Using canopy indices to quantify the economic optimum nitrogen rate in spring wheat

Abstract

In-season N applications to spring wheat (Triticum aestivum L.) may increase profits and improve N fertilizer accuracy. The objectives were to develop a calibration tool employing normalized difference vegetative index (NDVI) and SPAD 502 chlorophyll meter (SPAD) measurements for calculating the differential from the economic optimum N rate (dEONR) at growth stages Z22, Z24, and Z31 to Z39 and provide N rate algorithms for use in applying N fertilizer at a variable rate. Sensing was conducted trials over 3 yr encompassing 10 site-years across Southeastern Buenos Aires Province, Argentina. The relationship between sensor indices and dEONR was evaluated by fitting quadratic plateau (QP) regression models. Statistically significant QP models were determined at the Z24, Z31, and Z39 growth stages. Relative SPAD (rSPAD) and relative NDVI (rNDVI) reduced variation and improved the calibration of measured N stress with the dEONR. For Z31 and Z39, the rSPAD had the best goodness of fit statistics when compared to rNDVI [adjusted R2 (adjR2)= 0.67 and 0.57 at Z31 and 0.68 and 0.52 at Z39, respectively]. However, adjustment at Z24 was higher for rNDVI (adjR2 = 0.53 and 0.61 for rSPAD and rNDVI, respectively). A single QP model to estimate the dEONR with 58% confidence was adjusted for the Z31 and Z39 growth stages. This indicates that the same calibration for N rate determination based on rSPAD or rNDVI values can be used during stem elongation in spring wheat. This model can be used as an N rate algorithm for applying N fertilizer in-season.