Development of a model using the nitrogen nutrition index to estimate in-season rice nitrogen requirement

In-season nitrogen (N) nutrition diagnosis and N requirement (NR) estimation are critical to precision management of crop fertilization. Although many studies focus on the diagnosis of in-season crop N nutrition, few of them associated the N status with crop NR or topdressing N fertilizer. The objective of this study was bridging the gap between crop N nutrition status and crop NR for easy implementation of precision fertilization. Field experiments were conducted with different N levels and six rice (Oryza sativa L.) varieties in eastern China during 2010-2014. Rice in-season NR and N nutrition index (NNI) were calculated with plant chemical analysis and pre-established critical N dilution curve. Results showed a significant linear relationship between in-season rice NR and NNI based on the long-term field experiment, and the NR-NNI regression slope changed with days after transplanting (DAT) in a regular pattern. Based on this data changing pattern, a novel NR estimation model was established with three easy-obtained independent variables (DAT, N-fertilizer recovery efficiency and NNI). Independent model validation results showed very good performance on different rice varieties and growth stages, with root mean square error (RMSE) of 5.93, 6.96, 8.74 and 6.43 kg ha(-1) for three japonica rice and one indica rice varieties, respectively. This model will significantly facilitate the estimation of in-season rice NR and provide strong technical support for the precision management of N fertilization.