the latest instances of crystal growth, occurring at temperatures below the initial appearance of quartz in the crystallizing assemblage and above the solidus, just before eruption. These quartz grains might have been stored in a mush for a considerably longer duration than what is reflected in Ti-in-quartz diffusion, preceding remobilization and dissolution, succeeded by swift regrowth. Also, distinct high-Ti rims are sometimes observed in quartz, which may indicate an increase in temperature shortly before the eruption. Ti diffusion timescales across boundaries of Nizaka quartz vary within 0.6-22 years of eruption. Ti timescales from Yanagisawa quartz vary within 0.2-21 years of eruption. The ages peak at ∼1-3 years prior both eruptions.
Both caldera-forming deposits consist of plagioclase, quartz, orthopyroxene and Fe-Ti oxides. Low-Al Mg-Hornblende occurs as a phenocryst in Yanagisawa deposit but is absent in Nizaka. Naruko rhyolite magmas has been stored at temperatures of 777–832ºC and 790–838ºC for the Nizaka and Yanagisawa, respectively. Both magmas stored at a depth of 4.5–7.7 km under the pressure of 116–197 MPa.
Phenocrysts of plagioclase and orthopyroxene from deposits of both eruptions are mostly zoned, with a prevalence reverse-zoned crystal. Reverse zoning is caused by high-Ca or high-Mg mantle zones or rims in plagioclase and orthopyroxene, respectively. High-Mg mantle zones in orthopyroxene enclosed in the crystal at a distance from Mg-depleted rim and core. Thus, we consider that the dark mantle zones correspond to the interaction of the hot magma through underplating at the base of the mush column, essentially primed a volume of material for eruption through thermal rejuvenation and/or volatile exchange, rather than being the eruption trigger itself. The ubiquitous presence of such zoning patterns in Opx suggests that magmatic rejuvenation might have a system-wide influence. Based on Fe-Mg diffusion, this process occurred within 15-930 years (centuries are predominant). At the same time, Opx often have narrow high-Mg rims at the contact with groundmass, which is also caused by interaction with a less evolved magma. However, the position strictly at the rim implies that such interaction may be the trigger of an eruption. In the case of both eruptions outermost high-Mg rims yield 2-54 years, with the ages peak at ∼2-6 years before eruption.
CL imaging of quartz reveals oscillatory and complexly zoned grains with the greyscale intensity inferred to reflect Ti composition. Crystals from both eruptions have a multiple zoning pattern and abundant melt embayments. Crystals with dark (low-Ti) cores predominate in both eruptions, which agrees with the observed patterns in Opx (low-Mg cores and high-Mg rims or mantles). Naruko quartz grains exhibit signs of multiple resorption episodes, indicating that quartz undergoes cycles of dissolution and growth over relatively short geological periods. We deduce that the timeframes represented by most quartz occurrences likely capture only the latest instances of crystal growth, occurring at temperatures below the initial appearance of quartz in the crystallizing assemblage and above the solidus, just before eruption. These quartz grains might have been stored in a mush for a considerably longer duration than what is reflected in Ti-in-quartz diffusion, preceding remobilization and dissolution, succeeded by swift regrowth. Also, distinct high-Ti rims are sometimes observed in quartz, which may indicate an increase in temperature shortly before the eruption. Ti diffusion timescales across boundaries of Nizaka quartz vary within 0.6-22 years of eruption. Ti timescales from Yanagisawa quartz vary within 0.2-21 years of eruption. The ages peak at ∼1-3 years prior both eruptions.