Dependency of Long-Term Soil Quality Controls on Summer Fallow Tillage and Soil Layers for Dryland Winter Wheat in Loess Plateau

Abstract

The capacity for winter wheat to produce sufficient yield may be influenced by soil tillage practices and soil quality. However, determining how to quantify the impact of long-term tillage on soil quality is crucial. Here, we address this issue by comparing soil properties and wheat yield under four tillage systems during summer fallow in the Loess Plateau. Twenty-two soil properties were explored to estimate soil quality. Results showed that a rotational tillage (PT/ST) during summer fallow decreased soil bulk density (ρ_b) and improved soil gravimetric water content (θ_g), soil organic carbon (SOC), soil capillary porosity (P_c), and total porosity (P_t) in 0–50 cm soil layers. A minimum dataset (MDS) of thirteen indicators was selected to calculate soil quality index (SQI). Treatment of PT/ST had higher SQI value in soil layers of 0–10 cm and 20–50 cm, and PT/ST showed a significant increase in yield since the third year. No tillage (NT) during summer fallow decreased soil physical and chemical indicators, thus decreasing soil quality. These findings suggest that a rotation tillage of PT/ST during summer fallow could enhance soil chemical and biological properties concurrently, and PT/ST may provide a promising management strategy to sustain soil quality and grain yield for dryland winter wheat in the Loess Plateau.

1. Introduction

Soil plays an important role in nutrition cycling, carbon storage, and nutrition utilization for crop production. Different tillage practices have distinct impacts on soil characteristics and crop growth. The magnitude of soil quality in sustainable agriculture has gained increasing attention over decades, and tillage practice, reduced organic matter, and nonaggregated soil combine to influence soil vulnerability [1,2]. Winter wheat monoculture is a common cropping practice in the Loess Plateau of China. Given that it is inherently risky, caused by unpredictable precipitation, temperature fluctuations, and other variable conditions, the summer fallow period has always been a dominant stage influencing wheat production in this region. In addition, summer fallow could decrease energy input through abandoning production in one season, thus mitigating production reduction risk for winter wheat [3,4]. Usually, approximately 60% of annual precipitation falls in the summer fallow period; however, the precipitation stored in soil during summer fallow is stagnant at about 19%. Increasing soil water storage during summer fallow is a necessary way to enhance soil quality in the Loess Plateau.

Previous studies indicated that no tillage during summer fallow is positively related with precipitation storage, yet it accelerates soil organic nitrogen mineralization and nitrification and is accompanied by an increased accumulation in soil nitrate [5]. However, other studies showed that summer fallow management consisting of intensive tillage operations are effective ways to increase wheat growth and development [6,7]. In comparison to no tillage during summer fallow, adoption of tillage practices is more effective in maintaining soil moisture than no tillage in summer fallow [8,9]. It has been reported that subsoiling tillage significantly increases soil organic matter in comparison to deep plow tillage and no tillage during summer fallow [10]. In addition, deep plow tillage showed a larger enhancement in soil water storage and precipitation storage efficiency during two dry summer fallow seasons than in normal summer fallow season [11]. Nevertheless, soil is susceptible to erosion during summer fallow because mono-tillage operation degrades soil aggregates and buries crop residue [8]. Therefore, rotation tillage is an recommendable cultivation management to overcome some of the adverse impacts of mono-tillage, and it is necessary to assess the effect of adoption of rotational tillage for soil quality.

Soil properties are all interrelated; changes occurring to some properties might not influence soil environment, while changes in the properties might have a significant impact on an agro-ecosystem [12]. For analysis of soil quality, the most important task is to find suitable indicators that contribute to soil operation. In recent years, soil chemical and physical properties have become sensitive indicators responding to soil alteration caused by tillage management [12,13]. Soil physical indicators are responsible for soil nutrient distribution and are mainly defined by soil texture and structure [14,15,16]. And soil chemical indicators play dominant roles affecting soil fertility quality stability [15,17]. Soil quality is inconvenient to assess directly because of collective and multiple functional effects. Aziz et al. [12] suggested that soil quality could be evaluated through adaptation of change in a small dataset of soil indicators triggered by tillage. The challenge to confirm the usefulness of tillage technology and soil quality is to quantify a core set of soil characteristics. Indicator values could be combined into a quantifiable SQI through a flexible model, which integrates soil characteristics into a simplified format. A robust SQI could interpret soil degradation and provide an early indication of the need for taking remedial measures. The following four steps are required to determine the SQI: (i) to ascertain management objective; (ii) to screen the MDS of characteristics; (iii) to score the selected MDS characteristics based on their contribution in soil function; (iv) to aggregate the above individual scores into an indicator score.

