Estimating the Potential Value of Variable Rate Nitrogen Applications: A Comparison of Spatial Econometric and Geostatistical Models

Abstract

Site-specific crop response functions (SSCRFs) are useful for estimating the value of variable rate nitrogen applications (VRA), but appropriate statistical models are necessary. Problems estimating SSCRFs using experimental field data include region, spatial, treatment, and strip dependent heteroskedasticity and correlation. We develop a spatial autoregressive error (SARE) model for dealing with these problems and compare results with previous analysis based on a geostatistical (GEO) model. VRA value estimates for the two models differ notably for 1995 data from Southern Minnesota. Furthermore, findings show that the results of a comparison of model performance are location specific.

Keywords

• geostatistics
• precision agriculture
• site-specific crop response functions
• spatial autoregressive error
• variable rate nitrogen application