Leverhulme Wildfires PhD Studentship: Machine Learning and Data Assimilation for accurate wildfire predictions

This project aims to incorporate machine learning algorithms in data assimilation processes to improve efficiency and accuracy of wildfire estimations and predictions. The resulting models will assimilate data of wildfire variables (fire incidence, burnt area, fire duration, emissions etc) and wildfire drivers (a wide range, e.g. vegetation-related variables, climate related variables, human/society related variables etc). These data will be both qualitative (societal information) and quantitative (physical information and socio-economic information). The project will focus on the development of models to assimilate global data (e.g. from Earth observation or global socioeconomic databases), localization and domain decomposition techniques to assimilate existing local/in-situ data (e.g. field measurements of how certain types of fires burn and what they emit, country-specific socioeconomic information, data from surveys on how fire is perceived by local communities etc). The assimilated data will be used to: provide accurate estimation of the system (merging, cleaning and optimizing all the available data); understand human perceptions of fire (mainly assimilating and merging the available social media information); estimate optimal sensor positioning through the implementation of gaussian processes technologies. The project will analyse existing palaeoenvironmental databases and fire modelling and will contribute to the advancement of a wider model of human-climate-fire interaction being developed by the Leverhulme Centre for Wildfires, Environment and Society (https://centreforwildfires.org/).

Studentship Overview
The studentship will be supervised by Dr. Rossella Arcucci (https://www.imperial.ac.uk/people/r.arcucci) and co-supervised by Prof. Colin Prentice (https://www.imperial.ac.uk/people/c.prentice) at Imperial College London (https://www.imperial.ac.uk/). Dr Arcucci’s research focuses on numerical and parallel techniques for accurate and efficient data assimilation by exploiting the power of machine learning models. Prof Prentice's research focuses on understanding how plants react to and interact with changes in climate and other aspects of the physical environment.

The student will be affiliated with two of the six Global Institutes of Imperial College London, created to address some of the most important issues facing the world today – the Data Science Institute (http://www.imperial.ac.uk/data-science/ - where the student will be based) and Grantham Institute - Climate Change and the Environment (https://www.imperial.ac.uk/grantham/)t, both at Imperial’s South Kensington campus. The Data Science Institute hosts researchers who use data science techniques for applications in various fields, from healthcare to business, from environment to transport. The Grantham Institute is a leading authority on climate and environmental science, providing a vital global centre of excellence for research and education on climate change. The student will be aligned with the Science and Solutions for a Changing Planet Doctoral Training Partnership (SSCP DTP - https://www.imperial.ac.uk/grantham/education/science-and-solutions-for-a-changing-planet-dtp/news/) based at the Grantham Institute and will take part in the multidisciplinary training programme that the SSCP DTP provides to its PhD researchers.

The student will also join a vibrant interdisciplinary research community in the Leverhulme Centre for Wildfires, Environment and Society, which includes staff and PhD students from Imperial College London, King’s College London, the University of Reading and Royal Holloway, University of London, with a common vision of producing evidence-based understanding of the human-fire nexus that can help inform policy and practice.

How to apply
The applicant will have a good undergraduate degree (min 2.1) in environmental sciences or an allied field. They will either have, or be working towards, a Masters degree or equivalent in a relevant field. The successful candidate will have good quantitative skills and programming experience. They will have experience of writing to a high standard, and a willingness to work in interdisciplinary teams.

Applicants should submit:
i) A CV (max 2 A4 sides), including details of two academic references;
ii) A cover letter outlining their qualifications and interest in the studentship (max 2 A4 sides)
These should be sent by email to [Email Address Removed] and [Email Address Removed] by 15th June 2020 with “Leverhulme – Data Assimilation PhD” as the subject. Interviews will take place, virtually, at the end of June/ early July 2020.
For further information on the project, please contact [Email Address Removed]