Grain yield and protein responses in wheat using the N-Sensor for variable rate N application

Abstract

Soil types, cereal crop growth and grain yields are typically variable across many paddocks in the cropping regions of South Australia. In this study the value of a variable rate nitrogen fertiliser application, using the Yara N-Sensor, was compared with the standard practice of a uniform application, at crop growth stage 31, on the grain yield and protein content of wheat. These comparisons were made using the same total amount of fertiliser in paired variable and uniform rate treatments in commercial crops at a total of 10 sites over two years in the medium to higher rainfall areas of the Mid North and Yorke Peninsula of South Australia. The mean increase in wheat grain yield for the variable rate treatment was only 40 kg/ha, or 0.8%, when compared with the uniform rate treatment averaged over these 10 sites and two years. Grain yield differences ranged from 160 kg/ha more to 60 kg/ha less for the variable rate treatment when compared with the uniform rate treatment. Wheat grain yields with the uniform treatments ranged from 2.53 t/ha to 5.68t/ha and with a mean grain yield of 4.24 t/ha. The mean wheat grain protein content with the variable rate treatment was 11.0%, compared with 10.5% with the uniform rate treatment, a relative increase of 5.1%. Where grain yield responses to the variable rate treatments were compared between different biomass areas within a paddock, the greatest grain yield increases to a variable rate of N compared with a uniform rate were in the areas with the lowest 20% of crop biomass whereas grain yield differences were negligible in areas with the highest 60% of crop biomass. These low biomass areas also had the greatest grain yield response to the applied post emergent nitrogen fertiliser when compared with areas with no post emergent nitrogen fertiliser. N-Sensor outputs (biomass and N-rate) were compared with measurements of plant biomass, N uptake (kg N/ha) and %N content at points of contrasting biomass and N-rate within paddocks. There was a high correlation between the N-Sensor biomass and N-rate values and actual plant biomass and N uptake but not with the %N content. Crop biomass maps made using sensors such as the N-Sensor could provide useful data layers, which in combination with other datasets such as grain yield maps or elevation maps, be used to produce zone maps for further analysis or for variable rate input treatments. The N-Sensor could also be used in some situations to map variations in weed biomass for possible site specific weed management.