Researchers at Loughborough University have developed artificial intelligence (AI) tools that offer insights into how greenhouse gas (GHG) emissions associated with UK livestock farming and land use can be reduced.
The tools are hosted on an online digital platform and created as part of research funded by the UK Research and Innovation (UKRI) and the Engineering and Physical Sciences Research Council (EPSRC)
The aim is to provide farmers, farming organisations, and government bodies with valuable data on how changes in livestock practices and land use can help the UK achieve its 2050 ‘net zero’ goal.
Developed by a team led by Prof. Baihua Li, and Prof. Qinggang Meng, key features of the platform include machine learning models designed to estimate methane emissions from livestock farming, predict milk productivity, and ammonia emissions from dairy farms.
Achieving net zero by 2050 requires balancing GHG emissions with their removal and storage in ‘carbon sinks’ natural systems like forests, oceans, plants, and soil that absorb more carbon than they release.
The researchers also said that reducing farming’s environmental impact is challenging, as emissions, carbon storage, and farm productivity are shaped by multiple interacting factors, such as animal breed, feed, pasture, and climate.
These vary across farms, making a one-size-fits-all approach ineffective.
AI for livestock farms
The AI tools developed for livestock farms allow farmers to input details about their specific animals and practices to estimate their current annual GHG emissions.
Farmers can easily explore potential changes to their practices – simply by selecting options from drop-down menus or entering variable values.
One tool is designed specifically for dairy farmers, helping them estimate how their current practices affect individual cow milk yield and ammonia levels in waste.
Monitoring ammonia is crucial, as it interacts with soil microbes to produce nitrous oxide and may also indicate dietary imbalances.
These adjustments provide immediate insights into their potential impact on both emissions and farm productivity.
Another tool, developed for beef farmers, predicts methane emissions for individual cows based on farm-specific data.
It also helps farmers understand emissions in context by offering relatable comparisons – such as the number of trees needed to offset a cow’s annual emissions, the equivalent emissions from flights between London and New York, or the months of energy use in an average UK household.
The team has also developed a livestock emissions calculator based on Intergovernmental Panel on Climate Change (IPCC) guidelines, the global standard for climate reporting.
Suitable for farmers worldwide, it presents a user friendly format that simplifies complex government formulas helping farmers compare their emissions to official baselines.
Beyond farm-level tools, the research team has harnessed artificial intelligence to develop a user-friendly, web-based platform referred to as a ‘digital twin’ to provide detailed insights into how different types of land use affect methane emissions across the UK.
The digital twin features heatmaps of grazing livestock distribution, land cover types (such as agriculture, urban areas, and woodland), and methane emission concentrations across the UK.
It integrates real-time satellite methane observations from Sentinel-5P TROPOMI, a satellite from the European Space Agency’s Copernicus programme, AI models, datasets, and various intuitive visualisation tools.
Users will be able to adjust things like location, lan coverage percentages, seasons and years to track historical changes and model future emission scenarios based on climate and land use projections.
Researchers said that they hoped the tool will be used by policymakers, government bodies, and farming organisations to deepen understanding of how environmental factors influence emissions and enable data-backed decisions to be made to reduce emissions.
Before the digital platform hosting the tools can be deployed to users, the researchers said that they need to further refine and test their AI models, which requires additional data.
The team is calling for the following collaborators:
- Individual farmers who can share data on their specific practices, e.g., livestock details and feeding strategies;
- Farming cooperatives and organisations that have data from multiple farms;
- Governmental and regulatory bodies that manage compliance data, and can provide historical and geographical data;
- Agricultural research institutions;
- Retailers and supply chain stakeholders with data relevant to the agricultural industry.
Prof. Meng said: “We hope key stakeholders recognise the value of this platform, support efforts to achieve net zero emissions, and contribute essential data to help bring the technology to life, ultimately transforming our practises and ensuring a sustainable future for all.”
