Objective To explore the effects of different phosphorus fertilizers on phosphorus fractions and their availability in calcareous soil, and to provide a theoretical and practical basis for selecting phosphorus fertilizer varieties suitable for calcareous soil, improving crop yield, phosphorus fertilizer utilization efficiency and reducing soil phosphorus accumulation.

Methods A field experiment of winter wheat−summer maize rotation was carried out on calcareous Lou soil in Guanzhong, Shaanxi Province, October 2018. Six phosphate fertilizer treatments were set up: no phosphate fertilizer control (CK), superphosphate (SSP), calcium magnesium phosphate (CMP), monoammonium phosphate (MAP), diammonium phosphate (DAP) and ammonium polyphosphate (APP). The crop yield and phosphate fertilizer use efficiency were investigated. The content of soil phosphorus components was determined by the Tiessen−Moir phosphorus classification method, and the relationship between phosphorus components and crop yield and soil available phosphorus content was analyzed by the structural equation model.

Results Compared with CK, the yields of winter wheat, summer maize and winter wheat−summer maize crop systems treated with five phosphate fertilizers significantly increased by 56.8%−100.0%, 30.8%−50.8% and 39.6%−66.0%, respectively. The crop yield under the APP treatment was the highest. Furthermore, SSP, MAP and APP significantly increased the utilization rate of phosphate fertilizer as compared to CMP and DAP. Compared with CK, all phosphorus treatments significantly increased soil available phosphorus ( Olsen−P ), but the dynamics were changed with time. The soil available phosphorus is: APP>SSP>MAP>CMP>DAP>CK. Moreover, soil Olsen−P content was significantly positively correlated with crop yield and phosphorus utilization rate. The application of different phosphorus sources significantly affected the content of labile and moderately labile phosphorus in soil, but had no significant effect on occluded phosphorus. Compared with CK, the NaHCO₃−Po of phosphorus treatment was significantly reduced. The contents of Resin−P, NaOH−Pi and NaOH−Po in SSP and APP treatments were significantly higher than those in other treatments. The soil NaHCO₃−Pi content of SSP was the highest, followed by MAP and APP. The content of D.HCl−Pi in MAP treatment was the highest. Compared with CK, SSP and APP treatments significantly increased the proportion of soil labile phosphorus (LP) and decreased the proportion of moderately labile phosphorus (M−LP). SSP and MAP treatments significantly reduced the proportion of soil occluded phosphorus (OP). The contents of Resin−P, NaHCO₃−Pi, NaOH−Pi, D.HCl−Pi and LP were significantly positively correlated with soil Olsen−P, and Resin−P, NaHCO₃−Pi, D.HCl−Pi and M−LP were significantly positively correlated with crop yield. In the structural equation model, soil LP, M−LP and OP had direct or indirect positive effects on soil Olsen−P and crop yield through mutual transformation. The path coefficients of OP and M−LP to LP were 0.329 and 0.548, respectively. The total path coefficients of LP, M−LP and OP to crop yield were 0.829, 0.846 and 0.571, respectively.

Conclusion Phosphorus application increased crop yield and soil phosphorus availability mainly through the labile and moderately labile phosphorus content. The application of SSP, MAP and APP in calcareous soil increased the yield and efficiency better than CMP and DAP, because the application of MAP significantly increased the content of moderately labile phosphorus (D.HCl−Pi), whereas the application of SSP and APP significantly increased labile phosphorus (Resin−P and NaHCO₃−Pi) and moderately labile phosphorus (NaOH−Pi and NaOH−Po). Therefore, SSP, MAP and APP are recommended for calcareous soil, and APP is better.