In this study, a four-year field experiment was conducted to assess the effect of tillage during summer fallow on soil quality for dryland winter wheat in the Loess Plateau. A total of 22 soil characteristics in 0–50 cm soil layers under four tillage managements during summer fallow were detected. Through a flexible model, SQI was quantified, taking into account the above soil indicators to detect the dependency of long-term soil quality controls on summer fallow tillage. The specific objectives of this research were (i) to identify the prominent characteristics that contribute the most to the soil quality in each soil layer; (ii) to assess the interactive impact of different tillage during summer fallow and soil layers on soil quality; and (iii) to establish an evaluation system for soil quality relating to cultivation practices of dryland winter wheat in the Loess Plateau.

2. Materials and Methods

2.1. Experimental Design

In the study, winter wheat cv. Yunhan618, a common wheat variety, obtained from an agricultural materials company, was applied in monoculture. In the Wenxi experimental station of Shanxi Province, located in the Loess Plateau, a four-year randomized block design was set up with four tillage patterns during summer fallow, including four years of no tillage (NT), three years of plow tillage followed by one year of subsoiling tillage (3PT + ST), three years of subsoiling tillage followed by one year of plow tillage (3ST + PT), and a rotation tillage of plow tillage and subsoiling tillage in four years (PT/ST). These tillage practices during summer fallow were conducted by a ploughing and fertilizer integrated machine, which had a depth of 25–30 cm for plow tillage and a depth of 35–40 cm for subsoiling tillage. Each pattern had four plots as duplicates (size: 5 × 60 m). Simultaneously, the field was applied with 1500 kg hm⁻² of commercial organic fertilizer and crushed stubble during summer fallow. In addition, an additional 150 kg hm⁻² of N, 150 kg hm⁻² of P₂O₅, and 150 kg hm⁻² of K₂O were applied before sowing. The sowing rate was 97.5 kg hm⁻². At maturity stage, wheat from all plot fields with a total of 1200 m² was harvested to determine grain yield. The precipitation distribution in the four-year experiment is shown in

Table 1. The distribution of precipitation (mm) at the experimental site in four years.

Year	PF (mm)	PG (mm)		PT (mm)
1st year	283.7	126.8	63.7	474.2
2nd year	365.6	133.5	17.6	516.7
3rd year	165.4	195.9	53.4	414.7
4th year	94.7	172.7	125.6	393.0

Data source: Meteorological Observation Station of Wenxi County, Shanxi Province, China. P_F (mm): Precipitation during summer fallow period; P_G (mm): precipitation during growing season for wheat; P_T (mm): total precipitation.

.

2.2. Soil Sampling and Analysis

After the fourth year of wheat maturity, soil samples were taken and collected at a depth of 50 cm at 10 cm increments from five locations in each plots. These five samples from the same soil layer in each plot were mixed into one sample. Soil physical properties, including soil bulk density (ρ_b), gravimetric water content (θ_g), total porosity (P_t), and capillary porosity (P_c), were measured by core method using a stainless-steel cylinder with 100 cm³ volume [18]. Here, soil ρ_b was calculated using dry soil and stainless-steel cylinder volume by Equation (1):

$${\rho }_b=\frac{M_3-M_0}{V}$$

(1)

where ρ_b is soil bulk density (g cm⁻³), M₃ is the weight of dry soil and the stainless-steel cylinder (g), M₀ is the initial weight of the stainless-steel cylinder (g), and V is the volume of the stainless-steel cylinder (100 cm³).

The soil θ_g was measured by Equation (2):

$${\theta }_g=\frac{M_1-M_3}{M_3-M_0}\times 100\%$$

(2)

where θ_g is gravimetric water content (%) and M₁ is the weight of the fresh soil and the stainless-steel cylinder (g).

