John C. Eichelberger

Magmas in collision: Rethinking chemical zonation in silicic magmas

The heterogeneity of eruptions attributed to protracted fractionation in subvolcanic magma chambers may instead result from chamber recharge. The case of mafic magma intruding andesitic slush is recognized as giving rise to hybrid effusive eruptions such as at Unzen volcano, Japan, and Soufriere Hills volcano, Montserrat. We propose that zoned andesite-rhyolite explosive eruptions are the complement to mafic recharge, where the resident magma is also andesitic but the intruding magma is silicic.

Crystallization history of Obsidian dome, Inyo domes, California

Samples obtained by U.S. Department of Energy research drilling at the 600-year-old Obsidian Dome volcano provide the rare opportunity to examine the transition from volcanic (dome) to plutonic (intrusion) textures in a silicic magma system. Textures in the lavas from Obsidian Dome record multiple periods of crystallization initiated in response to changes in undercooling (ΔT) related to variable degassing in the mag-ma. Phenocr)ysts formed first at low ΔT. A drastic increase in ΔT, related to loss of a vapor phase during initial stages of eruption, caused nucleation of microlites. All of the lavas thus contain phenocrysts and microlites. Extrusion and subsequent devitrification of the dry (0.1 wt% H2O) magma crystallized spherulites and fine-grained rhyolite at high ΔT. A granophyric texture, representing crystallization at a moderate ΔT, formed in the intrusions beneath Obsidian Dome. Textures in the intrusion apparently represent crystallization of hydrous (1–2 wt% H2O) rhyolitic magma at shallow depths.

Gas transport and bubble collapse in rhyolitic magma: an experimental approach

A series of experiments was conducted to test concepts of porous flow degassing of rhyolitic magma during ascent and of the subsequent collapse of vesicles in degassed magma to form obsidian. Dense, synthetically hydrated, natural glasses were pressurized under water-saturated conditions and then decompressed to achieve a range of porosities in the presence of a tracer vapor, D2O. Rapid isotopic exchange indicative of vapor transport rather than of simple diffusion occurred at a porosity >60 vol.%, in accord with earlier gas permeability measurements on cold natural samples. In another series of experiments, natural and synthetic pumices, vesiculated by degassing to atmospheric pressure, rapidly collapsed to dense glass on repressurization to the modest pressures prevailing in lava flows. No relict bubble textures remained. These results support the hypothesis that effusive eruptions result from the syneruptive escape of gas from permeable magmatic foam, and that a process analogous to welding yields dense lavas when such foams are extruded.

Syneruptive mixing, degassing, and crystallization at Redoubt Volcano, eruption of December, 1989 to May 1990

Abstract The 1989–1990 eruption of Redoubt Volcano included initial energetic vertical venting of pumiceous tephra followed by shedding of dense pyroclastic flows from successive unstable domes. The suite of quenched magmatic materials thus produced, coupled with detailed seismic observations, provides a record of establishment and subsequent evolution of an active magma column in the upper crust. The initial eruptive event involved a dacite magma with rhyolite melt and an andesite magma with dacite melt. The dacite magma contains reaction rim-free hornblende, high Cl concentrations (0.12±0.02 wt.% of melt), and is microlite-free: all evidence for rapid ascent from depth. In contrast, the andesite component is microlite-rich, though this characteristic may have arisen by mixing between the two magmas rather than by slow ascent. Abundant dense juvenile lava-like clasts, with low H2O (0.1–0.3 wt.%), were ejected in the initial eruption event, implying that the magma column had reached the near surface and its upper portion had undergone degassing and bubble collapse prior to explosive disruption. Later eruptive events contain evidence of progressive blending of the two magmatic components, as well as mixing between fresh reservoir-derived and old conduit-resident magma. Magma apparently ascended to the surface from the 6–10-km-deep source chamber in a series of pulses, rather than by steady flow or gradual emptying of a shallowly intruded body. Increased shallow long-period seismicity, which may indicate subsurface degassing from the flowing magma column, occurred ≤5 days before each major eruptive event. Petrologic evidence indicates that each eruption contained some magma that ascended quickly from the chamber. With time, the proportion of new magma in eruptive batches may have decreased. However, even the final dome contains magma that left the chamber no more than a few days before extrusion. Together, the pulses apparently record passage of an andesite plume through Redoubts dacite magma reservoir.

