Introduction
Global food supply is increasingly unstable under rapid climate and socio-economic change.
Even small rises in global temperature have been linked to major disruptions, including prolonged droughts, severe flooding, heatwaves, and soil degradation, all of which affect agricultural productivity (IPCC, 2023). Food security is further threatened by heavy reliance on imports, over 50% of staple crops in some countries, such as the UK, leaving them more vulnerable to climate extremes (FAO, 2023). Monitoring and forecasting agricultural health is therefore critical. This includes major crops (e.g., cereals, vegetables, fruits) and livestock systems, which also contribute substantially to greenhouse gas emissions.
The growing availability of satellite Earth observation data – several petabytes generated annually – offers an unprecedented opportunity to observe these systems at scale. However, these multimodal datasets (imagery, hyperspectral data, radar, time series) are complex, making rapid and consistent insight extraction challenging.
This PhD project will apply & develop a physics-informed but low-cost AI framework to monitor and forecast global agricultural health. You will combine satellite remote sensing, field measurements, and process-based ecosystem models to investigate how crop and livestock management influence carbon storage, greenhouse gas fluxes, and food supply. Fieldwork will be based at Rothamsted Research’s North Wyke farm platform, with opportunities to collaborate internationally – including sites in Germany, the ICOS data centre in Sweden, United Unions, and agricultural monitoring sites in China. You will gain hands-on experience with eddy covariance flux towers, GPS surveys, and land surface monitoring.
On the computational side, you will apply advanced AI methods, including computer vision and large language models, to extract information from hyperspectral and high-resolution satellite imagery. You will combine AI with process-based ecosystem modelling and data assimilation to improve carbon and nitrogen cycle simulations, scaling findings from local sites to regional and global levels. Beyond ecosystem processes, you will integrate climate projections to assess how agricultural production affects sustainability and food security under future scenarios.
This interdisciplinary project will give you cutting-edge skills in Earth observation, AI, and ecosystem science, preparing you for careers in research, policy, and industry. You will work with leading AI and environmental science groups at the University of Southampton and global academic & industry partners.