THE VALUE OF INFORMATION CONCERNING NEAR-OPTIMAL NITROGEN-FERTILIZER SCHEDULING

A methodology for estimating the economic value of perfect information concerning the optimum application schedule of nitrogen fertilizer is illustrated using CERES-Maize, a simulation model of the growth, deve lopment, and yield of maize for a location in Florida, USA. The optimu m region of a response function in 11 dimensions was found using an ef ficient and robust hill-climbing technique. The optimum nitrogen appli cation schedule in any season was found to be highly dependent on weat her. For the years considered in the analysis, 1978 to 1987, maximum g ross margins ranged from -$15 to $526 ha-1, with amounts of nitrogen a pplied (as urea) varying from 49 to 240 kg ha-1. Ex-post analysis show ed that the value of perfect information compared with the best ex-ant e strategy considered was approximately $65 ha-1. The possibility of u sing the crop model in a yield forecasting mode at the farm level, to attempt to capture even a portion of these benefits ex ante, is briefl y considered. In farming areas of comparatively high levels of weather -related risk, the use of appropriate forecasting and input optimizati on techniques could result in moderate economic benefits over the long term.