On-farm comparison of variable rates of nitrogen with uniform application to maize on canopy reflectance, soil nitrate, and grain yield

Recent development in canopy optical-sensing technology provides the opportunity to apply fertilizer variably at the field scale according to spatial variation in plant growth. A field experiment was conducted in Ottawa, Canada, for two consecutive years to determine the effect of fertilizer nitrogen (N) input at variable- vs. uniform-application strategies at the V6-V8 growth stage, on soil mineral N, canopy reflectance, and grain yield of maize (Zea mays L.). The variable N rates were calculated using an algorithm derived from readings of average normalized difference vegetation index (NDVI) of about 0.8 m × 4.6 m, and N fertilizer was then applied to individual patches of the same size of NDVI readings (0.8 m × 4.6 m) within a plot (2184 m2). Canopy reflectance, expressed as NDVI, was monitored with a hand-held spectrometer, twice weekly before tasseling and once a week thereafter until physiological maturity. Soil mineral N (0-30cm depth) was analyzed at the V6 and VT growth stages. Our data show that both variable and uniform-application strategies for N side-dressings based on canopy-reflectance mapping data required less amount of N fertilizer (with an average rate of 80kg N ha-1 as side-dressing in addition to 30kg N ha-1 applied at planting), and produced grain yields similar to and higher nitrogen-use efficiency (NUE) than the preplant fully fertilized (180kg N ha-1) treatment. No difference was observed in either grain yield or NUE between the variable- and uniform-application strategies. Compared to unfertilized or fully fertilized treatments, the enhancements in grain yield and NUE of the variable-rate strategy originated from the later N input as side-dressing rather than the variation in N rates. The variable-rate strategy resulted in less spatial variations in soil mineral N at the VT growth stage and greater spatial variations in grain yield at harvest than the uniform-rate strategy. Both variable- and uniform-application strategies reduced spatial variations in soil mineral N at the VT stage and grain yield compared to the unfertilized treatment. The variable-rate strategy resulted in more sampling points with high soil mineral N than the uniform-rate strategy at the VT stage. © 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.