Understanding particle size distributions to improve ash dispersal modelling

  • Sara J Osman

Student thesis: Master's ThesisMaster of Science (MSc)

Abstract

Volcanic eruptions can generate large volumes of ash and cause far-reaching air traffic disruption. To mitigate against aircraft encounters with ash clouds Volcanic Ash Advisory Centres (VAACs) forecast the expected location of the ash in the atmosphere. The size distribution of particles making up an ash cloud can vary significantly with eruption, but it is important to initialise dispersion models with an appropriate particle size distribution (PSD) because sedimentation rates are controlled by particle size. This project aims to better understand the range of PSDs generated from volcanic ash eruptions and consider the use of PSDs in ash dispersion modelling.
To understand the process of grain size analysis, I collected and analysed samples from the Minoan eruption on Santorini, compiled a total grain size distribution (TGSD) for the 1919 eruption of Kelut and calculated median grain size for a sample from the Askja 1875 eruption.
I compiled published grain size data and found that for large phreatomagmatic eruptions, grain size remains relatively constant with distance. This suggests that TGSDs for these eruptions could be compiled from fewer samples than are required for typical magmatic eruptions, which in turn could provide a larger dataset for dispersion modelling and studies on controls of eruption intensity.
To test the sensitivity of modelled ash concentrations to the input size distribution, I ran NAME with a range of PSDs from different types of eruption. I found clear differences between modelled mass loadings and the extent of the plume for mafic and silicic eruptions. The default PSD used by the London VAAC is most similar to the finest (silicic) test eruptions and I recommend that a second default PSD should be considered for operational forecasting, suitable for coarse-grained, mafic eruptions. This could be compiled from ground and airborne samples or based on a suitable statistical distribution.
Date of Award23 Jan 2020
Original languageEnglish
Awarding Institution
  • The University of Bristol
SupervisorAlison C Rust (Supervisor), Frances Beckett (Supervisor) & Katharine V Cashman (Supervisor)

Keywords

  • Volcanic ash
  • Particle size distribution
  • Dispersion modelling

Cite this

'