Objectives To investigate how long-term N application gradients and manure addition regulate phosphorus fraction transformations in calcareous soils under rainfed winter wheat on the Loess Plateau, and how these changes affect available phosphorus (AP) supply and its coupling with grain yield formation, thereby providing a scientific basis for synergistic N–P management and optimized integrated organic–inorganic fertilization in dryland wheat systems.

Methods A long-term fertilization experiment was conducted in a rainfed winter wheat field on the Loess Plateau from autumn sowing in 2014 to harvest in 2024. A two-factor split-plot design was employed, with nitrogen (N) application rate as the main-plot factor (N0, N75, N150, N225, and N300) and manure application as the subplot factor (M0 and M1), resulting in 10 treatments with three replicates. After the 2024 harvest, soil samples (0–20 cm) were collected, and inorganic and organic phosphorus fractions were determined using the Hedley sequential fractionation method. Regression analysis and Pearson correlation heatmaps were further used to evaluate the relationships among P fractions, available phosphorus (AP), and crop yield.

Results (1) Compared with mineral fertilizer alone, the combined application of organic and inorganic fertilizers increased winter wheat grain yield by an average of 15.6%. The M1N150 treatment produced the highest yield (6,500.9 kg/hm²), which was 64.5% higher than M0N0.(2) Relative to M0, the combined application of organic and inorganic fertilizers (M1) increased labile P fractions (H₂O-P_i, NaHCO₃-P_i, and NaHCO₃-P_o), moderately labile P fractions (NaOH-P_i and NaOH-P_o), stable P fractions (HCl-P_i, and Residual-P), and Olsen-P by 73.4%, 36.7%, 6.1%, and 311.5%, respectively. The increase in the labile pool exceeded that of the stable pool, consistent with the overall rise in AP. Across different N rates, labile inorganic P and AP generally increased and then plateaued, whereas the stable pools changed only slightly. Grain yield changed mirrored the increases in AP and labile inorganic P. (3) Correlation analysis showed that available phosphorus (AP) was extremely significantly and positively correlated with labile inorganic P fractions (H₂O-P_i, NaHCO₃-P_i, and NaOH-P_i) (P < 0.001), indicating that labile inorganic P is the key pool sustaining AP supply. Data from 2015–2024 further showed a significant positive relationship between grain yield and AP. Two-way ANOVA indicated significant interaction effects between N application and manure addition on NaOH-P_i and total phosphorus (TP) (P < 0.05).

Conclusions Based on the 2014–2024 long-term field experiment, winter wheat grain yield increased with increasing N input and then plateaued; manure addition showed more pronounced yield gains and P-pool optimization under the moderate N rate, with M1N150 performing the best overall. Long-term N application combined with manure application promoted the redistribution of soil P among pools of different lability, enhanced the P-supplying potential of labile inorganic P (H₂O-P_i and NaHCO₃-P_i) and moderately labile P (NaOH-P_i and NaOH-P_o), and inhibited the fixation of P into less extractable forms. Overall, yield improvement and available phosphorus (AP) accumulation were mainly driven by labile inorganic P, identifying it as the key P pool supporting crop production.