Combined Controlled-Release and Common Fertilizer Application Increases Apple Productivity by Optimizing Soil Nutrient and Microbial Communities

Abstract

Apples in China are planted mainly in nutrient-poor mountain soil, and a large amount of fertilizer input results in resource waste and a decrease in nutrient utilization efficiency. Controlled-release fertilizer (CRF) has been shown to be environmentally friendly and increase crop yield, but nutrient release cannot be precisely synchronized with apple demand. Here, a suitable secondary fertilization method was established by a two-year apple field experiment with CRF and common compound fertilizer (CF) at various ratios under a 25% reduction in application. The application of CF and CRF changes the temporal and spatial distributions of soil NPK nutrients, decreasing NPK losses and NH₃ emissions. The NH₃ emissions under CF and CRF decreased by 17.98–44.86%, as N loss decreased by 11.59–29.81% and by 4.45–8.19%, with respect to those under CF alone, while the soil pH and electrical conductivity increased by 8.28–17.12% and 10.73–18.29%, compared with those under CF alone. The increase in soil P and K also decreased losses by 8.28–17.12% and 10.73–18.29%. The combined application of CF and CRF can increase soil microbial diversity and functional taxa and nutrient cycling genes, resulting in efficient nutrient transformation and supply for apple trees. The regulation of nutrients and microbes by the secondary application of CF and CRF drives an increase in apple yield of 23.71–54.32%, resulting in high economic benefits. In total, the application ratio of CF and CRF at 3:7 in March and July was an effective way to balance apple productivity and the soil ecological environment, providing a sustainable solution for mountainous orchard ecosystems globally.

1. Introduction

The apple (Malus pumila Mill.) The planting area in China has increased to 1,996,632 hectares by 2023, accounting for 43% of the total global apple planting area [1]. The average N fertilizer usage in conventional Chinese apple orchards has reached 600–800 kg ha⁻¹ to maintain high yields, exceeding growth needs [2]. In Shandong Jiaodong Peninsula, the use of annual amounts of 528 kg N ha⁻¹, 252 kg P₂O₅ ha⁻¹ and 528 kg K₂O ha⁻¹ has become a common nutrient management method, as apple trees grow in nutrient-poor mountainous and hilly soil [3,4]. More than 58% of N and 67% of P₂O₅ were not used, resulting in the nutrient uptake efficiency of apple trees being lower than that of wheat and maize [5,6,7,8].

Fertilizer volatilization [9], leaching [10], and runoff loss [11] and fixation [12] are the main ways to threaten the environment. The “Sustainable Development Goals (SDGs)” related to the environmental sustainability of the UN will be overcome [13]. This can also damage the soil microbial ecosystem, resulting in a decrease in microbial diversity and inhibition of the abundance of functional genes related to nutrient cycling, affecting crop nutrient uptake and productivity [14]. Optimizing fertilization can increase soil microbial taxa to optimize the soil ecological environment, resulting in interactions between nutrient supply and microbial function. Given, the Chinese government has proposed a series of policies to decrease fertilizer usage to accelerate clean production in apple orchards [15].

Controlled-release fertilizers (CRFs) can slowly release nutrients and have become a promising alternative to solve the above problems [16]. CRFs can increase crop yield and reduce the environmental pressures caused by fertilizer usage, while there are fixed nutrient release patterns during production. The nutrient needs of apple trees fluctuate with environmental changes such as temperature, rainfall, and soil moisture [17,18]. The use of the CRF cannot match the dynamic nutrient needs of apple trees, as they have a complex growth period that lasts for a year [19,20]. Optimizing the nutrient supply for apple trees by using CRF and common fertilizer (CF) is vital for achieving stable nutrient release and rapid supplementation to increase agronomic efficiency [21]. CRF can decrease N losses in orchard soil, whereas excessive CF can cause nutrient surpluses to increase the risk of environmental pollution [22]. How to coordinate the application of CRF and CF to meet the nutrient needs of apple trees during the budding and fruit expansion period, thereby decreasing environmental pressure, is a key issue during apple tree growth.

2. Materials and Methods

2.1. Experimental Site

A field experiment was performed at an apple orchard in Laishan Town, Laishan District, Yantai, Shandong, China (37°25′44″ N, 121°24′26″ E), from November 2020 to October 2022. The soil has a loamy texture and is classified as a Lithic Udic-Orthic Primosol on the basis of the Chinese Soil Resources system. The basic physicochemical properties of the soil were as follows: pH 6.65 with 1:5 soil/deionized water extract, 25.31 mg/kg NO₃⁻-N with 1 mol/L KCl solution extract, 13.54 mg/kg NH₄⁺-N with 1 mol/L KCl solution extract, 21.27 mg/kg available P with 0.5 mol/L NaHCO₃ solution extract, and 65.84 mg/kg available K with 1 mol/L NH₄OAc solution extract. The field experiment site has a temperate monsoon climate, with an average annual rainfall of 690.2 mm and a temperature of 17.0 °C.

