Objectives  To provide scientific basis for phosphorus fertilizer reduction in the middle and lower reaches of the Yangtze River Basin, we studied the effect of phosphorus fertilizer reduction on wheat yield and soil quality.

Methods  The field experiments were conducted in Hubei and Zhejiang Provinces. The treatments included: non-phosphorus application control (CK), farmers’ conventional phosphorus application (FP), 80% of farmers’ conventional phosphorus application (P80), 60% of farmers’ conventional phosphorus application (P60), and P60+phosphate-solubilizing bacteria application (PB60). Wheat yield data were collected from 2019 through 2020. Following wheat harvest in 2020, we assessed the physical and chemical properties, microbial biomass phosphorus content, enzyme activity and phospholipid fatty acid content within 0—20 cm horizon of the cultivated soils.

Results  In Hubei wheat region, the wheat yields under FP, P80 and P60 treatments were similar in the first year of phosphorus fertilizer reduction. However, wheat yields under P80 and P60 treatments were less than that under FP in the second year of phosphorus fertilizer reduction. In Zhejiang province, the wheat yields under P80 and FP treatments were similar in two years, while PB60 increased the wheat yield compared with P60 in 2020. In Hubei Province, available P contents under CK, P80 and P60 treatments were less than that under FP, whereas P80 and P60 treatments produced higher soil microbial biomass P (MBP) content than FP. In Zhejiang Province, P80 and FP treatments had similar soil available P content, while under P60 and CK treatments soil available P content was significantly lower than that under FP treatment. Compared with FP treatment, P80 and P60 treatments decreased soil phosphorus activation coefficient and ratio of bacteria to fungi, but they increased soil NH₄⁺-N content, NO₃^–-N content, soil N-cycling enzyme activities and relative fungi abundance in the two provinces. FP, P80 and P60 treatments had similar soil pH value, organic carbon content, total N content, readily available K content, soil phosphatase activity, relative abundance of bacteria and relative abundance of arbuscular mycorrhizal fungi in the two provinces. Compared to P60, PB60 treatment elevated soil available P content, soil phosphorus activation coefficient, and relative abundance of arbuscular mycorrhizal fungi. Redundancy analysis indicated that soil microbial biomass P and soil NH₄⁺-N were the dominant factors affecting soil microbial community structure change in Hubei and Zhejiang provinces, respectively.

Conclusions  We found that soil available P, microbial biomass P, and wheat yield had different responses to phosphorus fertilizer reduction in the two wheat fields. Continuous reduction of phosphate fertilizer could result in decrease in wheat yield in Hubei Province due to low soil available P content; thus, it is not recommended to reduce phosphorus fertilizer rate in Hubei. However, in Zhejiang Province where soil available P content is high and P reduction doesn’t decrease the wheat yield and soil quality, 80% of farmers’ conventional phosphorus application is recommended. In addition, phosphate fertilizer reduction does not significantly affect soil phosphatase activity, but significantly affects soil microbial community structure, increases the relative abundance of soil fungi and decreases the bacterial/fungal ratio. The application effect of phosphate-solubilizing bacteria in Zhejiang wheat area is better than that in Hubei wheat area.