A novel approach for determining nitrogen requirement based on a new agronomic principle—sugarcane as a crop model

There are growing evidence that nitrogen (N) recommendation based on the expected yield concept developed by Stanford in 1973 lacks in agronomic principles, despite it widespread use worldwide. In Brazil, the main sugarcane producer worldwide, for example, a fixed N factor of 1 kg N per Mg −1 of stalk is used. However, literature demonstrates that sugarcane responsiveness to N is much higher in sandy soils rather than in clayey soils, whereas the recommended N is usually higher in clayey soils because of its improved yield potential. We investigated 146 response curves of sugarcane ( Saccharum spp.) to N to define a better approach for determining N rates instead of using exclusively the expected yield concept. First, we found no correlation between the economically optimal N rate (N opt , kg ha −1 ) and the yield (Mg ha −1 ) obtained in N opt . Second, we calculate the N requirement (N req , kg N Mg −1 stalk) as the quotient between N opt and the yield, for each response curve. There was a negative correlation between yield at control plot and N req , demonstrating that higher N rates are required to maximize yield in sandy (low yield) sites. The N req was 1.3, 0.9, 0.7 and 0.6 kg N Mg −1 stalk for expected yields < 69, 69–84, 84–102, and > 102 Mg ha −1 , respectively, differing from the standard 1 kg N Mg −1 stalk currently used. The method proposed here is based indirectly on the N-supplying power of the soil and should be tested for other crops to deliver an improved N recommendation system.