Effects of different intercropping patterns on soil organic carbon content, enzyme activity, and aggregate stability in Apple orchards on the loess Plateau

Abstract

This study aimed to investigate the impacts of distinct intercropping patterns on soil quality in apple orchards located on the Loess Plateau. Despite the recognized benefits of intercropping, systematic research on apple-onion and apple-rapeseed systems remains limited, particularly regarding their combined influence on soil organic carbon fractions, aggregate stability, and enzyme activities under arid conditions. Four distinct planting configurations were implemented in the orchard: conventional tillage between rows (CT), natural grassing between rows (NG), intercropping onions between rows (IA), and intercropping rapeseed between rows (IR). A three-year field trial was conducted to thoroughly examine alterations in soil physical characteristics, carbon and nitrogen concentrations, enzymatic activities, and aggregate stability. Our findings demonstrate that, enhance soil quality. Notably, the NG, IA, and IR treatments notably increased soil organic matter content (up to 40.04% under IA) and a reduction in soil bulk density (13.19% under IA) within the top 0–20 cm soil layer. The IA treatment recorded the highest levels of total organic carbon and its fractions, likely due to the dense root system of onions enhancing microbial activity and organic matter input. Concurrently, the IR prominently bolstered the mechanical and hydraulic stability of soil aggregates (e.g., 160.72% increase in WR0.25), thus enhancing overall soil structural integrity. Furthermore, various intercropping configurations differentially influenced soil enzymatic activity, with NG demonstrating optimal β-glucosidase (βG) and cellobiohydrolase (CBH) activities, IA excelling in N-acetyl-β-D-glucosaminidase (NAG) and leucine aminopeptidase (LAP) activities, and IR standing out in xylanase activity. This study provides the first comprehensive evaluation of intercropping effects on soil carbon sequestration and structural stability in arid apple orchards, offering novel insights into optimizing sustainable soil management practices. In summary, the IA and IR intercropping models demonstrate a pronounced capacity to improve orchard soil quality, highlighting their potential as a scientific foundation for enhancing soil fertility and ensuring sustainable development of dryland apple orchards in ecologically fragile regions.

Introduction

The Northwest Loess Plateau, characterized by a typical temperate semi-arid climate with distinct continental monsoon features, faces numerous agricultural challenges due to its low annual precipitation and inherent soil nutrient deficiencies, thereby limiting the development of local agriculture^1,2. Despite these limitations, the dryland orchard industry has expanded rapidly in recent years, emerging as a pivotal economic driver in the region³. However, traditional orchard management practices, such as extensive clean tillage and heavy reliance on inorganic fertilizers and pesticides, have inadvertently worsened issues like soil erosion and nutrient leaching, ultimately leading to soil fertility degradation, stunted fruit tree growth, yield reduction, and compromised quality⁴. Recognizing these challenges, the Chinese government has prioritized agricultural sustainability, fostering global interest in green, practical, and cost-effective apple orchard management strategies, particularly intercropping^5,6.

Intercropping, the practice of cultivating multiple crops in the same field, is a widely used agricultural strategy^7,8. Research indicates that this approach can enhance soil moisture, enzyme activity, and fertility, while mitigating the effects of soil-borne pests and diseases, thereby improving soil quality and crop yield and quality^9,10. For instance, Yang et al. (2019)¹¹ found a significant increase in apple root and white clover biomass in an apple-white clover intercropping system. Similarly, Zhang et al. (2019)⁵ reported that apple-rye intercropping elevated soil levels of total organic carbon and soluble organic carbon, enhancing soil fertility and quality. Allium fistulosum L., a perennial herb of the Amaryllidaceae family, is nutritionally beneficial due to its protein, carbohydrate, vitamin, and mineral content. Brassica napus L., known as rapeseed, is a cruciferous plant cultivated for its edible oil and vegetables. Studies have shown that sunflower and onion intercropping can notably boost rhizosphere soil phosphatase activity and enhance sunflower’s average dry biomass and phosphorus uptake¹². Zhou et al. (2019)¹³ also demonstrated that intercropping Chinese milk vetch and rapeseed could enhance soil moisture, nitrogen content, and microbial diversity, while effectively managing the soil nitrogen pool. However, research specifically focusing on apple-onion and apple-rapeseed intercropping systems are still limited—particularly those that quantitatively assess their combined effects on soil organic carbon components, aggregate stability, and soil enzyme activities under arid and semi-arid conditions.

Soil aggregates are the basic structural units of soil structure, and their resilience to external environmental perturbations is a critical measure of soil quality and health, encompassing water stability, mechanical stability, chemical stability, biological stability, and pH stability¹⁴. The quantity and size distribution of soil aggregates influence both crop growth and a range of soil physical, chemical, and biological processes. Internal properties such as soil texture and moisture can lead to the degradation of aggregate structure, resulting in ‘invisible’ soil degradation¹⁵. Extensive research has highlighted the pivotal role of soil aggregates in the transformation of organic carbon¹⁶. Aggregates physically shield organic carbon from microbial contact, slowing its mineralization¹⁷. Moreover, the content of macroaggregates is closely linked to soil organic carbon levels¹⁸. Microaggregates, rich in colloids and inorganic substances, form tight bonds with organic carbon, providing a stable sequestration site¹⁹. Agricultural practices including planting patterns, tillage, and fertilization strategies, have a significant impact on soil aggregate formation and organic carbon content. In intercropping systems, crop diversity introduces variability in root exudates and residue inputs, impacting soil aggregate development and organic carbon dynamics, and thus the characteristics of soil aggregates and organic carbon pools^20–22. While intercropping is known to enhance soil aggregate stability and organic carbon content²³the specific effects of different intercropping patterns on soil aggregate organic carbon remain unclear. Water is a key factor in aggregate fragmentation, and the distribution of water-stable aggregates is commonly used to assess aggregate quality^24,25. However, in the drylands of Longdong, where rainfall is absent for 9–10 months annually and irrigation is lacking over 80% of the area, the soil remains dry for extended periods. This study, conducted at the Pingliang Comprehensive Experimental Station of the National Apple Industry System in Jingning County, Gansu Province, involves a three-year intercropping model positioning test to systematically analyze changes in soil enzyme activity, aggregate structure, and organic carbon content under various intercropping patterns. We hypothesize that, compared with conventional tillage (CT), intercropping patterns (IA and IR) will significantly improve soil quality in arid apple orchards by increasing soil organic carbon content, enhancing aggregate stability, and boosting enzyme activities. Specifically, apple-onion intercropping (IA) may enhance microbial activity and organic matter input through the dense root system of onions, while apple-rapeseed intercropping (IR) may significantly improve soil stability by enhancing soil structure and water stability. The findings aim to elucidate the mechanisms of soil organic carbon sequestration and inform the adoption of intercropping practices in apple orchards across Northwest China.