The soil P_t was determined by soil ρ_b and particle density (P_d) using Equation (3):

$$P_t=\left(1-\frac{{\rho }_b}{P_d}\right)\times 100\%$$

(3)

where P_t is soil total porosity (%) and P_d is a constant of 2.65 g·cm⁻³ for mineral soil [18].

The soil P_c was calculated using Equations (4) and (5) based on the relationship between ρ_b and soil capillary water content (θ_c) [19].

$${\theta }_c=\frac{M_2-M_3}{M_3-M_0}\times 100\%$$

(4)

$$P_c=\frac{{\theta }_c\times {\rho }_b}{V}\times 100\%$$

(5)

where θ_c is soil capillary water content (%) and P_c is capillary porosity (%).

According to the method of Hong et al. [20], the soil samples with particle size of >0.25 mm were applied to determine soil water-stable aggregates (WSAs). The mean weight diameter (MWD) and geometric mean diameter (GMD) were calculated using Equations (6) and (7):

$$MWD={\sum }_{i=1}^n\left(\overline{x_i}·w_i\right)$$

(6)

$$GMD=exp\left({\sum }_{i=1}^n{w}_i·\mathit{ln}\overline{x_i}{\sum }_{i=1}^n{w}_i\right)$$

(7)

where $\overline{x_i}$ is the average WSAs diameter, and w_i represents the WSAs contents of each class (>5 mm, 1–5 mm, and 0.25–1 mm).

In addition, 5 g of dry soil sample was extracted with 25 mL of 2 M KCl solution, and the amount of soil nitrate and ammonium nitrogen were measured using an automatic ion analyzer (Easy Chem Plus, Anagni, Italy) following the methods of [21]. The potassium dichromate volumetric method was used to measure soil organic carbon (SOC) [22]. The molybdenum–antimony anti-colorimetric method was used to measure soil available phosphorus content (AP) [23]. The method of alkaline-hydrolysis diffusion was used to analyze soil alkali–hydrolyzable nitrogen (AN) amount [24]. Soil microbial biomass carbon was measured with chloroform fumigation and K₂SO₄ extraction with 0.45 of conversion coefficient K [25]. In addition, soil enzyme activities were determined as described by Gu et al. [26]. Activity of soil catalase was determined by measuring the reduction in H₂O₂ according to the method of KMnO₄ titration. Activity of soil urease was measured using indophenol blue colorimetry. Activity of soil phosphatase was determined through a disodium phenyl phosphate method, and activity of soil invertase activity was measured by 3,5-dinitrosalicylic acid colorimetry method using sucrose as the substrate.

2.3. The Assessment of SQI

In order to reduce redundancy of variables, Pearson’s correlation coefficients (p < 0.05) and principal component analysis (PCA) were applied to select appropriate indicators that had the highest loading values into a minimum dataset (MDS) [27,28]. Depending on whether a higher value was considered “good” or “bad” in terms of soil function, the retained indicators were ranked in increasing or decreasing order [29,30]. The linear scoring technique and SQI were calculated from the following scoring function [31]:

$$f\left(x\right)\left\{\begin{array}{c}0.1 x\le L\\ 0.9\times \frac{x-L}{U-L}+0.1 L\le x\le U\\ 1 x\ge U\end{array}\right.$$

(8)

$$f\left(x\right)\left\{1-\begin{array}{c}1 x\le L\\ 0.9\times \frac{x-L}{U-L} L\le x\le U\\ 0.1 x\ge U\end{array}\right.$$

(9)

$$SQI={\sum }_{i=1}^n\frac{s_i}{n}$$

(10)

where each characteristic was normalized in a range from 0 to 1 and calculated under Equation (8) (“more is better” pattern, M) and Equation (9) (“less is better” pattern, L); x was value of the soil indicators; L was lower threshold of each indicator and U was upper threshold of each indicator; S_i was the indicator score for each variable i; and n was the number of variables in MDS.

2.4. Statistical Analysis

In order to achieve a more meaningful and interpretable solution, the relationships between soil physical and chemistry factors were explored by redundancy analysis (RDA) through R version 4.2.0. All data are expressed as means ± standard deviation (n = 4), and the statistical significance of differences was determined through one-way and two-way ANOVA to determine statistically homogenous groups followed by Duncan test (p < 0.05) through the software packages SPSS 20.0.

