Cattle and degraded peatlands together produce 11% of UK greenhouse gas emissions, and a new AI system aims to help farmers cut that figure. This matters because UK agricultural emissions, after decades of decline, began rising again in 2020 and 2021. The government has committed to net zero by 2050, but farmers and land managers currently lack a practical tool to test how specific changes—such as altering grazing patterns or restoring peat bogs—would affect their actual emissions. The project builds a "self-learning digital twin": a computer model that ingests real-time satellite images and field measurements, then updates itself as new data arrives. Farmers can ask it questions about their land and get answers grounded in current conditions. If the system works, it could save farmers time and money by identifying which interventions actually reduce emissions on their specific land, rather than relying on generic advice. The project also creates a "Community of Practice in AI for Net Zero", bringing computer scientists together with social and behavioural researchers to address who might benefit or be disadvantaged by such technology. This is applied research with a clear practical endpoint: a decision-support tool for land managers.
View original technical description
Greenhouse gas emissions from agriculture and land use in the UK contribute to global climate change. The UK is committed to achieving net zero greenhouse gas emissions by 2050. Since 1990, greenhouse gas emissions from agriculture and land use have fallen, but in 2020 and 2021 they started rising again. 11% of UK GHG emissions stem from cattle and sheep grazing (7%) and degraded peatlands (4%). This research project is developing an Artificial Intelligence algorithm called a 'Self-Learning Digital Twin' for sustainable land management. A Digital Twin applies computational modelling, environmental measurements and an Artificial Intelligence algorithm to provide new environmental insights into the functioning of a system. Farmers and land managers can ask questions that the Digital Twin can answer. In a nutshell, it is a digital model of the physical environment and is updated from real-time data, so that it mirrors the environment at all times. Digital Twins can support farmers and environmental managers to achieve better outcomes for their greenhouse gas emission reductions, ultimately saving time and resources. The self-learning digital twin learns from real-time satellite images, greenhouse gas measurements from field instruments and other data. Its underlying model improves over time as new data are becoming available. The project will promote sustainable cattle and sheep farming practices and peatland restoration. We will prepare the ground for an ethical and socially responsible application of artificial intelligence for achieving net zero greenhouse gas emissions. An important part of our work is to build a 'Community of Practice in AI for Net Zero' that brings together computer scientists with environmental, behavioural and social science researchers to develop a common approach. We will incorporate the social and ethical dimensions of digital twins, including who they may benefit or disadvantage.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know