The objective of this study is to determine the potential for using spatial econometric analysis of
combine yield monitor data to estimate the site-specific crop response functions. The specific case
study is for site-specific nitrogen (N) application to corn production in Argentina. Spatial structure
of the yield data is modeled with landscape variables, spatially autoregressive error and groupwise
heteroskedasticity. Results suggest that N response differs by landscape position, and that site-specific
application may be modestly profitable. Profitability depends on the model specification used, with all
spatial models consistently indicating profitability, whereas the nonspatial models do not.