Materials and methods

Description of study site

The experimental site, situated at the Pingliang Comprehensive Experimental Station of the National Apple Industry System in Jingning County, Pingliang City, Gansu Province (35.40° N, 105.72° E; 3.3 hectares), is at an altitude of 1,561 m. The area experiences an annual precipitation of 450 mm, primarily from July to September, and an annual evaporation rate of 1,531 mm, with a dryness index of 2.53. The region enjoys 2,238 h of annual sunshine, has a frost-free period of 159 days, and an average annual temperature of 7.1 °C, characteristic of the typical rain-fed agricultural regions of the Northwest Loess Plateau. The soil is loessial, silty loam with a pH range of 7.6 to 8.9. It features a deep soil layer, low organic matter content, poor deep soil water regulation, weak aggregation, and is prone to erosion, leading to significant soil erosion.

The experimental materials were 7-year-old apple trees in the same orchard, the variety was “ Liquan Duanfu ”, the dwarfing interstock was M9T337, the base stock was Malus robusta, and the row spacing was 2 m × 4 m. The experimental design included four treatments: conventional tillage (CT), natural grass (NG), intercropping with Allium fistulosum L. (IA), and intercropping with Brassica campestris L. (IR). Prior to the experiment, the plots were plowed to eliminate row-space weeds. Throughout the experiment, no weeding or irrigation was conducted. Weeds within a 0.5 m strip adjacent to the tree rows were removed. In the CT treatment, complete rotary tillage was conducted to a depth of 20 cm using a rotary cultivator three times per year (March, June, and September) to maintain a bare soil surface. No cover crops or vegetation were allowed to grow. The NG treatment area was maintained at a 3 m width. Grass was mowed and the clippings were left in place when it reached over 40 cm in height, with the stubble height kept between 5 and 10 cm. Allium fistulosum L. was broadcast-sown at a density of approximately 160 plants/m² in a 3 m wide band. It was sown in early August, mowed and covered before winter, and this practice was repeated for three consecutive years. Winter oilseed rape (Brassica campestris L.) was sown in rows at a density of about 40 plants/m² with a 3 m wide band. Sown in early August, it is cold and saline-alkali tolerant, allowing it to thrive in the local environment. In late April of the following year, it was mowed and the biomass was incorporated into the soil for three consecutive years. All treatments involved the timely removal of pernicious weeds, such as deep-rooted wheatgrass and high-dry wild castor. Trees with similar crown dimensions and uniform growth (crown diameter 1.8–2.2 m, tree height 2.5–3.0 m, trunk height 0.8–1.0 m) were selected. Each treatment was replicated 30 times, with each tree serving as a single replicate. Protective rows were established between treatments, and cultivation management practices were uniform across the experiment.

Test methods and determination.

In September 2021, the aboveground vegetation and surface litter were meticulously cleared from the soil surface. Soil samples from the 0–20 cm and 20–40 cm layers were collected at five randomly selected points within each plot (20 m × 15 m). The samples were then homogenized to form approximately 1 kg of soil samples, which were placed in a rigid container for transportation back to the lab for air-drying. During air-drying, animal and plant debris were removed, and the soil was crushed into small blocks approximately 1 cm in diameter along natural fissures. Once adequately dried, the samples were prepared for soil aggregate analysis. Soil aggregates were separated using both dry and wet sieving techniques, following standardized procedures^26,27. For dry sieving, approximately 100 g of air-dried soil was placed on a nest of sieves with mesh sizes of 2 mm, 0.5 mm, and 0.25 mm, and sieved for 2 min at an oscillation frequency of 60 Hz using a mechanical shaker (Retsch AS200). For wet sieving, 50 g of soil aggregates (2–0.25 mm fraction) was placed on the same sieve set and submerged in deionized water for 5 min to pre-wet. Then, the samples were oscillated vertically for 30 min at 30 cycles per minute (cpm) using a Yoder-type wet-sieving apparatus. Aggregates retained on each sieve were then carefully collected, dried at 40 °C, and weighed to determine the distribution of macro-aggregates (> 2 mm, 0.5–2 mm, 0.25–0.5 mm) and micro-aggregates (< 0.25 mm). For each treatment, three independent samples of soil aggregates were analyzed, and the average values of the measurements were used for further statistical analysis. Concurrently, soil samples were analyzed for mechanical composition using a laser particle size analyzer (Mastersizer 2000, Malvern Instruments, Malvern, England), with each sampling point being repeated three times.

