Optimizing Phosphorus Application for Winter Wheat Production in the Coastal Saline Area

Abstract

Phosphorous (P) fixation in saline soils is a concern worldwide. To investigate optimization strategies for P fertilizer application that improve P use efficiency (PUE) and crop yield in saline soil, in terms of P sources and rates, we conducted a two year field experiment in the coastal saline area of China to investigate the effects of P rates and sources, including superphosphate (SSP), monoammonium phosphate (MAP) and ammonium polyphosphate (APP) on yield, aboveground P uptake, agronomy efficiency (AE_P), and soil available P of winter wheat (Triticum aestivum L.). Wheat yield, under the three P sources, increased with P rates and reached a plateau under 20 kg P ha⁻¹ SSP,18 kg P ha⁻¹ MAP, and 17 kg P ha⁻¹ MAP, respectively. The application of SSP increased the wheat yield by 9–11% compared to MAP and APP. The aboveground P uptake of winter wheat under SSP was 14% and 13% higher than MAP and APP, respectively, under the optimal P application rate. The AE_P under SSP was higher than the other two P sources under the same P rate. SSP increased the soil Ca²⁺ concentration by 20–42%, but decreased the Na⁺ concentration by 14–18% at the P rate of 26 kg P ha⁻¹ in all soil layers, including 0–20, 20–40, 40–60 cm, compared to CK (0 kg P ha⁻¹). The soil Olsen-P concentration under APP was higher than the other P sources. This study suggests that optimizing P sources and rates can improve wheat yield, PUE, and soil quality in the coastal saline soil.

1. Introduction

Soil salinization is a global problem and recognized as one of the main threats to environmental resources as over 950 mha of land in over 100 countries are salt affected, covering approximately 10% of the total land surface of the globe [1,2]. Saline soil formed by the calcareous parent material is rich in basic salts (Ca²⁺, K⁺, Mg²⁺ and Na⁺) and soluble CO₃²⁻ and HCO₃⁻, which produce a lot of OH⁻ and lead to a higher soil pH. Phosphorous (P) is sparingly mobile in soil and soil pH is considered as one of the major soil properties to govern the P uptake efficiency of plants [3,4]. On one hand, P can only be effectively utilized by plants in inorganic form (HPO₄²⁻ or H₂PO₄⁻ and H₃PO₄); however, due to the precipitation of orthophosphate ions in calcareous soil with a high pH and its adsorption with calcium carbonate, plants cannot utilize P effectively [5,6]. On the other hand, the high pH and electrical conductivity (EC) in saline soils reduces the activities of soil microbial biomass and alkaline phosphatase [7,8], which further affects the transformation of P in soil and reduces the chemical and biological availability of P. It is estimated that if the fixed P in global soils becomes plant available, it will be enough to sustain crop productivity for 100 years [6]. Therefore, improving soil P levels and applied P use efficiency (PUE) through soil P management is of great significance to ensure food security in salinized areas.

Many farmers have applied large amounts of P in excess of crop requirements to increase soil P fertility and crop yield due to the low soil Olsen P content [9]. However, only 20% of P fertilizer is used by crops during the first growing season in China [10], which suggests that a large amount of applied P fertilizer was accumulated as legacy P in the soil or was lost to the environment. The overuse of P by farmers caused low PUE and environmental risks [10,11]. Several strategies for P management have been proposed, including the widespread adoption of reduced tillage systems, the introduction of new crops and high-yielding cultivars, the intensification and extension of crop rotations to increase PUE and decrease environmental risks caused by excess P [12,13,14,15]. However, optimizing the P fertilizer application rate by matching P supply to crop demand can provide maximum P efficiency and achieve long-term yield sustainability [16]. The P application rate was based on the location field experiment to maintain the soil available P within the critical level of yield response to the P application rate [17]. The critical level for crop yield depends on the characteristics of the different plant species and cropping systems. The critical level for maize (Zea mays L.) ranged between 12.1 and 17.3 mg P kg⁻¹, and for winter wheat it ranged between 12.5 and 19 mg P kg⁻¹, based on three 15 years long-term fertilizer experiments [18]. This approach helps maintain soil P at an optimal level, as well as meeting the needs of crops, increasing profitability and achieving maximum economic efficiency, and reducing the environmental risks of P leaching [4,19,20].

