The study area (Figure 1) is located in Shuifu Township of Gaolan County, Gansu province, China, located in the hilly and gully region of Loess Plateau (36°16' N, 103°49' E), with an average altitude of 1,825 m. The region has a temperate semi-arid climate with an annual average temperature of 7.2°C. The annual average precipitation is about 266 mm, the annual average evaporation is 1,660 mm, the annual average sunshine is 2,768 h, and the frost-free period is 144 d. Rainfall is concentrated between July and September. The main soil types are calcareous soil, red clay, and alluvial soil, with mainly calcareous soil, with loose soil aggregate structure, poor water and fertilizer holding capacity, and poor fertilizer and water storage capacity. The main perennial vegetation includes: Reaumuria songarica, Caragana, Zygophyllum mucronatum, and Peganum harmala. The main annual vegetation is Setaria viridis, Enneapogon borealis, and Polypogon fugax.

In late August 2019, we used GPS to locate and select the sampling sites. We selected three hills (hill A, B, and C) as the spatial repetitions with similar landscape type, elevation (the relative elevation of three hills no more than 50 m), and slope aspect, which enabled us to control the effect of elevations on the SOC, STN, and STP and to focus on the variation of SOC, STN, and STP along the slope aspect gradients. The cross sections of the southern slope aspect (azimuth Angle 180°), southwestern slope aspect (225°), western slope aspect (270°), northwestern slope aspect (315°), and northern slope aspect (360° or 0°) of each hill were examined. In the middle of each hill (the altitude about 1,800 m) five soil sample points were randomly sited. Soil samples were collected for each soil profile using a soil auger (3.5 cm in diameter) at 0–10, 10–20, 20–40, 40–60, 60–80, and 80–100 cm, three replicates were excavated from each layer, and a total of 450 soil samples were collected at 75 soil profiles. Soil profiles were excavated at each standard sample point, and undamaged soil was collected on the left- and right-side walls and front surfaces of each layer with a volume of 100 cubic centimeters for the determination of soil bulk density. The collected soil samples were taken back to the laboratory, removed from the impurities, dried naturally, ground through 0.25 mm fine sieve, and stored for analysis of SOC, STN, STP, and other contents indicators.

The soil water content (θ, %) was measured gravimetrically by weighing the soil sample, drying it in an oven at 105°C for 24 h, and then reweighing the sample. The bulk density of soil (ρ, g/cm³) was determined from undisturbed core segment as the mass of soil per unit volume. The pH of a 2:5 soil: water suspension represents the pH of the soil, measured with pH acidometer. Fresh above-ground biomass (BM_ag) samples were oven-dried at 80°C to a constant weight (48 h), and biomass was expressed in dry weight (g/m²). The SOC (g/kg) of each soil layer was determined by titration with ferrous sulfate after oxidizing and boiling 0.5 g of airdried and sieved soil samples with concentrated sulfuric acid-potassium dichromate (15). STN content was measured with the Kjeldahl method (22). STP content was measured by molybdenum-antimony and scandium colorimetric method (23). These indexes (θ, ρ, pH, SOC, STN, and STP) were determined by referring to the data standards of soil Agricultural Chemical Analysis (24). The 50% parallel samples were set for each batch of samples, and two blank samples were added at the same time to control the accuracy of sample determination and eliminate the errors caused by reagents, instruments, and environment that may occur in the sample determination process.

We calculated the content (g/kg) of SOC, STN, and STP for each of the six soil layers for each site. We then calculated the SOC, STN, and STP densities (SOCD, STND, and STPD) of each layer using Equation (1).

In Equation (1), BD and G represent the bulk density (g/cm³), and the fractional percentage (%) of gravel larger than 2 mm in diameter of layer n, respectively.

Following Jobbágy and Jackson (25), we applied the following equation to describe the general vertical distribution patterns of SOC, STN, and STP.

which can be log-transformed to

In Equation (3), Y represents the content or density of SOC, STN, and STP, and D represents soil depth measured as the midpoint depth of each layer, i.e., 5, 15, 30, 50, 70, and 90 cm. The coefficient γ stands for the rate of decrease in SOC, STN, and STP with depth, and a larger value indicates a larger difference between surface and bottom soils in nutrient contents or densities. We applied Equation (3) with all sites pooled together to calculate the overall rates of decrease in SOC, STN, and STP on each slope aspects, and we also applied this equation to every site where soil depth reached 100 cm to obtain site-specific γ values.

