Method for Precision Nitrogen Management in Spring Wheat: I Fundamental Relationships

Wheat (Triticum aestivum L.) fields in the semi-arid Northern Great Plains are spatially variable in soil N fertility and crop productivity. Consequently, there is interest in applying variable, rather than uniform rates of fertilizer N across the landscape. Intensive soil sampling as a basis for variable-rate fertilizer management is too costly when compared to the value of wheat in this region. The objective of this research was to determine relationships between yield and protein, and protein and available N as needed to develop a cost-effective variable-rate N fertilizer strategy for spring wheat. A three-year study (1996–1998) was carried out at a site near Havre, Montana, USA (48°30N, 109°22W). Treatments consisted of three water regimes, four cultivars, and five fertilizer N levels per water regime in a randomized complete block design with four replicates. Scatter diagrams of relative yield vs. grain protein were consistent with earlier investigators, and indicated protein concentrations at harvest provided a method for indexing N nutrition adequacy (deficiency vs. sufficiency) in wheat. A critical protein concentration of 13.2% was defined using a graphical Cate-Nelson analysis. This value appeared to be consistent across the three water regimes and four cultivars as 159 (88%) of the 180 water×cultivar×N level episodes were in positive quadrants. No correlation could be found between relative yield and protein for episodes below the critical level (r²=0.1). Hence, grain protein concentrations could not be used to predict the magnitude of yield losses from N deficiency. Grain protein content would be useful for prescribing fertilizer recommendations where N deficiency (<13.2% protein) reduces grain yield under semi-arid conditions. Inverse slopes (dy/dx) of the protein-available N curves reveal that it takes 12–18 kg N/ha to change protein 1% (e.g., 12% vs. 13%) where wheat is under water stress during grain fill. The total N requirement could then be computed by summing the N required for raising protein and the N removed by the crop in the year when the grain was harvested.