Previous studies have focused on optimizing the P application rate to achieve high crop yield and PUE; however, the effects of differences in physical and chemical reaction processes after the application of different P types on soil available P and yield are largely neglected. The predominant commercial P products produced from rock phosphate, such as monoammonium phosphate (MAP), diammonium phosphate (DAP) and ammonium polyphosphate (APP), have rapidly increased and have been widely used in agricultural production in recent years. Although the P in these products is water soluble, there are significant differences in the solubility and mobility of P due to the mobility of accompanying ions [21]. Moreover, the transformation process and PUE of different P fertilizers vary in different soil conditions [22]. Therefore, it is essential to improve our understanding of the potential effects of P fertilizer type on soil P availability and grain yield.

In the study, we investigated the effects of P fertilizer types and rate on PUE and yield in saline soil. The changes of yield, aboveground P uptake, AE_P and soil available P levels with different application rates of different P fertilizer types was analyzed. We hypothesized that: (1) the aboveground P uptake and yield of wheat varied in response to different P fertilizer types, which determined the differences in the optimal P fertilizer application rate; (2) SSP would have a higher advantage in improving PUE and yield compared to MAP and APP because the replacement of Ca²⁺ for Na⁺ in saline soil would reduce the P fixation.

2. Materials and Methods

2.1. Study Site Characterization

The experiment was conducted in the Kenli county (37°35′ N, 118°35′ E, elevation 8.8 m asl), Shandong Province, China, between June 2017 and June 2019. The study area was located in the east of the Yellow River Delta (YRD) and the south coast of Bohai Gulf, which has a warm continental monsoon climate with an average annual temperature and mean annual precipitation of 14.1 °C and 584 mm, respectively. The annual precipitation was 665.7 and 433.1 mm during the winter wheat season in 2017–2018 and 2018–2019, respectively.

Table 1. Physicochemical properties of the top 0–20 cm and 20–40 cm soil profile in the experiments before field trial sowing in 2017.

Characteristics	Soil Depths
	0–20 cm	20–40 cm
pH	7.77	7.72
Bulk density (g cm−3)	1.24	1.2
EC (μS cm−1)	766	578
Salt content (g kg−1)	2.14	2.2
Organic carbon (g kg−1)	6.82	4.08
Total N content (g kg−1)	0.95	0.68
Olsen-P (mg kg−1)	7.06	5.81
Exchangeable K (mg kg−1)	102	126
Soluble Ca2+ (mg kg−1)	348	198
Soluble Na+ (mg kg−1)	472	270

shows the basic physicochemical properties of the soil before the field trial sowing in 2017. For more detail about measurements see Section 2.3.

The field trial land was reclaimed in 2006. Since then, cotton has been planted as a pioneer crop at the beginning of reclamation under flood irrigation. Prior to the experiment, the field was under a typical winter wheat–summer maize rotation between 2014 and 2016. Both winter wheat and summer maize, in each season, was fertilized with ~600 kg ha⁻¹ of diammonium phosphate (18% N and 46% P₂O₅) and ~300 kg ha⁻¹ of urea (46% N), which were mixed into the upper 0–20 cm soil layer by rotary tiller with a widespread device. The field was irrigated twice in the winter wheat season, once after sowing (100 mm) and once at the stem elongation stage (100 mm). No irrigation was provided during the maize growing season.

2.2. Experimental Design

The experiment was conducted in order to investigate the yield performance and P utilization of wheat (Triticum aestivum L.) under different amounts and types of phosphate fertilizer conditions within a winter wheat—summer maize cropping system. The experiment was designed as a split-plot randomized complete block with four replicates. Each block was composed of five plots with different P application rates: 0, 13, 26, 39, and 78 kg P ha⁻¹ (recorded as CK, P1, P2, P3, and P4, respectively). Each plot was split into three individual subplots, which received three P sources: calcium superphosphate (SSP, 16% P₂O₅), monoammonium phosphate (MAP, 11% N and 49% P₂O₅), and ammonium polyphosphate (APP, 23% N and 44% P₂O₅). The SSP was provided by the Yunnan Yuxi Xinghai Fertilizer Co., Ltd. (Yuxi, Yunnan, China). The MAP was provided by the Hubei Xinyangfeng Fertilizer Co., Ltd. (Jinmen, Hubei, China). The APP was provided by the Shindoo Chemical Industry Co., Ltd. (Chengdu, Sichuan, China).