The Yanfu-3 apple trees grafted on Pingyi sweet tea (Malus hupehensis Rehd.) rootstocks that needed to grow for 7–8 years to begin fruiting were used in the field experiment, with a planting distance of 2 × 4 m. Apple trees were trained to grow radially around branches with vertical growth and weight gain bands. Tree vigor was controlled by pruning branches in spring and summer. Drip irrigation was adopted with 3–4 supplementary irrigations during the growing period to maintain the soil moisture content at 60–80% of the field water-holding capacity. Pest and disease control, including physical and low-toxicity chemical pesticides, was carried out, and artificial weeding was performed 2–3 times in fertilized trenches and interrow areas during the growing period. The experiment was conducted in the 9th year without fertilizer application in the 7th and 8th years.

2.2. Experimental Design

2.2.1. Fertilizer Preparation

In accordance with the local traditional fertilizer application amount, controlled-release fertilizer (CRF) and common fertilizer (CF) were used in the field experiment, with NPK application decreasing by 25% to 400 kg ha⁻¹ (N), 200 kg ha⁻¹ (P₂O₅) and 400 kg ha⁻¹ (K₂O). CRFs were prepared from a mixture of 63 kg of urea, 35.87 kg of diammonium phosphate ((NH₄)₂HPO₄) and 66 kg of potassium sulfate (K₂SO₄) with a 4% coating of castor oil biobased material in the laboratory, achieving a 3-month nutrient release time (Figure S1). CF was prepared by blending 63 kg of urea, 35.87 kg of diammonium phosphate ((NH₄)₂HPO₄) and 66 kg of potassium sulfate (K₂SO₄). Urea (46% N), (NH₄)₂HPO₄ (18% N, 46% P₂O₅) and K₂SO₄ (50% K₂O) were provided by Shandong Nongda Fertilizer Sci&Tech. Co., Ltd. (Tai’an, China).

2.2.2. Field Experimental Design

Seven fertilizer application treatments, each replicated three times with 10 trees per replicate, were arranged in a randomized design in the field (

Table 1. Experimental fertilization treatments.

Treatment	Fertilization Period and Application Rate (2021/2022)
	Budding Period (March)	Fruit Expansion Period (July)	Fruit Expansion Period (August)
Unf	-	-	-
J-CF	30% CF	70% CF	-
A-CF	30% CF	-	70% CF
J-30CF70CRF	30% CF	70% CRF	-
J-50CF50CRF	50% CF	50% CRF	-
A-30CF70CRF	30% CF	-	70% CRF
A-50CF50CRF	50% CF	-	50% CRF

Note: Unf, without fertilizer; J-CF, application of 30% common fertilizer in March and 70% common fertilizer in July; A-CF, application of 30% common fertilizer in March and 70% common fertilizer in August; J-30CF70CRF, application of 30% common fertilizer in March and 70% controlled-release fertilizer in July; J-50CF50CRF, application of 50% common fertilizer in March and 50% controlled-release fertilizer in July; A-30CF70CRF, application of 30% common fertilizer in March and 70% controlled-release fertilizer in August; A-50CF50CRF, application of 50% common fertilizer in March and 50% controlled-release fertilizer in August.

). The total planting area of the experimental site was 2016 m². A total of 1.64 kg CRF or 1.58 kg CF was applied to each apple during the growth period. Four fertilization trenches (20 cm width × 20 cm depth) ran north–south and were positioned between the center of the trunk and the outer edge of the crown projection. The fertilizers were uniformly mixed with 0–20 cm surface soil, backfilled into the trenches and lightly compacted.

2.2.3. Nutrient Loss

The nutrient loss behavior of plants in response to the application of fertilizer in March and July was simulated and studied by the use of pots on the basis of the fertilization time in the field and high-yield treatments. No plants were grown in the pots to avoid plant-related effects with nutrient loss. Six treatments were repeated three times (

Table 2. Nutrient loss experiment.

Treatment	Budding Period (March)	Fruit Expansion Period (July)
CRF	100% CRF	-
30CF70CRF	30% CF	70% CRF
50CF50CRF	50% CF	50% CRF
30CF70CF	30% CF	70% CF
50CF50CF	50% CF	50% CF
CF	100% CF	-

Note: CRF, application of 100% controlled-release fertilizer in March; 30CF70CRF, application of 30% common fertilizer in March and 70% controlled-release fertilizer in July; 50CF50CRF, application of 50% common fertilizer in March and 50% controlled-release fertilizer in July; 30CF70CF, application of 30% common fertilizer in March and 70% common fertilizer in July; 50CF50CF, application of 50% common fertilizer in March and 50% common fertilizer in July; CF, application of 100% common fertilizer in March.

