Spectral Reflectance Models for Characterizing Winter Wheat Genotypes

Abstract

Optimum wheat (Triticum aestivum L.) yield can be achieved by developing and growing the best genotypes in the most suited environments. However, exhaustive field measurements are required to characterize plants with desirable traits in breeding plots. Remote sensing tools have been shown to provide relatively accurate and simultaneous measurements of plant characteristics without destructive sampling, and at low cost. The aim of this research was to develop and evaluate spectral reflectance-based models for characterizing winter wheat genotypes in the semiarid U.S. Southern Great Plains (SGP). Field experiments were conducted at Bushland, TX, during the 2011–2012 growing season. The spectral behavior of 20 wheat genotypes with wide genetic background was analyzed in relation to leaf area index (LAI) and yield under irrigated and dryland conditions. Reflectance-based models were developed and evaluated using three approaches: the maximum correlations, the optimum multiple narrow band reflectance (OMNBR), and the vegetation indices (VIs). Results indicated that the combinations of two to four bands in OMNBR models explained most of the variability (65% to 89% and 51% to 95% for dryland and irrigated conditions, respectively). Spectral regions in visible (VIS: 350–700 nm), near-infrared (NIR: 700–1,300 nm), and mid-infrared (MIR: 1,300–2,500 nm) were sensitive to LAI and yield, most commonly the MIR region. Models developed in this study are expected to assist in developing rapid and reliable methods for germplasm screening and selection of winter wheat genotypes.