Advancing Machine Learning for Extreme Wind Speed Prediction, Engineering – PhD

The University of Exeter’s is inviting applications for a fully funded PhD studentship to commence on on September 2025 (or sooner if appropriate). For eligible students the studentship will cover Home or International tuition fees plus an annual tax-free stipend of at least £19,237 for 3.5 years full-time, or pro rata for part-time study. The student would be based in Engineering (but work closely with Computer Science) in the Faculty of Environment, Science and Economy at the Streatham Campus in Exeter.

Climate change is amplifying the frequency and intensity of extreme weather events, creating significant risks for infrastructure, economies, and communities. It highlights the urgent need for innovative methods to accurately forecast extreme weather events and reduce their impacts. Nevertheless, the unpredictable and infrequent nature of extreme weather events poses distinct challenges, requiring the development of advanced, data-driven solutions that can handle their complexity and rarity. 

This PhD project aims to pioneer ML methodologies tailored to short-term spatiotemporal extreme wind speed prediction, addressing limitations in conventional models. 

The candidate will: 

• Analyse Extreme Data Distributions: Apply advanced statistical techniques, such as extreme value theory, to understand and model rare high-impact wind events. 

• Innovate Model Architectures: Develop novel model structures such as spatiotemporal graph neural networks, physics-informed neural networks, and Bayesian approaches to address evolving weather patterns in both spatial and temporal dimensions. 

• Enhance ML Predictability: Maintain model predictability under data distribution shifts caused by extreme wind events, exploring strategies such as data augmentation and novel loss functions. 

• Collaborate and Validate: Work with experts from the Met Office to validate models using real-time meteorological data, ensuring actionable and reliable forecasting tools. As part of this transformative project, the PhD candidate will: 

• Contribute to addressing one of the most urgent challenges posed by climate change. 

• Work with the Met Office who will provide expertise, data and links to important stakeholders to support this project. 

• Work under a multidisciplinary supervision team. 

• Develop expertise in advanced ML techniques with real-world applications, positioning themselves as a leader in climate-resilient technologies.