), resulting in a total of 18 plastic pots. A total of 3.75 kg of air-dried orchard soil from the 0–20 cm topsoil of the unfertilized area in the field experiment site was added to a plastic pot for the top 35 cm × bottom 25 cm × height 15 cm, with a 2 mm hole at the bottom. Soil (1.25 kg) was then added after fertilizer was applied to 30CF70CRF, 50CF50CRF, 30CF70CF and 50CF50CF. The remaining fertilizer was added at the same location in the plastic pots as it was in the field.

The soil water content remained at 100% of the soil water holding capacity for one week. The leachate was collected by adding 5, 12, 21, 23, 52, 50, 29 and 12 mL of water with a dropper to carry out leaching on the basis of rainfall in the field from March to October. The concentrations of NO₃⁻-N, NH₄⁺-N, P, and K in the leachates were determined by using flow analysis, Mo–Sb colorimetry and flame photometry. Leaching and measurements were carried out once a month from March to October. The concentrations of NO₃⁻-N, NH₄⁺-N, P, and K were calculated as follows:

$${\mathrm{N}\mathrm{O}}_3^{-}\text{-}\mathrm{N}\mathrm{loss}={\mathrm{V}}_1\times {\displaystyle \frac{\mathrm{C}\mathrm{i}\left({\mathrm{N}\mathrm{O}}_3^{-}\text{-}\mathrm{N}\right)}{m\times \mathrm{C}\left(\mathrm{N}\right)}}\times 100\%$$

$${\mathrm{N}\mathrm{H}}_4^{+}\text{-}\mathrm{N}\mathrm{loss}={\mathrm{V}}_1\times {\displaystyle \frac{\mathrm{C}\mathrm{i}\left({\mathrm{N}\mathrm{H}}_4^{+}\text{-}\mathrm{N}\right)}{m\times \mathrm{C}\left(\mathrm{N}\right)}}\times 100\%$$

$$\mathrm{P}\mathrm{loss}={\mathrm{V}}_1\times {\displaystyle \frac{\mathrm{C}\mathrm{i}\left(\mathrm{P}\right)}{m\times \mathrm{C}\left(\mathrm{P}\right)}}\times 100\%$$

$$\mathrm{K}\mathrm{loss}={\mathrm{V}}_1\times {\displaystyle \frac{\mathrm{C}\mathrm{i}\left(\mathrm{K}\right)}{m\times \mathrm{C}\left(\mathrm{K}\right)}}\times 100\%$$

where V₁, C_i(${\mathrm{N}\mathrm{O}}_3^{-}$-N), C_i(${\mathrm{N}\mathrm{H}}_4^{+}$-N), C_i(P) and C_i(K) represent the concentrations of ${\mathrm{N}\mathrm{O}}_3^{-}$-N, ${\mathrm{N}\mathrm{H}}_4^{+}$-N, P and K, in the leaching solution during the i-th month and m, C(N), C(P) and C(K) represent the contents of added CF and CRF, of N, P and K.

A continuous air-flow enclosure method was used to conduct an NH₃ emission test that was consistent with the treatment of nutrient loss experiments [23]. The pot for testing nutrient leaching was placed in a transparent Plexiglas chamber with an inner diameter of 6.5 cm × height of 12.5 cm. NH₃ emissions were collected in 200 mL of 2% boric acid solution prepared with methyl red, bromocresol, and ethanol as indicators. NH₃ emissions were collected for 4 h at 8:00 a.m.–10:00 a.m. and 14:00 p.m.–16:00 p.m. once a month from March to October. The NH₃ concentration was determined by titrating the ammonium captured in the boric acid solution with 0.01 mol/L H₂SO₄. The NH₃ emissions were calculated as follows:

$${\mathrm{NH}}_3 \mathrm{emissions} (\mathrm{kg} {\mathrm{ha}}^{-1} {\mathrm{d}}^{-1})={\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_4^{+}\text{-}\mathrm{N}}\times {10}^{-6}\times \mathrm{V}\times {\displaystyle \frac{{10}^4}{\mathsf{\pi }\times {\mathrm{r}}^2}}\times {\displaystyle \frac{24}{\mathrm{t}}}$$

where ${\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_4^{+}\text{-}\mathrm{N}}$, V, r² and t represent the concentration of ${\mathrm{N}\mathrm{H}}_4^{+}\text{-}\mathrm{N}$ in the NH₃ absorption liquid, the volume of the NH₃ absorption liquid, the sealed room radius and the NH₃ absorption time.

