Abstract

The critical nitrogen (N) dilution curve (CNDC) (Nc = A(1)W(-A2)) and the N nutrition index (NNI) have shown to be a robust diagnostic tool to assess crop N status. Several CNDCs were established for potatoes (Solanum tuberosum L.) in various conditions. However, it is difficult to know if these different CNDCs reflect a true variability across genotype x environment x management (GxE x M) scenarios, or are rather linked to estimation errors. Therefore, the aims of this study were to (i) estimate and compare CNDCs via a Bayesian statistical approach for different GxE x M scenarios in tropical environments and (ii) assess the uncertainty in the relationship between tuber yield and NNI. Field experiments were conducted in three site-years (SYs) in the southeastern region of Brazil. Treatments consisted of two potato cultivars (Agata and Electra) and five N fertilization managements [differing in the N rates (0 -320 kg N ha(-1)) and/or application timings]. Based on 95% credibility intervals, we found no significant differences between CNDCs across GxE x M scenarios assessed and thus derived a single generic CNDC for both cultivars and all SYs defined by Nc = 3.86 W-0.34. This global CNDC was used to compute NNI and discriminate tuber yields obtained under N-limiting and non-N-limiting conditions. An uncertainty analysis revealed that critical NNI threshold values for the tuber yield response to N management were not statistically different among GxE x M scenarios. Therefore, these results support the use of a unique CNDC for estimating the N status (NNI) of potato crop in tropical environments. We found no evidence of the need for SY- or cultivar-specific CNDCs. Future studies should evaluate N dilution curves and NNI for potato across different production environments worldwide.