Hyperspectral diagnosis of nitrogen status in arbuscular mycorrhizal inoculated soybean leaves under three drought conditions

Precision diagnosis of leaf nitrogen (N) content in arbuscular mycorrhizal inoculated crops under drought stress, using hyperspectral remote sensing technology, would be significant to evaluate the mycorrhizal effect on crop growth condition in the arid and semi-arid region. In this study, soybean plants with inoculation and non-inoculation treatments were grown under severe drought, moderate drought and normal irrigation conditions. Leaf spectral reflectance and several biochemical parameters were measured at 30 d, 45 d and 64 d after inoculation. Correlation analyses were conducted between leaf N content and the original and first derivative spectral reflectance. A series of first-order differential area indices and differential area ratio indices were proposed and explored. Results indicated that arbuscular mycorrhizal fungi improved leaf N content under drought stresses, the spectral reflectance in visible to red edge regions of inoculated plants was lower than that of non-inoculated plants. The first-order differential area index at bands of 638-648 nm achieved the best estimation and prediction accuracies in leaf N content inversion, with the determination coefficient of calibration of 0.72, root mean square error of prediction and relative error of prediction of 0.46 and 11.60%, respectively. This study provides a new insight for the evaluation of mycorrhizal effect under drought stress and opens up a new field of application for hyperspectral remote sensing.