2.3. Sample Collection and Analysis

2.3.1. Soil Nutrients

Soil samples from 0–20 cm, 20–40 cm and 40–60 cm depth with a 10 cm diameter for 3 random sampling points in the radial fertilization trench area of each treatment in field experiments were collected in 2021 and 2022 on the basis of the apple harvest. Each sampling point had a volume of 200 cm³ per soil layer, with a total of 126 samples collected over two years. After being air dried, ground, and passed through 2 mm sieves, the soil samples were placed in paper bags. The soil pH, NO₃⁻-N, NH₄⁺-N, available P and available K were subsequently measured by using the conductance method, flow analysis method, Mo–Sb colorimetry and flame photometry.

2.3.2. Apple Yield and Quality

All the apples from each fertilization treatment were harvested once in the optimal ripening period, and the total yield was measured for one year by using an electronic scale. Fifteen apples were randomly selected to determine quality from each replicate experiment site in the upper, middle and lower canopies, with a total of 45 apples analyzed per treatment. The single-fruit weight of the apples was determined by using a small electronic scale. Fruit hardness was determined by using a GY-1 fruit hardness tester. The fruit shape index was calculated as the length/diameter ratio of the apple. After extraction, 5.0 g of fresh apple pulp extract was reacted with anthrone reagent in a boiling water bath, after which it was centrifuged and filtered. The sugar content was analyzed by using a spectrophotometer to determine the absorbance at 620 nm with a glucose standard curve.

2.3.3. Economic Benefit

The economic benefits of the different fertilization treatments were evaluated on the basis of the annual apple yield and local market prices. The apple income, net apple income, and increased revenue compared with Unf were calculated as follows:

Apple income = Yield × Average price of apples

The average price of local apples during the experimental years (2021–2022) was 5.0 ¥/kg.

Net apple income = Apple income − Fertilizer Cost − Labor Cost

The labor cost (¥/ha) refers to the annual manual operation cost per hectare in the field, including that associated with fertilization, irrigation, pruning, weed control, and pest and disease control. It was calculated on the basis of the local actual labor price and total working hours during the apple tree growth period every year.

Increased revenue compared with Unf (10³¥/ha) = Net apple income (Treatment) − Net apple income (Unf)

$$\mathrm{Increased}\mathrm{revenue}\mathrm{compared}\mathrm{to}\mathrm{Unf}(\%)={\displaystyle \frac{\mathrm{N}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{p}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{e}(\mathrm{T}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t})-\mathrm{N}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{p}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{e}(\mathrm{U}\mathrm{n}\mathrm{f})}{\mathrm{N}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{p}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{e}(\mathrm{U}\mathrm{n}\mathrm{f})}}\times 100\%$$

2.4. Soil DNA Extraction and Sequencing of Bacteria and Fungi

A total of 0.5 g of rhizosphere soil from the apple trees in the fertilization area was collected and stored in a −80 °C freezer, after which DNA was extracted with a DNA Extraction Kit (Life Technologies, Carlsbad, CA, USA). Quantification was carried out by using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and 1 % agarose gel electrophoresis. The universal primers 338 F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806 R (5′-GGACTACHVGGGTWTCTAAT-3′) were used to amplify the V3–V4 region of the bacterial 16S rRNA gene. The primers ITS1F (5′-GGAAGTAAAAGTCGTAACAAGG-3′) and ITS2R (5′-GCTGCGTTCTTCATCGATGC-3′) were used to amplify the fungal ITS gene. Each sample underwent three independent PCR amplifications to decrease potential bias. The PCR products were extracted from a 2% agarose gel and purified by using a PCR Clean-Up Kit (YuHua, Shanghai, China) according to the manufacturer’s instructions and then quantified by a Qubit 4.0 (Thermo Fisher Scientific, Waltham, MA, USA). Each PCR product was normalized to equimolar concentrations and subjected to paired-end sequencing (2 × 300 bp) on the Illumina NextSeq 2000 platform (Illumina, San Diego, CA, USA). The 16S rRNA sequencing data were processed by the MOTHUR MiSeq pipeline.

The Chao1, Shannon, Simpson and Observed_species indices represent species change, evenness and abundance, and are used to reflect the response of microbial diversity to CRF. They were analyzed by using Mothur v1.30.1 to calculate alpha and β diversity indices on the basis of operational taxonomic unit (OTU) information. The differences in microbial communities under the CRF treatment were analyzed with nonmetric multidimensional scaling (NMDS) by using the “vegan” package in R software (4.4.0), which is based on Bray–Curtis dissimilarity. The stress value was used to represent the true pairwise dissimilarities between samples in the original high-dimensional data and was calculated by using the “vegan” package in R software.