Table 2. Chemical properties of different P fertilizers.

Characteristics	pH	TN	TP	Ions
				K+	Ca2+	Mg2+	Na+
		g kg−1	g kg−1	g kg−1
SSP	3.54	1	142	5.3	236.4	113.7	8.2
APP	4.47	232	195	2.1	0.6	0.4	0.8
MAP	4.24	109	405	4.1	11.3	7.8	7.1

SSP, superphosphate; MAP, monoammonium phosphate; APP, ammonium polyphosphate; TN, total nitrogen; TP, total phosphorous.

lists the details of the chemical properties of the different P fertilizers. The main plot and subplot sizes were 224 m² (16 m × 14 m), and 70 m² (14 m × 5 m), respectively.

The same amount of N and K fertilizer was applied to each plot. All treatments were fertilized with 200 kg N ha⁻¹, 60 kg K₂O ha⁻¹ for winter wheat. The chemical K fertilizer was applied in the form of potassium sulfate (50% K₂O), and urea (46% N) was used to match the same total application rate of N. For the winter wheat, an amount of 70 kg N ha⁻¹ was applied before sowing, whereas 130 kg N ha⁻¹ was applied at the stem elongation stage (GS31). All phosphorous (P) and potassium (K) fertilizers were incorporated into the soil before sowing, accompanied by the application of the basic N fertilizer. All basal fertilizers were broadcasted evenly, by hand, and incorporated into the upper 0–20 cm soil layer by rotary tiller. The N fertilizer was top-dressed into both sides of each inter-row soil with a furrowing machine at a depth of 10 cm at the GS31 stage of wheat. Straw from previous crops was also incorporated with the rotary tiller before planting, for each crop.

The typical winter wheat variety “Jimai22” was used in the experiment. The winter wheat seeding rate was 300 kg ha⁻¹, with a row spacing of 16.7 cm, and the sowing dates were 13 October 2017 and 15 October 2018. All treatments received only one irrigation (100 mm) at the GS31 stage of wheat (after topdressing N fertilizer), there was a 30 cm ridge to eliminate the interference from neighboring plots. Weeds, diseases and insect pests were controlled by applying the appropriate herbicides and insecticides, and no weed and pest problems were observed during the entire growing seasons.

2.3. Sampling and Laboratory Measurements

Soil samples, at 20 cm increments to a depth of 60 cm, were taken before wheat planting in 2017 and after the wheat harvest in 2019. In each plot, six subsamples of soil were obtained using a stainless-steel auger, then composited into a single sample. To weaken the soil sample variation, the sampling site was following an ‘S’ route. All soil samples were air-dried and sieved through a 2 mm mesh. The 10 g soil was extracted with a soil: water ratio of 1:5 and shaken at 180 rpm for 30 min, and the filtered soil solution (through filter paper until clear) was used to measure the soluble Ca²⁺ and Na⁺ by inductively coupled plasma optical emission spectroscopy (ICP-OES, OPTIMA 3300 DV, PerkinElmer, Waltham, MA, USA) and the electrical conductivity (EC) by a EC meter (F3-Meter, Mettler Toledo, Shanghai, China). The 10 g soil was used to determine the pH using a pH meter (S2-Meter, Mettler Toledo, Shanghai, China) at a soil: water ratio of 1: 2.5. The soil Olsen-P was measured by the ammonium molybdate-ascorbic acid method based on the extraction of aired-dried soil with 0.5 mol L⁻¹ NaHCO₃ at pH 8.5 (25 °C).

Fertilizer samples of approximately 500 g were selected before application, and 0.1995–0.2005 g samples were used to determine the cations content. The samples were digested with a mixture of HNO₃ (2.5%) in a microwave-accelerated reaction system (CEM, Matthews, NC, USA). The K, Ca, Na and Mg concentrations in the digested solutions were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, OPTIMA 3300 DV, PerkinElmer, USA).

