Cryptic young zircon and young plagioclase in the Kaharoa Rhyolite, Tarawera, New Zealand: Implications for crystal recycling in magmatic systems

We measured in-situ 238U-230Th zircon and bulk plagioclase 238U-230Th-226Ra disequilibria in rhyolite lava and tephra from the ~1315 AD Kaharoa eruption of Tarawera Volcano, New Zealand in order to constrain its history of chemical evolution. These data suggest that zircon records a protracted history (10s of kyr) whereas plagioclase is dominantly young (few kyr), but both phases crystallize up to the eruption. The Kaharoa eruptive period at Tarawera consists of ~2.5 km3 of crystal-rich rhyolite (74-75 wt% silica) lava and ~5 km3 of coeval tephra deposits, making it the largest silicic eruption in New Zealand in the last 1,000 years. 238U-230Th disequilibria measurements of zircon determined via SHRIMP-RG analyses produce an array of ages, with three main populations: (1) within error of eruption age; (2) 15-80 k.y.; (3) 100-175 k.y. Very few analyzed zircon fall within error of eruption age and little difference is seen in the age distribution of zircon between the lava and tephra. In contrast, 238U-230Th and 230Th-226Ra plagioclase ages appear to be within error of eruption age however this age is complicated zircon inclusions in the plagioclase. This contamination by zircon is seen in the 238U-230Th disequilibria and trace element data, where the addition of zircon pushes the bulk plagioclase separate towards more U-enriched values and high Zr values. However, the (230Th)/(232Th) ratios for the separates are the same as the whole rock values, indicating that any zircon in the bulk separate must be young (eruption age). This finding is also borne out in 230Th-226Ra disequilibria, where zero-age zircon contamination is reflected in increased (230Th)/[Ba] with no change in (226Ra)/[Ba]. In both cases, as little as 1 ppm of zero-age zircon contamination is needed to create these patterns. This signal of young plagioclase and zircon growth is in contrast to the protracted history seen in the SHRIMP-RG zircon data. This suggests that young zircon growth in the Kaharoa Rhyolite is "masked" in SHRIMP-RG analyses because the young zircon is present either inclusions in plagioclase excluded from zircon separate or as thin rims on older zircons that can't be effectively analyzed. The Kaharoa data implies that zircon records a longer history than plagioclase, and older zircon ages represent the mush/plutonic roots of the volcano. The existence of old zircon with little evidence for old plagioclase suggest that remobilization of the plutonic roots by new heating events destroys most of the old plagioclase crystals. Plagioclase and zircon both crystallize during the short period between remobilization and eruption. However, young zircon growth is cryptic, suggesting that in-situ zircon ages may in general be biased toward older ages. Overall, plagioclase and zircon offer potential to unravel a large time period of the history of a magmatic system, especially silicic magma with long histories.