2.5. Real-Time Fluorescence Quantitative PCR (qPCR)

Quantitative PCR was used to determine the absolute abundance of N- and P-cycling functional genes with the primer information listed in Table S1. The N cycling functional genes included narG, nasA, nirS2, nirK1, UreC, nosZ1, hzsA, amoA1, amoB, nxrA, napA, nirS1, nirS3, nirK2, gdh, nifH, niK3, nosZ2, hzo, hzsB, amoA2 and hao. The P cycling functional genes included cphy, phnK, ppk3, phoD, bpp, pqqC, ppx and phoX. Three replicates were established for all detections to ensure reproducibility, with the gene copy numbers calculated on the basis of Ct (threshold cycle) values.

2.6. Statistical Analyses

A partial least squares path model (PLS–PM) was applied to evaluate the relationships among treatments, bacteria, fungi, N cycle genes, P cycle genes, soil nutrient properties, nutrient loss, NH₃ emissions and apple yield. A bootstrap method (1000 iterations) was used to validate the estimates of the path coefficients and the coefficients of determination (R²). The goodness-of-fit index was used to evaluate the overall predictive performance of the model. The statistical analysis was performed using the IBM SPSS Statistics 22.0 software for Windows (SPSS, Inc., Chicago, IL, USA). Mean comparisons were performed using one-way ANOVA, and post hoc tests were conducted when appropriate (LSD, p < 0.05). The figures were prepared using Origin 2022 (OriginLab, Northampton, MA, USA).

3. Results

3.1. Apple Yield and Economic Benefit

The combined application of CF and CRF increased apple yields and economic benefits, with differences among treatments and years (p < 0.05) (

Table 3. Effects of the combined application of CF and CRF on the yield and economic benefits of apple.

Year	Treatments	Single Fruit Weight (g)	Fruit Hardness (kg/cm)	Fruit Shape Index	Soluble Sugar Content (%)	Yield (kg/ha)	Fertilizer Cost (¥/ha)	Labor Cost (¥/ha)	Apple Income (¥/ha)	Net Apple Income (103¥/ha)	Increased Revenue Compared to Unf
											(103¥/ha)	%
2021	Unf	153.51 d	6.68 d	0.89 b	11.76 c	25,320 c	0	0	126,600 d	126.60 c	-	-
	J-CF	229.26 c	7.52 c	0.82 c	13.19 b	26,424 c	4512.48	1500	132,120 c	126.11 c	−0.49	−0.39
	A-CF	281.15 a	8.30 a	0.86 b	13.88 b	25,896 c	4512.48	1500	129,480 c	123.47 c	−3.13	−2.47
	J-30CF70CRF	229.13 c	7.77 b	0.81 c	17.38 a	32,688 a	7139.66	1500	163,440 a	154.80 a	28.20	22.28
	J-50CF50CRF	251.78 b	7.97 ab	0.83 c	17.51 a	30,516 b	6389.04	1500	152,580 b	144.69 b	18.09	14.29
	A-30CF70CRF	220.49 c	8.13 a	0.88 b	15.82 ab	30,900 b	7139.66	1500	154,500 b	145.86 b	19.26	15.21
	A-50CF50CRF	222.38 c	8.43 a	0.96 a	10.90 c	29,820 b	6389.04	1500	149,100 b	141.21 b	14.61	11.54
2022	Unf	178.32 e	5.62 a	0.85 a	16.42 a	24,036 d	0	0	120,180 e	120.18 d	-	-
	J-CF	221.97 b	5.83 a	0.82 a	16.36 a	27,888 c	4512.48	1500	139,440 d	133.43 c	13.24	11.02
	A-CF	250.21 a	5.87 a	0.83 a	17.10 a	29,412 bc	4512.48	1500	147,060 c	141.05 b	20.87	17.36
	J-30CF70CRF	202.7 c	7.15 a	0.83 a	18.16 a	37,092 a	7139.66	1500	185,460 a	176.82 a	56.64	47.13
	J-50CF50CRF	201.74 c	6.19 a	0.85 a	16.23 a	33,096 b	6389.04	1500	165,480 b	157.59 b	37.41	31.13
	A-30CF70CRF	234.84 ab	6.80 a	0.83 a	18.04 a	31,164 b	7139.66	1500	155,820 b	147.18 b	27.00	22.47
	A-50CF50CRF	180.47 d	6.17 a	0.86 a	17.21 a	31,452 b	6389.04	1500	157,260 b	149.37 b	29.19	24.29
ANOVA
Years		**	***	***	***	***	-	-	-	***	-	-
Treatments		***	***	**	***	***	-	-	-	***	-	-
Years × Treatments		**	**	***	***	**	-	-	-	*	-	-

Note: Unf, no fertilizer; J-CF, 30% CF in March and 70% CF in July; A-CF, 30% CF in March and 70% CF in August; J-30CF70CRF, 30% CF in March and 70% CRF in July; J-50CF50CRF, 50% CF in March and 50% CRF in July; A-30CF70CRF, 30% CF in March and 70% CRF in August; A-50CF50CRF, 50% CF in March and 50% CRF in August. Values followed by different lowercase letters in the same column indicate significant differences at the 5 % level. *, significance at the 0.05 probability level (p < 0.05); **, significance at the 0.01 probability level (p < 0.01); ***, significance at the 0.001 probability level (p < 0.001).