Experimental and textural constraints on mafic enclave formation in volcanic rocks

Abstract We have used experiments and textural analysis to investigate the process of enclave formation during magma mixing at Southwest Trident volcano, Alaska. Andesite enclaves are present throughout the four dacite lava flows produced by the eruption, and resemble mafic enclaves commonly found in other volcanic rocks. Our experiments replicate the pressure–temperature–time path taken by enclave-forming andesite magma as it is engulfed in dacite during magma mixing. Pressure and temperature information for the andesite and dacite are from [Coombs et al., Contrib. Mineral. Petrol. 140 (2000) 99–118]. The andesite was annealed at 1000°C, and then cooled to 890°C at rates of 110°C h−1, 10°C h−1, and 2°C h−1. Once cooled to 890°C, andesite was held at this lower temperature from times ranging from 1 to 40 h. The andesite that was cooled at the slower rates of 2°C h−1 and 10°C h−1 most resembles enclave groundmass texturally and compositionally. Based on simple thermal calculations, these rates are more consistent with cooling of the andesite groundmass below an andesite–dacite interface than with cooling of enclave-sized spheres. If enclaves do crystallize as spheres, post-crystallization disaggregation must occur. Calculations using the MELTS algorithm [Ghiorso and Sack, Contrib. Mineral. Petrol. 119 (1995) 197–212] show that for incoming andesite to become less dense than the dacite ∼34 volume % of its groundmass must crystallize to undergo ∼18 volume % vesiculation; these values are similar to those determined for Southwest Trident enclaves. Thus such crystallization may lead to ‘flotation’ of enclaves and be a viable mechanism for enclave formation and dispersal. The residual melt in the cooling experiments did not evolve to rhyolitic compositions such as seen in natural enclaves due to a lack of a decompression step in the experiments. Decompression experiments on Southwest Trident dacite suggest an average ascent rate for the eruption of ∼2–3 MPa h−1. An andesite experiment that was cooled and then decompressed at this rate contains melt that matches that of the natural enclaves. It is apparent that decompression (ascent)-induced crystallization occurs in enclaves, but not in the form of microlites as happens in the dacite host, due either to insufficient residence time at chamber temperatures or to the pre-existing microphenocrysts which act as sites for new growth.

Eruption of andesite triggered by dyke injection: contrasting cases at Karymsky Volcano, Kamchatka and Mt Katmai, Alaska

Arc volcanoes often erupt andesite that appears to have been stored in reservoirs at shallow depth for protracted periods. As crystal-rich andesite is close in density to upper crust, such storage may be quite stable. Petrological evidence, and occasionally geological and geophysical evidence as well, suggests that the immediate trigger for eruption of the stored magma is injection of new magma into the reservoir, presumably through dykes rising from depth. When the dyke magma is more mafic than the stored andesite, effusive eruption typically results. When the dyke magma is voluminous and more silicic, the results are catastrophic, with production of discontinuously zoned tephra deposits and caldera collapse. Contrasting end-members are illustrated by the eruptions of Karymsky Volcano in 1996 and of Mt Katmai in 1912.

Calcic cores of plagioclase phenocrysts in andesite from Karymsky volcano: Evidence for rapid introduction by basaltic replenishment

Calcic cores in plagioclase of Karymsky andesite of the 1996–2000 eruptive cycle texturally and compositionally (both trace and major elements) mimic the plagioclase phenocrysts of basalt erupted 6 km away at the onset of the cycle. These observations support the view that simultaneous eruption of andesite and basalt at Karymsky in the beginning of the cycle represents an example of replenishment and eruption triggering of an andesitic reservoir. Homogeneity of andesitic output occurred within two months. This suggests to us that blending of injected basalt into reservoir magma was thorough and rapid.

Rhyolite intrusions in the intracaldera Bishop Tuff, Long Valley Caldera, California

Abstract Drilling of the Long Valley Exploratory Well on the resurgent dome in the Long Valley Caldera revealed > 300 m cumulative thickness of granophyric intrusions within the 1180-m-thick, 760 ka intracaldera Bishop Tuff. The intrusions are aphyric to sparsely plagioclase-phyric, high-silica, high-barium and low-strontium rhyolites. They resemble the lavas of the Early Rhyolite, the first phase of post-caldera volcanism. A mean 40 Ar/ 39 Ar age of 590 ± 17 ka from a part of a shallow intrusion is coeval with Early Rhyolite volcanism. A second mean age of 454 ± 17 ka from the same intrusion may reflect either younger Early Rhyolite activity with no external equivalent or hydrothermal resetting of the argon system. Hydrothermal alteration of the intrusions is characterized by introduction of quartz, calcite and pyrite and formation of illite/smectite. High CO 2 content of fluids apparently inhibited zeolite formation. Alteration varies locally within intrusions and intrusive groups and does not vary systematically with depth. Oxygen shows consistent depletion of the 18 O isotope from an initial magmatic composition of + 6.0 to + 8.5‰ to values ranging from +1.4 to −0.4‰. The constant oxygen isotope depletion most likely reflects alteration of intrusions due to local emplacement-induced hydrothermal circulation rather than a caldera-scale hydrothermal system. In contrast, 18 O depletion of the host Bishop Tuff increases regularly with depth (except at an intrusive contact). A pre-Early Rhyolite geothermal gradient of approximately 70 °C/km was inferred. This is substantially higher than the current gradient but substantially lower than expected for the case of a conductive regime over a shallow residual magma chamber. Either the intrusions were fed from a deep chamber, or a cool hydrologic recharge regime was established early in caldera history. The age, thickness and suspected lateral extent of these shallow intrusions are such that emplacement of the intrusions, rather than inflation of a shallow chamber, is responsible for resurgence of the central Long Valley Caldera. Similar intrusions occur in another well on the resurgent dome ( LV13-21 ) but not in wells located off the dome.

