Nitrogen fertilizer recommendations based on plant sensing and Bayesian updating

Abstract

Methods are available to predict nitrogen needs of winter wheat based on plant sensing, but adoption rates by producers are low. Current algorithms that provide nitrogen recommendations based on plant sensing implicitly assume that parameters are estimated without error. A Bayesian updating method was developed that can incorporate precision plant sensing information and is simple enough that it could be computed on-the-go. The method can consider producers prior information and can account for parameter uncertainty. Bayesian updating gives higher nitrogen recommendations than plant sensing recommendations using a plug-in method. These recommendations increase net returns over the previous recommendations, but not enough to make plant sensing profitable in this scenario.