Statistical methods for predicting responses to applied nitrogen and calculating optimal nitrogen rates

Models of response to applied N ran he useful for deriving improved N dose recommendations. Here we show how response model parameters can be estimate d and how model predictions and model N dose recommendations can be evaluat ed. For parameter estimation, we use a statistical approach based on random parameter models, Two methods for evaluating models are applied, The first method is to calculate mean squared error of prediction (MSEP) by cross va lidation, and the second is to perform nonparametric regressions to evaluat e the profitability of calculated optimal N rates. Tile proposed methods ar e used with a data set consisting of 37 winter wheat (Triticum aestivum L.) experiments. Different functions taking into account end-of-winter mineral soil N are evaluated. The results show that the different functions all ha ve similar MSEP values for predictions of yield and grain protein content a nd lead to N recommendations of similar profitability. However, there are s ubstantial differences in MSEP for residual mineral N at harvest. One of th ese models is then compared with a model that does not include any site-yea r characteristic and with a model that does not have random parameters. We find that using the model without a site-year characteristic leads to predi ctions that are less accurate and optimal N rates that are less profitable by F 17 to F 105 ha(-1). Another result is that tile gross margin obtained with tire optimal N rates calculated using the model without random paramet ers is lower by F 438 to F 550 ha(-1).