The objective of this research was to determine if canopy reflectance measured by an Unmanned Aerial Vehicle (UAV) equipped with a 5-band multi-spectral camera can differentiate between water and nitrogen (N) deficiency in irrigated maize. Crop reflectance was used to generate a Normalized Difference Red Edge (NDRE), Green Leaf Index (GLI), and a Blue Reflectance Index (BRI). These indices were then used in combination to categorize N and water stressed experimental units into a Combined Index (CI) indicating water-stressed, N-stressed, or non-stressed crops. The CI generated at blister (R2) successfully identified 90% of experimental treatments to the correct group but only identified 60% of treatments when generated at the 14th leaf stage (V14). The CI methodology was subsequently applied to two independent site-years where only N deficiency gradients were imposed. The CI was not successful at separating treatments at the validation sites, incorrectly identifying water stress where there was none. Among individual indices investigated, NDRE had the strongest relationship to grain yields (r2 = 0.62, p < 0.001) but a weaker linear relationship compared to the CI (r2 = 0.68, p < 0.005) where deficit irrigation treatments were imposed. At sites where irrigation was sufficient to meet crop water demand, NDRE (r2 = 0.63, p < 0.05) had a stronger relationship to grain yield compared to the CI (r2 = 0.41, p = 0.31). This study found that, under narrow cropping system circumstances, N and irrigation-induced differences in maize productivity can be differentiated in-season by a combination of reflectance indices, but that NDRE alone provides superior information under broader contexts.
Leaf Canopy Layers Affect Spectral Reflectance in Silver Birch
