Continuous estimates of heat emission at Mt. Ruapehu using the Unscented Kalman Smoother

Abstract

Volcanic lakes often capture a significant amount of volcanic heat emission and thus provide a unique opportunity to monitor changes inside the volcano. We present a Bayesian inversion method to automatically infer changes in volcanic heat emission over time at the base of a volcanic lake from lake monitoring data using a non-linear Kalman Smoother. Our method accounts for the, sometimes large, uncertainties in observations and the underlying physics-based model to generate probabilistic estimates of heat emission. We verify our results using a synthetic test case and then estimate the daily heat input rate into Mt. Ruapehu’s Crater Lake, New Zealand, between 2016 and 2022. Time-frequency analysis of the heat input rate shows dominant periods of heating cycles ranged between 100 - 250 days. The period between 2017 and 2020 was dominated by shorter cycles and greater-than-average heat input rate which points to changes in the magmatic heat supply and the hydrothermal system during this time.