Skip to content

Probabilistic Modelling of Volcanic Ash: Eruption Source Parameter Uncertainty

Charlie Bates1, Frances Beckett1, Nina Kristiansen1, Jeremy Phillips2, Shannon Williams2 and Mark Woodhouse2

Abstract

Volcanic ash clouds represent a significant hazard to aviation. To mitigate the risk, a global network of Volcanic Ash Advisory Centres (VAACs) is responsible for providing guidance on the presence of ash in the atmosphere. Forecasts are generated using atmospheric transport and dispersion models, underpinned by observations and driven by meteorological data. To initialise simulations the release of ash into the atmosphere must be defined, these are referred to as Eruption Source Parameters (ESPs). Notably, the height over which the ash resides and Mass Eruption Rate (MER), are leading order parameters, yet remain challenging to constrain during real-time event response. This study applies two new tools, MERPH (Woodhouse, 2024) and PVA (Williams et al., Submitted) to constrain uncertainty in ESPs and meteorological data for simulations of ash clouds using the Numerical Atmospheric-dispersion Modelling Environment (NAME). MERPH uses large historical datasets and Bayesian methods to explore the relationship between plume height and MER. PVA provides a computationally efficient statistical framework for propagating uncertainties on ESPs in atmospheric dispersion model simulations using ensemble meteorology, generating probabilistic forecasts. We analyze the eruptions of Raikoke (June 2019) and Shiveluch (April 2023), aiming to improve uncertainty representation in ash dispersion simulations, focusing on plume height and MER. Using MERPH and PVA, we seek to quantify these uncertainties to enhance probabilistic forecast reliability. We compare probabilistic and deterministic outputs, considering the influence of the uncertainty on ESPs and meteorological data. Our results highlight the importance of representing ESP uncertainty for operational forecasts for the aviation industry.