The wheat shoot samples were collected at the stem elongation stage (GS31), the flowering stage (GS65), and the maturity stage (GS92). Entire shoots were randomly selected from two 0.5 m lengths of four adjacent rows in each plot. At maturity, the grain yield and aboveground biomass were measured from an area of 3 m² (1.5 m in width × 2.0 m in length) in the middle of each plot. In this study, the term “shoot” refers to all aboveground parts of the wheat plants, including straw and grain. The thousand-kernel weight was calculated by measuring the mass of three groups of 500 wheat kernels per plot. The kernel per ear was calculated for 30 randomly selected wheat ears from each plot. All of the plant samples were oven-dried at 60–65 °C to a constant weight and subsequently weighed, and then were ground using a stainless-steel grinder. The shoot samples were digested with a mixture of HNO₃ and H₂O₂ in a microwave-accelerated reaction system (CEM, Matthews, NC, USA). The P, Ca and Na concentrations in the digested solutions were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, OPTIMA 3300 DV, PerkinElmer, USA).

2.4. Data Calculations

The agronomic use efficiency of P (AE_P, kg kg⁻¹) is the yield increase per unit of P applied and the higher AE_P values indicate more efficient use of P fertilizer. The AE_P was calculated as the following equation:

$${\mathrm{AE}}_{\mathrm{P}}=(Y_P-Y_0)∕P$$

where Y_P and Y₀ are grain yield in P applied plots and without P application plots (kg), respectively; P is the P application rate (kg).

2.5. Statistics Analysis

The wheat yield, aboveground P uptake and AE_P response curves to P application rates, as well as the soil Olsen-P concentration response curves to P application rates for the different P sources, were generated using the NLIN procedure in the SAS software (SAS v9.3; SAS Institute, Cary, NC, USA) [23]. Three response models (quadratic, quadratic-plateau, and linear-plateau) were generated in SAS. The linear-plateau model produced the best fitting results for characterizing the response of the wheat yield and P uptake to P application rates. The quadratic model produced the best fitting results for characterizing the response of the AE_P and soil Olsen-P concentration to P application rates. Descriptive statistical analyses were performed to calculate the average and standard deviation. A one-way ANOVA model was applied to assess the soluble ions (Ca²⁺, Na⁺) in soils applied with different P sources. To compare the differences in the wheat grain yield, shoot biomass, spike number, grain number per spike, and 1000-kernel weight of wheat yield, the multivariate ANOVA model was used, with the P rates and P sources serving as the fixed effects and years as the random effects. Multiple comparisons of the means were examined with Tukey’s K test for statistically significant differences.

3. Results

3.1. Yield under Different P Rate and Source in Two Years

The phosphorus application rate significantly affected the winter wheat yield through the yield composition, including grain yield, shoot biomass, spike number, grain number and 1000-kernel weight (p < 0.001;

Table 3. The wheat grain yield (14.0% moisture), shoot biomass, spike number, grain number per spike, and 1000-kernel weight of maize as affected by the P rates, P sources, and years.

Treatment	Grain Yield	Shoot Biomass	Spike Number	Grains Number	1000-Kernel Weight
	(Mg ha−1)	(Mg ha−1)	(m−2)	(Spike−1)	(g)
P rate (R)
CK	3.06 c	4.57 c	324 c	22.7 c	33.4 c
P1	5.65 b	8.24 b	503 b	38.0 b	36.1 b
P2	6.64 a	9.70 a	631 a	39.8 a	37.9 a
P3	6.67 a	9.64 a	637 a	39.7 a	37.8 a
P4	6.74 a	9.83 a	642 a	39.6 a	37.8 a
P source (S)
SSP	6.13 a	9.13 a	544 a	36.2 a	37.9 a
MAP	5.51 b	8.02 b	558 a	34.8 b	35.5 c
APP	5.62 b	8.04 b	540 a	36.8 a	36.4 b
Year (Y)
2017–2018	5.61	7.92	605	34.4	35.2
2018–2019	6.07	8.87	489	37.5	38.0
Source of variation
R	<0.001	<0.001	<0.001	<0.001	<0.001
S	<0.001	<0.001	ns	<0.001	<0.001
Y	<0.001	<0.001	<0.001	<0.001	<0.001
R × S	<0.05	<0.05	ns	<0.001	ns
R × Y	ns	ns	ns	<0.05	<0.01
S × Y	ns	ns	ns	ns	<0.05
R × S × Y	ns	ns	ns	<0.001	<0.01

The P rates include CK, 0 kg P ha⁻¹; P1, 13 kg P ha⁻¹; P2, 26 kg P ha⁻¹; P3, 39 kg P ha⁻¹; P4, 78 kg P ha⁻¹. The P sources include superphosphate (SSP), ammonium polyphosphate (APP), and monoammonium phosphate (MAP). Means followed by the different lowercase within each line denote significant differences in the same N rate (p < 0.05). ns: not significant.