). The application of the CF in March and the CRF in July at a 3:7 ratio consistently resulted in the highest apple yield and net income for the two years, with yields increasing by 23.7% and 32.9%, compared with those resulting from the application of only the CF in March and July. Although the different CF and CRF treatments resulted in varying fertilizer costs, compared with Unf, J-30CF70CRF maintained the highest net revenue increase rates, which were 22.28% and 47.13%. The application of CRF in August can increase apple yield and revenue, but compared with July, there was no advantage. Compared with Unf, only CF application results in relatively low yield increases, with even a negative net income increase in 2020. Compared with Unf, CRF improved apple quality, resulting in a single-fruit weight increase of 1.21–64.02%, a fruit hardness increase of 9.79–27.22% and a soluble sugar content increase of 4.81–48.89%.

3.2. Soil Nutrient Content

The soil NPK available nutrient content was vertically stratified across the two growth periods, with higher concentrations in the 0–20 cm topsoil than in the 20–40 cm and 40–60 cm subsoils (Figure 1). Compared with the application of only CF and Unf, the combined application of CF and CRF increased soil nutrient contents after the annual apple harvest. The soil pH tended to slightly increase from 5.99 to 7.01, which is suitable for apple growth. Although there were no differences between the treatments, compared with CRF, CF had a lower soil pH with an acidifying effect. Compared with Unf and CF, CRF can increase nutrient accumulation and regulate nutrient distribution, resulting in efficient utilization by apple trees and the potential for nutrient supply. Compared with Unf, J-30CF70CRF had the highest NO₃⁻-N content, with values of 29.53 and 41.19 mg/kg in the two years, increasing by 52.6% and 81.9%, in the 0–20 cm topsoil layer. J-30CF70CRF also maintained higher levels of available K content in the 0–20 cm topsoil, with an average of 144.27 mg/kg after two years. With respect to the NH₄⁺-N and available P contents, J-50CF50CRF achieves higher levels than Unf and CF, with values of 27.61 and 37.37 mg/kg, in the experimental years. In the deeper layers, the NH₄⁺-N and NO₃⁻-N contents were lower than those in the topsoil by 42.3–68.9%, whereas compared with those in the other layers, the NO₃⁻-N, available K, NH₄⁺-N and available P contents were higher in the J-30CF70CRF and J-50CF50CRF soils.

3.3. Soil Nutrient Loss

There was an increase in nutrient loss in the second month after the application of fertilizer, with the trend of the CF being greater than that of the CRF (Figure 2). The losses of NO₃⁻-N, NH₄⁺-N, available P and available K in the CRF treatment were lower than those in the CF combined with CRF and only the CF treatment. During the apple growth period, NO₃⁻-N loss displayed a bimodal distribution, and compared with the CRF treatment, the CF treatment resulted in the greatest NO₃⁻-N loss, which was 18.94 mg/L and 21.96 mg/L at 2 and 5 months, which increased by 55.18% and 19.24%. Compared with that of the same period, the total loss of NO₃⁻-N in the 30CF70CRF treatment decreased by 7.74%. NO₃⁻-N loss increased by 3.58 mg/L as the CF application rate increased from 30% to 50%. Although NH₄⁺-N loss also displayed a bimodal distribution, the leaching concentration was less than 0.21 mg/L. The loss of available P and available K displayed a highly unimodal distribution 6 months later, with obvious differences over the 8-month experimental period. The average P and K losses of 50CF50CF and 30CF70CF were 2.54 mg/L and 35.96 mg/L, the highest of which reached 13.91 mg/L and 170.13 mg/L. The application of CRF resulted in a lower loss content, decreasing the loss of available P and K by 17.12% and 18.29%, for 30CF70CRF. These results indicate that the combination of CF and CRF can optimize nutrient release while decreasing nutrient loss to minimize environmental pollution.

3.4. NH₃ Emissions

NH₃ emissions in the CF were higher than those in the CRF, with an increase of 163.36% during the whole apple growth period (Figure 3), and 30CF70CRF decreased by 4.19% compared with 50CF50CRF. NH₃ emissions peaked at 2 months after the application of fertilizer, and the concentration of CF reached 6.59 kg ha⁻¹ d⁻¹, which was higher than that of CRF, 30CF70CRF, 50CF50CRF, 30CF70CF and 50CF50CF by 3.16-, 1.60-, 1.39-, 1.57- and 1.46-fold. Although there were no differences when CF or CRF was applied twice, the NH₃ emissions of CRF and 30CF70CRF were stable, with a low coefficient of variation of 18.3–22.5%. Overall, the combined application of CF and CRF can decrease NH₃ emissions while balancing loss reduction and practicality, offering a sustainable strategy for minimizing N loss in apple orchards.

