Delivering Reliable Crop Yield Forecasts for Crop Nutrition Planning.

Anyone who has ever had to consistently deliver reliable yield forecasts to producers can well appreciate the famous saying from Neils Bohr: "Prediction is very difficult, especially about the future". Measurements are always inadequate, models are always wrong, and “the future ain’t what it used to be”.  Nevertheless, useful and increasingly accurate forecasts can be obtained by incorporating existing knowledge in appropriately-scaled models. The PRS™ Nutrient Forecaster is a knowledge-based model that predicts crop yields as a function of environmental conditions and fertilizer rate.  Soil nutrient supplies are estimated from a standardized measurement of soil nutrient flux to ion exchange membranes [Plant Nutrient Simulator (PRS^TM)-probes].  A crucial component of this approach is the evaluation of forecasts with backcasts of actual yields.  This approach has been used to improve fertilizer decisions by farmers in western Canada since 1997.