A new BGS platform to model volcanic ash in the atmosphere

Fabio Dioguardi has been working as a volcanologist at BGS since 2015. His fundamental and applied research focuses on the physical processes and impacts of explosive volcanic eruptions by means of an integrated approach combining field data, experiments and modelling.

Volcanic tephra in the atmosphere and on the ground produced during explosive volcanic eruptions represents a threat to human health and infrastructure and possible serious disruption to various sectors of economy and business. For example, volcanic ash (tephra less than 2 mm in size) in the atmosphere can seriously damage aircrafts, leading to preventive measures like airspace closure in the event of an explosive volcanic eruption, similar to the 2010 eruption of Eyjafjallajökull in Iceland.

In order to mitigate the potential impact of these natural events, simulation tools to predict the advection and sedimentation of tephra in the atmosphere have been developed and are currently used by volcano observatories, Volcanic Ash Advisory Centres (VAAC) and researchers. The simulation of such complex phenomena is not an easy task, but it consists of two fundamental steps.

The first step is characterisation of the source, i.e. modelling of the volcanic plume by quantifying parameters like the source mass flux of emitted volcanic material (also known as mass eruption rate (MER)), the spatial evolution of the plume (e.g. top height; wind-blown trajectory), particle characteristics (grain size; density; shape), and duration of the eruption.

The second step involves simulation of the advection-diffusion-sedimentation of the emitted volcanic tephra cloud.

Both steps are coupled to numerical weather prediction model outputs, since weather conditions like temperature, humidity and wind influence both processes. For example, temperature and humidity influence the plume rise in the atmosphere, while wind can bend the plume and plays a major role in the near to far-field transport and sedimentation of volcanic ash.

The workflow leading to dispersion simulation outputs can be time-consuming and complicated, so BGS’s Innovation Flexible Fund supported the creation of a new modelling platform (BGS automatic ash dispersion modelling, or BGS-AADM) designed to simplify this workflow for both real-time and reanalysis (past eruption) applications.

This platform:

• automatically downloads and processes data of the NOAA-GFS numerical weather prediction model
• characterises the source conditions, particularly the top plume height and mass eruption rate, using both the BGS-USGS eruption source parameters database and REFIR, a semi-automatic tool for the quantification of top plume height and mass eruption rate (and associated uncertainty) based on observational data

We have used REFIR in collaboration with the Istituto Nazionale di Geofisica e Vulcanologia (INGV, Italy) to quantify these source parameters on selected test past eruptions of Mount Etna in the framework of the Horizon 2020 EUROVOLC project activities (Figure 1).

The platform also:

• runs dispersion simulations with two dispersion models, HYSPLIT and FALL3D
• automatically produces contour plots of the dispersion simulations results with a new Python package (Ash Model Plotting) developed at BGS in the framework of the Innovation Flexible Fund project

Graphical outputs include maps of total column mass loading in the atmosphere, i.e. the total mass of tephra integrated from the ground to the top of the volcanic cloud, tephra concentration at various altitudes and total tephra deposit (mass loading) on the ground (Figure 2).

BGS-AADM is designed for both research applications on explosive volcanic eruptions producing ash clouds (e.g., the analysis of past events) and real-time applications during volcanic unrest and eruptions. In particular, it will be used as a further source of information to support BGS response during a future volcanic eruption in Iceland, in particular for better assessing the uncertainty of eruption source parameters and dispersion model outputs.