Establishing critical nitrogen dilution curve based on whole plant or leaf dry matter for drip-irrigated spring wheat: A Bayesian analysis

Abstract

The critical nitrogen dilution curve (CNDC) serves as a practical tool for nitrogen (N) diagnosis and plays a significant role in fertilization decision-making. However, the feasibility of establishing CNDC based on leaf dry matter (LDM) using Bayesian hierarchical approach and its performance in the spring wheat (Triticum aestivum L.) N nutrition diagnosis and prediction system remain to be investigated. Field experiments were conducted across three regions in Northern Xinjiang, involving four widely cultivated wheat varieties with varying N fertilizer rates. We employed the Bayesian hierarchical approach to develop CNDCs based on plant aboveground dry matter (PDM) and LDM of wheat, the established model was Nc= 3.71PDM^−0.30 and Nc= 3.54LDM^−0.39. We analyze the sources of differences in model parameters and evaluate the diagnostic and predictive performance of the established models. Throughout the study period, the N nutrition index (NNI) calculated from the PDM and LDM derived CNDCs showed a significant positive correlation with SPAD values, with correlation coefficients (r) of 0.80 and 0.76 respectively. Compared with the PDM based CNDC, the LDM based CNDC demonstrated slightly inferior performance in N nutrition diagnosis, but exhibited excellent performance in predicting the critical N concentration (Nc) and yield. Validation results showed that the coefficient of determination (R²) for yield predicted by the LDM based CNDC was 0.936. This study provides novel ideas for the theoretical basis and development potential of CNDC driving variables, offering valuable insights for promoting precision fertilization and agricultural sustainability in farming systems.