). The grain yield of the winter wheat increased first and then stabilized with the increase in the P application rate, regardless of which P source was applied (Figure 1); this sharply increased under P1 and P2 compared to CK, but indistinctively increased under P3 and P4 compared to P2 (Table 3). The grain yield increased by 0.85 and 1.2 times under P1 and P2, respectively, compared to CK. The two year results showed that all of the P sources reached a maximum grain number and 1000-kernel weight under the P2 level (Table 3). The grain yield, shoot biomass, grain number and 1000-kernel weight were significantly affected by the P source (p < 0.001; Table 3), and the highest values were observed under SSP compared to other P sources. All of these yield and biomass indicators were significantly affected by the number of years (p < 0.001; Table 3). The grain yield, shoot biomass, grain number and 1000-kernel weight in the 2018–2019 season were higher than in the 2017–2018 season, while the spike number showed the opposite result (Table 3). The interactive effect of the P rate and source was significant for the grain yield, shoot biomass, and grain number. The interactive effect of the P rate and number of years was significant for the grain number and1000-kernel weight. The interactive effect of the P source and year was significant for the 1000-kernel weight. The interactive effect of P source, rate and year was significant for grain number) and 1000-kernel weight (Table 3).

3.2. Estimated Optimal P Fertilizer Rate of Three P Sources Based on the Wheat Yield, Aboveground P Uptake and P Agronomy Efficiency

The application of the phosphorus fertilizer increased the winter wheat yield but did not further increase the yield when the P application rate exceeded 20 kg P ha⁻¹ SSP, 18 kg P ha⁻¹ MAP, and 17 kg P ha⁻¹ APP (Figure 1). The yield increase (slope of linear response curve) per unit P fertilizer input for SSP, MAP and APP was 0.21, 0.18 and 0.20 Mg, respectively. The calculated optimal P rates based on the wheat yield response curve were 20 kg P ha⁻¹ for SSP, 18 kg P ha⁻¹ for MAP and 17 kg P ha⁻¹ for APP, and the average optimal P application rate for the winter wheat was 18 kg P ha⁻¹. Although the optimal P application rates of the three P fertilizers were close, the yields under the optimal P application were quite different. The grain yield of the winter wheat under the optimal P rate of SSP (7.2 Mg ha⁻¹) was higher than MAP (6.4 Mg ha⁻¹) and APP (6.5 Mg ha⁻¹).

The agronomic use efficiency of P (AE_P) decreased gradually with the P application rate and showed different responses to the applied P sources (Figure 1), which were simulated using the quadratic model. The AE_P of SSP decreased from 175 kg kg⁻¹ (P fertilization rate was 13 kg ha⁻¹) to 44 kg kg⁻¹ (P fertilization rate was 78 kg ha⁻¹). The AE_P of MAP decreased from 142 kg kg⁻¹ (P fertilization rate was 13 kg ha⁻¹) to 35 kg kg⁻¹ (P fertilization rate was 78 kg ha⁻¹). The AE_P of APP decreased from 154 kg kg⁻¹ (P fertilization rate was 13 kg ha⁻¹) to 36 kg kg⁻¹ (P fertilization rate was 78 kg ha⁻¹). The AE_P of SSP was higher than that of MAP and APP under the same P application rate.3.3. Effects of P fertilizer Rate and Sources on Soil Available P at Different Soil Depths.

