Search for a command to run...
Correlations between plant canopy reflectance and aboveground biomass can possibly be used for early prediction of crop yield. Field experiments were conducted in 1998 and 1999 on two soil types to assess whether measurements of canopy reflectance at given stages of development could be used to discriminate high from low potential yields among genotypes with known differences in potential grain yield and whether a consistent relationship between yield and canopy reflectance could be used for screening and predicting soybean [ Glycine max (L.) Merr.] yield in a variety trial. A 3‐by‐42 factorial experiment, arranged in a randomized complete block design with three replications, was used on each soil type for both years. Three population densities (25, 50, and 75 seeds m −2 ) represented low, optimum, and high levels. Forty‐two historical varieties represented nearly six decades (1934–1992) of soybean yield improvement in Canada. Canopy reflectance was measured with a hand‐held multispectral radiometer on three sampling dates (approximately R2, R4, and R5 stages) for each site. Grain yield at harvest was measured. Soybean grain yield was highly positively correlated with canopy reflectance, expressed as normalized difference vegetation index (NDVI), at all sampling dates. Regression analyses showed a positive relationship between NDVI and grain yield, with R 2 up to 0.80 ( P < 0.01) and progressive improvement from R2 to R5 stages. Population density did not affect the yield–NDVI relationship at the development stages studied. Our data suggest that canopy reflectance measured nondestructively between R4 and R5 stages adequately discriminates high‐ from low‐yielding genotypes and provides a reliable, fast, repeatable indicator for screening and ranking soybean genotypes based on the relationship between NDVI and grain yield ( R 2 ranged from 0.44–0.80).