Premium
The Impact of Eruption Source Parameter Uncertainties on Ash Dispersion Forecasts During Explosive Volcanic Eruptions
Author(s) -
Dioguardi Fabio,
Beckett Frances,
Dürig Tobias,
Stevenson John A.
Publication year - 2020
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
eISSN - 2169-8996
pISSN - 2169-897X
DOI - 10.1029/2020jd032717
Subject(s) - plume , volcano , atmospheric dispersion modeling , wind speed , environmental science , atmosphere (unit) , volcanic ash , meteorology , explosive eruption , dispersion (optics) , atmospheric sciences , panache , geology , air pollution , seismology , geography , magma , chemistry , physics , organic chemistry , optics
Volcanic ash in the atmosphere is a hazard to aviation. To predict which areas of airspace are most likely to be affected by the presence of ash, Volcanic Ash Advisory Centers (VAACs) use observations and atmospheric dispersion models. These models are initialized with, among other parameters, a mass eruption rate ( MER ), which quantifies the emission rate into the atmosphere at the source. This influences the predicted spatial–temporal evolution and concentration of the ash cloud. Different models are available to estimate MER from the volcanic plume height and some models also include the weather conditions (e.g., wind speed). The REFIR software tool uses time‐series of plume height estimated from observations and weather data to provide estimates of MER through time. Here we present an updated version of REFIR that can now be used also to calculate MER for past eruptions and produce output parameters in a format suitable for use with the NAME dispersion model (UK Met Office—London VAAC). We also investigate how uncertainty in input parameters is propagated through to dispersion model output. Our results show that a +/−1 km uncertainty on a 6 km high plume can result in the affected area ranging by a factor of three between the minimum and maximum estimates. Additionally, we show that using wind‐affected plume models results in affected areas that are five times larger than using no‐wind‐affected models. This demonstrates the sensitivity of MER to the type of plume model chosen (no‐wind‐ vs. wind‐affected).