Nitrogen (N) rate recommendation tools are utilized to help producers maximize grain yield production. Many of these tools provide recommendations at field scales but often fail when corn N requirements are variable across the field. This may result in excess N being lost to the environment or producers receiving decreased economic returns on yield. Canopy reflectance sensors are capable of capturing within-field variability, although the sensor algorithm recommendations may not always be as accurate at predicting corn N needs compared to other tools. Therefore, the integration of within-field canopy reflectance sensor tools with field-scale N recommendation tools may help account for yield variability from N applications, and improve N rate recommendations by utilizing the strengths of multiple tools. Research was conducted to determine which N rate recommendation tool was most effective at recommending economical optimal N rates (EONR) under varying soil and weather conditions across the Corn Belt. A second objective using a canopy reflectance algorithm was evaluated to by changing the base N rate of the algorithm which was determined by these tools. Research was conducted on N response plots across eight U.S. Midwest states in 2014 and 2015. Two sites from each state totaling 32 site years, resulting in a range of historically productive areas, were used to evaluate differences in soil and weather environments. Field-scale tools that were compared included pre-plant soil nitrate test, pre-sidedress soil nitrate test, maximum return to N (MRTN), yield goal based calculations, and the Maize-N crop growth model. These tools were also compared to N recommendations form a canopy reflectance sensor using the Holland and Schepers algorithm. Tools were evaluated for an at-planting and/or sidedress N application. Each tool’s performance was evaluated using the root mean square errors (RMSE) and the average difference of the tool’s N recommendation to the measured site’s EONR, and the percentage of sites where the N recommendations were within 30 kg N ha-1 of EONR. A second objective was to determine if the Holland and Schepers algorithm could be improved by integrating the best performing N recommendation tools that were previously evaluated. Tools were integrated by replacing the base N rate, or the farmer’s historical N rate, with the N recommended from the best performing tools. Results of comparing the performance of all tools showed that for recommendations made at-planting the Wisconsin PPNT, MRTN and state-specific yield goals performed the best. For sidedress recommendations MRTN and state-specific yield goal recommendations were the best performing tools. The canopy reflectance sensor using the farmer’s N rate as the base N rate for the algorithm recommendation did not perform as well as the farmer’s N rate as a standalone recommendation. For the second objective of replacing the farmer’s N rate as the base N rate in the algorithm with MRTN or the state-specific yield goal showed minimal improvement. The canopy reflectance sensor performed better when using the scaling factor to increase the MRTN and state-specific yield goal calculations by a factor of 1.75 and 1.65, respectively. The canopy reflectance sensor was best improved by adding 56 and 70 kg N ha-1 to the overall Holland and Schepers algorithm when using MRTN and the state-specific yield goal as the base N rate, respectively. Overall, using these tools as the base N rate for this algorithm is not appropriate as it caused under-recommendations of EONR and growers would be required to speculate on the scaling factor or how much extra N to apply to the recommendation in order to maximize the performance.
