Observing and modelling volcanic cloud evolution through satellite data and data assimilation

Project Details

This is a proposed CASE Project (to be confirmed in May) in association with the Met Office

Key Questions

How do volcanic clouds evolve?


Volcanic clouds are composed of a mixture of solid volcanic ash particles, volcanic gases (e.g. sulphur dioxide; SO2) and different aerosols that form from the gaseous components (e.g. sulphate aerosols). The composition and structure of a volcanic cloud evolves over time as the cloud disperses in the atmosphere. This affects which methods are best used to observe and model the clouds. Close to the volcano, volcanic clouds can extend many kilometres vertically and contain very large volcanic ash particles affected by eruption column processes, while volcanic clouds transported thousands of kilometres from the source are typically present as very thin layers (~a kilometre thick) and may contain a mixture of gas, very-fine-grained ash and aerosols

Aims of the Project

To develop satellite techniques that will feed into the development of the operational volcanic ash forecasts at the London VAAC and help in delivering more accurate quantitative volcanic ash data.

Project Description

The first part of the PhD project will look at the evolution of volcanic clouds by simultaneously analysing satellite observations of volcanic ash, volcanic SO2 and sulphate. The interplay between the species and the changes in the cloud structure and composition (e.g. the particle size) will be studied using new generation satellite instruments. The first is the AHI instrument on the Himawari geostationary satellite situated at 140.7° East. The instrument provides high temporal and spatial resolution multi-spectral observation of the large number of active volcanoes situated around the Pacific rim. The student could investigate the impact of the choice of ash microphysical model (differing ash compositions and a coating of ice or sulphuric acid) on the plume properties retrieved using the Optimal Retrieval of Aerosol and Cloud algorithm (github.com/ORAC-CC/orac/wiki). A complementary study could be conducted on retrievals using very high spectral resolution measurements from the next generation of IASI instruments to be launched in 2023. Here the objective would be to optimise the quantification of the ash, sulphate and SO2 components in the volcanic plume. The second part of the PhD project will improve the linkages between satellite observations and dispersion modelling by exploring data assimilation (DA) methods. Here we refer to DA as a method where the model simulated concentration fields are optimised according to observations (and not source term estimation via inversion methods). The application of DA techniques to volcanic clouds is still a relatively new application with ongoing research, and it has the potential to greatly improve volcanic ash forecasts. An in-depth understanding of the satellite observations and related uncertainties is crucial in designing a robust DA system. An ensemble DA system based on an Ensemble Kalman Filter (EnKF) will be integrated and tested with the Met Office Numerical Atmospheric-Dispersion Modelling Environment (NAME) dispersion model, building on open source EnKF software successfully applied to other models (Pardini et al., 2020; Mingari et al., 2021). The general idea of Ensemble-based filters is that the state of a system is described in a statistical way by an ensemble of numerical simulations and not by a single simulation. By applying the filter, the forecast state obtained by the ensemble simulations is corrected with observations and a new analysed state is produced, which is then used as initial conditions for the subsequent forecast. A current challenge with DA of volcanic clouds is that the volcanic cloud height and thickness are often poorly constrained. Two key components of the project will be to seek the best way to assimilate satellite derived vertical information such as cloud height and thickness information, and to explore how uncertainties in the observations data should best be incorporated.

Methods to be used

See project description

Specialised skills required

Undergraduate degree in physics, maths or chemistry.

Please contact Roy Grainger on r.grainger@physics.ox.ac.uk if you are interested in this project