3.5. Microbial Diversity Under CRF Fertilization

Compared with Unf, the combined application of CF and CRF increased the bacterial and fungal Chao1, Shannon and Observed_species indices (p < 0.05), while there was no effect on the Simpson index (Figure 4a,b). The Chao1 indices of bacteria and fungi also increased (p < 0.05) with the application of CRF compared with those with the application of only CF, and the observed species index increased for fungi. Soil NO₃⁻-N, available K and pH were the key factors driving microbial diversity, and the bacterial Shannon index increased in soils with high NO₃⁻-N contents (p < 0.01) (Figure 4c). There was no obvious effect of soil nutrient content on fungal diversity (p > 0.05). The results of nonmetric multidimensional scaling (NMDS) further suggested that there was clear separation of microbial communities among Unf, 30CF70CF and 30CF70CRF, indicating that CRF regulated bacterial and fungal diversity (p < 0.05) (Figure 4d). Compared with 30CF70CF, CRF can increase bacterial β diversity but can decrease fungal β diversity without obvious effects (Figure 4e).

3.6. Microbial Community Succession Under CRF Fertilization

The combined application of CF and CRF drove distinct succession of soil bacterial and fungal communities (Figure 5a,b). The relative abundance of Actinobacteria, Chloroflexi and Firmicutes in 30CF70CRF was greater than that in Unf and 30CF70CF, whereas the abundance of Proteobacteria was lower. The application of CF or CRF decreased the abundance of Ascomycota in fungi, whereas the abundances of Basidiomycota and Mortierellomycota were greater than those in Unf and 30CF70CF. There was no obvious effect on soil nutrient properties or bacteria or fungi, but the relative abundances of Actinobacteria, Chloroflexi, Firmicutes and Basidiomycota were positively correlated with soil nutrients, while those of Proteobacteria and Ascomycota were negatively correlated (Figure 5c). Taxonomic dendrogram analysis indicated that 30CF70CRF was enriched in key functional taxa involved in nutrient cycling with N fixing bacteria and P solubilizing fungi (Figure 5d), including Caenimonas, Cupriavidus, Cunninghamellaceae, Gibellulopsis and Cunninghamella. Compared with Unf and 30CF70CF, 30CF70CRF is the main driver of microbial community succession, with 30CF70CRF forming a unique and stable microbial community structure.

3.7. Response of Soil Functional Gene Abundance to CRF Fertilization

The combined application of CF and CRF increased the abundance of functional genes related to N and P cycling (Figure 6a). The relationships between N and P cycling gene abundances and Unf, 30CF70CF and 30CF70CRF indicated that the N cycling genes of nirK1 and the P cycling gene of pqqC in 30CF70CRF were most strongly divergent from those in Unf (Figure 6b). These results indicated that CRF can drive the enrichment of nutrient cycling genes. 30CF70CRF was distinctly separated from Unf and 30CF70CF along the MDS1 axis, indicating a unique functional gene community structure shaped by the combined application of CRF and CF (Figure 6c). The greatest relative abundance of all the tested N and P cycling genes was detected for 30CF70CRF (Figure 6d,e). The relative abundances of the UreC, nosZ1, gdh, nifH, amoA2, cphy, ppk3, pqqC and ppx genes in 30CF70CRF were clearly greater than those in Unf and 30CF70CF. The abundance of N- and P-cycling functional genes increased with increasing bacterial and fungal α diversity (Figure 6f).

3.8. Effect to Yield from PLS–PM

The relationships between treatment, the microbiome, functional genes, soil nutrient content, soil nutrient loss, NH₃ emissions and apple yield were elucidated using PLS–PM analysis (Figure 7). Treatment had a direct positive effect (p < 0.05) on the bacterial and fungal communities (Figure 7a,b). Compared with fungi, bacteria were more closely correlated with N and P cycling genes. N cycling genes and soil nutrient content had direct effects (p < 0.05) on apple yield, whereas the effects of soil nutrient content on soil nutrient loss and NH₃ emissions were negative as N cycling genes mediated. The standardized total effects indicated that soil nutrient content, treatment, and N and P cycling genes contributed the most to apple yield (Figure 7c).

