Jessica F. Larsen

Eruption of Alaska volcano breaks historic pattern

In the late morning of 12 July 2008, the Alaska Volcano Observatory (AVO) received an unexpected call from the U.S. Coast Guard, reporting an explosive volcanic eruption in the central Aleutians in the vicinity of Okmok volcano, a relatively young (∼2000-year-old) caldera. The Coast Guard had received an emergency call requesting assistance from a family living at a cattle ranch on the flanks of the volcano, who reported loud “thunder,” lightning, and noontime darkness due to ashfall. AVO staff immediately confirmed the report by observing a strong eruption signal recorded on the Okmok seismic network and the presence of a large dark ash cloud above Okmok in satellite imagery. Within 5 minutes of the call, AVO declared the volcano at aviation code red, signifying that a highly explosive, ash-rich eruption was under way.

Experimental study of plagioclase rim growth around anorthite seed crystals in rhyodacitic melt

Abstract The purpose of this study was to replicate experimentally the growth of new rims around highly anorthitic plagioclase “core”. phenocrysts, analogous to the incorporation of xenocrysts into a silicic magma body through magma mixing. Aniakchak rhyodacite forms the bulk starting composition, and phase-equilibria experiments constrained the pre-eruption magma conditions to be ~110 MPa and 870.880 °C. The experimental runs were seeded with Great Sitkin anorthite (An91-95) crystals. New rim growth of An28-38 plagioclase occurred at rates between 3.5 (±0.3) × 10.10 to 60.6 (±20.0) × 10.10 cm/s at pressures and temperatures from 50 to 150 MPa and 825 to 880 °C. The values in parentheses are ±1σ standard deviation. Microlite crystallization (An27-41) occurred in all experiments within the plagioclase stability field, and their growth rates varied from 4.4 (±1.3) × 10.10 to 65.7 (±10.1) × 10.10 cm/s. The rim and microlite growth rates are similar to one another within each experiment, and microlite number density (NV) is correlated approximately inversely with rim growth rates. Microlite crystallinities increased from 4.2 to 49.7 vol% as a function of increasing ΔTeff up to 95 °C. The results indicate growth-dominated crystallization at low ΔTeff, and nucleation dominated crystallization at high ΔTeff, in agreement with previous studies. Assuming the experiments apply to nature, the rim growth rates can provide a minimum estimate on how fast magma mixing can occur. Rims that are 10 to 100 μm wide can grow in ~10 days to 4 months, recording fast mixing timescales as long as eruption occurs shortly after mixing. The growth rate estimates presented here generally agree with those derived from sodic rims growing around anorthite cores after mixing between basalt and andesite prior to the 1996 eruption of Karymsky volcano, Kamchatka.

Experimental constraints on bubble interactions in rhyolite melts: implications for vesicle size distributions

We have studied interactions between bubbles of two distinct size classes in rhyolite melts experimentally decompressed between 200 and 80 MPa. The first set of ‘decompression’ bubbles has a size range (Rdec) of 1–11 μm and is formed from nucleation and growth upon isothermal decompression of the melt. The larger populations of ‘hydration’ bubbles are on average 30–40 μm in radius (Rhyd) and are formed from pore spaces present that were filled with water vapor during the saturation runs prior to the decompression experiments. The first type of interaction results in the elongation of decompression bubbles oriented radially around the larger hydration bubbles. The degree of elongation increases both as a function of distance and with increasing ratio of hydration to decompression bubble size (Rhyd/Rdec). The second type of interaction studied results in a reduction of the size of decompression bubbles located within a range of distances from 10 to 65 μm from a hydration bubble surface, relative to the modal size of the unaffected bubbles in the same sample. In addition, within an average distance of 10 μm, melt next to the hydration bubble surface is depleted in decompression bubbles. Our results indicate that concentration gradients in the melt are probably responsible for bubble size reduction and the depleted zones, because the predicted time scales for Ostwald ripening are much longer than those of the experiments. These effects persist even to the lowest ending pressures studied (80 MPa), which indicates that size distributions of small bubbles may be affected by concentration gradients in the depleted melt shell surrounding large bubbles. Large bubbles present in an ascending magma, prior to a subsequent nucleation event, could therefore affect the growth of the smaller bubble population occurring within the depleted melt shell of the larger bubbles, and produce a bimodal vesicle size distribution. Elongated decompression bubbles may be strained as a result of melt flowing away from the much larger hydration bubbles as they grow. Estimates of capillary number (Ca) plotted against deformation (Df) indicate that bubbles in water-rich rhyolite melts are deformable, even at small sizes (1 μm) and small values of Ca. Our results show a different trend of Df with Ca than previous studies in non-geological systems predict, indicating that viscosity effects may be important. The preservation of deformation textures depends strongly on relaxation time, explaining the lack of deformation textures in less viscous natural lavas.