Aggregate stability was assessed using the mean weight diameter (MWD), geometric mean diameter (GMD), percentage of aggregate decomposition (PAD), and water stability coefficient (WSC). For aggregates larger than 0.25 mm, the index R_0.25 is utilized for evaluation. The calculation method for this index is as follows:

1

Mean weight diameter (MWD):

2

In the formula, MWD is the average mass diameter, mm; X_i is the average pore size of the i th pore size sieve and the i + 1 th pore size sieve, mm; i is the mass percentage of dried aggregates on the sieve of pore size I; mean geometric diameter (GMD):

3

X_i is the median diameter of aggregates in group i, and Wi is the content of aggregates in group i;

Aggregate destruction rate (PAD):

4

Water Stability Coefficient (WSC):

5

Soil organic carbon (SOC) was determined by the potassium dichromate external heating method after air-drying of the screened soil aggregates²⁸.

Soil sampling

When assessing soil organic carbon and carbon metabolism enzyme activity, soil samples were collected as follows: the aboveground vegetation and surface litter were meticulously cleared. Samples from the 0–20 cm and 20–40 cm soil layers were obtained for each plot (20 m × 15 m), with five random sampling points selected per plot. The collected samples were homogenized to form a composite sample of approximately 1 kg. These samples were sealed in plastic bags, refrigerated during transport, and promptly returned to the laboratory for analysis. A portion of the samples was stored at 4 °C for the assessment of active organic carbon, while the remainder was air-dried, ground, and sieved for total soil organic carbon determination. The soil organic (SOC) matter was gauged through the oxidation process with potassium dichromate²⁹. Total organic carbon (TOC) and light fraction organic carbon (LFOC) were measured using the potassium dichromate-external heating method³⁰. The refractory organic carbon (ROC) was quantified by the potassium dichromate oxidation method³¹. Microbial biomass carbon (MBC) was assessed by the chloroform fumigation-extraction method with a fumigation time of 24 h, followed by extraction with 0.5 M K₂SO₄ and applying a kEC conversion factor of 0.45³³. Dissolved organic carbon (DOC) was extracted using deionized water at a 1:5 (w/v) ratio and measured via TOC analyzer. Particulate organic carbon (POC) was isolated using 5 g/L sodium hexametaphosphate (Na₆(PO₃)₆) as a dispersing agent, followed by wet sieving through a 53 μm mesh³³and soil enzyme activities were analyzed via the microplate fluorescence method³⁴.

In this study, the soil sample amount for each treatment was approximately 1 kg. The data for soil organic carbon (SOC), total organic carbon (TOC), refractory organic carbon (ROC), and particulate organic carbon (POC) are expressed as content (g/kg). The data for dissolved organic carbon (DOC) and microbial biomass carbon (MBC) are expressed as content (mg/kg).

6

Data processing.

Data collation was conducted using Excel 2019, while statistical analysis was performed using SPSS 21.0. A one-way analysis of variance (ANOVA) followed by the LSD test was employed to assess the significance of soil indices at the 0.05 level for the 0–20 cm and 20–40 cm layers under various treatments. Pearson correlation analysis was applied to examine relationships among soil texture, physical properties, active organic carbon components, aggregate composition, and soil enzyme activity. The data presented in the figures are expressed as mean ± standard deviation.

Principal Component Analysis (PCA) was conducted to comprehensively evaluate the effects of different intercropping treatments on soil quality. Prior to PCA, all soil indicators (including SOC, aggregate stability, and enzyme activities) were standardized using the formula:

where X_ij​ is the value of the jth indicator for the ith sample, Xˉ_j​ is the mean of the jth indicator, and S_j​ is the standard deviation of the jth indicator.

The PCA was performed using SPSS 21.0 software. The correlation coefficient matrix was calculated, and principal components with eigenvalues greater than 1 were retained. Three principal components were extracted, accounting for 52.684%, 34.972%, and 11.078% of the variance, respectively, with a cumulative variance contribution of 98.914%. The scores for each principal component were calculated based on the factor loadings, and a comprehensive score for each treatment was derived using the formula:

where F₁​, F₂​, and F_3​ are the scores of the first, second, and third principal components, respectively.

Results and discussion

Effects of different intercropping patterns on soil physical properties

Table 1 indicates that in the 0–20 cm soil layer, the soil organic matter content for the NG, IA, and IR treatments exceeded that of the CT treatment to varying extents, with the ON treatment also showing higher organic matter content than the CT treatment(P < 0.05). Among them, the IA treatment showed the highest increase in soil organic matter content, reaching 40.04%. Conversely, the soil pH and bulk density in the NG, IA, and IR treatments were reduced compared to the CT treatment. Notably, the IR and IA treatments exhibited a significant reduction in soil bulk density, with decreases of 11.81% and 13.19%, respectively, compared to the CT treatment(P < 0.05). Varied intercropping patterns induced changes in the proportions of clay, silt, sand, and physically bound clay within the orchard soil. In both the 0–20 cm and 20–40 cm soil layers, the NG treatment had the highest ratios of clay, silt, and physically bound clay, which were 1.03 ( 1.09 ) times, 1.05 ( 1.04 ) times and 1.08 ( 1.09 ) times of CT treatment, respectively—while the sand ratio was the lowest, at only 83.33% and 82.39% of the BF treatment(P < 0.05).

Table 1.

Changes of soil physical properties in the vertical direction of Gansu Longdong orchard under different intercropping treatments.

