Objectives  We studied the appropriate P fertilizer application rate, the critical Olsen-P value for crop yield and the response of Olsen-P to soil P balance under long-term P application in sandy fluvo-aquic soil. We aimed to provide a basis for P nutrient management in agricultural fields.

  Methods  A long-term experiment was conducted since 2008 in Langfang, Hebei Province. The planting system is winter wheat-summer maize rotation on sandy fluvo-aquic soil, with phosphorus fertilizer (P₂O₅) levels of 0, 45, 90, 135, 180, and 225 kg/(hm²·a), and expressed as P0, P45, P90, P135, P180, P225. Crop yield and P uptake were analyzed every year, and soil Olsen-P was measured after harvesting winter wheat and summer maize.

  Results  P application increased yield and P uptake in winter wheat and summer maize, but they did not always increase with the increased P application level. For the highest annual yield (14627 kg/m²), the application rate of 152 kg/hm² was required. The appropriate application rate was 90 kg/hm² if the actual production target was 90% of the highest yield. Over time, the efficiency of P fertilizer use in sandy fluvo-aquic soil increased. For the 12-year average, P fertilizer’s apparent and cumulative use efficiency was 36.98%–98.10% and 26.26%–71.85%, respectively. P application did not affect crop P absorption when the annual application rate exceeded 90 kg/hm², resulting in lower apparent and cumulative P fertilizer use efficiencies. The apparent P balance for treatments P0 to P225 was –11.30, –7.38, 0.94, 20.05, 37.21, and 57.68 kg/hm², respectively, over the 12-year period. –144.92, –88.57, 11.33, 240.56, 446.48, and 692.15 kg/hm² were the cumulative P balances from 2010 to 2020, respectively. The linear relationship between Olsen-P and P accumulation was shown in two sections, with a turning point of 218.81 kg/hm² for P accumulation. Every 100 mg/hm² of P cumulative increment resulted in 0.48 mg/kg increase in Olsen-P, which increased to 3.37 mg/kg after the point. The Olsen-P agronomic thresholds for winter wheat and summer maize were 10.20 mg/kg and 5.93 mg/kg, respectively. When the P₂O₅ application rate was 90 kg/hm², the soil available P content in both seasons was close to the agronomic threshold.

  Conclusions  The appropriate annual P₂O₅ application rate in sandy fluvo-aquic soil under a winter wheat-summer maize rotation system with the return of crop stalks to the field is 90 kg/hm². This results in the soil available content required for the agronomic threshold to attain a high crop yield. The storage P threshold of sandy fluvo-aquic soil was 218.81 kg/hm². When the cumulative P amount is less than the threshold, the available soil P will be increased less by P application. On the other hand, when the cumulative P amount exceeds the threshold, the soil available P will greatly increase.