Synchronizing nitrogen supply with crop demand through postponed and reduced topdressing improves nitrogen use efficiency and maintains yield in wheat

Background Nitrogen (N) management is essential for wheat productivity; however, the traditional practice of excessive N fertilization frequently leads to low N use efficiency and increased environmental risk. Although numerous studies have shown that postponed or split N topdressing can enhance wheat production, it remains unclear whether it is possible to maintain grain yield and increase N use efficiency by moderately reducing N fertilizer input when coupled with postponed N topdressing. Results A two-year field experiment was carried out during the 2022 and 2023 growing seasons. Seven N fertilizer treatments were established: a conventional N treatment (NS37: basal + topdressing at 3‑leaf stage and 7‑leaf stage), basal N only (NS), zero‑N control (N 0 ), a postponed N topdressing treatment (NS58: basal+ topdressing at the 5‑leaf stage and 8‑leaf stage), and three postponed‑reduced N topdressing treatments (NS5 R 8: basal + 50% N at the 5‑leaf stage + full N at the 8‑leaf stage; NS R 5 R 8: 50% basal N + 50% N at the 5‑leaf stage + full N at the 8‑leaf stage; NS 0 5 R 8: no basal N + 50% N at the 5‑leaf stage + full N at the 8‑leaf stage). Grain yield, tiller traits, spike characteristics and grain quality were measured. N 0 and NS treatments significantly reduced grain yield, total tiller number, effective spike number and grain quality parameters. In contrast, the postponed N topdressing (NS58) and three postponed-reduced N topdressing treatments (NS5 R 8, NS R 5 R 8, and NS 0 5 R 8) maintained yield, had a higher proportion of effective spikes, lower number and proportion of degenerated spikelet degeneration compared with NS37. Among these, NS5 R 8 (50% reduction at the 5-leaf stage; total N reduced by 20%) achieved the highest yield, slightly higher than that of NS37 in both growing seasons. The yield stability was associated with three key physiological improvements: (1) Tiller quality was improved, as reflected by an increased proportion of effective spikes. In 2023, although NS5 R 8 significantly reduced total tiller number from 1060 × 10 4 ha − 1 in NS37 to 780 × 10 4 ha − 1 , effective spike number were maintained at a level comparable to NS37 (426 × 10 4 ha − 1 in NS37 vs. 396 × 10 4 ha − 1 in NS5 R 8). Consequently, the proportion of effective spikes increased from 40% in NS37 to 51% in NS5 R 8, indicating improved tiller quality rather than quantity. (2) Spike development was optimized through reduced spikelet degeneration. In 2022, compared with NS37, NS5 R 8 reduced degenerated spikelet number per spike from 3.6 in NS37 to 3.1 without compromising total spikelet number per spike (22.9 in NS37 vs. 23.0 in NS5 R 8), and increased grain number per spike (40.3 in NS37 vs. 46.8 in NS5 R 8). (3) Enhanced grain filling contributed to higher thousand-grain weight. In 2022, thousand-grain weight under NS5 R 8 was 41.4 g, higher than that under NS37 (38.2 g). Therefore, the agronomic nitrogen use efficiency (ANUE) was significantly improved under NS5 R 8. In addition, NS58 had no effect on the grain quality parameters, whereas N 0 and NS significantly reduced all the measured quality indicators, the postponed‑reduced treatments (NS5 R 8, NS R 5 R 8, and NS 0 5 R 8) did not substantially affect flour yield or bulk density, but reduced stability time, formation time, settlement value, grain protein content, wet gluten content, and hardness value. Conclusions Overall, within precision agriculture systems, the proposed approach of combined postponed with reduced N topdressing stabilized wheat yield while enhancing nitrogen use efficiency by improving proportion of effective spikes in tiller, reducing spikelet degeneration, and increasing thousand-grain weight.