Dr Gautham Das, University of Lincoln; Prof Michael Lengden, University of Strathclyde; Prof Astley Hastings, University of Aberdeen; Tom Maidment, Hilton Foods

Scientific Background
Methane emissions from grazing livestock are a major contributor to agricultural greenhouse gases and play a significant role in national and global climatechange discussions (Niloofar et al., 2021). Because these emissions are highly variable across space and time—driven by animal behaviour, pasture conditions, and microclimate—they are difficult to measure and predict with accuracy. Existing methods often rely on extrapolated estimations from subsample measurements and lack the resolution required for targeted mitigation or reliable reporting.
Recent advances in drone sensing (Menon et al, 2026), IoT animalmounted sensors, environmental monitoring systems (Zhang et al., 2023), and machine learning provide a timely opportunity to address this challenge. Datadriven methods have shown strong potential for modelling complex, nonlinear agricultural emission systems. The MethAIr project leverages these developments generate scientifically robust, highresolution methane emission maps for pastoral systems.
Research Methodology
The project takes a multimodal, sensingrich, datadriven approach. The student will collect information from dronebased platforms—including multispectral imagery, thermal imagery, and aerial methane concentration measurements—to capture livestock distribution, pasture biomass, and spatiotemporal emission patterns. Groundtruthing will be conducted through established systems such as GreenFeed units, chambers, or micrometeorological towers. All data streams will be fused using advanced AI techniques, including spatiotemporal deep learning models (e.g. ConvLSTM, transformers) and uncertaintyaware methods (Bayesian neural networks). The student will build prediction models capable of generating methane emission maps that resolve diurnal and seasonal dynamics.
Skill Development
The student will gain comprehensive interdisciplinary understanding and skill development in environmental sensing, drone operations, IoT data handling, and machinelearning model development. They will learn to integrate heterogeneous datasets, build uncertaintyaware AI systems, and conduct fieldbased environmental measurements.
Through work with supervisors from multiple institutions and industry partners, the student will experience crosssector collaboration and applied research translation. The project prepares the student for careers in Applied AI and digital agriculture.