3. Results

3.1. Variation of Soil Indicators

Different summer fallow tillage induced significant changes in soil indicators (Figure 1). The results showed that the treatment of PT/ST had the highest values of θ_g, P_c, P_t, soil water-stable aggregates of 0.25–1 mm, soil nitrate nitrogen content, SOC, AP, AN, and soil microbial biomass carbon content in 0–50 cm soil layers, while NT had the lowest values. In addition, the treatment of NT had the lowest values of soil catalase activity, soil urease activity, soil phosphatase activity, soil invertase activity, MWD, and GMD. However, the values of ρ_b, soil particle size of 1–5 mm, and pH were highest under NT compared with other treatments in 0–50 cm soil layers.

3.2. Assessment of SQI by PCA and RDA

The results showed that values of KOM for factor analysis were greater than 0.5, and the significance of Bartlett’s test of sphericity were lower than 0.05 (

Table 2. KMO and Bartlett’s test for factor analysis.

Soil Layers		0–10 cm	10–20 cm	20–30 cm	30–40 cm	40–50 cm
Kaiser–Meyer–Olkin Measure of Sampling Adequacy		0.616	0.607	0.586	0.700	0.658
Bartlett’s Test of Sphericity	Approx. Chi-Square	1436.643	1387.412	1466.251	1902.125	1393.412
	df	231	231	231	231	231
	Sig.	0.000	0.000	0.000	0.000	0.000

The sampling is acceptable and sufficient if the value of Kaiser–Meyer–Olkin (KMO) is 0.5 and above.

). These results demonstrated that correlation matrix was appropriate for factor analysis. In

Table 3. Load matrix of quality evaluation index in 0–50 cm soil layers.

