Petrology and Fe–Ti oxide reequilibration of the 1991 Mount Unzen mixed magma
Introduction
The 1991–1995 eruption of Mount Unzen provides an opportunity to examine the petrologic effects produced by a magma mixing event and to use this data to determine the timing of magma mixing. The eruption consisted primarily of lava extrusion and dome growth, with intermittent dome collapse that produced numerous pyroclastic flows. There was also a vulcanian explosion on June 11, 1991 (Nakada et al., 1997). The eruptive cycle began with volcano-tectonic earthquakes beneath the Tachibana (Chijiwa) Bay (20 km west of Mt. Fugen) in November 1989 (Nakada and Shimizu, 1997). From November 1989 to May 1991 the earthquake epicenters moved toward the summit craters, culminating with a swarm of shallow high-frequency earthquakes 8 days before the first period of dome growth, May 20, 1991 (Shimizu et al., 1997). Prior to dome growth, there were phreatic explosions beginning on February 12, 1991, which included juvenile magma by April 9, 1991. Dome growth and pyroclastic flows continued through 1995. Samples from the 1991–1995 Mount Unzen eruption contain dacitic composition phenocrysts of plagioclase, hornblende, biotite, magnetite, ilmenite, and quartz (Nakada and Fujii, 1993; Nakamura, 1995) and a groundmass of plagioclase, pyroxene, magnetite, ilmenite, and pargasite (Fig. 1). Many of the phenocrysts contain evidence of a magma-mixing event in the form of reverse zoning, partial dissolution, and reaction rims. In addition, there is a high-Al pargasite in the groundmass that is in sharp compositional contrast with the low-Al hornblende phenocrysts (Nakada and Fujii, 1993; Nakada et al., 1995). The reaction rims around many of the biotite phenocrysts also contain pargasite (Fig. 2). The crystallization of pargasite is interpreted to have been caused by an injection of a higher temperature magma that is preserved as aphyric enclaves averaging 56 wt.% SiO2 (Nakada et al., 1997). These enclaves, however, commonly contain scattered (1–5%) xenocrysts of plagioclase, and/or embayed quartz, from the low-temperature dacite.
The Mount Unzen magnetite and ilmenite phenocrysts contain prominent zoning making it possible to estimate temperatures at different times in the magma history (Nakamura, 1995). These temperature vs. time estimates assume that diffusion in the magnetite and ilmenite occur at the same rate. Temperatures and oxygen fugacity calculations based on coexisting titaniferous magnetite and ferroan ilmenite began with the original work of Buddington and Lindsley (1964). Since then there have been numerous refinements to both the mineral reformulation model and the temperature–oxygen fugacity algorithm (Carmichael, 1967; Powell and Powell, 1977; Spencer and Lindsley, 1981; Stormer, 1983; Andersen and Lindsley, 1988; Ghiorso and Sack, 1991; Andersen et al., 1993) that have increased our ability to determine these parameters. To obtain original magmatic temperatures and oxygen fugacities, it is assumed that the magma's phenocryst assemblage is in equilibrium and that the Fe–Ti oxides have retained their preeruption compositions. For explosively erupted and rapidly cooled volcanic samples we expect the Fe–Ti oxide compositions to reflect the conditions in the preeruption magma storage region prior to eruption. Many rapidly cooled volcanic samples, however, preserve a range in Fe–Ti oxide compositions, indicating a lack of equilibrium (Nakamura, 1995; Venezky and Rutherford, 1997).
