Effects of N application rate on N remobilization and accumulation in maize (Zea mays L.) and estimating of vegetative N remobilization using hyperspectral measurements

Hyperspectral data are useful for the nondestructive and timely monitoring of the physiological status and photosynthetic performance of crops. The objectives of the present study were (i) to analyze the effect of N application rate on the post-silking N uptake and vegetative N remobilization status and (ii) to indicate the optimum vegetation index (VI) for the assessment of the N remobilization status in maize. Two field experiments were conducted with five N treatments with varying rates and three maize (Zea mays L.) cultivars in 2016 and 2017, and time-course hyperspectral measurements were taken of collected plant samples. The present study considered the following three approaches: commonly used vegetation indices, novel indices of all possible two-band combinations as either simple ratios (SR) or normalized differences (ND) and spectral differences between the periods 63 and 75 (or 90) day after sowing (DAS). The commonly used VIs, such as the Green normalized difference vegetation index (GNDVI) at 75 DAS and Green model (G-M) at 90 DAS, were successfully used to assess the leaf N remobilization. The VIs computed with all possible two-band combinations ranging from 350 to 1350 nm were related to vegetative N remobilization. The results show that novel indices, such as SR (764 nm/716 nm), SR (762 nm/700 nm), ND (756 nm/550 nm) and ND (900 nm/750 nm), were successfully used to assess leaf N remobilization during the filling stage but were not suitable for monitoring culm N remobilization. Moreover, the spectral differences (i.e., between reflectance at 940 and 799 nm) between the corresponding 63 and 75 (90) DAS could stably estimate leaf N remobilization (R2 > 0.47, RMSE < 6.59, RE (%) < 20.7%). The idea of using remote sensing to calculate leaf N remobilization was feasible because the method calculates nitrogen uptake during critical periods and thus quantifies N deficiency and N management.