Indicator	Soil layer (cm)	CT	NG	IA	IR
Organic matter content/ SOC(g/kg)	0–20	14.26 ± 0.017d	15.64 ± 0.014c	19.97 ± 0.046a	17.12 ± 0.018b
	20–40	10.82 ± 0.057c	13.87 ± 0.085a	12.29 ± 0.077b	12.65 ± 0.082b
pH	0–20	8.86 ± 0.091a	8.78 ± 0.102a	8.62 ± 0.115a	8.58 ± 0.223a
	20–40	8.88 ± 0.099a	8.80 ± 0.145a	8.60 ± 0.167a	8.55 ± 0.198a
Soil bulk density/ (g/cm3)	0–20	1.44 ± 0.015a	1.40 ± 0.011a	1.25 ± 0.009b	1.27 ± 0.011b
	20–40	1.46 ± 0.014a	1.41 ± 0.011a	1.43 ± 0.012a	1.39 ± 0.015a
Clay content/ (%)	0–20	10.71 ± 0.093a	11.01 ± 0.068a	10.53 ± 0.089a	10.72 ± 0.092a
	20–40	10.78 ± 0.085b	11.73 ± 0.044a	11.37 ± 0.049a	10.83 ± 0.039b
Silt content/ (%)	0–20	68.54 ± 0.881b	71.65 ± 0.356a	68.28 ± 0.897b	68.17 ± 0.882b
	20–40	69.16 ± 0.954b	71.74 ± 0.415a	69.75 ± 0.895b	69.15 ± 0.913b
Sand content/ (%)	0–20	20.81 ± 0.321a	17.34 ± 0.126b	21.26 ± 0.315a	21.11 ± 0.308a
	20–40	20.05 ± 0.071a	16.52 ± 0.034b	18.89 ± 0.055a	20.03 ± 0.049a
Physical clay particles/ (%)	0–20	47.09 ± 0.056b	50.66 ± 0.025a	46.89 ± 0.034b	46.74 ± 0.039b
	20–40	47.51 ± 0.363b	51.71 ± 0.152a	48.11 ± 0.214b	47.14 ± 0.285b

Note: There was a significant difference in the 0.05 level between different lowercase letters in the same peer (P < 0.05). CT, clean tillage between rows; NG, inter-row natural grass; IA, inter-row onion; IR, inter-row rape.

Effects of different intercropping patterns on soil active organic carbon fractions

Different intercropping patterns significantly influenced soil organic carbon content. Soil organic carbon (TOC), dissolved organic carbon (DOC), resistant organic carbon (ROC), particulate organic carbon (POC), heavy fraction organic carbon (HFOC), light fraction organic carbon (LFOC), and microbial biomass carbon (MBC) levels decreased with increasing soil depth across all intercropping patterns (Fig. 1). In the 0–20 cm layer, the IA treatment exhibited the highest levels of SOC, DOC, ROC, POC, HFOC, and LFOC, surpassing the CT treatment by 40.02%, 3.38%, 16.87%, 32.99%, 31.79%, and 90.35%, respectively (P < 0.05). Meanwhile, the MBC content was highest in the NG treatment at 87.02 mg·kg^− 1, which was 4.29% greater than that in the CT treatment (P < 0.05). In the 20–40 cm layer, the NG, IA, and IR treatments exhibited significantly higher levels of TOC, POC, ROC, and MBC compared to the CT treatment. However, the IA and IR treatments had significantly lower contents of DOC and LFOC than the CT treatment, with reductions of 11.40-17.91% and 10.14–21.74%, respectively (P < 0.05).

Fig. 1.

Effects of different intercropping patterns on soil active organic carbon fractions. CT, clean tillage between rows (conventional tillage without intercropping); NG, inter-row natural grass; IA, inter-row onion; IR, inter-row rape. Different uppercase letters above the histogram indicate that the difference between different soil layers of the same treatment reaches a significant level (P < 0.05), and different lowercase letters indicate that the difference between different treatments of the same soil layer reaches a significant level (P < 0.05). A TOC; B DOC; C ROC; D POC; E HFOC; F LFOC; G MBC.

Effects of different intercropping patterns on soil mechanical aggregate size fractionation

The impact of various intercropping patterns on soil mechanical aggregate size distribution is depicted in Fig. 2. Within the 0–20 cm soil layer (Fig. 2A), the shares of > 2 mm and 0.5-2 mm soil mechanical aggregates were markedly reduced in the NG, IA, and IR treatments compared to the CT treatment, with reductions of 12.04–46.52% and 4.55–32.42%, respectively (P < 0.05). In contrast, the NG and IA treatments exhibited a 29.78% and 68.20% increase in the proportion of 0.25–0.5 mm soil mechanical aggregates relative to the CT treatment, while the IR treatment showed a 32.92% decrease in this aggregate size, all statistically significant (P < 0.05). The proportions of < 0.25 mm soil mechanical aggregates were significantly elevated in the NG, IA, and IR treatments by 23.37%, 140.58%, and 46.35% compared to the CT treatment, respectively (P < 0.05). In the 20–40 cm soil layer (Fig. 2B), soil mechanical aggregate content diminished with decreasing soil particle size, with the > 2 mm fraction being the most abundant. The respective proportions for the CT, NG, IA, and IR treatments were 42.55%, 36.74%, 31.86%, and 36.62%. Conversely, the smallest proportion was observed in the 0.25–0.5 mm aggregate size, with the CT, NG, IA, and IR treatments recording 5.77%, 10.12%, 11.11%, and 9.62%, respectively.

Fig. 2.

Effects of different intercropping patterns on particle size classification of soil mechanical aggregates. CT, clean tillage between rows (conventional tillage without intercropping); NG, inter-row natural grass; IA, inter-row onion; IR, inter-row rape. Different uppercase letters above the histogram indicate that the difference between different soil layers of the same treatment reaches a significant level (P < 0.05), and different lowercase letters indicate that the difference between different treatments of the same soil layer reaches a significant level (P < 0.05). A Particle size classification of mechanical aggregates in 0–20 cm soil layer; B Particle size classification of mechanical aggregates in 20–40 cm soil layer.

