A Decision Support System for Optimum Use of Fertilizers

The Decision Support System for Agriculture (DSS4Ag) is an expert system being developed by the Site-Specific Technologies for Agriculture (SST4Ag) precision farming research project at the INEEL. DSS4Ag uses state-of-the-art artificial intelligence and computer science technologies to make spatially variable, site-specific, economically optimum decisions on fertilizer use. The DSS4Ag has an open architecture that allows for external input and addition of new requirements and integrates its results with existing agricultural systems’ infrastructures. The DSS4Ag reflects a paradigm shift in the information revolution in agriculture that is precision farming. We depict this information revolution in agriculture as an historic trend in the agricultural decision-making process. Using this decision support system, in 1997 we generated a variable-rate fertilizer recommendation recipe for a 54.6 ha wheat field with the goal of optimum economic return, not maximum yield. The field was split into blocks, alternately fertilized with the variable-rate recipe and with the uniform application method used by the farmer. The DSS4Ag fertilizer recipe reduced fertilizer cost 39.7% and yield 3.3%, which resulted in a net economic gain of US $14.31/ha as compared to the uniform application used by the farmer. In 1998 the DSS4Ag was tested on a 64.7 ha wheat field and a 89 ha potato field. The DSS4Ag wheat recipe reduced fertilizer cost by US $15.81/ha, a saving of 30.6% between DSS4Ag and the recipe developed from recommendations by Kephart et al. (1991). The DSS4Ag potato recipe increased fertilizer cost by US $82.83/ha, a 70.1% cost increase compared to the conventional fertilization. Four test plots showed that the tested fertilizer management methodologies, while applying very different fertilizer rates, produced crops of similar yield and quality. The plots also demonstrated the DSS4Ag’s ability to characterize and identify management zones of common and different productive potential.