Spatial patterns in hard red winter wheat quality and basis

Abstract

Chapter I:

Hard red winter wheat is broadly grown in Great Plains states. Knowing how wheat quality is correlated across space may help wheat buyers know where to find wheat with the characteristics they need. The mean of wheat quality characteristics is estimated for each location first. A variogram is used to represent the spatial correlation of these local expected values. Variograms are also estimated for the residuals by year. Such information could be used to determine how large an area a wheat sample represents. The result shows that expected local wheat quality has a strong spatial correlation even over a large distance. The spatial correlation of residuals changes across years. Thus, the conclusion is that having a survey each year provides new information to wheat buyers. The results can explain why Plains Grains Inc., who provided the data, conducts a wheat quality survey every year.

Chapter II:

This chapter looks at the spatial patterns of hard red winter wheat protein and basis. Additionally, a hedonic model between wheat basis and protein is built to determine if the protein premium varies across space. The spatial regression models are estimated using Bayesian Kriging so that coefficients can vary across space. The theoretical variogram model is fitted in the covariance matrix so we can quantify the spatial variation. Local basis and protein premium are highly correlated across space, and protein premium changes relatively large every year. The hedonic model shows that protein premiums are largest in the western part of the Southern Great Plains. The Pacific Northwest region shows no protein premium, which is presumably because protein premiums are paid directly through price in these areas.