Effects of different intercropping patterns on soil Water-Stable aggregate size fractionation

The influence of various intercropping systems on the distribution of water-stable soil aggregate sizes is illustrated in Fig. 3. Within the 0–20 cm soil layer (Fig. 3 A), water-stable aggregate content rose as particle size diminished. Notably, for the 0.5–2 mm and 0.25–0.5 mm fractions, the IR treatment had the highest water-stable aggregate content, which was 2.35 and 2.69 times greater than that of the CT treatment, respectively (P < 0.05). In contrast, the IR treatment had the lowest water-stable aggregate content in the < 0.25 mm fraction, constituting only 89.03% of the CT treatment’s level. In the 20–40 cm soil layer (Fig. 3B), a comparable pattern emerged, with water-stable aggregate content increasing as particle size decreased. The IA treatment featured the highest content of 0.5–2 mm water-stable aggregates at 5.75%, significantly exceeding the CT treatment’s level (P < 0.05). In the 0.25–0.5 mm category, the IR treatment’s water-stable aggregate content peaked at 8.51%, which was 2.16 times that of the CT treatment. Although the NG treatment exhibited the highest content of < 0.25 mm water-stable aggregates at 93.43%, this increase was not statistically significant compared to the CT treatment. Conversely, the IR treatment had the lowest content in this fraction, at 95.37% of the CT treatment.

Fig. 3.

Effects of Different Intercropping Patterns on Soil Water-Stable Aggregate Size Fractionation. CT, clean tillage between rows (conventional tillage without intercropping); NG, inter-row natural grass; IA, inter-row onion; IR, inter-row rape. Different uppercase letters above the histogram indicate that the difference between different soil layers of the same treatment reaches a significant level (P < 0.05), and different lowercase letters indicate that the difference between different treatments of the same soil layer reaches a significant level (P < 0.05). A Particle size classification of water-stable aggregates in 0–20 cm soil layer; B Particle size classification of water-stable aggregates in 20–40 cm soil layer.

Effects of different intercropping patterns on soil aggregate stability

DR_0.25, the mean weight diameter (DMWD), and the geometric mean diameter (DGMD) of mechanically stable aggregates can characterize the mechanical stability of aggregates (Fig. 4A, C, D). In the 0–20 cm soil layer, the IR treatment exhibited the highest DR0.25 at 79.13%, surpassing the CT treatment by 17.14% (P < 0.05). The IA treatment achieved the highest MWD and GMD values, recording 1.23 and 0.95, respectively, which were 7.89% and 17.28% greater than the CT treatment’s values. Within the 20–40 cm soil layer, the IR treatment maintained the highest DR_0.25 at 84.10%, whereas the NG treatment had the highest MWD and GMD, reaching 1.31 and 1.02, respectively. The water stability of aggregates is assessed by the proportion of water-stable aggregates larger than 0.25 mm (WR_0.25), the mean (MWD), and geometric (GMD) mean diameters of these aggregates. In the 0–20 cm layer, the IR treatment’s WR_0.25, MWD, and GMD were notably higher than the CT treatment’s, with increases of 160.72%, 11.11%, and 7.69%, respectively (P < 0.05) (Fig. 4B, E, F). In the 20–40 cm layer, the IR treatment’s WR_0.25 was the highest at 11.23%, and the IA treatment’s MWD and GMD were the highest, at 0.31 and 0.28, respectively, both significantly exceeding the CT treatment’s values. These findings suggest that both the IR and IA treatments significantly enhance the mechanical and water stability of soil aggregates, thereby improving soil structure.

Fig. 4.

Differences in stability of surface soil aggregates under different intercropping patterns. CT, clean tillage between rows (conventional tillage without intercropping); NG, inter-row natural grass; IA, inter-row onion; IR, inter-row rape. Different uppercase letters above the histogram indicate that the difference between different soil layers of the same treatment reaches a significant level (P < 0.05), and different lowercase letters indicate that the difference between different treatments of the same soil layer reaches a significant level (P < 0.05). A DR_0.25; B WR_0.25; C DMWD; D DGMD; E WMWD; F WGMD.

The percentage of aggregate decomposition (PAD) and the water stability coefficient (WSC) serve as key indicators of aggregate stability, where a lower PAD and a higher WSC denote greater aggregate stability. As depicted in Fig. 5A, within the 0–20 cm soil layer, the IR treatment exhibited the lowest PAD at 0.79%, which was significantly reduced by 12.22% compared to the CT treatment (P < 0.05). In the 20–40 cm layer, both the IA and IR treatments demonstrated significantly lower PAD values than the CT treatment, with reductions of 4.40% and 5.49%, respectively (P < 0.05). Figure 5B illustrates that, in both soil layers, the IA and IR treatments achieved significantly higher WSC values than the CT treatment (P < 0.05). Notably, the IR treatment achieved the highest WSC values, exceeding the CT treatment’s values by 119.19% in the 0–20 cm layer and by 57.63% in the 20–40 cm layer. These results indicate that both the IA and IR treatments provide effective enhancement of topsoil stability in the orchard.

Fig. 5.

Differences in the destruction rate and water stability coefficient of surface soil aggregates under different intercropping patterns. CT, clean tillage between rows (conventional tillage without intercropping); NG, inter-row natural grass; IA, inter-row onion; IR, inter-row rape. Different uppercase letters above the histogram indicate that the difference between different soil layers of the same treatment reaches a significant level (P < 0.05), and different lowercase letters indicate that the difference between different treatments of the same soil layer reaches a significant level (P < 0.05). A PAD; B WSC.

Effects of different intercropping patterns on soil enzyme activities

Figure 6 illustrates the impact of various intercropping patterns on soil enzyme activities. Within the 0–20 cm soil layer, the NG treatment displayed the highest β-glucosidase (βG) and cellobiohydrolase (CBH) activities, recording 20.42 nmol·g⁻¹·h⁻¹ and 3.09 nmol·g⁻¹·h⁻¹, respectively. These values were 69.04% and 104.64% greater than those observed in the CT treatment (Figure. 6B, E). The IA treatment showed the most pronounced activities for N-acetyl-β-D-glucosaminidase (NAG) and leucine aminopeptidase (LAP), surpassing the CT treatment by 13.27% and 19.82%, respectively (P < 0.05) (Fig. 6C, D). Conversely, the IR treatment had the highest xylanase (β-XYL) activity, 1.56 times greater than the CT treatment (P < 0.05). In the 20–40 cm soil layer, the NG treatment again exhibited the highest activities for urease, βG, NAG, LAP, and CBH, with respective values of 116.08 µg·g⁻¹·h⁻¹, 15.36 nmol·g⁻¹·h⁻¹, 1.68 nmol·g⁻¹·h⁻¹, 24.99 nmol·g⁻¹·h⁻¹, and 2.57 nmol·g⁻¹·h⁻¹ (Fig. 6A, E). Simultaneously, the IR treatment retained the highest β-XYL activity, peaking at 2.98 nmol · g⁻¹·h⁻¹.

