Magma mixing and high fountaining during the 1959 Kīlauea Iki eruption, Hawai‘i

Abstract

The 1959 Kīlauea Iki eruption provides a unique opportunity to investigate the process of shallow magma mixing, its impact on the magmatic volatile budget and its role in triggering and driving episodes of Hawaiian fountaining. Melt inclusions hosted by olivine record a continuous decrease in H 2 O concentration through the 17 episodes of the eruption, while CO 2 concentrations correlate with the degree of post-entrapment crystallization of olivine on the inclusion walls. Geochemical data, when combined with the magma budget and with contemporaneous eruption observations, show complex mixing between episodes involving hot, geochemically heterogeneous melts from depth, likely carrying exsolved vapour, and melts which had erupted at the surface, degassed and drained-back into the vent. The drained-back melts acted as a coolant, inducing rapid cooling of the more primitive melts and their olivines at shallow depths and inducing crystallization and vesiculation and triggering renewed fountaining. A consequence of the mixing is that the melts became vapor-undersaturated, so equilibration pressures cannot be inferred from them using saturation models. After the melt inclusions were trapped, continued growth of vapor bubbles, caused by enhanced post-entrapment crystallization, sequestered a large fraction of CO 2 from the melt within the inclusions. This study, while cautioning against accepting melt inclusion CO 2 concentrations "as measured" in mixed magmas, also illustrates that careful analysis and interpretation of post-entrapment modifications can turn this apparent challenge into a way to yield novel useful insights into the geochemical controls on eruption intensity.

Magma decompression rates during explosive eruptions of Kīlauea volcano, Hawaii, recorded by melt embayments

“…To limit our modeling to realistic pre-eruptive H 2 O contents we therefore impose an upper limit on the initial H 2 O concentration of 0.8 wt%, slightly above the maximum observed in melt inclusions from this eruption (and from Kīlauea in general e.g. Wallace and Anderson, 1998;Sides et al, 2014a). We find that the best-fitting ascent model is then obtained with an initial H 2 O concentration of 0.75 ± 0.005 wt% (close to the maximum melt inclusion value), an initial CO 2 of 400 ± 100 ppm, and an initial S concentration of 1410 ± 20 ppm, with the latter similar to the maximum value measured in the embayment.…”

“…This was the initial episode of the archetypal Hawaiian-style eruption in 1959, which involved fountain heights of up to 370 m (Eaton et al, 1987), the highest observed in the modern era. We chose the first episode of the 1959 eruption because subsequent phases were affected by the drain-back of degassed lava into the conduit (Wallace and Anderson, 1998;Sides et al, 2014a). Our samples are taken from layer ρ17 of Stovall et al (2011).…”

“…In the case of embayment IkiE1, which is too short to preserve the initial H 2 O concentrations, we also include constraints from analyses of melt inclusions (Sides et al, 2014a). We assume that the embayments were at equilibrium prior to eruption, implying that H 2 O, CO 2 and S concentrations were initially constant along the length of the embayment.…”

Water, fluorine, and sulfur concentrations in the lunar mantle

“…Data sources: Lunar samples "H74220 inclusions" and "H74220 glass beads" are from Hauri et al (2011). "MORB inclusions" are from Saal et al (2002) at water depth of 3600-3900 m. "Samoa inclusions" are from Workman et al (2006) at water depth of 780-4170 m. "Galapagos inclusions" are from Koleszar et al (2009) at water depth of about 2750 m. Hawaii inclusions and matrix glass are from Kilauea lava fountains (Sides et al, 2014). Other data are from this work.…”

The 2011 eruption of Nabro volcano, Eritrea: perspectives on magmatic processes from melt inclusions

“…Melt inclusion data for silicic lavas from Dabbahu volcano indicated a shallow magma reservoir, with crystallisation driven by cooling rather than decompression (Field et al 2012b). More broadly, basaltic melt inclusion studies of intraplate and riftrelated volcanic systems have focussed primarily on Hawaii [e.g., (Edmonds et al 2013;Norman et al 2002;Sides et al 2014)], Iceland (Maclennan 2008;Neave et al 2014;Winpenny and Maclennan 2011) and Piton de la Fournaise (Bureau et al 1998) in addition to studies of Mid-Ocean Ridge Basalts [MORB; e.g., (Font et al 2007;Nielsen et al 1995)]. These studies reveal that there is considerable variation in mantle volatile content and provide evidence that mantle plumes may be enriched in volatiles compared to MORB (Dixon and Clague 2001;Dixon et al 2002;Wallace 1998).…”