Speaker
Description
Biomass crops contribute to climate change mitigation through both fossil fuel substitution and soil carbon sequestration. Among these, Miscanthus a fast-growing C4 perennial grass native to East Asia is well suited to UK growth conditions owing to its high productivity, perenniality, and climatic adaptability. Perennial energy crop planting in the UK is projected to expand by 30,000 ha annually to support net-zero targets, creating a growing need for scalable, non-destructive yield monitoring. Unmanned aerial vehicle (UAV) remote sensing offers flexible, high-resolution observations bridging the gap between spatially limited ground surveys and costly airborne/satellite platforms. This study evaluates the Radaz RD350, the world's first tri-band UAV synthetic aperture radar (SAR) system, for estimating Miscanthus biomass yield, with simultaneous acquisition of C-, L-, and P-band data allowing for a direct comparison of wavelength sensitivity to crop stem structure. Yields were estimated for three Miscanthus genotypes: Miscanthus × giganteus (MxG) and two novel hybrids, Miscanthus 10 (GNT10) and Miscanthus 43 (GNT43). Comparative band analysis showed L-band backscatter provided the strongest sensitivity to stem structure and the highest predictive performance. Destructive biomass sampling provided ground-truth measurements, which were used to train a Random Forest regression model relating L-band quad-polarisation backscatter to biomass yields; the model predicted biomass yield with an RMSE of 133.35 g DM m⁻² and an R² of 0.87. Among genotypes, GNT43 produced the highest mean yield (704.59 ± 1.20 g DM m⁻²), followed by GNT10 (669.55 ± 1.21 g DM m⁻²), with MxG recording the lowest mean yield (620.53 ± 1.21 g DM m⁻²)