In-Season Nitrogen Status Assessment and Yield Estimation Using Hyperspectral Vegetation Indices in a Potato Crop

The rate and timing of N applications are important issues in precision agriculture because of the within-field spatial and temporal variability of soil N availability. In-season assessment of potato (Solanum tuberosum L.) crop N status (CNS) is required to better match N fertilizer supply to crop N demand and improve N use efficiency. The objective of this study was to investigate the ability of hyperspectral vegetation indices (HVIs) to assess the CNS and tuber yield of irrigated 'Russet Burbank' potato at different growth stages. A 2-yr field experiment was conducted near Quebec City, QC, Canada, on plots receiving five different N rates ranging from 0 to 280 kg N ha(-1), with 40% applied at planting and 60% at hilling. Entire plant samples were collected biweekly for determination of the N nutrition index (NNI) as the N status reference method. In-field hyperspectral reflectance derived from a handheld spectroradiometer and using two fields of view (FOV; 7.5 degrees and 25 degrees) was obtained on several dates during both growing seasons. The sensitivity of the five HVIs most correlated to the NNI was evaluated by analyses of variance and least significant differences. It was found that HVIs computed from reflectance in the red-edge spectral region and using a wider FOV were the most appropriate indices to detect potato crop N stress. Among these indices, the CI1(red-edge) (red-edge chlorophyll index 1) was the most sensitive to potato N content and could explain 76% of the variability in total tuber yield at 55 d aft er planting (DAP).