Soil Layers	0–10 cm			10–20 cm			20–30 cm			30–40 cm			40–50 cm
Principal Components	PC1	PC2	PC3	PC1	PC2	PC3	PC1	PC2	PC3	PC1	PC2	PC3	PC1	PC2	PC3
% of Variance	66.556	19.865	7.126	66.297	17.318	8.675	58.433	18.721	12.268	58.761	18.209	16.44	57.831	21.098	15.581
Cumulative %	66.556	86.421	93.547	66.297	83.615	92.29	58.433	77.154	89.422	58.761	76.97	93.41	57.831	78.929	94.51
Eigenvalues	14.642	4.370	1.568	14.585	3.81	1.908	12.855	4.119	2.699	12.927	4.006	3.617	12.723	4.642	3.428
Gravimetric water content (θg)	0.996	−0.055	0.005	0.962	0.238	0.033	0.961	−0.068	0.263	0.960	−0.121	−0.239	0.935	−0.224	−0.133
Bulk density (ρb)	−0.874	0.334	0.304	−0.966	0.099	0.185	−0.969	0.098	0.180	−0.916	0.335	−0.104	−0.955	0.214	−0.160
Capillary porosity (Pc)	0.902	−0.412	−0.062	0.985	0.030	−0.131	0.958	−0.195	0.076	0.927	−0.313	−0.039	0.960	−0.226	0.058
Total porosity (Pt)	0.976	0.039	−0.118	0.904	0.347	−0.128	0.861	0.306	0.300	0.837	0.111	−0.297	0.832	0.085	−0.355
Soil particle size of >5 mm	−0.190	−0.863	0.423	−0.848	0.458	0.239	−0.606	0.576	0.487	0.715	0.667	0.097	0.321	0.931	0.095
Soil particle size of 1–5 mm	0.239	0.958	−0.053	−0.563	0.616	−0.484	−0.121	−0.126	0.931	−0.575	0.622	−0.471	−0.947	0.137	−0.161
Soil particle size of 0.25–1 mm	0.701	0.623	−0.291	0.935	−0.211	0.097	0.909	0.051	−0.302	−0.259	−0.468	0.770	0.460	0.694	0.440
Soil water-stable aggregates of 1–5 mm	0.503	0.773	−0.230	0.355	0.892	−0.194	0.326	0.868	−0.270	0.589	0.518	0.610	0.692	0.695	0.143
Soil water-stable aggregates of 0.25–1 mm	0.928	0.217	0.261	0.670	0.668	0.256	0.724	0.225	0.629	0.607	0.764	−0.169	0.515	0.594	−0.612
Mean weight diameter (MWD)	0.790	0.237	0.468	0.656	−0.272	0.332	0.530	0.770	−0.231	0.747	0.219	0.617	0.699	0.473	0.496
Geometric mean diameter (GMD)	0.785	0.304	0.458	0.732	−0.270	0.391	0.607	0.757	−0.176	0.809	0.316	0.459	0.792	0.482	0.329
Soil nitrate nitrogen content	0.913	−0.251	0.295	0.920	0.091	0.373	0.930	0.258	0.206	0.948	0.135	0.066	0.926	−0.265	0.017
Soil ammonium nitrogen content	0.957	0.128	−0.081	0.695	0.408	−0.352	0.521	−0.450	−0.357	−0.292	−0.636	0.591	−0.211	−0.204	0.936
Soil organic carbon content	0.844	−0.141	0.373	0.707	0.350	0.506	0.767	0.428	0.300	0.743	0.376	0.451	0.874	0.280	−0.105
Soil available phosphorus content	0.866	−0.476	−0.061	0.979	−0.090	0.002	0.983	−0.049	0.034	0.966	−0.218	0.107	0.993	−0.091	−0.011
Soil alkali-hydrolyzable nitrogen content	0.917	−0.173	−0.192	0.910	0.099	0.313	0.817	0.284	0.216	0.910	−0.389	0.002	0.812	−0.386	0.307
pH	−0.906	0.252	−0.015	−0.909	0.237	−0.005	−0.587	−0.327	0.417	−0.747	−0.241	−0.239	−0.949	−0.207	0.127
Soil microbial biomass carbon content	0.940	−0.213	−0.254	0.931	0.144	−0.315	0.872	−0.464	0.145	0.848	−0.386	−0.350	0.757	−0.526	−0.377
Soil catalase activity	0.242	−0.771	−0.509	0.689	−0.440	−0.574	0.708	−0.180	−0.255	0.688	0.415	−0.555	−0.155	0.122	−0.925
Soil urease activity	0.718	0.372	0.047	0.291	0.934	0.022	0.688	−0.614	0.352	−0.531	0.374	0.735	−0.903	0.090	0.287
Soil phosphatase activity	0.955	−0.159	0.092	0.860	−0.353	−0.172	0.746	−0.523	−0.230	0.879	−0.387	−0.178	0.343	−0.835	0.397
Soil invertase activity	0.968	0.135	−0.185	0.938	−0.140	−0.312	0.923	−0.332	−0.091	0.816	−0.535	−0.117	0.714	−0.666	−0.204

The underline in the table represent the chosen indicators in principal components.

, PCA analysis showed that three principal components (PCs) with eigenvalue greater than 1, explaining 93.547% of the total variance for all soil indicators, were selected. According to loading value, soil quality indicators of θ_g, P_t, soil ammonium nitrogen content, soil phosphatase activity, and soil invertase activity possessed relatively higher loading values in PC1. Similarly, soil particle size of 1–5 mm and soil catalase activity were selected in PC2 and PC3, respectively. However, results from Pearson’s correlation coefficients showed a significant correlation between θ_g and other indicators (Figure 2). Thus, in order to reduce redundancy of variables, only θ_g was retained into an MDS. In addition, RDA analysis showed that soil physical structure clearly distinguished four treatments in 0–50 cm soil layers, which indicated the dependency of soil quality controls on summer fallow tillage and soil layers (Figure 3). In the 0–10 cm soil layer, axes 1 and 2 explained 98.18% and 1.51% of the variation, respectively. And ρ_b and θ_g explained 90.8% and 5.4% of soil chemistry indicators in RDA ordination, respectively. Overall, indicators of ρ_b and θ_g were performed to calculate SQI in the 0–10 cm soil layer.

