Validating Weather and Sensor-based N Prediction Models for Michigan Corn Production

Differences in weather, particularly the timing and amount of rain, cause variability in corn response to N from one year to another. The regional Maximum Return to Nitrogen (MRTN) approach used in seven midwestern states provides recommendations based on documented yield responses to N, and allows for economic optimization, but does not adjust for year-specific weather events. This trial compared two year-specific approaches as alternatives to MRTN. One was the use of a deterministic daily weather-based crop and soil model (Adapt-N), and the other was a crop canopy sensor (Greenseeker®) using an algorithm developed in Minnesota. The trial was implemented at the Michigan State University research station near Lansing, Michigan in 2015. A series of N rates ranging from 33 to 167% of the MRTN rate were applied with 40 lb N/A as starter and the remainder as sidedress urea ammonium nitrate (UAN) injected into the soil at the V4 growth stage. In comparison, the weather-based rates were determined just before the V8 growth stage, and applied at V8 using a surface band of UAN with urease inhibitor.

The series of N rates applied at V4 showed an optimum N rate of 173 lb/A, producing a corn yield of 165 bu/A. The MRTN rate was 140 lb/A, producing a yield of 150 bu/A. Owing to greater than normal rainfall before the V8 growth stage, the model recommended a rate of 171 lb/A. Applied at V8, this rate produced a yield of 154 bu/A, similar to but slightly below the yield expectation from the N response curve derived from V4 application timing. The crop canopy sensor recommended a rate of only 121 bu/A, resulting in a significantly lower yield of 138 bu/A. The results from this trial showed the need and potential for weather-based prediction methods. More site-years are needed, however, to evaluate these tools for long-term performance over a range of soil and weather conditions.