Regional Investigation on Interaction of Nitrogen Management, Hybrid Selection, and Population on Corn Production

IPNI-2011-GBL-47

25 Mar 2014

2013 Annual Interpretive Summary


Making sound recommendations for N fertilizer rate and timing for optimal corn yield and minimal N loss can be complex and challenging, especially considering seasonal and locational variability. Splitting N applications between pre-plant and in-season allows room for adjustment to specific seasonal conditions. Tools such as crop sensors and the Maize N model have been developed to help fine-tune corn N management. The general objective of this study, initiated in 2012, is to evaluate these two approaches for determining in-season application N rates for corn over a three state region (Nebraska, North Dakota and Missouri). Other factors such as plant population, hybrid drought score and soil productivity were also evaluated.

Two experimental sites were selected in each state, making a total of 6 sites. Sites within states were in close proximity, each state having a high and a lower soil productivity site. At each site, a high and an average seeding rate were evaluated for both low and high drought score hybrid for Nebraska and Missouri; the two hybrids used in North Dakota in 2013 were not selected based on drought score. Four basic N treatments were used: unfertilized check, N-rich strip (preplant), sensor-based approach and model-based approach. The two latter treatments involved in-season application. Sensor-based treatments were determined by canopy reflectance using a Handheld Crop Sensor (Holland Scientific). The initial and in-season N application method and source varied by state.

In-season fertilizer application N rates for the model-based treatments were higher than in-season N rates for the sensor-based treatments at 4 of the 6 sites in 2013. Sensor-based treatments had a significantly lower yield than model-based treatments at 2 of the 6 sites, while model-based treatments had a significantly lower yield than sensor-based treatments at 1 of the 6 sites. Overall, it appears from 2013 results that yield is better protected using the model-based approach than the sensor-based approach. However, the sensor-based approach generally produced higher N-use efficiency than the model-based approach. This project will continue in 2014.