Objectives This study investigated the effects of phosphorus (P) application rates on the annual grain yield of wheat-maize rotation systems and soil phosphorus budget, and the relationships among phosphorus budget, soil total phosphorus (TP), available phosphorus (Olsen-P), phosphorus activation coefficient (PAC), phosphorus fractions, and phosphorus pools. The findings will provide a theoretical support for improving phosphorus fertilizer efficiency in calcareous cinnamon soils.

Methods A long-term field experiment was initiated since 2008 (the no-P treatment started since 2015) under a winter wheat-summer maize rotation system in Shijiazhuang, Hebei Province, China. Five P₂O₅ application rates were setup: 0, 75, 120, 165, and 210 kg/hm², denoted as P0, P75, P120, P165, and P210, respectively. Annual crop above-ground yields and phosphorus uptake were assessed to calculate soil phosphorus budget from 2016 to 2022. Soil samples (0−20 cm depth) were collected after maize harvest in 2016, 2019, and 2022 for analysis of soil TP, Olsen-P, PAC, phosphorus fractions (using the Hedley sequential extraction method), and phosphorus pools.

Results Compared to P75, the P120, P165, and P210 treatments significantly increased annual grain yields by 9.54%, 11.81%, and 12.46%, and grain P uptake by 16.64%, 30.09%, and 39.58%, respectively. A linear-plus-plateau model analysis indicated the optimal annual P₂O₅ application rate of 112 kg/hm² for stable yields. The average annual P surpluses for P0, P75, P120, P165, and P210 treatment were −18.54, −0.84, 13.22, 27.93, and 43.91 kg/hm², respectively, with significant differences among treatments. A model fitting analysis indicated that soil phosphorus began to accumulate when the P₂O₅ application rate exceeded 76.64 kg/hm². Cumulative P surplus increased over the experimental period. For every 100 kg/hm² of soil P surplus, soil TP, Olsen-P and PAC increased by 29.70 mg/kg, 3.88 mg/kg and 0.31%, respectively; the contents of HCl-Pi, NaHCO₃-Pi, NaOH-Pi, H₂O-Pi, NaOH-Po, NaHCO₃-Po increased by 20.90, 4.09, 1.53, 0.57, 0.53 and 0.44 mg/kg, respectively; Labile P, moderately labile P and stable P increased by 5.37, 2.58 and 21.61 mg/kg, respectively; inorganic and organic P increased by 26.96 and 1.03 mg/kg, respectively. Soil phosphorus surplus accumulated primarily in the surface layer, accounting for 77.22% of the total-P.

Conclusions In a wheat-maize rotation system on calcareous cinnamon soil, soil P began to accumulate when the P₂O₅ application rate exceeded 76.64 kg/hm². Subsequently, the contents of total soil P, available P, P activation coefficient, and various forms of P (excluding Residual-P) increased, with HCl-Pi (a form of inorganic phosphorus) showing the most pronounced rise. The response of soil phosphorus pools to P surplus demonstrated that stable phosphorus > active phosphorus > moderately active phosphorus, and inorganic phosphorus > organic phosphorus. Considering both yield effects and environmental risks, the recommended reasonable range for P₂O₅ application is 112–135 kg/hm², which not only ensures stable annual crop yields but also keeps P surplus within environmentally safe limits.