A SPATIAL MODEL FOR HIGH PLAINS AGRICULTURE.

A random field model, consisting of a regional trend component and a local interaction component, describes the spatial pattern of corn and wheat yields in the High Plains. The presence of spatial autocorrelation in the local component raises interesting statistical estimation problems. Temporal variations in the slope and interaction parameters are a function of climatic fluctuations. The local component reflects intercounty spatial autocorrelation in precipitation patterns. [ABSTRACT FROM AUTHOR]