Band selection for analysing wheat water status under field conditions using relative depth indices (RDI)

Abstract

In this study, the hypothesis that optimum wavebands could be selected for calculation of RDI values was tested. The canopy reflectance data from 350 nm to 2500 nm were collected at various growth stages of wheat irrigated in six amounts before stem elongation. Leaf water contents (%LWC) were insensitive to the irrigation level, while the leaf area index (LAI) and dry phytomass (DM) appeared to be sensitive to the irrigation level. The irrigation levels did not cause considerable spectral response within a specific growth stage, but spectral differences between the growth stages were evident, resulting in strong relationships between the spectral data and the biophysical parameters based on the pooled data of different stages and poor and inconsistent relationships between the spectral data and the biophysical parameters based on data within a specific growth stage. Under field conditions, the peak absorption positions of the wheat canopy were found to shift away from the theoretical wavelengths in the water absorption regions. When using the pooled data covering all the growth stages, the RDI values performed better than reflectance minimum and large differences were found between the RDI values on different selected wavebands in correlating the spectral data with the biophysical parameters. The 965–1085 nm and 1192–1282 nm spectral regions appeared to be the optimum bands for relating RDI to %LWC.

Acknowledgment

This research was supported by China National High Technology Research and Development Program of China (2002AA243011) and Beijing Demonstration Project for Precision Agriculture (A00300100584) from State Development Planning Commission, PR China.