For the calculation formula of Grey Correlation the SOC, STN, and STP contents of soil layers 0–10, 10–20, 20–40, 40–60, 60–80, and 80–100 cm were used as the reference sequence. X₀ was used for correlation degree analysis, and other soil physical and chemical properties were used as the comparison sequence X_i. Since the factors are inconsistent with the target results, the mean value of the data is compared before calculating the correlation degree, and the formula is as follows:

In which av (x₀) and av (x_i) are the average values of X₀ and X_i, respectively.

The formula for calculating the correlation coefficient is as follows:

In Equation (5), ρ is the discrimination coefficient and range of 0 < ρ < 1, generally ρ is 0.5, I (k) is the absolute difference between X₀ and X_i.

Correlation degree R_i:

As is shown in Figure 2, the contents of SOC, STN, and STP at 0–100 cm varied from 1.65–12.87, 0.45–1.53, and 0.07–0.78 g/kg, and the variations were significant (p < 0.05). The SOC content showed a significantly decreasing trend with soil depth at all sampling positions, and the SOC content at 0–10 cm was about 2–3 times larger than that at 80–100 cm. STN content decreased first and then increased with depth, and 80 cm was the cut-off point. STP content increased significantly (p < 0.05) with depth, and STP content at 80–100 cm was 4–6 times that of surface layer. The STN and STP content showed an increasing trend along the southern to northern slope aspect, except for SOC content (Figure 2).

As is shown in Figure 3, SOCD, STND, and STPD ranges between 3.9–9.5, 0.58–1.35, and 0.33–0.56 kg/m² in the 0–100 cm soil layer for all the slope aspects. There were significant differences among different slope aspects (p < 0.05). In the 0–60 cm, the SOCD and STND accounted for more than 75 and 64% of the total soil layer (0–100 cm). Below 40 cm, the STPD accounted for more than 81% of 0–100 cm and showed an overall increasing trend with the increase of soil depth.

Significant differences between the two slope aspects were found in soil properties and vegetation (Table 1). θ, SOC, SOCD, STN, STP, R_i, VC, and BM_ag on the northern slope aspect were higher than those on the southern slope aspect except for ST, and their average values were 1.35, 1.70, 1.78, 2.32, 1.69, 1.75, 2.12, and 1.93 times higher than those on the southern slope aspect. ST on the northern slope aspect was 6.77 °C lower than the southern slope aspect. Although the value of pH and ρ on the southern slope aspect was greater than on the northern slope aspect, the difference was not significant (Table 1).

On the northern slope aspect, θ (0–0.1 m) was significantly greater than θ in the other soil layers. On the southern slope aspect, a decreasing trend in θ occurred, where θ (0.2–0.4 m) was minimum, followed by an increasing trend in θ with a further increase in depth. On both the southern and northern slope aspects, soil pH showed an increasing trend with soil depth. Soil pH on the southern slope aspect were all greater than on the northern slope aspect. On both the southern and northern slope aspects, ρ decreased with soil depth from the surface to 100 cm. For all soil layers there was no significant difference in ρ between the two slope aspects. In terms of nutrient content, the contents of SOC, STN, and STP in each soil depth were all higher on the northern than southern slope aspect. The contents of SOC and STN decreased gradually with soil depth, while the content of STP increased gradually. The SOC and STN contents of 0-40 cm northern slope aspect were significantly different from those of the southern slope aspect. There was no difference in STP content between the two slope aspects at 0–20 cm, but the difference between 40 and 100 cm was more and more significant (Figure 4).

With all slope aspects sites pooled, the average decreasing rates γ for SOC, STN, and STP contents were 0.38, 0.36, and −0.47, respectively (Figure 5), and γ for SOC, STN, and STP densities were 0.39, 0.35, and −0.52, respectively (Figure 6). As the STP content gradually increases with soil depth, the γ_STP is negative. After calculating specific-slope aspect γ for SOC, STN, and STP, we found γ_SOC > γ_STN > γ_STP for both content and density (Figure 7). This meant that the difference between deep soil and surface soil was larger in SOC. Although γ_STP is negative, the size of the value indicates that there is a great difference between surface soil and deep soil both content and density. For a single soil nutrient, we found that both the content and density of γ_N > γ _All > γ_S (Figure 7), which also indicated that there were significant differences between the southern and northern slope aspect in terms of both the content and the decreasing rates.

