This experiment was conducted at the Water Conservation and Irrigation Laboratory at Shihezi University, Shihezi City, Xinjiang, China (44°20′ N, 88°30′ E). Arid climatic conditions were typical for the region, with cold winters (average temperature of -12.08 °C and minimum temperature of -30.90 °C) and hot summers (average temperature of 23.88°C and maximum temperature of 39.30 °C), annual rainfall of 110 - 200 mm, annual evaporation rate of 1000 - 1500 mm, and a 30 years average annual temperature of 7.9 °C. Rainfall and air temperature during the study period are summarized in Figure 1.

The soil type of the experiment site was grey desert soil, with basic physical and chemical properties, in the cultivated layer (0 - 20 cm), as follows: total N 1.18 g kg⁻¹, total P 0.53 g kg⁻¹, available P 19.30 mg·kg⁻¹, available K 119.80 mg·kg^-1, bulk density 1.49 g·cm^-3, organic matter 21.56 g·kg^-1, and soil pH 8.17.

The experiment was conducted as a two-factor randomized block design with four N application gradients: 0 (N₀), 60 (N₁), 120 (N₂), and 180 (N₃) kg·ha^-1; and two P (P₂O₅) application gradients: 0 (P₀), and 100 (P₁) kg·ha^-1, for a total of eight nitrogen and phosphorus treatments: N₀P₀, N₁P₀, N₂P₀, N₃P₀, N₀P₁, N₁P₁, N₂P₁, N₃P₁, with three replications per treatment.

Grass plots 24 m² (4 m × 6 m) were established in April 2019, with a 1 m wide isolation strip between the plots to prevent water and nutrient movement between the plots. Plots were deep-tilled and cleared of debris prior to planting. Seed of alfalfa WL366HQ, an adaptable, high yielding and quality variety, was manually strip-seeded at a rate of 18 kg·ha^-1, with planting rows spaced 20 cm apart, at a depth of 2 cm. Inlaid drip irrigation tapes were buried 10 -15 cm below the soil surface and spaced 60 cm apart, with drip heads spaced 20 cm apart.

Urea (46% N) and monoammonium P (52% P, 12.2% N) were used as the N and P source, respectively. In order to ensure that the test was only affected by the P fertilizer, the effect of N in the monoammonium P was counteracted by applying urea. Fertilizers were added to the fertigation tanks, and was drip applied with water 3–5 d after cutting (first flowering stage) on 9 May, 30 May, 6 July, and 12 August 2020.

After each crop of alfalfa was cut, a five-point (“S” pattern) soil sampling method was used, l samples (0–20 cm) were taken using a soil auger (5 cm in diameter), composited, mixed thoroughly, and then passed through a coarse sieve (2 mm) to remove roots and debris. Soil samples were processed using the quadratic method and duplicate samples were retained; with one sample air-dried for the determination of soil physicochemical properties and P fraction content, and the remaining sample was stored in a refrigerator (-20°C) for reference.

Total N content was determined by Kjeldahl method (Lu, 2000). Total P content was determined by sulfuric acid-perchloric acid decoction method (Lu, 2000). Available P content was extracted with NaHCO₃ and determined using the molybdenum antimony colorimetric method (Lu, 2000). Soil pH was measured using a portable pH meter, with a soil-to-water ratio of 2.5:1. (LEICI, INC, Shanghai, China).

Inorganic P fractions were determined by the Sequential fractionation method (SFRM) (Gu and Jiang, 1990), which extracts inorganic P in the order of Ca₂-P, Ca₈-P, Al-P, Fe-P, O-P, and Ca₁₀-P. Organic P fractions were determined by the Bowman-Cole method (Bowman and Cole, 1978), which categorizes organic P into LOP, MLOP, MROP and HROP. The extraction reagents were shown in Table 1.

Data was compiled using Microsoft Excel 2016 (Microsoft Corp., Redmond, WA, USA), SPSS 22.0 (SPSS Inc., Chicago, IL, USA) was used for ANOVA, and Duncan’s method was used for significance of difference analysis (α = 0.05). Redundancy analysis and Pearson correlation analysis of date were conducted using Canoco 5.0 (Microcomputer Power, Ithaca, NY, USA) and Origin 2019b (OriginLab Corp., Northampton, MA, USA), respectively.

Both N and P single fertilization had significant (p < 0.05) effects on alfalfa total soil N, total P, available P content, and pH. Combined N and P had significant (p < 0.05) effects (on soil total P and available P content (except for the first crop available P content) (Table 2).

