2:00 PM - 2:15 PM
[SVC30-02] Constraining residence times of plagioclase phenocrysts at the latest activity of Zao volcano
Keywords:Diffusion modeling, Plagioclase, Zoning profile, Zao volcano
1. Introduction
Time scales of magmatic processes are critical information to reveal eruption mechanisms and evaluate volcanic hazards. To estimate magmatic time scales, diffusion modeling of crystals has been used recently. Plagioclase, a ubiquitous phenocryst in volcanic rocks, provides the benefit of detecting magmatic processes. We report the residence times of plagioclase phenocrysts and constrain the pre-eruptive processes at the latest activity of Zao volcano.
2. Samples and analytical methods
Juvenile bomb samples are collected from Okama activity products (OKP), which were erupted between 12 and 19C. These samples are medium-K and calc-alkaline series basaltic andesite to andesite (56-57% SiO2), including plagioclase, orthopyroxene, clinopyroxene, and rare olivine as phenocryst.
Mg and CaAl-NaSi in plagioclase were used to estimate residence times. Zoning profiles of plagioclase were measured by electron probe microanalysis. High spatial resolution An zoning profiles are required to detect the nm-scale diffusion of CaAl-NaSi. An zoning profiles were also obtained by calibrated BSE images using the ImageJ software.
3. Processes recorded by plagioclase textures and compositions
Textures and compositions of plagioclase phenocrysts record various crystallization histories. Four populations of plagioclase phenocrysts are present in OKP samples. Type A has oscillatory zoned cores. The oscillatory zoning reflects cyclic changes of the chemical and physical conditions induced by injections of hotter magma into the shallow reservoir. Type B and C have patchy zoned cores of low- and high- An, respectively. The patchy texture was formed by partial dissolution. The former was formed by injection of hotter magma, whereas the latter was formed by decompression at calcic and H2O-undersaturated melt conditions. Type D shows the honeycomb texture, which was formed by skeletal growth under supercooling conditions. The dusty zoned or high-An rim, which is observed in type A, B, and C, reflects the interaction of hotter and calcic melt before eruptions. The low-An outermost rim observed in all types was produced by supercooling or an increase in the liquidus temperature due to water exsolution induced by magma ascent.
4. Residence times of plagioclase phenocrysts
Residence times of plagioclase phenocrysts were calculated using Mg and CaAl-NaSi diffusion modeling. The former one includes the effects of coupling with An content and provides more precise determinations than the other steady-state diffusion models. However, this model failed to calculate residence times of type A and B plagioclase phenocrysts because of strong changes of the An composition in small steps. To avoid this problem, we used type C and D phenocrysts which have a lower degree of partial dissolution. The initial Mg concentrations were estimated from measured An content. The boundary conditions were assumed to be in equilibrium with the groundmass composition. The diffusion equation was calculated numerically by a finite-difference scheme.
We first used the one-step model, which assumes that diffusion starts only after the entire crystal grew. The calculated profiles at the crystal cores reach the measured Mg concentrations within 50-300 yrs. However, this model could not obtain good fits in the dusty and high-An rim. We used the multi-step model, which provides the time scales of different processes recorded in different parts of the crystals. An example of a type C plagioclase gave first step residence time of 100 yrs for the core profile resulted from diffusion in a high-temperature melt and gave second step residence time of < 5 yrs for the rim profile by the diffusion in a low-temperature melt. This model obtained better fits to the dusty and high-An rim.
In contrast, the CaAl-NaSi diffusion time scales indicate shorter residence times (10-50 yrs) than Mg diffusion. In a simple case, we evaluated the effect of multi-dimensional diffusion in the case of Mg diffusion and found this becomes significant when diffusion progresses to the center part of the crystal. Compared to the two-dimensional diffusion model, the one-dimensional one gives several times longer results. The results of the two-dimensional model are in good agreement with those of CaAl-NaSi diffusion modeling. These results indicate that the shallow magma reservoir had been mixed continuously with the newly injected high-temperature magma just before the eruption.