According to the gray correlation degree results (Tables 2–4), the correlation degree between soil physical and chemical properties and SOC, STN, and STP contents on different slope aspects was between 0.59–0.88, 0.57–0.83, and 0.52–0.89, respectively. On the whole, soil physical and chemical properties have a high correlation degree with SOC and STN contents, with a high possibility of comprehensive influence, but a low degree of overall correlation with STP content. The correlation degree of SOC and STN content with soil physical and chemical properties was the highest on the northern slope, and the correlation degree of SOC content with electrical conductivity (EC), ST, ρ, θ, and STN content was above 0.8. The correlation degree of STN content with EC, ST, ρ, and SOC content was above 0.8. There was a high correlation between STP content and soil physical and chemical properties on the southern slope. From the perspective of single soil factor, the correlation degree of EC, ρ, STN, and SOC content on five slope aspects is above 0.73. The correlation degree of ST, EC, ρ, SOC, and STN content on five slope aspects were all above 0.7. On the five slope aspects, only pH and STP content had a correlation degree above 0.81, and other influencing factors had no significant effect on STP content.

The results show (Figures 2, 3) that the SOC, STN, and STP contents and SOCD, STND, and STPD on the northern slope aspects are significantly higher than other slope aspects, and STN and STP gradually increase from the southern to northern slope aspects, which is consistent with the research results of Lenka et al. (26) and Huang et al. (3). This distribution pattern was probably due to the heterogeneity of the ST on the slope aspects. Topographic obstructions and aspects affect the radiation distribution (12). The potential annual direct incident radiation on the southern slope aspect was about 8, 37, and 93% higher than on the southwestern, western, and northern slope aspect, respectively (27). Higher radiation resulted in higher ST, which led to faster decomposition of soil organic matter, and consequently less accumulation of the SOC in soil on the southern slope aspect compared with the northern slope aspect (7), the northern slope aspect received less solar radiation, humidity, and low ST, which inhibited the decomposition of organic matter by soil microorganisms, leading to increased accumulation of SOC, STN, and STP (15). Chen et al. (7) found lower SOC and STN contents on the southern slope aspect rather than the relatively more shaded northern slope aspect. A relatively high P content was observed in the soils on the relatively cool northern slope aspect, suggesting that vegetation on the southern slope aspect absorbed a relatively higher amount of P in order to obtain a richer biomass (28).

The current research found that the SOC and STN contents decreased with soil depth (Figure 2), and the decrease of STN content was related to organic matter (29). STP content increased with soil depth (Figure 2), the results indicated that the vegetation litter had little influence on STP in the topsoil, and the main source of STP in soil is mineral weathering. The content of STP on the surface is too low, which is mainly due to the uptake of P by the growth of surface plants, because the roots of plants are mainly distributed in the 0–30 cm soil layer. Vegetation under semi-arid climatic conditions is principally controlled by soil water storage capacity and soil depth (30). With the deepening of soil layers, the soil environment becomes relatively closed, the variation range of soil density decreases, the input amount of organic matter and microbial activities weaken, the content of organic carbon in deep soil decreases and the change is more stable than that in surface soil, and the SOC, STN, and STP in soil with different slope aspects show obvious vertical distribution characteristics. And we found that the 0–60 cm SOCD and STND accounted for more than 75 and 64% of 0–100 cm, indicating that SOC and STN had similar surface aggregation phenomenon (31). The important reason for this phenomenon is that the vertical distribution pattern of plant roots and the accumulation of organic residues and microbial activities occurred in the first 30 cm of the soil layer (29, 32). Compared with the deep soil, the surface soil has a large amount of plant litter, which makes the organic matter returned to the soil increase, and the surface soil has a better structure, which is conducive to water and air permeability, which provides the basic conditions for the decomposition of litter and indirectly improves the quality of surface SOC (33), this is also consistent with the research results of Li et al. (34) and Quan et al. (35). In this study, STP content increased with soil depth, while STP content in surface layer was low, which may be due to the absorption of P by roots of surface vegetation (36, 37). Meanwhile, it also reflected that SOC, STN, and STP contents had no significant interaction in terms of soil depth.

