Satellite-based monitoring of forage quality in grasslands of the United Kingdom using sentinel-2 data and random forest regression

Summary

This study developed and validated satellite-based models using Sentinel-2 optical remote sensing and random forest regression to predict forage quality across United Kingdom grasslands. Calibration utilised over 9,500 georeferenced observations from permanent and improved pastures at the North Wyke Farm Platform (2020–2022), with forage quality measured via near-infrared sensors on agricultural machinery. The calibrated models demonstrated strong performance for predicting crude protein, water-soluble carbohydrates, neutral detergent fibre, and acid detergent fibre, with spatial variations between paddocks more pronounced than seasonal changes, offering a scalable alternative to labour-intensive destructive sampling methods.

Regional applicability

This study was conducted directly within United Kingdom grassland systems (southwest England), making the findings directly applicable to UK pasture management and forage monitoring. The authors note that the calibrated models are suitable for forage systems with species composition and quality ranges similar to the North Wyke dataset but should not be directly transferred to other forage systems without recalibration.

Key measures

R² values (0.77–0.86), RMSE, crude protein (CP), water-soluble carbohydrates (WSC), neutral detergent fibre (NDF), acid detergent fibre (ADF), red-edge and NIR wavelength predictors

Outcomes reported

The study predicted four critical forage quality attributes (crude protein, water-soluble carbohydrates, neutral detergent fibre, and acid detergent fibre) using Sentinel-2 satellite data and random forest regression models. Model performance achieved R² values between 0.77 and 0.86 across all forage quality measures, demonstrating high predictive accuracy for large-scale forage monitoring.

Topic tags