In general, under the same N treatment, with increasing P fertilizer application, soil total N, total P, and available P content increased. P₁ was increased total N, total P and available P content by 3.25 - 18.55, 8.92 - 74.79, 34.67 - 192.71 and 0.24 - 1.34%, respectively, when compared to P₀. In general, under the same P fertilizer treatments, with increased N application, total soil N increased gradually (except for the 1st crop of P₁, which increased initially and then decreased) and reached a maximum value under the N₃ treatment, and was significantly (p < 0.05) higher than the rest of the N application treatments. Total soil P and available soil P content and soil pH showed a gradual decreasing trend (except for the 3rd crop of P₂, which had an increasing - decreasing - raising trend). Nitrogen application under P₀ conditions elevated soil total N by 17.22 - 89.63%, and decreased soil total P, available P content, and pH by 0.28 - 18.45, 2.24 - 28.51 and 0.73 - 2.98%, respectively. Nitrogen application under P₁ conditions elevated soil total N content by 17.71 - 76.51%, and soil total P, available P content, while pH decreased by 0.34 - 8.58, 4.76 - 19.38 and 0.65 - 3.29%, respectively.

Nitrogen and P single fertilization and N-P combined had highly significant (p < 0.01) effects on soil total inorganic P and total organic P content (Figures 2A, C). Under the same N application treatment, soil total inorganic P content showed an upward trend with increasing P application. P₁ increased by 11.82 - 86.50%, when compared with P₀ (Figure 2A). Total organic P content decreased, and P₁ decreased by 4.77 - 24.97%, respectively, when compared with P₀ (Figure 2A). Under the same P treatment, increasing N fertility resulted in decreased total soil inorganic P and total organic P contents of alfalfa showed a gradual decreasing trend and reached the lowest value under N₃ treatment (Figures 2B, D). Nitrogen application treatments were all significantly (p < 0.05) smaller than the no N application treatments (except for the total organic P content at the level of P₁ in the 2nd crop). Total inorganic P and total organic P contents of alfalfa were reduced by 4.66 - 19.51 and 6.90 - 17.58%, respectively in N application treatments when compared with no N application treatments under P₀ conditions, and 1.11 - 9.05 and 6.64 - 28.20% under P₁ conditions (p < 0.05) (Figures 2A, C).

Nitrogen application, P application, and N-P rationing all had significant (p < 0.05) effects on the content of inorganic P fractions in soils under alfalfa production (Figure 3). Under the same N application conditions, with increased P application, the Ca₂-P, Ca₈-P, Al-P, Fe-P, O-P, Ca₁₀-P contents of soil under alfalfa production showed an upward trend. Phosphorus application increased when compared with no applied P by 27.06 - 114.89, 17.92 - 132.90, 16.76 - 146.09, 4.44 - 62.61, 4.77 - 94.95, and 5.3 - 51.23%, respectively, and the upward trend increased with the increase of cutting stubble. Under P₀ conditions, the Ca₂-P, Ca₈-P, Al-P, Fe-P, O-P, and Ca₁₀-P contents of soils showed a gradual decreasing trend (except for the first crop where Fe-P showed a decreasing and then increasing trend, and the first and third crops where O-P showed an increasing and then decreasing trend) with the increase in N application, and reached the minimum value under N₃ treatment. Nitrogen application reduced 13.73 - 33.01, 6.67 - 18.07, 2.98 - 28.02, 2.80 - 22.72, 2.96 - 20.42, and 5.54 - 20.11%, respectively, when compared to no application. Under P₁ conditions, soil Ca₂-P and Fe-P showed a gradual decreasing trend with increased N application and reached the minimum value under N₃ treatment, and N application reduced 14.42 - 26.79 and 2.07 - 17.64%, respectively, when compared with no applied N treatment. Soil Ca₈-P, Al-P, and Ca₁₀-P in the N treatment was significantly (p < 0.05) smaller than that of the unapplied N treatment [except for the third and Ca₈-P in the 4th crop was significantly (p < 0.05) larger than that in the unapplied N treatment; the Ca₁₀-P in the 2nd crop under N₁ treatment and Al-P under N₂ treatment were not significant; the O-P showed a trend of increasing and then decreasing except for the 2nd crop, which showed a trend of decreasing and then increasing], and reached the maximum value under N₂ treatment.