The response of the soil available P to the P fertilizer was related to the differences in the P fertilizer rate, sources and soil depth (Figure 2). The soil Olsen-P concentration under APP was higher than MAP and SSP at soil depths of 0–20 cm, 20–40 cm, and 40–60 cm. The soil Olsen-P concentration under all of the P sources increased with the P application rate at the soil depth of 0–20 cm (Figure 2a). There was no significant difference in the soil available P concentration between SSP and MAP in the range of the low P application rate, between 0 and 39 kg P ha⁻¹; however, the soil Olsen-P concentration of MAP was higher than that of SSP with the increase in the P application rate, between 39 and 78 kg P ha⁻¹ (Figure 2a). The soil Olsen-P concentration of APP increased gradually with the P application rate, while MAP and SSP increased first and then decreased with the P application rate at the soil depth of 20–40 cm (Figure 2b). The P application rate of MAP and SSP under the highest soil Olsen-P concentration were 69 kg P ha⁻¹ and 64 kg P ha⁻¹, respectively, at the depth of 20–40 cm (Figure 2b). At the depth of 40–60 cm, the soil Olsen-P concentration of the three P sources increased first and then decreased with the P application rate. The P application rates of APP, MAP and SSP under the highest soil Olsen-P concentration were 59, 56 and 59 kg P ha⁻, respectively (Figure 2c).

3.3. Effects of P Fertilizer Sources on Soil Alkaline Cations at Different Soil Depths

The soil water soluble Ca²⁺ content decreased with the soil depths (Figure 3). The soil Ca²⁺ content increased under P2-SSP and P2-MAP but decreased under P2-APP compared to under CK, at the depth of 0–20 cm (p < 0.05; Figure 3a). Similar to the depth 0–20 cm, the soil Ca²⁺ content under P2-SSP, at the depth of 20–40 cm and 40–60 cm, was higher than that under other treatments. However, the soil Ca²⁺ content under the P2-APP treatment in the 20–40 and 40–60 cm soil layers were higher than, and no difference from, CK, respectively. There was no significant difference in the soil Ca²⁺ content between P2-MAP and CK at depths of 20–40 cm and 40–60 cm (Figure 3a). The soil water soluble Na⁺ under P2-SSP was lower than the other treatments at all soil depths, and the soil Na²⁺ content under P2-MAP was lower than that of CK at a depth of 0–20 cm (p < 0.05; Figure 3b).

4. Discussion

The relationship between the P fertilizer application rate and yield showed that P could increase the winter wheat yield, but it did not increase significantly beyond the optimal application rate (approximately 18 kg P ha⁻¹) (Figure 1), which was similar to previous reports [24,25]. The yield increased significantly from CK (0) to P2 (26 kg P ha⁻¹), whilst it barely changed from P2 (39 kg P ha⁻¹) to P4 (78 kg P ha⁻¹) (Table 3). The P application affected the winter wheat yield through yield composition and the increase in the grain yield was mainly attributed to the spike number, grain number and 1000-kernel weight. The grain number and 1000-kernel weight of each type of P fertilizer reached the maximum under P2 (Table 3). The optimized P application rate can allocate more available photosynthates to the effective population of winter wheat and achieve a high yield [26,27]. The grain number and 1000-grain weight of wheat in this study reached the maximum when applying 13–26 kg P ha⁻¹, which was different from the previous result reported by Zhang et al. [26]. The fact that the grain number and 1000-grain weight were reached when applying 50 kg P ha ha⁻¹ may be attributed to the following factors: (1) the yield level of this study was low (6.5 vs. 8.6 Mg ha⁻¹), therefore, the P requirement of the crop was reduced; (2) the sand content of the soil in this study was relatively high, and the P fertilizer applied to the soil in season was converted into more available P. Although the yield of this experiment was higher than the average yield of the YRD region (6.5 vs. 5.5 Mg ha⁻¹), the yield potential of this region was still large [28,29].

