Temporal-volume probabilistic hazard model for a supervolcano: Taupo, New Zealand

Abstract

Supervolcano eruptions are very low probability, but extremely high impact, geohazards. Taupo volcano hosted the youngest known supereruption (VEI8), the c. 1100 km3 Oruanui (Kawakawa) eruption at 25.4 ka BP. Eruptions from Taupo have had regional to global environmental effects, and a supereruption is acknowledged as one of the greatest risks to the sustainability of our modern civilization. However, unlike less extreme volcanic systems, there has been little quantification of the hazard from individual supervolcanoes. Since the Oruanui eruption, we have a record of 28 smaller (between 0.01 and 45 km3) eruptions, of which the largest was the Taupo eruption of CE 232. Hence eruptions from a supervolcano can be of a size that can be mitigated against, and forecasting the magnitude of future events is vital. Most probabilistic supereruption forecasts aggregate many volcanoes in order to use a time-invariant (Poisson process) probability of eruption occurrence. The few studies of individual supervolcanoes have assumed that eruption volumes are independent of the volcano history. Here we apply more sophisticated temporal-volume models to identify and capture possible long-term process controls on when and how large an eruption will be. We find that the volume history of the volcano has a strong modulating effect on the likelihood of eruption, and that the magnitude is positively correlated with the likelihood of an eruption. The volumes of the largest eruption(s) from Taupo may not be consistent with the model based on the output of the single volcano. Widening the spatial limits to the central Taupo Volcanic Zone, and considering caldera-forming eruptions from the last 1.6 My, we find that the same model can satisfactorily explain the eruption record. However, the most significant influence on these large volumes is the elapsed time since the previous caldera-forming event rather than the likelihood of eruption, implying that the magma bodies required to support such eruptions can take significant time to assemble. The probability of an eruption at least as large as the Taupo eruption within the next 500 years is estimated at between 0.5% and 1.3%.