Carl R. Thornber

Geothermometry of Kilauea Iki lava lake, Hawaii

Data on the variation of temperature with time and in space are essential to a complete understanding of the crystallization history of basaltic magma in Kilauea Iki lava lake. Methods used to determine temperatures in the lake have included direct, downhole thermocouple measurements and Fe-Ti oxide geothermometry. In addition, the temperature variations of MgO and CaO contents of glasses, as determined in melting experiments on appropriate Kilauean samples, have been calibrated for use as purely empirical geothermometers and are directly applicable to interstitial glasses in olivine-bearing core from Kilauea Iki. The uncertainty in inferred quenching temperatures is ±8−10° C. Comparison of the three methods shows that (1) oxide and glass geothermometry give results that are consistent with each other and consistent with the petrography and relative position of samples, (2) downhole thermo-couple measurements are low in all but the earliest, shallowest holes because the deeper holes never completely recover to predrilling temperatures, (3) glass geothermometry provides the greatest detail on temperature profiles in the partially molten zone, much of which is otherwise inaccessible, and (4) all three methods are necessary to construct a complete temperature profile for any given drill hole. Application of glass-based geothermometry to partially molten drill core recovered in 1975–1981 reveals in great detail the variation of temperature, in both time and space, within the partially molten zone of Kilauea Iki lava lake. The geothermometers developed here are also potentially applicable to glassy samples from other Kilauea lava lakes and to rapidly quenched lava samples from eruptions of Kilauea and Mauna Loa.

Vapor transfer prior to the October 2004 eruption of Mount St. Helens, Washington

Dome lavas from the 2004 eruption of Mount St. Helens show elevated Li contents in plagioclase phenocrysts at the onset of dome growth in October 2004. These cannot be explained by variations in plagioclase-melt partitioning, but require elevated Li contents in coexisting melt, a fact confirmed by measurements of Li contents as high as 207 µg/g in coexisting melt inclusions. Similar Li enrichment has been observed in material erupted prior to and during the climactic May 1980 eruption, and is likewise best explained via pre-eruptive transfer of an exsolved alkali-rich vapor phase derived from deeper within the magma transport system. Unlike 1980, however, high Li samples from 2004 show no evidence of excess (210Pb)/(226Ra), implying that measurable Li enrichments may occur despite significant differences in the timing and/or extent of magmatic degassing. Diffusion modeling shows that Li enrichment occurred within ∼1 yr before eruption, and that magma remained Li enriched until immediately before eruption and cooling. This short flux time and the very high Li contents in ash produced by phreatomagmatic activity prior to the onset of dome extrusion suggest that vapor transfer and accumulation were associated with initiation of the current eruption. Overall, observation of a high Li signature in both 1980 and 2004 dacites indicates that Li enrichment may be a relatively common phenomenon, and may prove useful for petrologic monitoring of Mount St. Helens and other silicic volcanoes. Lithium diffusion is also sufficiently rapid to constrain vapor transfer on similar time scales to short-lived radionuclides.

Chronology of the episode 54 eruption at Kilauea Volcano, Hawaii, from GOES-9 satellite data

The free availability of GOES satellite data every 15 minutes makes these data an attractive tool for studying short-term changes on cloud-free volcanoes in the Pacific basin. We use cloud-free GOES-9 data to investigate the chronology of the January 1997, episode 54 eruption of Kilauea Volcano, Hawaii. Seventy-six images for this effusive eruption were collected over a 60-hour period and show the opening and shutdown of active fissures, the draining and refilling of the Pu‘u ‘O‘o lava lake, and the cessation of activity at the ocean entry.

Morphology and composition of spinel in Pu'u 'O'o lava (1996-1998), Kilauea volcano, Hawaii

Abstract The morphology and composition of spinel in rapidly quenched Pu’u ’O’o vent and lava tube samples are described. These samples contain glass, olivine phenocrysts (3–5 vol.%) and microphenocrysts of spinel (∼0.05 vol.%). The spinel surrounded by glass occurs as idiomorphic octahedra 5–50 μm in diameter and as chains of octahedra that are oriented with respect to each other. Spinel enclosed by olivine phenocrysts is sometimes rounded and does not generally form chains. The temperature before quenching was calculated from the MgO content of the glass and ranges from 1150°C to 1180°C. The oxygen fugacity before quenching was calculated by two independent methods and the log fO2 ranged from −9.2 to −9.9 (delta QFM=−1). The spinel in the Pu’u ’O’o samples has a narrow range in composition with Cr/(Cr+Al)=0.61 to 0.73 and Fe2+/(Fe2++Mg)=0.46 to 0.56. The lower the calculated temperature for the samples, the higher the average Fe2+/(Fe2++Mg), Fe3+ and Ti in the spinel. Most zoned spinel crystals decrease in Cr/(Cr+Al) from core to rim and, in the chains, the Cr/(Cr+Al) is greater in the core of larger crystals than in the core of smaller crystals. The occurrence of chains and hopper crystals and the presence of Cr/(Cr+Al) zoning from core to rim of the spinel suggest diffusion-controlled growth of the crystals. Some of the spinel crystals may have grown rapidly under the turbulent conditions of the summit reservoir and in the flowing lava, and the crystals may have remained in suspension for a considerable period. The rapid growth may have caused very local (μm) gradients of Cr in the melt ahead of the spinel crystal faces. The crystals seem to have retained the Cr/(Cr+Al) ratio that developed during the original growth of the crystal, but the Fe2+/(Fe2++Mg) ratio may have equilibrated fairly rapidly with the changing melt composition due to olivine crystallization. Six of the samples were collected on the same day at various locations along a 10-km lava tube and the calculated pre-collection temperatures of the samples show a 5°C drop with distance from the vent. The average Fe2+/(Fe2++Mg) of the spinel in these samples shows a weak positive correlation with decreasing MgO in the glass of these samples. The range in Cr2O3 (0.041–0.045 wt.%) of the glass for these six samples is too small to distinguish a consistent change along the lava tube. The spinel in the Pu’u ’O’o samples shows a zoning trend in a Cr–Al–Fe3+ diagram almost directly away from the Cr apex. This compares with a zoning trend in rapidly quenched MORB samples away from Cr coupled with decreasing Fe3+. The trend away from Cr displayed by spinel in rapidly quenched samples is in marked contrast to the trend of increasing Fe3+ shown by spinel in slowly cooled lava.