Objectives To assess the effects of the cultivation models on nitrogen (N) use efficiency and grain yield of hybrid indica rice under nitrogen-reduction.

Methods  The hybrid indica rice cultivar Chengyou 981 and Yixiangyou 2115 were selected as the planting materials. The conventional high-yielding nitrogen application rate (187.5 kg/hm²) with wet irrigation and 20.0×10⁴/hm² of plant density was set as the control cultivation model or treatment (T0). Other three cultivation models/treatments in the experiment were N-reduction cultivation models: 10% reduction of N application rate with wet irrigation and 20.0×10⁴/hm² plant density (mono-nitrogen reduction, T1), 10% reduction of N with wet irrigation and 24.0×10⁴/hm² of plant density (increased density under nitrogen-reduction, T2), and 10% reduction of N with alternate wetting and moderate drying irrigation, and 24.0×10⁴/hm² of plant density (controlled irrigation and increased density under nitrogen-reduction, T3). The root growth, N absorption and utilization, and grain yield of hybrid indica rice were investigated.

Results 1) Compared with the conventional high-yielding treatment (T0), the root dry weight, root α-NA oxidation and root bleeding intensity of rice in T1 were significantly reduced; the root α -NA oxidation and root bleeding intensity decreased significantly in T2; and the root growth indices in T3 were similar to T0 at jointing and heading stages, but significantly higher than T0 at maturity stage. 2) Compared with T0, the N accumulation of rice at jointing, heading and maturity stages in T1 significantly decreased, while the N accumulation at jointing and heading stages in T2 and T3 were slightly different from T0. The N accumulation during heading and maturity stages and N allocation in panicle at maturity stage were T3>T0>T2>T1. The recovery efficiency, agronomic efficiency and partial factor production of N fertilizer were T3>T2>T1, and the indices of T3 were also higher than T0. 3) The rice yield was in the order of T3>T0>T2>T1. T1 decreased yield by 6.37% compared with T0, owning to the significant decrease in effective panicle number and spikelets per panicle. T2 had a lower yield than T0, owing to the decreased effective spikelet number and 1000-grain weight; T3 had a higher yield than T0 by 1.78%, owning to the increased seed setting rate and 1000-grain weight.

Conclusions Reduction of N fertilizer input would significantly inhibit root growth and activity, affect the N uptake and accumulation of rice, and result in significant yield reduction. Increasing densities under nitrogen-reduction could effectively increase the root biomass and N accumulation of rice, but could not offset the yield loss caused by decreased effective spikelet number and 1000-grain weight. Controlled irrigation combined with increased densities and reduced N input could significantly improve root physiological activity at the middle to late stage, N accumulation and translocation from vegetative organs to the panicle of rice, nitrogen use efficiency and yield.