摘要

As a plant-based N diagnostic tool, critical N curves can be used to adjust the N input in a field according to the crop demand. Because the existing critical N curve determination methods may overestimate the critical N point, making the designed N curve above the actual one, this study presents an improved method to overcome this drawback. An N experiment was conducted during 2 yr (2010-2012) using different winter wheat (Triticum aestivum L.) cultivars in the northwestern plain of Shandong Province, China. Shoot biomass and N concentration were sampled weekly during the growing season. A new method for designing critical N curves was developed based on the existing N curve determination method and the t-test inversion. Using the above data set, the new and existing methods for designing a critical N curve were compared, and the effects of the different cultivars on the proposed N curve were evaluated. The results showed that the critical N curves generated using existing methods can successfully classify N-limited plants but fail to classify non-N-limited plants, however, the critical N curve determined using the new method was able to classify both N-limited and non-N-limited plants satisfactorily. Cross-validation showed that the new method exhibited a highest total correct classification rate of 76%, while the existing methods displayed total correct classification rates below 64%. Additionally, the proposed critical N curve is robust within species within the same region. Thus, the new method may be useful for N management in the field.