2008 Joint Annual Meeting (5-9 Oct. 2008): Crop Sensors for Variable-Rate Nitrogen Application to Cotton.

671-2 Crop Sensors for Variable-Rate Nitrogen Application to Cotton.



Tuesday, 7 October 2008: 8:45 AM
George R. Brown Convention Center, 360F
Peter C. Scharf1, Luciane Oliveira1, Earl Vories2, David Dunn3 and Gene Stevens3, (1)University of Missouri, Columbia, MO 65211
(2)USDA-ARS Cropping Systems & Water Quality Research Unit, USDA-ARS, 269 Agricultural Engineering Bldg, Columbia, MO 65211
(3)University of Missouri, PO Box 160, Portageville, MO 63873
Cotton yield can be very responsive to nitrogen fertilizer.  However, over-application of N can result in excess vegetative growth, which can delay maturity and increase the need for growth regulator, defoliant, and insecticide, in addition to wasting money on fertilizer that does not produce a return.  Our objective was to calibrate canopy reflectance sensors to predict the amount of N fertilizer needed by a cotton crop.  Six N rate experiments were carried out over a two-year period on widely varying soil types.  Three sensor types (Greenseeker, Crop Circle, and Cropscan) were used at three stages (early square, mid square, and first bloom) and three heights above the canopy (10, 20, and 40 inches).  The most profitable N rates for these six experiments were 0, 45, 60, 80, 175, and 200 lb N/acre.  Applying these rates would have increased profitability by $43/acre relative to a typical producer rate of 100 lb N/acre on every field.  All three sensors were able to give fairly good predictions of the most profitable N rate at mid square and first bloom, but not at early square.  The visible/near-infrared ratio from a height of 20 inches predicted the best N rate with R2 > 0.5 for all three sensors and was the most consistent predictor.  Our results support the feasibility of applicator-mounted sensors to control variable-rate N applications to cotton.  Initial field-scale demonstrations are planned for 2008.