Fig. 6.

Effects of Different Intercropping Patterns on Soil Enzyme Activities. CT, clean tillage between rows (conventional tillage without intercropping); NG, inter-row natural grass; IA, inter-row onion; IR, inter-row rape. Different uppercase letters above the histogram indicate that the difference between different soil layers of the same treatment reaches a significant level (P < 0.05), and different lowercase letters indicate that the difference between different treatments of the same soil layer reaches a significant level (P < 0.05). A Urease; B βG; C NAC; D LAP; E CBH; F β XYL.

Correlation analysis

To elucidate the interconnections among soil physical properties, active organic carbon components, enzymatic activities, and aggregate stability under various intercropping treatments, a correlation analysis was conducted on 28 pertinent indicators (Fig. 7). The findings indicated that organic matter had a highly significant positive association with TOC (r = 1, P < 0.01) and HFOC (r = 1, P < 0.01). Additionally, soil pH showed a highly significant positive correlation with soil bulk density (r=−1, P < 0.01) and a significant negative correlation with WGMD (r=−1, P < 0.01). It also exhibited significant negative correlations with POC (r = 0.−95, P < 0.05), DR_0.25(r=−0.97, P < 0.05), and DMWD (r=−0.97, P < 0.05).

Fig. 7.

Correlation Analysis. X1-X28 represent Organic matter, pH, Soil bulk density, Clay, Silt, Sand, Physical clay particles, SOC, DOC, ROC, POC, HFOC, LFOC, MBC, DR0.25, WR0.25, DMWD, DGMD, WMWD, WGMD, PAD, WSC, Urease, βG, NAG, LAP, CBH, and βXYL, respectively.

Principal component analysis

To thoroughly assess the effects of various intercropping practices on orchard soil quality, a principal component analysis (PCA) was applied to the processed data from 28 indicators. The analysis yielded three principal components with eigenvalues exceeding 1, specifically 14.802, 9.792, and 3.406 (Table 2). These components accounted for 52.684%, 34.972%, and 11.078% of the variance, respectively, culminating in a cumulative variance contribution of 98.914%, which satisfies the criteria for analysis. Table 2 details the composition of the principal components: the first principal component (PC1) includes parameters such as pH, soil bulk density, DOC, ROC, POC, DR_0.25, WR_0.25, DMWD, DGMD, WMWD, WGMD, PAD, WSC, andβXYL. The second principal component (PC2) encompasses organic matter, clay, silt, sand, physical clay particles, SOC, HFOC, LFOC, MBC,βG, NAG, LAP, and CBH. The third principal component (PC3) consists solely of urease activity.

Table 2.

Principal component analysis and variance interpretation.

Index	Load
	PC1	PC2	PC3
Organic matter	0.614	0.779	−0.126
pH	−0.965	−0.258	−0.044
Soil bulk density	−0.929	−0.368	−0.049
Clay	−0.455	0.842	0.29
Silt	−0.604	0.683	0.412
Sand	0.58	−0.709	−0.402
Physical clay particles	−0.617	0.679	0.397
TOC	0.569	0.797	−0.204
DOC	−0.978	0.206	−0.039
ROC	−0.608	0.552	−0.571
POC	0.933	0.251	−0.259
HFOC	0.61	0.79	−0.064
LFOC	0.496	0.731	−0.468
MBC	−0.438	0.895	−0.089
DR0.25	0.94	0.19	0.285
WR0.25	0.873	−0.186	0.45
DMWD	0.947	0.163	−0.276
DGMD	0.919	0.353	−0.173
WMWD	0.842	0.465	−0.275
WGMD	0.966	0.257	0.018
PAD	−0.931	0.159	−0.33
WSC	0.929	−0.127	0.347
Urease	0.105	0.427	0.898
βG	−0.358	0.808	0.468
NAG	−0.241	0.806	−0.541
LAP	−0.409	0.912	−0.04
CBH	−0.121	0.99	0.073
β XYL	0.935	0.017	0.354
Eiges values	14.802	9.792	3.406
Proportion of variance /%	52.684	34.972	11.078
Cumulative variance /%	52.684	87.836	98.914

The PCA analysis offers a succinct overview of the intricate relationships among various soil quality indicators under different intercropping treatments. It reveals that the first principal component accounts for a significant proportion of the variability in soil physical properties, carbon fractions, and enzyme activities linked to soil stability and nutrient cycling. The second principal component primarily reflects organic matter content, soil texture, and specific enzyme activities. Meanwhile, the third principal component highlights the importance of urease activity, which is essential for nitrogen cycling. Collectively, this PCA analysis provides valuable insights into how different intercropping systems affect soil quality and aids in selecting the best intercropping practices for sustainable orchard management.

Comprehensive ranking

To assess the overall impact of different intercropping treatments on the quality of orchard soil, a comprehensive score (F) was calculated for each treatment. This score was derived by summing the product of each principal component score and its corresponding variance contribution rate, as follows: F = F1 × 52.864% + F2 × 34.792% + F3 × 11.078%. Table 3 presents the principal component scores and comprehensive scores for each treatment, along with their corresponding rankings. According to the results in Table 3, the comprehensive scores for the different intercropping treatments are as follows: −0.89089 (CT), −0.04327 (NG), 0.41031 (IA), and 0.52386 (IR). Therefore, the ranking of these treatments in terms of their impact on orchard soil quality is: IR > IA > NG > CT.