By the same token, indicators of θ_g, ρ_b, P_c, content of soil available phosphorus, soil urease activity, soil catalase activity, and SOC were selected in three PCs in the 10–20 cm soil layer, while θ_g, soil urease activity, and soil catalase activity were retained in MDS (Table 3, Figure 2). RDA analysis showed that axes 1 and 2 explained a large proportion of the variation, and P_t and soil particle size of 1–5 mm were the most significantly discriminating variables, accounting for 91.9% and 4.8% of soil chemistry indicators, respectively (Figure 3). And θ_g, soil urease activity, P_t, soil catalase activity, and soil particle size of 1–5 mm were performed to calculate the SQI.

In the 20–30 cm soil layer, only θ_g, soil water-stable aggregates of 1–5 mm, and soil particle size of 1–5 mm were retained in MDS (Table 3, Figure 2). And the indicators of P_c and soil water-stable aggregates of 1–5 mm explained 87.9% and 7.4% of soil chemistry indicators in RDA analysis, respectively (Figure 3). Thus, P_c, soil water-stable aggregates of 1–5 mm, θ_g, and soil particle size of 1–5 mm were performed to calculate the SQI.

In the 30–40 cm soil layer, θ_g, soil water-stable aggregates of 0.25–1 mm, and soil particle size of 0.25–1 mm were retained in MDS (Table 3, Figure 2). In the meantime, RDA analysis showed that θ_g and soil particle size of >5 mm explained 87.2% and 9.6% of the variation in soil chemistry indicators, respectively (Figure 3). Overall, indicators of θ_g, soil water-stable aggregates of 0.25–1 mm, and soil particle size of 0.25–1 mm were performed to calculate the SQI.

In the 40–50 cm soil layer, P_c, soil particle size of >5 mm, AN, and soil catalase activity were retained in MDS (Table 3, Figure 2). RDA results showed that θ_g and GMD were correlated with soil chemistry indicators, explaining 77.1% and 16.8% of the variation, respectively. Ultimately, P_c, soil particle size of >5 mm, AN, soil catalase activity, θ_g, and GMD were performed to calculate the SQI in the 40–50 cm soil layer.

3.3. Comprehensive Evaluation of Soil Quality by SQI and Yield

According to threshold distribution, the above selected indicators were calculated through SQI scoring technique (

Table 4. Threshold values and standardized scoring functions used for soil quality indicators.

Soil Quality Indicator	Unit	Function	Lower Threshold	Upper Threshold	Reference
Gravimetric water content (θg)	%	M	9.68	22.81	-
Bulk density (ρb)	g cm−3	L	0.80	1.47	[32]
Capillary porosity (Pc)	%	M	15.00	105.00	[33]
Total porosity (Pt)	%	M	20.00	70.00	[33]
Soil particle size of >5 mm	%	M	10.84	18.95	-
Soil particle size of 1–5 mm	%	M	22.37	37.82	-
Soil particle size of 0.25–1 mm	%	M	32.77	43.77	-
Soil water-stable aggregates of 1–5 mm	%	M	0.00	85.00	[34]
Soil water-stable aggregates of 0.25–1 mm	%	M	0.00	85.00	[34]
Mean weight diameter (MWD)	mm	M	0.76	0.84	-
Geometric mean diameter (GMD)	mm	M	0.69	0.77	-
Soil nitrate nitrogen content	mg kg−1	M	12.47	40.72	-
Soil ammonium nitrogen content	mg kg−1	M	6.86	17.25	-
Soil organic carbon content	g kg−1	M	10.00	45.00	[35]
Soil available phosphorus content	mg kg−1	M	5.02	30.61	-
Soil alkali-hydrolyzable nitrogen content	mg kg−1	M	20.61	74.70	-
pH	-	L	5.5	6.5	[36]
Soil microbial biomass carbon content	mg kg−1	M	75	700	[33]
Soil catalase activity	mL of (0.1 mol L−1 KMnO4) g−1	M	1.16	1.44	-
Soil urease activity	mg NH3–N g−1 h−1	M	123.30	234.70	-
Soil phosphatase activity	mg phenol g−1 h−1	M	71.70	95.09	-
Soil invertase activity	mg glucose g−1 h−1	M	257.71	520.65	-

Since there was no information about the upper and lower thresholds for some soil quality indicators, the minimum and maximum observational values of these variables were considered as lower and upper thresholds.