The rate at which Fe–Ti oxides reequilibrate has become a focus of study because nonequilibrated oxides are found in many volcanic rocks, and it has been recognized that they contain potential rate data for eruption processes. Hammond and Taylor (1982)showed that for slowly cooling systems, Fe–Ti oxides give lower temperatures than would be expected due to reequilibration during cooling. Fe–Ti reequilibration was interpreted to occur by the oxidation of titanomagnetite and reduction of ilmenite with the latter step as the rate-limiting process. Frost and Lindsley (1991)and Lindsley et al. (1991)suggested that Fe–Ti oxide pairs in the Bishop Tuff pyroclastic flows reequilibrated during cooling at the surface. Frost and Lindsley (1992)suggest an alternative explanation for the Bishop Tuff involving the mixing of two magmas with similar magnetite compositions and different ilmenites. They were able to demonstrate that this is a viable solution using the QUILF model, which utilizes equilibrium between Fe–Ti oxides and several silicate minerals (Andersen et al., 1993). Gardner et al. (1995)found that the rims of Fe–Ti oxides could be experimentally reequilibrated in 4 days at 850°C. Nakamura (1995)used D values for Fe–Ti diffusion from Freer and Hauptman (1978)to determine the time required to develop composition profiles (20 μm in length with a 6 mol% ulvospinel difference) in titaniferous magnetite. The profiles and diffusion data indicate that mixing was a continuous process throughout the 1991–1995 Mount Unzen eruption (Nakamura, 1995).
The present study adds to the work of Nakamura (1995)by examining in more detail the compositions of different size magnetite and ilmenite phenocrysts found in the Mount Unzen samples, and by comparing them to Fe–Ti oxide reequilibration experiments using a similar composition magma. The comparison of Fe–Ti oxide pairs in the experimental and natural samples allow for the development of a time–temperature history of the Mount Unzen eruption. Additionally, an experimental phase equilibrium study of the Mount Unzen June 11, 1991 mixed magma is used to corroborate the Fe–Ti oxide temperatures.
Section snippets
Methods
Mount St. Helens samples from two different eruptions were used for experiments on Fe–Ti oxide reequilibration because they bracket the Mount Unzen bulk composition and they contain unzoned magnetite and ilmenite (Rutherford and Devine, 1988; Gardner et al., 1995). The hydrothermal experiments on crushed dacitic pumices from the Mount St. Helens May 18, 1980, (MSH-18) and MSH Yn (Yn) deposits were carried out at different temperatures, pressures, and times in cold seal pressure vessels. The
Petrology of Mount Unzen magmas
As previously described (Nakada and Fujii, 1993; Nakamura, 1995; Sato et al., 1997), the magma erupted at Mount Unzen in the 1991–1995 event consists of a coarse-grained dacite containing phenocrysts (<1 cm) of plagioclase, hornblende, biotite, quartz, magnetite, and ilmenite, in order of decreasing abundance. The large hornblende phenocrysts are slightly zoned with 6.9 wt.% Al2O3 cores and have little to no breakdown (reaction rim) where in contact with the groundmass. Mafic enclaves of
Reequilibration process
The reequilibration of Fe–Ti oxides following a temperature change is divided into three different processes; oxide–oxide reequilibration, intraoxide reequilibration, and oxide–silicate reequilibration (Frost, 1991). In oxide–oxide reequilibration there is an Fe2+Ti=2Fe3+ exchange: Fe2TiO4 (in-mag)+Fe2O3 (in-ilm)=FeTiO3 (ilm)+Fe3O4 (mag). If the reequilibration occurs only by the Fe–Ti exchange with the magnetite–ilmenite boundary remaining in the same place, then we can use data from Fe–Ti
Mount Unzen magma temperatures
The Fe–Ti oxide phenocrysts in the Mount Unzen magma, as described above, are chemically zoned so that the cores yield temperatures of 780–800°C whereas the magnetite–ilmenite contacts give temperatures of 900±20°C. Rims of magnetite adjacent to melt are somewhat more variable, giving 840 to 890°C temperatures. One possible explanation for the lower temperatures here is that the large grains originally began to dissolve and that reequilibration by diffusion at the melt–oxide interface occurred
Mount Unzen phase equilibria
In order to further determine the conditions during mixing in the Unzen magma, we calculated the pressure and the depth of the rhyodacitic magma prior to mixing using the Al-in-hornblende geobarometer (Johnson and Rutherford, 1989), and we also determined PH2O–T phase equilibria for this magma composition. The Al-in-hornblende geobarometer is used with the large hornblende phenocrysts because this hornblende, quartz, and other phases required for the geobarometer were apparently part of the
Magma ascent rate
The new 850° to 870°C estimate for the temperature of the Unzen mixed magma based on Fe–Ti oxides, and supported by the phase equilibria, indicate that the ascent of this magma was somewhat faster than previously suggested (Nakamura, 1995; Nakada et al., 1995). We assume that the magma ascent began at the time of magma mixing to obtain a maximum estimate of the ascent rate. The diffusion calculations (Table 4) indicate that the time to produce the 20-μm profiles seen in the Unzen titaniferous
Conclusions
The 1991–1995 Mount Unzen eruption contained magma which appears to be a recent mixture of a phenocryst-rich dacite and an aphyric andesite (e.g., Nakada et al., 1995, Nakada et al., 1997). The following conclusions are made regarding the magmatic conditions and the magma mixing process based on the experiments and analyses done in this study:
(1) The preeruption rhyodacite magma was at a depth of 6 to 7 km (160 MPa) based on Al-in-hornblende geobarometry, and at a temperature of 790±20°C. The
Acknowledgements
We would like to thank Don Lindsley, Jim Gardner, Dick Yund, and Michihiko Nakamura for their reviews of this manuscript, although responsibility for the conclusions remains with the authors. This research was supported by NSF grant EAR-97-06251.
