Estimating soybean seed protein and oil concentration before harvest

Abstract

AbstractTemperature and precipitation during the growing season have been shown to influence the protein and oil composition of the soybean [Glycine max L. Merr.] seed. A method based on these parameters was developed to estimate protein and oil concentrations of the seed before harvest. This method was developed with protein and oil data and temperature and precipitation data from the Uniform Soybean Tests, Southern Region, for the years 1975 to 1983. Classification and regression “tree‐based” analyses were used to determine the month and numeric value (“splitting point”) of the environmental variable that correctly classified the variation from median protein and oil composition for the 126 location‐years. Temperature in September was most influential in determining the splitting point for three of the four variables. Oil concentrations from the location‐years were separated into low vs. high median‐based boundary categories most readily by the September sum of minimum temperatures. Total protein and oil concentrations from the location‐years were classified best by September growing degree days. Protein‐to‐oil ratios were best separated by the September mean minimum temperature. The August mean maximum temperature best separated protein concentration. These data demonstrate that temperature during specific months of the crop year were useful in estimating the final concentration of protein and oil in the seed and could be used by seed processors to estimate seed composition before harvest.