Nitrogen concentration in crop plants is a key index for assessing plant growth status and predicting grain yield and quality. Non-destructive monitoring and diagnosis of plant nitrogen status is of significant importance for precise nitrogen management and productivity forecasting for field crops.
The present study was conducted to determine the quantitative relationships of leaf nitrogen concentration to canopy reflectance spectra in both wheat and rice. Ground-based canopy spectral reflectance and nitrogen concentrations in leaves were measured with six field experiments consisting of five different rice varieties and three different wheat varieties and varied nitrogen fertilization levels across six growing seasons. All possible ratio vegetation indices (RVI), difference vegetation indices (DVI), and normalized difference vegetation indices (NDVI) of sixteen wavebands from the MSR16 radiometer were calculated. Analyses were made to determine the relationships of seasonal canopy spectral reflectance and all possible vegetation indices to leaf nitrogen concentrations in rice and wheat under different nitrogen treatments and variety types.
The results showed, as expected, that nitrogen concentrations in wheat and rice leaves increased with increasing nitrogen fertilization rates. Canopy reflectance, however, was a more complicated relationship. In the near infrared portion of the spectrum (760～1100 nm), canopy reflectance increased with increasing nitrogen supply, whereas in the visible region (460～710 nm), canopy reflectance decreased with increasing nitrogen supply. For both rice and wheat, leaf nitrogen concentration was best evaluated at 610, 660 and 680 nm. Among all possible RVIs, DVIs, and NDVIs, the NDVI(1220, 610) was most highly correlated with leaf nitrogen concentrations in both rice and wheat. In addition, the correlation of NDVI(1220，610) to leaf nitrogen concentration was found to be higher than that of individual wavebands at 610, 660 and 680nm in both rice and wheat. These results indicated that leaf nitrogen concentrations in both wheat and rice could be monitored with common wavelengths and spectral parameters, and the integrated regression equation could be used to describe the dynamic change patterns of leaf nitrogen concentrations in rice and wheat with reflectance spectra parameters, although separate regression coefficients would slightly enhance the prediction accuracy.