Wednesday, November 15, 2006
286-9

Using Remote Sensing to Optimize In-Season Nitrogen in Winter Wheat.

Ravi Sripada1, Randy Weisz2, Ronnie Heiniger2, Jeffrey G. White2, and Dianne C. Farrer3. (1) North Carolina State Univ, Crop Science Dept, 4123A Williams Hall, Campus Box 7620, Raleigh, NC 27695-7620, (2) North Carolina State Univ, Crop Science Dept, Campus Box 7620, Raleigh, NC 27695-7620, (3) North Carolina State Univ, Campus Box 7620, Raleigh, NC 27695

Remote sensing in the form of aerial color infrared (CIR) photography has been shown as a tool for in-season N management in winter wheat (Triticum aestivum L.). The objectives of this study were (i) to develop a methodology for predicting in-season N requirement for winter wheat at growth stage (GS)-30 using aerial CIR photography, and (ii) to determine if the relationship between optimum N30 rates and spectral indices measured at GS-30 is sensitive to the density of the wheat crop (biomass). Field studies were conducted for 3 winter wheat growing seasons over a wide range of soil conditions across North Carolina. Different planting date seeding rate (PDSR) combinations were applied to create a range of biomass levels at GS-30. Different levels of N were applied (i) at GS-25 (N25) to create a range of N supply and winter wheat color and near infrared (NIR) radiance and (ii) at GS-30 (N30) to measure grain yield response to N30. Aerial CIR photographs were obtained at each site at GS-30 prior to N applications. Significant biomass response to PDSR and yield response to N25 and N30 was observed. Optimum N30 ranged from 0 to 124 kg ha-1 with a mean of 55 kg ha-1. Better prediction of optimum N30 rates was obtained with spectral indices calculated relative to high-N reference strips compared to absolute bands or indices. Biomass measured at GS-30 influenced the strength of the relationship between optimum N30 and spectral indices. The best predictor of optimum N30 (R2 = 0.85) was a quadratic model based on measured winter wheat radiance in the green (G) band expressed relative to high-N reference strips (Rel GS) when the GS-30 biomass was > 1000 kg ha-1.