Database-Driven N Decision Support: I. Corn Nitrogen Rate Calculator.

Nitrogen rate guidelines have a historical basis of interpretation from research trials. Conversion of response trial results to N rate recommendations was, beginning in the 1970s, often implemented using the rationale developed by Stanford (1973) who related expected yield (yield goal) to the rate of fertilizer N application. Vanotti and Bundy (1994) argued for the inclusion of soil-specific characteristics to refine N fertilizer rate guidance. In practice, this meant using data from multiple N response trials as the basis for fertilizer N recommendations. This alternative rationale, based directly on N rate response trials, was the underpinning for the regional Corn Belt development of the Maximum Return to Nitrogen (MRTN) approach utilized in the on-line Corn Nitrogen Rate Calculator (CNRC) decision support tool [Nafziger et al. (2004); Sawyer et al. (2006)]. There were several reasons for discarding the yield goal-based recommendation, but of most importance was the fact that across recent trials corn yields showed little correlation with economic optimum N rate (EONR). This disparity developed over time as corn yields in geographic areas with highly productive fields increased faster than EONR, while yields in other geographic areas with lower productivity required more N fertilization than indicated by the yield-based system. As a result, yield-based N rates for highly productive soils were generally too high, and in less-productive soils too low to maximize return to N. The MRTN rate and most profitable rate range recommendations from the CNRC are based on an extensive database of N rate response trials (currently totaling more than 1,400 trials) conducted in seven states. This database-driven decision support approach is unique in that recommended N application rates are derived dynamically from the database of response trials. This provides for discrete geographic and rotation specific rate guidelines, with the additional benefit of utilizing directly the research trial database that supports rate recommendations. The CNRC database-driven approach is applicable to optimization of any rate-dependent crop input, and so may find wider use in applications such as seeding rate decisions.