4. Discussion

4.1. Combined Application of CF and CRF Increases Apple Yield and Decreases Nutrient Loss

The growth of apple trees requires an adequate and precise nutrient supply during the germination and fruit growth periods to ensure maximum production efficiency. The CRF in this study had a sigmoid nutrient release curve with three stages: slow, accelerated and stable (Figure S1) [24,25]. CRFs can cumulatively release 72.94%, 61.15% and 67.01% of N, P and K, over a period of 3 months, achieving a stable and continuous supply of nutrients before fruit ripening and harvesting. 30CF70CRF achieved the highest apple yield and net income, as the CF can rapidly release nutrients to meet the needs of new organ differentiation in March, while CRF will continuously supply a large amount of nutrients to support fruit development in July [19]. The combined application of CRF and CF also increased the yields of tomato, rice and potato to achieve greater economic benefits, as reported by Qu et al. (2020), Souza et al. (2020) and Hou et al. (2020) [26,27,28]. The yield increase observed was attributed to the synergistic effects of CF and CRF, which increased with increasing single-fruit quality. The raw cost of the combined application of CF and CRF increases as the cost of the coating materials and coating processes increase, but the economic returns from higher yields surpass the fertilizer costs.

The application of fertilizers not only aims to increase crop yields but also aims to achieve the goal of balancing soil nutrients [29,30]. In this study, the application of CRF and CF achieved nutrient regulation to optimize the nutrient supply to meet apple growth requirements and decrease loss and NH₃ emissions in the context of CRF [31]. 30CF70CRF decreased the loss of NO₃⁻-N, NH₄⁺-N, P and K in the soil by 7.04–8.19%, 11.59–25.15%, 12.96–17.12% and 16.05–18.29% (Figure 1), but compared with only the application of CF during the experimental years, the losses increased by 4.41–13.34%, 34.11–47.39%, 16.82–37.04% and 9–24.68%. Optimizing N fertilizer application can optimize soil NO₃⁻-N and NH₄⁺-N contents to increase the productivity of apple trees [32,33], and the soil P and K contents also increase. The effectiveness of P easily decreased because P was fixed with certain cations in soil through the formation of insoluble complexes [34]. Applying the corresponding rate of P during apple growth periods is important. The long-term application of fertilizer decreases soil pH [35], and a suitable soil pH still range for apple growth is maintained for a short-term experimental year [36]. NH₃ emissions are mainly lost from N in the field [37,38], but they are affected by soil type, climatic conditions and fertilization practices. Here, the combination of CF and CRF decreased NH₃ emissions by 17.98–44.86% as CRF delayed nutrient release and met crop needs.

4.2. Combined Application of CF and CRF Regulates Soil Microbial Communities and Functions

Soil microbial communities and functions can drive nutrient cycling and crop productivity, which mediate the transformation, retention, and supply of soil nutrients [14]. In this study, the combined application of CF and CRF regulated the soil microbial communities and optimized their functions, which was associated with the synergistic nutrient supply of the two fertilizers. The relative abundances of Actinobacteria, Chloroflexi and Firmicutes in 30CF70CRF increased compared with those in Unf and only CF, enhancing the decomposition and transformation of soil organic nutrients to increase nutrient availability for apple trees [39]. Soil nutrient content was the main driver of microbial communities, and microbial diversity was affected by soil NO₃⁻-N, available K and pH, while soil nutrients also directly affected key functional taxa. The combined application of CF and CRF enriched the functional taxa of N- and P-cycling genes, such as UreC, nosZ1, gdh, nifH, amoA2, cphy, ppk3, pqqC and ppx. These changes form a unique and stable microbial functional network, mediating efficient nutrient cycling. This microbial regulatory mechanism further links soil nutrient supply to apple nutrient uptake, making it an important biological pathway for combined applications to achieve high yields and environmental friendliness [40].

5. Conclusions

A two-year field experiment was conducted to study the effects of combined application of CF and CRF on apple productivity and soil nutrient and microbial parameters. The combined application of CF and CRF can increase apple yield and quality, achieving the greatest net economic benefits by applying 30% common fertilizer in March and 70% controlled-release fertilizer in July (J-30CF70CRF) to 34,890 kg/ha and 165,810.3 ¥/ha. Compared with the combined application of only CF, the combined application of CF and CRF decreased the loss of NO₃⁻-N, NH₄⁺-N, P and K by 4.45–8.19%, 11.59–29.81%, 8.28–17.12% and 10.73–18.29%, while it decreased NH₃ emissions by 17.98–44.86%. A suitable soil pH level can also be maintained to ensure apple tree growth through the use of CF and CRF. The synergistic nutrient supply of CF and CRF increased bacterial and fungal alpha diversity and enriched beneficial functional taxa, thereby upregulating the abundance of N- and P-cycling genes and achieving efficient nutrient transformation and supply. This study provides a scientific basis for precision nutrient management in the apple orchards on the Jiaodong Peninsula and provides a scalable method for the green and sustainable development of similar regions with comparable soil and climatic conditions.