Canopy reflectance in cotton for growth assessment and lint yield prediction

A field experiment was conducted in 2001–2002 to investigate relationships between canopy spectral reflectance and leaf area index (LAI), aboveground biomass (ABM), and lint yield of irrigated cotton across four N fertilizer rates of 0, 56, 112, and 168 kg N ha−1. These N rates were used to generate a wide range of difference in canopy structure and lint yield. Measurements of canopy reflectance were made throughout the growing season using a hand-held spectroradiometer. Samples for LAI and ABM were obtained four (2001) or five (2002) times during squaring and fruiting. Mean reflectance values in red (Rred) and near infrared (RNIR) regions were obtained from canopy reflectance data based on the Landsat Thematic Mapper bands. The reflectance ratio vegetation index (RVI), normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), wide dynamic range vegetation index (WDRVI) as well as several hyperspectral reflectance indices were calculated. Most reflectance indices had exponential relationships with both LAI and ABM and reached saturation at high LAI and ABM, but were linearly correlated with log(LAI) and log(ABM). Relative lint yield was linearly correlated to the reflectance indices measured any time after the first square stage and the strongest correlation was obtained at the early flower stage with r2 of 0.56–0.89 (P< 0.01). Therefore, the canopy reflectance indices measured at early flower stage of cotton growth could serve as input to a crop growth model for predicting potential yield loss. These results indicate that the early flower stage is an appropriate time to collect canopy reflectance data for cotton yield estimation. © 2006 Elsevier B.V. All rights reserved.