The effects of N application, P application, and N-P combined the content of organic P fractions in soils under alfalfa production were highly significant (p < 0.01) (Figure 4). Under the same N application condition, with the increase of P application, LOP, MLOP, MROP, and HROP in soil showed a significant (p < 0.05) decreasing trend, and P application reduced 7.27 - 13.27, 5.13 - 16.35, 4.49 - 24.90 and 4.77 - 24.97% when compared with the no P application, respectively. Under P₀ conditions, with increased N application, the contents of LOP, MLOP and MROP in alfalfa soil showed a decreasing trend (except for the 1st crop where the MLOP content showed a decreasing and then increasing trend), and the N application treatments were significantly (p < 0.05) reduced by 5.44 - 14.99, 6.15 - 16.37 and 10.32 - 34.46%, respectively, when compared with that of the unapplied N treatment. The HROP content of soil under alfalfa production, except for the 3rd crop which showed a gradual decreasing trend and then increasing trend (reaching the minimum value under N₂ conditions), and its content of N application treatments was significantly (p < 0.05), smaller than that of the unapplied treatments, by 6.90 - 17.58%. The LOP, MLOP, and HROP contents of soils under alfalfa production showed a significant (p < 0.05) decreasing trend under P₁ conditions (LOP contents did not differ significantly between N₂ and N₃ treatments), and the N application treatments were reduced by 6.27 - 14.93, 3.56 - 21.74 and 6.64 - 28.20%, respectively, when compared with no applied N treatments (p < 0.05). The MROP content of alfalfa soil showed a gradual decrease with the increase of N application in the 1st and 2nd crops. In that, it showed a trend of decreasing and then increasing in the 3rd and 4th crops (reaching the minimum under N₂ treatment), but the differences between N₂ and N₃ treatments were not significant.

All N and P treatments were dominated by Ca₁₀-P for inorganic P and MROP for organic P (about 60% of the total) (Figure 5). Under the same N application conditions, P application increased the Ca₂-P, Ca₈-P, Al-P, and LOP proportions, decreased the Fe-P, MROP, and HROP proportions, and had a nonsignificant effect on the O-P and MLOP proportions (change of less than 5%), when compared with the no P application treatment. Under P₀ conditions, the Ca₂-P and MROP ratios tended to decrease with increasing N application and reached a minimum under N₃ treatment, while the rest of the P form ratios did not change significantly. Under P₁ conditions, the percentages of Ca₂-P, Fe-P, MROP, and HROP showed a decreasing trend with increasing N application, all of which reached the minimum value under N₃ conditions, the percentages of Ca₈-P, O-P, and LOP showed an increasing trend, and the percentages of Al-P, Ca₁₀-P, and MLOP showed insignificant changes.

Soil available P, total N, and pH were the main drivers affecting soil P fractions, explaining 67.3%, 25.3%, and 6.5%, of their variation, respectively (Figure 6). For the content of P fractions in soil (Figure 7), soil inorganic P fractions were significantly (p < 0.05) positively correlated with each other, and all of them were significantly (p < 0.05) positively correlated with soil total P, total inorganic P, and available P. Soil organic P fractions were significantly (p < 0.05) positively correlated with each other, all of them were highly positively correlated with soil total organic P (p < 0.05), and significantly (p < 0.05) negatively correlated with soil total N. pH and Total N were significantly (p < 0.05) negatively correlated with each other.

Soils are the main source of nutrients required for crop growth and N and P content, and thus their physicochemical properties are important for crop nutrient utilization (Seleiman et al., 2017, 2020a). Previous study have demonstrated that the application of N and P fertilizers effectively increases the levels of N and P in the soil during the growing season for alfalfa (Liu et al., 2022). In our study, the application of N and P fertilizers increased the total N content, while the application of P fertilizers significantly increased the total and available P content. At the same level of P fertilizer, total P and available P content of soil under alfalfa production showed a gradual decrease with increased N fertilization (Table 2). This increase indicates that N and P fertilizers are stored in the soil in a slow-acting form, while P fertilizer application promotes the development of alfalfa rhizobacteria (Qu et al., 2020), which in turn increases the amount of immobilized N in the soil. Nitrogen sufficiency increased alfalfa growth and root development, which in turn increasing soil P demand and uptake (Song et al., 2022). Increased N application, resulted in a smaller decrease in total and available P content under P fertilizer treatment when compared to no P application. This indicates that the regulation of N fertilizer on P uptake capacity of alfalfa may be related to the content of P in the soil. When soil P is deficient, the effect of N fertilizer in promoting P uptake by alfalfa is more pronounced (Tong et al., 2023).