Late Pleistocene and Holocene Caldera-Forming Eruptions of Okmok Caldera, Aleutian Islands, Alaska

Okmok volcano, in the central Aleutian arc, Alaska, produced two caldera-forming eruptions within the last ∼12,000 years. This study describes the stratigraphy, composition, and petrology of those two eruptions. Both eruptions initially produced small volumes of felsic magmas, followed by voluminous andesite and basaltic andesite. The Okmok I eruption produced >30 km 3 DRE of material on Umnak Island, and Okmok II ∼15 km 3 . However, a significant proportion of material not accounted for here was deposited into the oceans during both events. The Okmok I pyroclastic flow deposits contain evidence for interaction with snow/ice, particularly along the northern flanks of the caldera. Although both Okmok I and II eruptions involved a phreatomagmatic component, the accumulation of a large volume (>15km 3 ) of volatile-rich, mafic-intermediate magma in the shallow crust may provide the driving force for the catastrophic eruptions. Agglutinate deposits associated with Okmok II indicate energetic lava fountaining simultaneous with caldera-collapse, similar to other descriptions of mafic-intermediate caldera-forming deposits such as in the New Hebrides.

A micro-reflectance IR spectroscopy method for analyzing volatile species in basaltic, andesitic, phonolitic, and rhyolitic glasses

Abstract Volatile contents of geologic glasses are used to model magma chamber and degassing processes, thus, there is considerable interest in small-scale analytical techniques for analyzing volatiles in glasses. Infrared (IR) spectroscopy has the advantage of determining volatile speciation in glasses (e.g., OH-, molecular H2O, molecular CO2, and CO3 2-). However, sample preparation for the most common IR method used, micro-transmission IR spectroscopy, is complicated because glasses must be prepared as thin, parallel-sided wafers. Raman analysis, while valuable for Fe-poor samples, can be difficult to use for Fe-rich glasses. We have calibrated a micro-reflectance infrared method for determining volatile species using calculated Kramers-Kronig absorbance (KK-Abs.) spectra that requires that only one side of a glass be polished. The method is easier to use than other reflectance methods where it is difficult to determine the baseline for the IR bands. Total H2O wt% = m·(3600 cm-1 KK-Abs.), where m, is the slope of the calibration line that is obtained from a fit to the data. The m value is related to the calculated refractive index, n, for a range of aluminosilicate glass compositions allowing the technique to be applied to samples with unknown calibration slopes. For calc-alkaline andesite glasses we determined calibration slopes for micro-reflectance IR measurements of molecular H2O, molecular CO2, and CO3 2-. The method has been calibrated for glasses with up to 6.76 wt% total H2O (but is useful for glasses with more than 20 wt% total H2O) and has been calibrated for glasses with up to 0.575 wt% total CO2. This technique provides a means to analyze volatile abundances in samples that are not possible to analyze or prepare for analysis with transmission micro-IR techniques. We have determined volatile contents in fragile samples such as cracked, vesicular, or crystal-bearing glasses formed by volcanic or impact processes or in high-pressure bubble nucleation experiments and H diffusion experiments. We have monitored H uptake during weathering of basaltic glasses that cannot be polished and determined volatiles in melt inclusions and pumice

Pre-eruptive magma mixing and crystal transfer revealed by phenocryst and microlite compositions in basaltic andesite from the 2008 eruption of Kasatochi Island volcano

Abstract The August 7-8, 2008, eruption of Kasatochi Island volcano, located in the central Aleutians Islands, Alaska, produced abundant, compositionally heterogeneous basaltic andesite (52-55 wt% SiO2) that has been interpreted to result from pre-eruptive magma mixing. The basaltic andesite contains two populations of plagioclase phenocrysts. The first, volumetrically dominant population consists of oscillatory-zoned phenocrysts with an overall normal zonation trend toward comparatively sodic rims (An55-65), interrupted by dissolution features and spikes in calcium content (up to ~An85). The second population consists of phenocrysts with highly calcic compositions (~An90). These phenocrysts contain sharp decreases in calcium content close to their rims (reaching as low as ~An60), but are otherwise texturally and compositionally homogeneous. Groundmass plagioclase microlites are generally much more calcic than rims of the first phenocryst population, with more than 50% of measured microlites containing >An80. Major, minor, and trace element concentrations of plagioclase microlites and phenocrysts indicate that oscillatory-zoned phenocrysts derived from cooler (800-950 °C), more silicic mixing magma, while unzoned, calcic phenocrysts were associated with hotter (900-1050 °C), mafic magma. The mixing of these magmas just prior to eruption, followed by decompression during the eruption itself created high effective undercoolings in the mafic end-member, and lead to the nucleation of high-An microlites. MgO and FeO concentrations of plagioclase microlites and high-An phenocryst rims (up to ~0.4 and ~1.3 wt%, respectively) provide further evidence for high mixing- and eruptioninduced effective undercoolings.

