Use of corn height to improve the relationship between active optical sensor readings and yield estimates

Early in-season loss of N continues to be a problem in corn (Zea mays L.). One method to improve N use efficiency is fertilizing based on in-season crop foliage sensors. The objective of this study was to evaluate two ground-based, active-optical (GBAO) sensors and explore the use of corn height with sensor readings for improving relationships with corn yield. Two GBAO sensors (GreenSeeker(A (R)) (GS), Trimble, Sunnydale, CA, USA; and Holland Crop Circle (CC) ACS 470 Sensor(A (R)), Holland Scientific, Lincoln, NE, USA) were used within 30 established corn N-rate trials in North Dakota at the V6 and V12 growth stages in 2011 and 2012. Corn height was recorded manually at the date of sensor data collection. At the V6 growth stage, the GS relationship to yield and the INSEY (in-season estimate of yield) value was improved when the sensor reading was multiplied times corn height. At the V12 stage, using the GS, the INSEY relationship with yield was also generally increased when height was considered. The CC-based red/near-infrared INSEY relationship with yield was similar to the GS INSEY. The CC-based red edge/near infrared INSEY relationship was increased with height only at the first sensor date, but not with the second. The second CC-based sensor-INSEY relationship with yield was maximized using sensor reading only. Segregating the 30 site data set into sites with high clay surface textures and sites with medium texture improved all INSEY relationships compared to pooling all sites. Relationships between INSEY and corn yield at no-till sites were significant at the V12 stage in the wetter 2011 growing season, but not at the V6 stage either year, nor at the V12 stage in the very dry 2012 season. In the high clay and medium textured soils at the V6 stage, corn height improved the relationship between INSEY and yield often enough to suggest that incorporating corn height into an algorithm for yield prediction would strengthen yield prediction, and thus improve N rate decisions.