Estimating spring wheat nitrogen use efficiency via proximal and UAV sensing in Northwest Latvia
¹Institute of Agricultural Resources and Economics, Department of Plant breeding and Agroecology, “Dižzemes”, Dižstende, LV-3258 Talsi distr., Latvia
²Faculty of Agriculture and Food Technology, Latvia University of Life Sciences and Technologies, 2 Liela Str., LV-3001 Jelgava, Latvia
*Correspondence: zaiga.jansone@arei.lv
Abstract:
Phenotyping nitrogen use efficiency (NUE) is labour-intensive and time-consuming, often requiring destructive biomass sampling. Cost-effective sensing tools provide a promising alternative for rapid assessment of numerous wheat genotypes. In this study, sixteen spring wheat genotypes were evaluated in Latvia over three consecutive years (2021–2023) under two nitrogen fertilization levels (N75 and N150) in a split-split-plot design with two replicates, totaling 64 plots. NUE consistently differed between N rates and was strongly influenced by year-specific environmental conditions, providing contrasting scenarios for testing sensing approaches. To capture this variation, two platforms were tested for spectral estimation of NUE: a low-cost proximal phenomobile equipped with an RGB sensor, and an unmanned aerial vehicle (UAV) with a multispectral sensor. Canopy reflectance was measured at three growth stages (tillering, flowering, and milk development) to calculate 8 proximal and 9 UAV-based visible-spectrum vegetation indices (VIs). Although relationships between VIs and NUE were environmentally dependent, significant and robust correlations were found. Proximal sensing generally provided stronger prediction models, with the Normalized Green-Red Difference Index (NGRDI) and Green Area Index (GA) consistently most predictive across years. The milk development stage (GS75) proved optimal for NUE estimation. Comparisons of NGRDI between platforms demonstrated their compatibility, though UAVs offer higher throughput for large-scale phenotyping. These findings highlight the potential of integrating agronomic evaluation with canopy reflectance traits to support breeding and precision nitrogen management.
Key words:
canopy reflectance, correlation, NUE prediction, proximal and remote sensing, Triticum aestivum L.