Dynamics of an unusual cone-building trachyte eruption at Pu‘u Wa‘awa‘a, Hualālai volcano, Hawai‘i

The Pu‘u Wa‘awa‘a pyroclastic cone and Pu‘u Anahulu lava flow are two prominent monogenetic eruptive features assumed to result from a single eruption during the trachyte-dominated early post-shield stage of Hualālai volcano (Hawaiʻi). Puʻu Wa‘awa‘a is composed of complex repetitions of crudely cross-stratified units rich in dark dense clasts, which reversely grade into coarser pumice-rich units. Pyroclasts from the cone are extremely diverse texturally, ranging from glassy obsidian to vesicular scoria or pumice, in addition to fully crystalline end-members. The >100-m thick Pu‘u Anahulu flow is, in contrast, entirely holocrystalline. Using field observations coupled with whole rock analyses, this study aimed to test whether the Pu‘u Wa‘awa‘a tephra and Pu‘u Anahulu lava flows originated from the same eruption, as had been previously assumed. Crystal and vesicle textures are characterized along with the volatile contents of interstitial glasses to determine the origin of textural variability within Pu‘u Waʻawaʻa trachytes (e.g., magma mixing vs. degassing origin). We find that (1) the two eruptions likely originated from distinct vents and magma reservoirs, despite their proximity and similar age, (2) the textural diversity of pyroclasts forming Pu‘u Wa‘awa‘a can be fully explained by variable magma degassing and outgassing within the conduit, (3) the Pu‘u Wa‘awa‘a cone was constructed during explosions transitional in style between violent Strombolian and Vulcanian, involving the formation of a large cone and with repeated disruption of conduit plugs, but without production of large pyroclastic density currents (PDCs), and (4) the contrasting eruption styles of Hawaiian trachytes (flow-, cone-, and PDC-forming) are probably related to differences in the outgassing capacity of the magmas prior to reaching the surface and not in intrinsic compositional or temperature properties. These results further highlight that trachytes are “kinetically faster” magmas compared to dacites or rhyolites, likely degassing and crystallizing more rapidly.

Water-magma interaction and plume processes in the 2008 Okmok eruption, Alaska

Eruptions of similar explosivity can have divergent effects on the surroundings due to differences in the behavior of the tephra in the eruption column and atmosphere. Okmok volcano, located on Umnak Island in the eastern Aleutian Islands, erupted explosively between 12 July and 19 August 2008. The basaltic andesitic eruption ejected ∼0.24 km3 dense rock equivalent (DRE) of tephra, primarily directed to the northeast of the vent area. The first 4 h of the eruption produced dominantly coarse-grained tephra, but the following 5 wk of the eruption deposited almost exclusively ash, much of it very fine and deposited as ash pellets and ashy rain and mist. Meteorological storms combined with abundant plume water to efficiently scrub ash from the eruption column, with a rapid decrease in deposit thickness with distance from the vent. Grain-size analysis shows that the modes (although not their relative proportions) are very constant throughout the deposit, implying that the fragmentation mechanisms did not vary much. Grain-shape features consistent with molten fuel-coolant interaction are common. Surface and groundwater drainage into the vents provided the water for phreatomagmatic fragmentation. The available water (water that could reach the vent area during the eruption) was ∼2.8 × 1010 kg, and the erupted magma totaled ∼7 × 1011 kg, which yield an overall water:magma mass ratio of ∼0.04, but much of the water was not interactive. Although magma flux dropped from 1 × 107 kg/s during the initial 4 h to 1.8 × 105 kg/s for the remainder of the eruption, most of the erupted material was ejected during the lower-mass-flux period due to its much greater length, and this tephra was dominantly deposited within 10 km downwind of the vent. This highlights the importance of ash scrubbing in the evaluation of hazards from explosive eruptions.

Spatial point pattern analysis applied to bubble nucleation in silicate melts

Experimental bubble nucleation studies are used for determining the nucleation mechanism as a function of experimental conditions, the resulting bubble number density, and can also yield estimates of the melt-vapor surface tension. This provides important information on gas exsolution in silicate melts, which can be applied towards understanding magmatic degassing in volcanic conduits. At present, determination of nucleation processes in tiny experimental samples relies upon visual observations. To improve the characterization of the spatial distribution of bubbles, we present a new application of spatial point pattern analysis. This technique allows the quantitative description of the spatial distribution of nucleation sites and has the potential to distinguish between homogeneous, heterogeneous, and multiple nucleation events. Since point pattern analysis highlights clustering or spatial regularity among objects, it may improve our understanding of the melt structure underlying the spatial distribution of nucleation sites, as well as interactions between bubble populations resulting from different nucleation pulses within a single experimental sample.