MODELING LARGE-VOLUME FELSIC ERUPTIONS FROM TRACE-ELEMENT GEOCHEMISTRY

The existence of large bodies of felsic magma in the upper crust is inferred from ignimbrites estimated to represent eruptions of 100's to 1000's of km³, but the origin of the magmas is debated. One prominent model involves fractional crystallization of an intermediate magma, resulting in crystals and a complementary felsic melt. Compaction can segregate the liquid from the crystals once crystals account for ~50% of the system, and extraction of a fraction of the liquid from the system is hypothesized to yield silicic ignimbrites.

Felsic igneous rocks, including large felsic ignimbrites, have highly fractionated whole rock chemistry (e.g. negative Eu anomalies) that is not seen in intermediate igneous rocks envisioned to be their parent magmas. Intermediate igneous rocks are also not chemically complementary to the fractionated felsic magmas (e.g. positive Eu anomalies). Thus, a fractional crystallization model for the creation of felsic magma in the upper crust requires that a portion of the fractionated felsic liquid remain in the system after eruption. Crystallization of the resulting system, consisting of unerupted felsic liquid and crystals, produces intermediate plutonic rocks with neither a fractionated, nor complementary, chemistry. An important variable in this model is the fraction of liquid that must be left behind to mask the chemical signature of melt extraction from the crystal mush.

We used the Eu deficiency (conc. of the Eu anomaly in ppm) and erupted volume of felsic ignimbrites as starting parameters for predicting the Eu excess and volume of the associated plutonic rocks. Results suggest that a system with ~50% crystals that erupts more that 25% of the liquid fraction will leave behind a pluton with a chemical signature that is not consistent with data for plutonic rocks. Because >75% of the felsic liquid must remain in the system, magma fluxes > 0.02 km³/yr are required to build chambers capable of generating even moderately sized ignimbrites on timescales of 200-300 ka (U/Pb zircon age ranges of large ignimbrites). Calculated intrusive:extrusive ratios are >20, far greater than typical estimates. Although calculated magma flux rates are in agreement with rates calculated for super eruptions, at these flux rates it is difficult to fractionally crystallize a system due to the near constant heat input.