UK-wide Biodiversity Model Emulation
I am working on P3, a new Defra project focused on emulating a biodiversity model for the whole UK.
This is the first national-scale biodiversity model emulation task.
Research
My doctoral research in statistics at the University of Exeter focuses on Gaussian processes, Vecchia approximations, statistical computing, and high-performance computing. My applied work involves statistical emulators and machine-learning models for agricultural and environmental research.
Research interests
- Gaussian processes and Vecchia approximations
- Statistical computing and high-performance computing
- Model emulation and Bayesian optimisation
- Deep learning and time-series analysis
Selected projects
Agricultural Uptake and Biodiversity Modelling
Developed a modelling pipeline linking agricultural uptake and biodiversity models, with reproducible datasets and outputs for downstream analysis.
Fitted a Gaussian process emulator of the linked workflow to accelerate exploration of model inputs, scenarios, and behaviour. Automated health checks flag failed runs, data-quality issues, and unexpected outputs.
Farm Management Model Emulation
Built a Transformer–MLP encoder–decoder as a fast surrogate for a Bayesian optimisation model for farm management.
Developed scalable preprocessing, training, hyperparameter tuning, and evaluation workflows for time-series and spatially structured agricultural data. The work involved collaboration with farm scientists, economists, and biologists.
ADD-TREES: AI-elevated Decision-support via Digital Twins for Restoring and Enhancing Ecosystem Services
PhD Student (Maths & Statistics)
The project develops AI-powered digital twins and decision-support tools to inform land-use decisions, with a focus on woodland creation, ecosystem services, and the UK’s net-zero goals.
Project Website Video: ADD-TREES Overview Video: Vecchia Approximation Talk
Conferences and events
ISBA 2026 · Nagoya, Japan
Presented “Scalable deep kernel learning with Vecchia approximation”, my research on scalable Gaussian process models.
AI UK 2024 · London
Attended AI UK in London in 2024, representing the ADDTREES project.