The large differences in microclimate induced by aspect had resulted in large variability in vegetation patterns (Table 1), which in turn affected the patterns of SOC, STN, and STP (38). ST and θ are the most important driving factors affecting soil properties and soil structure in hill ecosystems, as well as affecting carbon cycle, nitrogen mineralization, and nitrification rates, and the conversion rate of organic matter to P, and thus play an important role in regulating soil SOC, STN, and STP balance (39). As shown in Table 1, the ST and θ of the northern slope aspect were significantly different from that of the southern slope aspect, which was an important reason for the significant difference in SOC, STN, and STP contents between the two slope aspects. Although Sonmez et al. (40) found that soil pH decreased with soil depth, the current study revealed a different pattern (Figure 4). This was likely due to the litterfall decomposition in the study area, which contributed organic acids to the topsoil, lowering soil pH (59). However, soil pH is greater in the deeper layers due to the breakdown of the organic acids in the topsoil (42). The decreasing trend of ρ with soil depth may be related to the issue of the declining effects of compaction and soil hardening from the surface (43).

The SOC content decreased with soil depth, as was found by most studies (20, 44, 45, 58). The greater SOC in upper soil layers was related to above-ground litter (46), along with plant roots and their exudates (47, 48). With an increase of soil depth, the influence of litter weakened, and root presence also decreased. However, below 0.6 m, SOC increased slightly. This may be due to the stability of deeper SOC. Chang et al. (49) found that the sub soil environment was not favorable to decomposition. Powlson et al. (50) and Ussiri and Lal (51) also showed that SOC in topsoil is more sensitive to soil perturbations than SOC in sub soil. The vegetation coverage and biomass of the northern slope aspect were higher than that of the southern slope aspect (Table 1), the plants had deeper roots, and the relatively rich vegetation coverage could reduce the loss of SOC, STN, and STP caused by the rainy season precipitation. The higher soil water content on the northern slope aspect can effectively improve the habitat of soil microorganisms, enrich the quantity and diversity of soil microorganisms, and create favorable conditions for litter decomposition and soil nutrient release. On the southern slope aspect, plants have shallow roots, sufficient illumination, high temperature, large evaporation and low water content, which affect the storage and accumulation of SOC, STN, and STP.

We found that the declining rates of SOC and STN on the southern slope were lower than the mean, indicating that the surface soil of the southern slope was more vulnerable to soil organic matter loss because of their lower plant coverage and more severe surface soil weathering (52). Since the main source of STP is parent material weathering, the P on the soil surface is used for plant absorption and natural weathering (53), resulting in a higher phosphorus content in deeper soil than the surface, resulting in a negative reduce rate.

The results show that the correlation degree between soil physical and chemical factors and SOC content was above 0.6 on different slope aspects, and it has influence on SOC in different degrees (Table 2). Soil density, porosity, and other properties will affect soil texture, change soil permeability and water holding capacity, change the growth and development of plant roots and microbial activity conditions, and then affect the SOC and STN contents (32). Wang et al. (54) found that SOC as the cementing agent of soil aggregates has a direct effect on soil capillary porosity and soil density by affecting soil texture and structure. The influence of EC and soil plant roots show higher response relationship with SOC content, which is similar to the conclusion drawn by Zhao et al. (55). STN mainly comes from the accumulation and decomposition of organic matter. SOC as an important energy source of microbial activities that affects the decomposition of organic matter and then affects STN content, especially the available nitrogen in the soil (56). STP comes from soil organic matter and soil-forming minerals. After organic carbon mineralization, most of the soil phosphorus finally exists in inorganic form in soil for plant use. Soil pH was closely related to soil microbial activity. Too high or too low pH values will affect the activity of soil microorganisms and reduce the conversion rate of organic carbon (57) and the rate of mineralization of soil-forming minerals. Therefore, soil pH is the main factor affecting STP content.