A prototype procedure is presented for developing subjective, probabilistic evaluations for future uncertainties about commodity yield and export expectations. The subjective uncertainty information is incorporated into a large multi commodity systems model in order to estimate future commodity price uncertainties that are useful for public policy and program decisions in agriculture.