Evaluating regional mean optimal nitrogen rates in combination with indigenous nitrogen supply for rice production

Nitrogen (N) is an essential nutrient that requires careful management in intensive rice systems, since insufficient amounts might result in yield losses, and excessive application might harm the environment. However, farmers often tend to apply a large excess of N fertilizer to ensure high rice yields, primarily because of the absence of reliable methods to estimate optimal N application rates. Therefore, a large-scale study comprising 514 field experiments for rice was conducted in seven rice regions (totaling 1.253 million ha) in Hubei province, Central China. We (i) evaluated yield responses to different N application levels, (ii) established indigenous soil N supply (INS) classification systems for different rice regions by identifying and using the preferable predictor of INS, and (iii) determined the optimal N application rates for each region based on regional mean optimal N rates (RMONR) in combination with INS. In all of the rice regions, rice yields were significantly higher in plots receiving N than in plots without N (termed no-N plots). The highest yields were obtained in plots receiving medium nitrogen (MN) treatment, where the average partial factor productivity (PFPN) and agronomic efficiency (AEN) were 50.4 kg grain kg−1 N and 12.5 kg grain kg−1 N, respectively. Yield responses of rice to N fertilizer were different among different rice regions because of regional variations in climatic conditions and soil fertility. A significant positive relationship between grain yield with and without N fertilizer also proved an important effect of INS on yield response to N fertilizer. On the basis of regression models, relative yields of 90%, 80%, 70%, 60%, and 50% were used as the critical values to obtain INS classifications, which were estimated by the yield of the no-N treatment but not by alkaline hydrolyzable-N. An obvious increasing trend of economic optimum N rate (EONR) with decreasing INS (from Class 1 to Class 6) was found for each rice region. Averaged across all rice regions of Hubei province, EONR was 150 kg ha−1, which was lower than the N application rate for the MN treatment of 171 kg ha−1. The results indicated that the N application rate recommended by local rice technicians could be cut by 12% (1–18% in different rice regions) without any loss of yield. In conclusion, the recommended N fertilizer application, based on RMONR in combination with INS, is feasible for regional rice production in China and other countries that have large numbers of small farmland areas and where agricultural testing equipment is absent or less modern.

► Yield responses of rice to N fertilizer were different among different rice regions. ► Yield response of rice to N fertilizer is closely associated with INS. ► The yield of the no-N plot is a better predictor of INS than alkaline hydrolyzable-N. ► The N rate recommended by local rice technicians is higher than the optimal N rate. ► The recommended N application based on RMONR in combination with INS is feasible.