Table 3.

Comprehensive ranking of the effects of different intercropping treatments on soil quality in orchards.

Treatment	Principal component score			Comprehensive score	Comprehensive score ranking
	PC1(F1)	PC2(F2)	PC3(F3)
CT	−0.81776	−1.17805	−0.43984	−0.89089	4
NG	−0.87205	0.96119	0.75209	−0.04327	3
IA	0.57746	0.68497	−1.20306	0.41031	2
IR	1.11236	−0.46811	0.89081	0.52386	1

Discussion

Strategic intercropping refers to the intentional selection and spatial arrangement of complementary crop species aimed at optimizing resource utilization and enhancing ecosystem benefits. It leverages interspecific interactions and nutritional niche differentiation to enhance soil physical and chemical properties, thus substantially improving the functionality of agroecosystems³⁵. Lu et al. (2019)³⁶ demonstrated that intercropping elevates soil organic matter content and significantly reduces bulk density in the topsoil compared to monocropping. Our findings indicate that soil bulk density and organic matter content undergo variable changes within the 0–20 cm layers under distinct treatments. Relative to the control treatment (CT), the NG, IA, and IR treatments consistently reduced soil bulk density and markedly increased soil organic matter content. This outcome is likely due to the intercropped plants’ root systems and exudates, which enhance soil porosity and aeration, subsequently improving water retention and fostering soil aggregate formation and stability³⁷. Moreover, the root activities and organic matter contributed by these crops supply rich nutrients and energy to soil microorganisms, stimulating microbial activity and further promoting the accumulation of soil organic matter and the enhancement of soil structure. Decreased tillage frequency and intensity, coupled with better soil moisture conditions, also help to minimize soil structural damage and preserve aggregate integrity³⁸. Furthermore, heightened soil biodiversity, including the activity of microbes and small soil fauna, bolsters aggregate formation and stability³⁹. Collectively, these factors contribute to the reduction in soil bulk density and the enrichment of organic matter content, ultimately improving soil fertility and the environment for crop growth. It is noteworthy that changes in soil bulk density in the 20–40 cm soil layer were not significant among treatments. This outcome may be attributed to the relatively shallow root systems of both intercropped species. Most root biomass and associated microbial activity are concentrated in the topsoil (0–20 cm), where organic matter inputs and aggregate transformations are more pronounced.

A stable soil pH is essential for crop growth due to its significant impact on soil nutrient availability⁴⁰. Our findings revealed that the pH levels in soils under NG, IA, and IR treatments were generally lower than those under CT, with variations associated with soil depth. This trend is corroborated by Li et al. (2023)⁴¹who observed a significantly lower soil pH under intercropping compared to conventional monocropping systems. Soil particles, integral to soil structure, dictate characteristics such as porosity, water retention, and gas exchange, which collectively influence soil fertility⁴². Our data indicated that the NG treatment significantly increased the proportions of clay, silt, and fine clay particles relative to CT, while decreasing the sand proportion. This shift is likely attributed to the natural grass’s roots and vegetation, which bolster soil resistance to erosion, thereby minimizing losses of fine particles. Additionally, the root activity and microbial metabolism associated with the grasses enhance organic matter accumulation and soil structure, promoting particle aggregation and stability. Decreased tillage and reduced soil compaction, coupled with enhanced water retention, further preserve soil particle stability, enriching the soil with a higher proportion of clay and silt particles. These factors collectively enhance the physical properties of orchard soils, thereby promoting root growth and the retention of nutrients and water.

Soil organic carbon (SOC) serves as a pivotal measure of soil quality and sustainable land management^43,44. Factors such as planting configurations, tillage strategies, and nutrient stewardship significantly affect SOC levels. Chapagain et al. (2014)⁴⁵ demonstrated that intercropping systems, specifically barley-pea combinations, enhance SOC and total nitrogen content compared to monocultures. Studies have shown that the quantity of MBC is primarily limited by the availability of organic carbon sources in the soil. It is also related to the reduction of the soil’s active carbon pool with increasing soil depth. Variations in aboveground biomass result in a significant difference in the amount of organic carbon entering the soil, leading to changes in the quality and quantity of substrates available to microbes. These changes, in turn, influence the dynamics of soil microbial biomass⁴⁶. Our research indicates that SOC, including TOC, DOC, ROC, POC, HFOC, LFOC, and MBC, diminishes with soil depth. This trend is presumably linked to variables such as organic matter decomposition, root systems, microbial activity, soil physical characteristics, and management practices. Within the 0–20 cm layer, the IA treatment exhibited superior levels of TOC, DOC, ROC, POC, HFOC, and LFOC, markedly surpassing those of the CT. This outcome is potentially attributed to the dense root system and robust microbial activity under the IA treatment, which likely enhance the input and transformation of SOC, thereby fostering its accumulation. The MBC was notably highest in the NG treatment, possibly owing to the grass roots and residues that may stimulate microbial activity. However, these beneficial impacts tend to diminish with increasing depth, leading to a decrease in organic carbon content in the subsoil. Interestingly, although TOC increased in the 20–40 cm layer under IA and IR treatments, the DOC and LFOC contents declined relative to CT. This apparent contradiction may reflect the complexity of deep-soil carbon dynamics, where soluble carbon compounds may leach downward or be rapidly utilized by microbes in subsoil environments with limited carbon replenishment.