References (31)
- et al.
A pascal program to assess equilibria among Fe–Mg–Ti oxides, pyroxenes, olivine, and quartz
Comp. Geosci.
(1993) - et al.
An experimental study of magnetite–titanomagnetite interdiffusion
Phys. Earth Planet. Inter.
(1978) - et al.
The ilmenite/magnetite assemblage: kinetics of re-equilibration
Earth Planet. Sci. Lett.
(1982) - et al.
Preliminary report on the activity at Unzen Volcano (Japan), November 1990–November 1991: dacite lava domes and pyroclastic flows
J. Volcanol. Geotherm. Res.
(1993) - et al.
Internally consistent solution models for Fe–Mg–Mn–Ti oxides: Fe–Ti oxides
Am. Mineral.
(1988) - et al.
Mg/Mn partitioning as a test for equilibrium between coexisting Fe–Ti oxides
Am. Mineral.
(1988) - et al.
Iron–titanium oxide minerals and synthetic equivalents
J. Petrol.
(1964) Iron–titanium oxides of salic volcanic rocks and their associated ferromagnesian silicates
Contrib. Mineral. Petrol.
(1967)- et al.
Comparison of microanalytical methods for estimating H2O contents of silicic volcanic glasses
Am. Mineral.
(1995) - et al.
Crystallization and fractionation trends in the system andesite–H2O–CO2–O2 at pressures to 10 kb
Geol. Soc. Am. Bull.
(1973)
Magnetic petrology: factors that control the occurrence of magnetite in crustal rocks
Mineral. Soc. Am. Rev. Mineral.
Occurrence of iron titanium oxides in igneous rocks
Mineral. Soc. Am. Rev. Mineral.
Equilibrium among Fe–Ti oxides, pyroxenes, olivine, and quartz. Part II: Applications
Am. Mineral.
Petrologic diversity in Mount St. Helens dacites during the last 4000 years: implications for magma mixing
Contrib. Mineral. Petrol.
Cited by (188)
Two-stage evolution of a bimodal reservoir: The case of Holocene lavas of the Lanín composite volcano, Southern Volcanic Zone, Chile
2024, Journal of South American Earth SciencesProlonged near-solidus and steady-state magma storage for the Youngest Toba Tuff: Evidence from TitaniQ thermometry and diffusion chronometry
2023, Earth and Planetary Science LettersMagmatic differentiation within the transcrustal magma reservoir of Develidağ volcano, Central Anatolia (Turkey)
2023, Journal of Asian Earth SciencesDiffusion-driven Zn and Mg isotope fractionation in magmatic Fe-Ti-Cr oxides and implications for timescales of magmatic processes
2023, Geochimica et Cosmochimica ActaTracking caldera cycles in the Aso magmatic system – Applications of magnetite composition as a proxy for differentiation
2023, Journal of Volcanology and Geothermal ResearchDating individual zones in phenocrysts from the 2016–2017 eruption of Bogoslof volcano provides constraints on timescales of magmatic processes
2023, Journal of Volcanology and Geothermal Research