Research Fellow in Physics-Informed Machine Learning for African Storms

View All Vacancies

Are you a Mathematician or Computer Scientist with expertise in Machine-Learning? Do you want to apply your skills to develop tools which will save lives and protect infrastructure in Africa? Do you want to further your career in one of the UK’s leading research-intensive universities?

We are seeking a specialist in machine learning to join our experienced team, to contribute to the improvement in storm predictions – nowcasting – for Africa. The NAIAR project – Nowcasting with Artificial Intelligence for African Rainfall – is conducting research to deliver new methods for storm prediction. Working with UK, African and international partner organisations, the results of your work will lead to nowcasting information that will benefit many people in Africa within minutes.

As postdoctoral research scientist in Physics-Informed Machine Learning for African Storms, you will conduct original research leading to an improvement in the fundamental science underpinning operational storm nowcasting for Africa. In particular, we aim to develop physics-informed machine-learning methods, in which reduced equation-sets or fluid-dynamical constraints are imposed into the system. Methods such as “transfer learning” will be studied, in order to efficiently train machine-learning systems on reduced dynamical constraints. A particular focus of the work will be the challenging problem of prediction of storm initiation: we will approach this problem by reference to simpler dynamical systems and the study of state-changes in these systems.

The project is led by the UK’s National Centre for Atmospheric Science (NCAS) at the University of Leeds and will be a collaborative enterprise between the Schools of Mathematics, Computing and Earth & Environment in Leeds, and the UK Centre for Ecology and Hydrology (UKCEH) in Wallingford. Our collaboration with UKCEH is long-standing and has led to many successful outputs over the years.

Tropical storms are inherently chaotic, on timescales of a few hours, and therefore daily forecasts of rainfall always have uncertainty. For this reason, “nowcasting” – the communication of real-time observations and short-range (0-6 hour) predictions, is vitally needed across the African continent. However, very few people in Africa currently benefit from nowcasting information.

Through the application of innovative meteorological research to harness satellite data, we are already delivering nowcasts of high-impact storms to Tropical Africa. The new methods you create will lead to improved nowcasting information, released in real time on existing online platforms (https://science.ncas.ac.uk/swift/ and https://eip.ceh.ac.uk/hydrology/sub-saharan-africa/nowcasting/ ) and the FASTA smartphone app being supported by the National Centre for Atmospheric Science (NCAS) at Leeds, and the UK Centre for Ecology and Hydrology (UKCEH). In this way the results of your research will be used to deliver real-time information to users in Africa.  

In delivering the real-time nowcasting services, our aim is to provide the capabilities needed for African national weather services and private sector companies to enhance their delivery of weather forecasts and ensure access to high-quality weather information for African populations. We have formal partnerships with a number of African weather services and other organisations. These partners will collaborate in this project, particularly in evaluation and implementation of methods, but also in our support for capacity-building in Africa. Ultimately, better preparedness and informed decision-making using nowcasting information will improve climate resilience and save lives and livelihoods.

We are open to discussing flexible working arrangements.

To explore the post further or for any queries you may have, please contact:

Professor Doug Parker , Science Co-ordinator for NCAS at Leeds

Email: [email protected]

Location:  Leeds - Main Campus
Faculty/Service:  Faculty of Engineering & Physical Sciences
School/Institute:  School of Mathematics
Category:  Research
Grade:  Grade 7
Salary:  £37,099 to £44,263 p.a.
Due to funding restrictions, an appointment will not be made higher than £39,347 p.a.
Working Time:  37.5 hours per week
Post Type:  Full Time
Contract Type:  Fixed Term (Until 31 August 2025 - To complete specific time limited work.)
Release Date:  Tuesday 20 February 2024
Closing Date:  Tuesday 19 March 2024
Reference:  EPSMA1103
Downloads:  Candidate Brief