Soil pH is directly related to the utilization of nutrients by the crop, and either too acidic or alkaline can result in the inability of the crop root system to absorb nutrients (Ringrose and Neilsen, 2005). Previous studies have shown that N application decreases soil pH, while P application has no significant effect on soil pH (An et al., 2022). Our study indicates that both N and P application have different abilities in lowering soil pH (Table 2). This could be attributed to the increased activity of rhizosphere microorganisms in alfalfa roots stimulated by N, resulting in more secretion from alfalfa roots and subsequently increasing soil acidity (An et al., 2022), and/or P fertilizer may lower pH by releasing H⁺ ions and promoting the binding of P ions with metal ions (Liu et al., 2022).

Inorganic P is the main source of P in the soil, and changes in its content can be used as an important assessment of P uptake by plants (Urlić et al., 2023). It has been shown that N application promotes the utilization of soil P by the plant and that an appropriate ratio of N to P helps to increase soil P content (Sun et al., 2022). In our study, soil TIP content showed a decreasing trend with increasing N application (Figure 2A). This trend may reflect enhanced photosynthesis and accelerated root growth in alfalfa from N fertilizer application, which in turn reduced the P content of the soil (Lu et al., 2019).

Soil reactive P Ca₂-P and Ca₈-P content was reported to decrease with increasing N application, while the proportion of Ca₂-P to inorganic P increased (Cui et al., 2019). In our study, when N fertilizer was applied alone, the content of each soil inorganic P fraction tended to decrease with increased N fertilizer application, while the proportion of Ca₂-P to inorganic P decreased. However, the proportion of Ca₈-P to inorganic P tended to increase with increasing N application under combined N and P conditions (Figure 5A). This indicates that the addition of N fertilizer not only promotes plant uptake of reactive P, but also facilitates the conversion and degradation of insoluble P to other P forms. Combined N and P application, optimizes soil quality conditions, provides a suitable environment for microbial growth, increases phosphatase activity, and ultimately facilitates the conversion of soil insoluble P to moderately stabilized forms of P (Cui et al., 2019; Wu and Yu, 2023). Differences between Ca₁₀-P and O-P contents of P application and no P application treatment were widened with increasing cutting stubbles and were higher in the P application treatment than in the no P application treatment (Figures 3C, F). This indicates that P fertilizer accumulates residually in the soil, while insoluble P remains the dominant form of P storage and the main direction of P transformation in the soil (Urlić et al., 2023).

Although soil organic P is one of the main fractions of the soil P pool, it is difficult for plants to utilized directly, leading to a serious underutilization of P resources. Therefore, fraction breakdown of organic P and research on methods to improve its utilization efficiency are key steps towards efficient utilization of soil P reservoirs (Sulieman and Mühling, 2021). It has been demonstrated that the application of organic acids promotes the mineralization of insoluble organic P and reduces the content of stable organic P fractions (Wang et al., 2023). The results of our study demonstrate that both N and P fertilizer application could reduce the proportion of MROP to organic P in the soil (Figure 5B). This indicates that the effect of N and P fertilizers on organic P in soil in alfalfa production is mainly in the mineralization and degradation of medium stable organic P. Application of P fertilizer alone and N-P combined increase organic P in LOP in addition to decreasing the percentage of stable organic P. This may be related to the interaction between N and P fertilizers. Phosphorus fertilizer decreases soil pH, providing a suitable acid-base environment for organic P mineralizing microorganisms and enzymes (Timofeeva et al., 2022). Nitrogen fertilizer promotes the growth of alfalfa roots and enhances the absorption of P by rhizobium, leading to increased production and activity of alkaline phosphatase and other mineralization factors. These factors, in turn, promote the activation of stable organic P in the soil, converting it into forms of P that can be readily absorbed by plants (Koester et al., 2021).

The physicochemical properties of the soil have an important effect on the transformation of P fractions (Zhang et al., 2021; Hei et al., 2023), and combined N and P fertilizers increases N content, lowers soil pH, and replenishes the soil’s available P content. Previous study indicates that soil organic matter content and total N content are important factors influencing the chemical forms of P in soil (Hei et al., 2023). The results of this study demonstrate that available P content, total N content, and pH were the main factors driving the conversion of P fractions in soils under alfalfa production (Figure 6). This could be due to the close correlation between the level of available P and alfalfa root growth, which in turn affects the regulation of the plant’s conversion of soil P components. Soil pH is an important environmental condition for maintaining the vitality of root microbiota and insoluble P mineralization enzymes, thus influencing the transformation of soil P (Cui et al., 2019).

Nitrogen in the soil is an essential element for plant development and microbial activity. In addition to enhancing microbial vitality, it also reduces the adsorption capacity of minerals, including Al and Fe on P, thereby decreasing the proportion of insoluble P in the soil and increasing the effectiveness of soil P for plants (Zhang et al., 2022).