This project was designed to evaluate the feasibility of using the GreenSeeker® RT200 mapping system to drive variable rate application (VRA) of plant growth regulators (PGRs), defoliants, and side-dress N on cotton in Georgia, USA. The study was conducted in producer fields using replicated strips to evaluate treatments. We observed a large amount of soil and yield variability in each of the fields that was not related to treatments.
VRA of Defoliants
NDVI appears to be an excellent tool for managing defoliant applications. Under typical growing conditions, cotton plants begin senescing as they approach maturity providing for a wide range of NDVI values and the opportunity to create management zones that benefit from VRA. In the three experimental fields which we studied, VRA resulted in 22.7% less, 18.2% less, and 15.6% less defoliant used per acre. In all 3 cases, defoliation effectiveness was at least as good as the constant rate. VRA of defoliants resulted in a net gain of $1.58/ac (savings in product). We believe that VRA of defoliants can be easily adopted by cotton producers.
VRA of Plant Growth Regulators (PGRs)
The results indicate that NDVI appears to be very good tool for differentiating PGR management zones early and late in the growing season. However, we observed that at mid-season, when the entire field is covered by a solid green canopy, NDVI values become saturated and are not useful for creating PGR management zones. In other words, NDVI is no longer able to discriminate a large plant from a smaller plant if both plants have enough canopy to completely fill the GreenSeeker’s optical field. To address this problem, we developed an ultrasonic sensing array which we used to collect plant height data. The ultrasonic sensors were mounted on the spray boom adjacent to the GreenSeeker sensors. We then developed decision rules for combining plant height and NDVI data to create biomass maps in late July and early August. The biomass maps were used to develop PGR management zones and subsequently, PGR prescription maps. In the three fields we studied, we used 14% less, 6.5% less and 1% less product for each of three PGR applications. These differences were primarily a result of the low biomass zones which received a zero rate during the first and second applications. Over the entire growing season, we used 7% less product in our VRA of PGR strips which resulted in an average savings of $0.12/ac.
VRA of Side-Dress N
In Georgia typically about 20-25% of the N is applied at planting with the remaining 75-80% of the N applied 7-10 weeks later as a liquid or granular side-dress treatment. Physiologically, side-dress N is applied from first square (flower bud) to early flower – a window of about 21 days. Because even a 10% reduction in N usage will save American cotton producers over $15 million, several researchers across the cotton producing region of the United States are developing algorithms for cotton nitrogen fertilization based on optical sensors (Arnall et al., 2008; Khalilian et al., 2008; Scharf et al., 2008). These algorithms calculate N requirements based on an in-season estimate of the potential or predicted yield estimated from VI response (Figure 1), determine the likely yield response to additional nitrogen fertilizer, and finally calculate N required for obtaining that additional yield (Raun et al., 2005). These algorithms tend to be region and variety specific.
During the 2010, 2011, and 2012 growing seasons, we conducted research to evaluate the feasibility of variably applying side-dress N on cotton using these algorithms. Our side-dress N experiment in 2010 on two producer fields resulted in between 6.7% to 11.7% reduction in side-dress N with an average reduction of 9.2% and with no yield reduction when compared to the uniform application rate of 80 1b N/ac. The average VRA rate in our two fields was 70.7 lb N/ac. Our financial savings were calculated to be $2.11/ac (savings in product).
During 2011 and 2012 we evaluated two different VRA algorithms for determining the amount of side-dress N to associate with measured NDVI values. In both years we found that algorithms which reduce side-dress N applications in high yield potential areas because of large NDVI values depress yield in those area. Apparent soil electrical conductivity (soil ECa) which serves as a surrogate for soil texture can be good way to identify areas of high yield potential. During 2011 and 2012, VRA of side-dress N was found to be more profitable than the uniform applications by as much as $70/ac ($173/ha). This is a result of both product savings and yield gains.
The table below summarizes our research group’s findings for VRA of agrochemicals on cotton. This is a summary of product savings only (does not include any yield gains). The final line of the table extrapolates to 2000 ac which is a typical area planted to cotton each year for many cotton farmers in Georgia.
Vellidis, G., B. Ortiz, G. Ritchie, A. Peristeropoulos, C. Perry and K. Rucker. (2009). Using GreenSeeker® to drive variable rate application of plant growth regulators and defoliants on cotton. In: van Henten, E.J., D. Goense, and C. Lokhorst (Eds.), Precision Agriculture ‘09 – Proceedings of the 7th European Conference on Precision Agriculture, Wageningen Academic Publishers, The Netherlands, pp. 55-72. Download PDF
Vellidis, G., H. Savelle, R. G. Ritchie, G. Harris, R. Hill, and H. Henry. 2011. NDVI response of cotton to nitrogen application rates in Georgia, USA. In: J.V. Stafford (Ed.), Precision Agriculture 2011 – Proceedings of the 8th European Conference on Precision Agriculture (8ECPA), Prague, Czech Republic, pp. 358-368. Download PDF
Liakos, V., G. Vellidis, G. Harris, R. Hill, and H. Henry. 2013. Variable rate application of side-dress nitrogen on cotton in Georgia, USA. In: J.V. Stafford (Ed.), Precision Agriculture 2013 – Proceedings of the 9th European Conference on Precision Agriculture (9ECPA), Lleida, Spain. Download PDF