Soil aggregates, integral to soil structure, serve as pivotal indicators for assessing soil quality and diagnosing land degradation⁴⁷. They are categorized based on particle size into macroaggregates (> 0.25 mm) and microaggregates (< 0.25 mm), with the former being a hallmark of superior soil structural integrity⁴⁸. Utilizing dry and wet sieving techniques, our study classified soil aggregates across the 0–20 cm and 20–40 cm soil layers. The findings revealed that various intercropping treatments facilitated the disintegration of macroaggregates into microaggregates. Specifically, the IA treatment markedly decreased the prevalence of mechanically stable macroaggregates in both layers, reaching only 72.55% and 94.02% of the levels observed under CT. The DR_0.25 and WR_0.25 values suggested that while the topsoil’s mechanical stability was satisfactory, its hydraulic stability was compromised, making it susceptible to water-induced disturbances. Notably, the IR treatment was particularly effective in bolstering soil aggregate stability, potentially owing to the enhanced physical structure and increased porosity from rape roots and their exudates, which fostered favorable conditions for aggregate formation⁴⁹. Moreover, the interplay of root activity and organic matter from intercropped crops stimulated soil microbial activity; microorganisms, in turn, produced substances that cemented aggregates, thereby reinforcing their stability. The diminished frequency and intensity of tillage under intercropping, coupled with enhanced soil moisture retention and heightened soil biodiversity, further bolstered aggregate formation and stability. Collectively, these factors enabled the IR treatment to markedly enhance both the mechanical and hydraulic stability of soil aggregates, ultimately elevating the overall soil quality within the orchard.

Soil enzymes are instrumental in catalyzing material metabolism and energy flow within soil ecosystems, with their activity levels serving as indicators of soil nutrient cycling rates and microbial activity⁴⁹. Bolinder et al. (1999)⁵⁰ have indicated that soil enzyme activities are highly sensitive to alterations in soil quality. Research indicates that traditional tillage practices lead to a decline in soil quality, affecting the structure of microbial communities and reducing soil enzyme activity⁵¹. Our findings revealed distinct trends in soil enzyme activities across various intercropping systems. Within the 0–20 cm soil layer, the activities of β-glucosidase and cellobiohydrolase peaked under the NG treatment, whereas N-acetyl-β-D-glucosaminidase and leucine aminopeptidase activities were most pronounced under the IA treatment. Notably, the IR treatment yielded the highest xylanase activity. In the 20–40 cm soil layer, aside from xylanase activity, which was most elevated under the IR treatment, the NG treatment fostered the highest activities for all other enzymes. These variations are likely due to the influence of root exudates on the soil microbial community’s structure and function, as well as their capacity to regulate soil organic matter input, hydration, aeration, and pH levels. Root exudates supply a wealth of carbon and energy sources for microorganisms, thereby enhancing specific enzyme activities. Moreover, the diversity and composition of the soil microbial community, in conjunction with soil management practices such as tillage and fertilization, can also exert an impact on enzyme activities⁵². The decomposition of crop residues releases organic matter, supplying substrates for enzymes and further stimulating enzyme activities⁵³. Consequently, these interrelated factors contribute to the variability in soil enzyme activities under different intercropping patterns, underscoring their regulatory functions in soil biochemical processes and their consequential effects on soil fertility and crop growth. Interestingly, although the enzyme activities (βG and CBH) were the highest under NG treatment, the content of organic carbon in the associated soil did not increase accordingly. This phenomenon may be attributed to the “priming effect”, where root exudate or litter input stimulates microbial communities and accelerates the decomposition of organic carbon in the primary soil. This enhanced mineralization can lead to a temporary net loss of carbon, explaining the difference between enzyme activity and carbon accumulation.

Although this study provides valuable insights into the effects of intercropping systems on soil quality in dryland apple orchards, it was conducted solely at a single experimental site in Jingning, Gansu Province. The site is representative of the semi-arid agroecosystems typical of the central Loess Plateau. However, the heterogeneity of climatic, edaphic, and topographic conditions across the broader Loess Plateau suggests that caution should be exercised when generalizing these findings. Future multi-site investigations across different ecological zones within the Plateau are warranted to validate the consistency and applicability of the observed intercropping effects.

Conclusion

This three-year study evaluated four intercropping patterns in apple orchards on the Loess Plateau: conventional tillage (CT), natural grassing (NG), intercropping with onions (IA), and intercropping with rapeseed (IR). The results showed that intercropping significantly improved soil quality compared to CT. The IA and IR treatments were particularly effective, with IA enhancing total organic carbon and IA and IR improving soil structure stability. Additionally, different intercropping patterns positively impacted soil enzyme activities. In summary, intercropping with rapeseed and onions was the most effective strategy for improving soil quality, as evidenced by increased organic carbon content, enhanced aggregate stability, and elevated soil enzyme activities. These improvements provide practical guidance for sustainable apple orchard management on the Loess Plateau. Although this study yielded valuable results from a three-year field experiment, it was conducted at a single site, which limits the understanding of the long-term effects of intercropping on soil succession and stability. In addition, the roles of crop root architecture, microbial communities, and their interactions in carbon and nitrogen cycling remain to be fully explored. Future studies should implement multi-site trials across different ecological zones of the Loess Plateau to verify the broader applicability of intercropping patterns. Moreover, integrating multi-dimensional indicators such as metagenomics, soil respiration, and nutrient turnover monitoring will help to more comprehensively elucidate the ecological mechanisms by which intercropping improves soil quality.

Author contributions

Conceptualization, M.M.; methodology, M.M. and W.S.; software, M.Z., validation, M.M. and W.S.; formal analysis, T.D., investigation, M.M., data curation, W.S.; writing—original draft preparation, M.M. and W.S.; writing—review and editing, M.M.; visualization, M.Z., supervision, M.M.; project administration, M.M.; funding acquisition, M.M. All authors discussed the design, analyses, and results of the study and participated in the writing of the manuscript. All authors have read and agreed to the published version of the manuscript.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author(s) on reasonable request” in the manuscript.

Declarations

Competing interests

The authors declare no competing interests.