Time scales of magmatic processes are critical information to reveal eruption mechanisms and evaluate volcanic hazards. To estimate magmatic time scales, diffusion modeling of crystals has been used recently. Plagioclase, a ubiquitous phenocryst in volcanic rocks, provides the benefit of detecting magmatic processes. We report the residence times of plagioclase phenocrysts and constrain the pre-eruptive processes at the latest activity of Zao volcano.
2. Samples and analytical methods
Juvenile bomb samples are collected from Okama activity products (OKP), which were erupted between 12 and 19C. These samples are medium-K and calc-alkaline series basaltic andesite to andesite (56-57% SiO2), including plagioclase, orthopyroxene, clinopyroxene, and rare olivine as phenocryst.
Mg and CaAl-NaSi in plagioclase were used to estimate residence times. Zoning profiles of plagioclase were measured by electron probe microanalysis. High spatial resolution An zoning profiles are required to detect the nm-scale diffusion of CaAl-NaSi. An zoning profiles were also obtained by calibrated BSE images using the ImageJ software.
3. Processes recorded by plagioclase textures and compositions
Textures and compositions of plagioclase phenocrysts record various crystallization histories. Four populations of plagioclase phenocrysts are present in OKP samples. Type A has oscillatory zoned cores. The oscillatory zoning reflects cyclic changes of the chemical and physical conditions induced by injections of hotter magma into the shallow reservoir. Type B and C have patchy zoned cores of low- and high- An, respectively. The patchy texture was formed by partial dissolution. The former was formed by injection of hotter magma, whereas the latter was formed by decompression at calcic and H2O-undersaturated melt conditions. Type D shows the honeycomb texture, which was formed by skeletal growth under supercooling conditions. The dusty zoned or high-An rim, which is observed in type A, B, and C, reflects the interaction of hotter and calcic melt before eruptions. The low-An outermost rim observed in all types was produced by supercooling or an increase in the liquidus temperature due to water exsolution induced by magma ascent.
4. Residence times of plagioclase phenocrysts
Residence times of plagioclase phenocrysts were calculated using Mg and CaAl-NaSi diffusion modeling. The former one includes the effects of coupling with An content and provides more precise determinations than the other steady-state diffusion models. However, this model failed to calculate residence times of type A and B plagioclase phenocrysts because of strong changes of the An composition in small steps. To avoid this problem, we used type C and D phenocrysts which have a lower degree of partial dissolution. The initial Mg concentrations were estimated from measured An content. The boundary conditions were assumed to be in equilibrium with the groundmass composition. The diffusion equation was calculated numerically by a finite-difference scheme.
We first used the one-step model, which assumes that diffusion starts only after the entire crystal grew. The calculated profiles at the crystal cores reach the measured Mg concentrations within 50-300 yrs. However, this model could not obtain good fits in the dusty and high-An rim. We used the multi-step model, which provides the time scales of different processes recorded in different parts of the crystals. An example of a type C plagioclase gave first step residence time of 100 yrs for the core profile resulted from diffusion in a high-temperature melt and gave second step residence time of < 5 yrs for the rim profile by the diffusion in a low-temperature melt. This model obtained better fits to the dusty and high-An rim.
In contrast, the CaAl-NaSi diffusion time scales indicate shorter residence times (10-50 yrs) than Mg diffusion. In a simple case, we evaluated the effect of multi-dimensional diffusion in the case of Mg diffusion and found this becomes significant when diffusion progresses to the center part of the crystal. Compared to the two-dimensional diffusion model, the one-dimensional one gives several times longer results. The results of the two-dimensional model are in good agreement with those of CaAl-NaSi diffusion modeling. These results indicate that the shallow magma reservoir had been mixed continuously with the newly injected high-temperature magma just before the eruption.