Grassland Modelling for Improved Utilisation

Grass is the most widely grown crop in the United Kingdom and globally, forming the foundation of livestock farming systems and rural economies. Yet managing grass growth remains one of the most time-consuming and uncertain aspects of farm management. Farmers traditionally rely on manual measurements, walking fields and recording grass height using plate meters. This process is labour-intensive, inconsistent, and often deprioritised during busy periods.

The Grassland Modelling for Improved Utilisation project, funded by Innovate UK, set out to demonstrate how Earth observation (EO) data could transform this process. By combining satellite radar imagery (C-SAR), atmospheric modelling, and machine learning algorithms, the project developed a system capable of measuring and predicting grass growth automatically across farms. The technology provides sub-weekly updates on pasture conditions and predicts grass growth up to two weeks in advance, enabling farmers to make informed grazing and feed management decisions.

A key innovation of the project was the mixed use of radar satellite data, which can operate reliably in cloudy conditions, and atmospheric based prediction models. This allows for proactive monitoring of grassland conditions regardless of the weather, providing farmers with consistent, timely information. EO data and atmospheric observations are combined with mechanistic models to produce an ensemble field-scale prediction supporting day-to-day farm management.

Trials conducted on farms in England demonstrated that the system could significantly improve decision-making and operational efficiency. Farmers reported time savings, reduced reliance on manual measurements, and improved confidence in managing grazing rotations. The technology enabled more precise use of fertilisers and feed resources, reducing costs while improving environmental sustainability.

The outputs of this project are now being commercialised into a service for farmers. Validation activities have recently begun in Ireland and across the EU, enabling access for a wider farming community and increasing the project's real-world impact.

Grass management is central to livestock farming, underpinning animal productivity while also delivering environmental benefits beyond agriculture through improved soil health, enhanced biodiversity, and carbon sequestration in grassland ecosystems.

Farmers who actively manage grass typically rely on traditional methods that involve manually walking fields and measuring grass height with a plate meter. This approach is unreliable, time-consuming and difficult to maintain consistently during busy periods, highlighting the need for more efficient solutions. As a result, farmers may miss opportunities to optimise grazing schedules or manage feed supplies effectively.

At the same time, farmers face increasing pressure to improve productivity while reducing environmental impact. Rising input costs, climate variability, and sustainability requirements have made efficient resource management more important than ever. Better information about grass growth can help farmers reduce waste by being proactive rather than reactive, improve feed efficiency, and make more sustainable decisions.

The “Grassland Modelling for Improved Utilisation” project was designed to address these challenges by providing frequent, automated information about grass growth. The project aimed to demonstrate how EO data could support practical farm decisions and improve the efficiency of grass-based livestock systems.

Farmers participating in the project were particularly interested in tools that could save time and provide reliable data without requiring additional labour. By delivering regular updates on pasture conditions, the system enabled farmers to make decisions based on current information rather than assumptions.

The project developed a grass growth prediction system that integrates multiple Earth observation data sources with predictive modelling techniques. The system continuously monitors grassland conditions and generates predictions that support farm management decisions. Satellite radar (C-Band SAR) imagery is used to estimate ground cover and vegetation conditions across grazing areas. Unlike optical satellite imagery, radar sensors can operate in cloudy weather, ensuring reliable data collection throughout the year. This capability is particularly important in regions where weather conditions frequently limit the effectiveness of optical remote sensing methods.

Atmospheric forecasting data is combined with satellite observations to provide medium-range (multiple weeks) predictions, which are beyond the forecast skill of traditional weather-based models. Weather variables, such as temperature, rainfall, and atmospheric conditions, influence plant growth and are incorporated into predictive models for short-range (roughly three days) predictions. By analysing these factors together, the system can estimate future grass growth rates and identify periods of rapid or limited growth.

The project also developed machine learning algorithms that analyse historical data and real-time observations to improve prediction accuracy by creating field-specific models. These algorithms continuously refine predictions as new data becomes available, allowing the system to adapt to changing environmental conditions. The resulting insights are delivered to farmers through “traffic light” based dashboards and reports that present information in a clear, accessible format. This allows farmers to quickly understand pasture conditions and make decisions without needing technical expertise in remote sensing or data analysis.

The grass growth prediction system relies on a combination of satellite and environmental data from the EU's Copernicus program and farm-level data sources.

Key observations include:

These data sources are processed using predictive machine learning algorithms that identify patterns and relationships between environmental conditions and grass growth. By combining multiple observations to create an ensemble-based model, the system produces more accurate predictions than traditional single-source, monolithic models. The integration of EO and ground-based data ensures that predictions remain reliable across different farm types and environmental conditions. This approach allows the system to be scaled to new regions while maintaining accuracy.

The “Grassland Modelling for Improved Utilisation” project has demonstrated clear benefits for farmers, agricultural businesses, and the wider agricultural sector.

One of the most immediate benefits has been time savings. Farmers participating in the project reported that automated grass growth predictions reduced the need for manual field measurements, allowing them to focus on other essential tasks. This improved efficiency has been particularly valuable during busy periods such as calving and harvesting seasons.

“One of our trial farmers, Will from Derby, currently spends 8 hours a week measuring grass, the equivalent of 1 whole working day. Agribot has reduced that to a few minutes”

The system has also strengthened resource management by providing accurate, timely information on grass availability. This enabled farmers to optimise grazing rotations, reduce reliance on purchased feed, and ultimately lower costs while improving livestock performance and overall farm financial sustainability. Environmental sustainability is a further key benefit. With more precise insights into grass growth, farmers can apply fertilisers and irrigation more efficiently, reducing waste and minimising environmental impact. Improved pasture management also supports soil health and biodiversity by preventing overgrazing and encouraging more balanced grazing patterns.

The project has also demonstrated broader benefits beyond individual farms. Accurate grass growth predictions can support supply chain planning, insurance risk assessment, and land management decisions. As the technology is adopted more widely, it has the potential to improve resilience across the agricultural sector.

Feedback from participating farmers has been overwhelmingly positive. Many reported increased confidence in managing their pastures and greater trust in data-driven decision-making. These experiences highlight the importance of practical, user-focused technologies in supporting modern agriculture.