Oxygen isotope compositions of intracaldera rocks: hydrothermal history of the Long Valley Caldera, California

The Bishop Tuff and associated rocks filling Long Valley Caldera, California represent a case where thick and relatively uniform geologic units of known initial compositions have been subjected to a strong geothermal fluid flux within an enclosed basin. Oxygen isotopic compositions within individual components of this ‘flux plate’ are used to characterize a magmatically driven paleohydrothermal system. Oxygen isotope ratios of silicate components analyzed with a laser-probe/mass-spectrometer system clearly illustrate the isotopic heterogeneity in the hydrothermally altered felsic volcanic rocks. The alteration resulted from moderately high-temperature hydrothermal activity. Hydrothermal activity is linked to periods of post-caldera-collapse volcanism and resurgence in the central caldera. In general, 18O depletion in the Bishop Tuff resulting from exchange with hydrothermal fluids proceeds as pumice matrix sanidine quartz due to progressive resistance to alteration. Samples from the Long Valley Exploratory Well (LVEW) centered on the resurgent dome show downward increasing exchange and disequilibrium over the wells 2000-m depth. A flanking well shows the opposite pattern over 1300 m depth. Oxygen isotope isopleths along a W-E vertical cross section of the caldera reveals convective circulation upwards and outwards from the central resurgent dome. A paleotemperature maximum of ~ 350 °C in LVEW and a geothermal gradient of ~ 130 °C/km are deduced from feldspar/water oxygen isotope fractionation equations using the pumice and sanidine components of the intracaldera volcanic rocks. This is consistent with the degree of alteration and the secondary mineralogy. Present-day temperature maximum in the well is ~ 100 °C, and the geothermal gradient is 50 °C/km. Fossil hydrothermal water, trapped in fluid inclusions in hydrothermal vein quartz hosted by intracaldera rocks, has a calculated oxygen isotope composition of − 10.2%. based on chemistry and trapping temperature. This is 4%. heavier than the average value for the present-day hydrothermal water and consistent with higher temperatures and water/rock ratios. Intrusive activity was initiated in the resurgent dome area soon after caldera collapse (760 ka), but migrated toward the western margin of the caldera, the locus of Holocene volcanism. Thus, the central hydrothermal system matured and waned as the heat sources from intrusions shifted westward. Geophysical evidence of renewed intrusive activity located beneath the resurgent dome may presage a future renewal of the cycle of hydrothermal activity in the central caldera region. Although the hydrothermal activity is complex and transitory, this history is normal for a large, relatively active caldera system such as Long Valley Caldera.

Comagmatic granophyre and dacite from Karymsky volcanic center, Kamchatka: experimental constraints for magma storage conditions

Abstract Despite a ∼30 000 years difference in age, two caldera-forming eruptions at Karymsky volcanic center, Kamchatka – Karymsky (7900 yr BP) and Academy Nauk (ca. 40 000 yr BP) – produced two-pyroxene dacites with the same composition and mineralogy. Granophyric xenoliths of the identical whole-rock chemistry were found in basalts erupted within Academy Nauk caldera in 1996. Unlike the dacites, however, the granophyres are holo-crystalline and contain biotite and amphibole. Large amphibole phenocrysts contain rare inclusions of clinopyroxenes, which compositionally overlap with clinopyroxenes in the dacites. The Al content of the amphibole suggests it grew at a pressure of about 100 MPa. Results of hydrothermal experiments and petrologic observations indicate that Academy Nauk and Karymsky dacites were last equilibrated at 883±19°C, 100±15 MPa and 871±19°C, 85±18 MPa, respectively, both at water-saturated conditions. The mineral assemblage of granophyre is reproduced by isobaric crystallization of the dacite at 100 MPa, implying that the granophyres were sampled from the crystallized silicic reservoir that produced the caldera-forming eruption of Academy Nauk. Similar chemical compositions of Karymsky and Academy Nauk dacites indicate that both were derived from the same crustal-level source. The eruptive history of the calderas can best be explained by two 10–12-km 3 dacitic batches that detached from a parental body situated in the lower crust, then ascended to 3–4 km depth, re-equilibrated, and erupted.