Many previous research results have shown that SSP could not achieve a good effect on crop P uptake in saline soil, particularly in high-pH saline soil, because its solubility decreased with the increase in soil pH and Ca²⁺ content, which ultimately resulted in low availability of soil P [22,30,31]. However, our results showed that under the same P application rate, before reaching the highest P uptake of wheat in coastal saline soil, the aboveground P uptake of the wheat under the SSP treatment was higher than that of MAP and APP (Figure 1). In addition, the aboveground P uptake under SSP was 14% and 13% higher than MAP and APP, respectively, under the optimal P application rate (Figure 1). In addition, the AE_P under SSP was higher than the other two P sources under the same P application rate. Moreover, the application of SSP increased the winter wheat yield by 9–11%, compared to MAP and APP. The advantage of SSP in PUE and yield may be related to its particular characteristics. On one hand, the object of this study is coastal saline soil in which the content of Na⁺ is high, and the application of SSP containing many Ca²⁺ (Table 2) increased the soil Ca²⁺ concentration, which replaced the high Na⁺ in the soil particles and promoted the leaching of Na⁺ from the soil layer [32,33]. This is evidenced in our results as SSP increased the soil Ca²⁺ concentration by 20–42% but decreased the Na⁺ concentration by 14–18% in all soil layers, including 0–20, 20–40, 40–60 cm, compared to CK (0 kg P ha⁻¹) (p < 0.05; Figure 3). Meanwhile, MAP only increased the soil Ca²⁺ concentration by 16% and decreased the Na⁺ concentration by 16% in 0−20 cm soil compared to CK (p < 0.05; Figure 3). APP only increased the soil Ca²⁺ concentration by 9.5% in 20–40 cm soil layers and did not affect the soil Na⁺ concentration at all soil layers (Figure 3). On the other hand, SSP can provide some nutrient elements conducive to crop growth, such as sulfur (S), silicon (Si) and magnesium (Mg) [34,35]. In addition, our result also proved that SSP has a higher Mg, which were scarce in the coastal saline soil (Table 2).

Notably, the coupling relationship between the P application rate and the P nutrient requirements for crop growth may be affected by the P fertilizer sources, which have different dissolution characteristics and soil responses in the process of optimizing the P supply to achieve high crop yield and PUE [22,36]. Our results indicated that the types of mineral P fertilizers played an important role in the soil available P (soil Olsen-P) concentration (Figure 2). Under the condition of the same P application rate, all depths of the soil (0–20, 20–40, 40–60 cm) had the highest soil available P concentration when APP was applied (Figure 2). This is mainly because the hydrolysis of pyrophosphate, which is the main component of APP, is a key step to determining the content of available P in soil [22]. Studies have shown that pyrophosphate hydrolyzed 50% in calcareous soil after 21 days, while only 30–35% in non-calcareous soil [37], indicating that the presence of calcium ions promoted the process of pyrophosphate hydrolysis. In this experiment, the large amount of calcium ions in the coastal saline soil promoted the process of hydrolysis and the hydrolysis rate was higher than the P fixation rate, and thus improved the availability of soil P.

Although our results indicate that appropriate rates and P sources (20 kg P ha⁻¹, SSP) resulted in high yield and PUE, most famers are unwilling to adopt the strategy. The main reason for this is that the nutrient content of superphosphate is too low, and therefore requires more time and effort to apply. In addition, the performance of SSP has been ignored by farmers and the farming industry. In the future, measures should be taken to encourage farmers to adopt the strategy by providing technical assistance through the science and technology publicity platform and developing derivative SSP products to facilitate application.

5. Conclusions

The results of improving the phosphorus use efficiency (PUE) and crop yield in salinized soils, from the perspective of optimizing the rate and type of phosphorus fertilizer, showed that the aboveground P uptake, AE_P, soil available P levels and grain yield of winter wheat were significantly affected by the optimal P application rate and source. P application increased the wheat yield, primarily, by the increasing spike number, grain number and 1000-grain weight. The P application rate, source, and year affected the yield of the winter wheat, and the source and rate of the P application fertilizer had an interactive effect on the yield. The grain yield of the winter wheat increased with the P application rate, but did not increase continuously when the P application rate exceeded 20 kg ha⁻¹ SSP,18 kg ha⁻¹ MAP and 17 kg ha⁻¹ APP, respectively. The aboveground P uptake increased continuously with the P application rate, but did not increase continuously when the P application rate exceeded 31 kg ha⁻¹ SSP, 30 kg ha⁻¹ MAP and 29 kg ha⁻¹ APP, respectively. The AE_P decreased gradually with the P application rate. Under the same P application rate, the soil available P level of APP was higher than that of SSP and APP in all soil layers. However, SSP had higher aboveground P uptake, AE_P and yield than the other two P sources, under the same P application rate, due to its higher Ca²⁺ levels, instead of Na⁺, in the soil. Overall, our results highlight the advantages of optimal P management strategies for reducing soil P fixation and improving winter wheat yield in salinized soils by optimizing P application rates and selecting P fertilizer sources tailored to local soil properties.