Investigating Melt Chemistries Trapped Within Crystals to Interpret Eruptive Processes in Caldera-Forming Eruptions

Abstract:

Determination of the pre-eruptive storage configuration of silicic magmas, and the initiation and progress of consequent large-scale eruptions can be evaluated through combined field and microanalytical geochemical studies. Here we present select results from studies that have coupled major, trace and volatile information with diffusion modeling from quartz-hosted melt inclusions and open melt pockets (reentrants) taken from stratigraphically controlled levels in the early eruptive sequences of two large, caldera-forming eruptions. The first involves fall deposits of the ~2.08 Ma, 2500 km³ Huckleberry Ridge eruption, Yellowstone. These deposits are multiply bedded and contain reworked intervals, indicating the opening phases of the eruption cumulatively lasted weeks to months. Although major element glass compositions are relatively uniform, several trace elements (e.g. Ba) serve to define statistically significant compositional clusters. These clusters are interpreted to represent multiple, discrete melt-dominant domains that were systematically tapped by multiple vents during eruption onset. Timescales of magma ascent feeding initial fall deposits are interpreted from scatter in measured H₂O concentrations from enclosed melt inclusions, implying ascent times as long as ~14 days from storage, reflecting highly variable and slow decompression conditions. The second example is from the ~25.4 ka, 530 km³ Oruanui eruption, New Zealand, where timescales of final ascent (i.e. hours) are determined by modeling H₂O and CO₂ gradients preserved in reentrants. Reentrants from the initial two phases of the Oruanui deposits indicate relatively slow magma ascent conditions (0.12-0.28 m/s). Intriguingly, these two eruption phases are associated with evidence for eruption initiation and modulation through rifting. A notable increase in ascent rates, however, is recorded in the phase three materials (2-5 m/s), which correlates with a field-inferred increase in eruption intensity.