Expanding the services of cereal/legume cover crop mixtures: From UAV-RGB species-dominance identification to precision-based pre-plant nitrogen decisions

Abstract Pre-plant nitrogen (N) fertilization is typically applied uniformly despite within-field variability in soil N availability, particularly following service crops (SC). In legume-cereal SC mixtures, competitive dominance reflects soil N conditions and may serve as an ecological indicator of N availability. This study assessed whether UAV-RGB species dominance mapping can support spatial pre-plant N decisions for a subsequent cash crop. The framework was developed in a controlled plot during one SC season with four N rates (0–120 kg N ha⁻¹) applied to create soil N variability. UAV-RGB imagery was acquired three times during the season and combined with biomass, soil N, and nitrogen nutrition index (NNI) measurements. A legume-to-cereal dominance ratio was derived from classified imagery and linked to NNI to define a threshold for binary pre-plant N decision mapping. The approach was subsequently evaluated in a 20-ha commercial field monitored over two seasons, using UAV imagery and biomass measurements. Soil N availability strongly influenced species dominance: legumes dominated under low N conditions, whereas cereals dominated under higher N levels. In the experimental plot, the dominance ratio derived from mid-season imagery (64–77 days after sowing (DAS)) corresponded with NNI measured later (91 DAS), enabling delineation of areas likely to require or not require pre-plant N fertilization. In the commercial field, classified canopy proportions aligned with measured biomass in both seasons. The experimentally derived threshold distinguished between an N-deficient and N-fertilized seasons, where most subplots required or did not require pre-plant N. These findings demonstrate that UAV-RGB mapping of species dominance in legume-cereal SC mixtures can indicate soil N availability before cash crop establishment. Although agronomic validation of variable-rate pre-plant fertilization was not conducted, the study presents a transferable framework that integrates SC-based soil management with precision nitrogen decision support, using accessible remote sensing tools. Graphical Abstract