). The SQI value was significantly affected by tillage practices, soil layers, and their interaction (

Table 5. The SQI as affected by tillage and 0–50 cm soil layers.

Tillage	SQI
	0–10 cm	10–20 cm	20–30 cm	30–40 cm	40–50 cm
4NT	0.18 d	0.56 c	0.18 c	0.13 c	0.50 c
3PT + ST	0.25 c	0.47 d	0.16 c	0.12 d	0.54 b
PT/ST	0.71 a	0.64 b	0.32 a	0.23 a	0.57 a
3ST + PT	0.45 b	0.71 a	0.23 b	0.15 b	0.53 b
ANOVA results
Tillage	**
Soil layers	**
Tillage × Soil layers	**

The different letters in the same soil layer indicate p < 0.05. ** Significance at p < 0.01.

). The SQI values showed an upward–downward–upward trend in all treatments in the soil profile. In the 0–10 cm soil layer, SQI in PT/ST was higher than that in other treatments, while NT had the lowest value. In the 10–20 cm soil layer, 3ST + PT had higher SQI value. In the 20–50 cm soil layers, PT/ST had higher SQI value than other treatments, and SQI values in the 0–10 cm and 40–50 cm soil layers were positively correlated with grain yield (Figure 4). In addition, grain yield as affected by tillage and precipitation are shown in

Table 6. The grain yield (kg hm⁻²) as affected by tillage and precipitation during summer fallow period (PF).

Tillage	Yield
	1 Year	2 Year	3 Year	4 Year
NT	3467.00 b	3956.22 d	4812.00 c	2668.85 d
3PT + ST	4719.00 a	5062.57 c	5471.90 b	4832.39 b
PT/ST	4760.77 a	5559.96 b	6009.75 a	5091.85 a
3ST + PT	4875.40 a	6176.55 a	5719.08 ab	3070.06 c
ANOVA results
Tillage	**
PF (mm)	*
Tillage × PF (mm)	**

The different letters of the same year denote significance at p < 0.05 (Duncan test). * Significance at p < 0.05, ** significance at p < 0.01.

, in which grain yield was positively correlated with tillage and P_F and their interaction. In the first year, grain yield from treatments of 3PT + ST, PT/ST, and 3ST + PT were significantly higher than that of NT, while no significant difference was found among these three treatments. The yield of 3ST + PT had the highest value in the second year. Since the third year, grain yield significantly increased in PT/ST, and the yield in NT had the lowest value.

4. Discussion

Soil quality is the comprehensive performance of diverse characteristics from soil physical, chemical, and biological indicators influenced by different land use cultivation [37,38]. In the present study, the chosen soil indicators calculating SQI were different in different soil layers and different treatments. First and foremost, PCA and RDA analysis confirmed that the indicator of θ_g was performed to calculate SQI in 0–50 cm. The result was consistent with the result produced by Shi et al. that θ_g played a dominant role in preserving soil quality [39]. Here, θ_g decreased with depth under all tillage practices, and PT/ST had the highest value of θ_g while NT had a minimum value. However, a previous study showed that NT is conducive to maintain soil moisture [5]. This may be because the effects of tillage on θ_g are complex, and depend on climate conditions and other soil properties [40]. For instance, soil ρ_b is routinely estimated to characterize the state of soil compactness in response to land use and determines soil water infiltration. In the present study, different tillage practices caused an increase in ρ_b with soil depth, and the ρ_b of PT/ST was significantly lower than that of other treatments. On the contrary, the treatment of NT had the highest ρ_b. The result precisely indicated that ρ_b is significantly negatively correlated with θ_g [41]. In addition, values of P_c and P_t were highest in PT/ST, while NT had the lowest P_c and P_t, which were similar to the results of López-Fando and Pardo [42]. Subsoiling tillage and deep plow tillage can loosen soil; these results suggest that PT/ST promoted rearrangement of soil particles and lengthened distances between soil aggregates, and resulted in more soil moisture in soil profiles [43].

Moreover, soil aggregate stability elucidates the effects of tillage practices on soil structural stability and is regarded as an essential influential indicator affecting soil quality [44,45,46]. Soil aggregates are dived into macro-aggregates (WSA > 0.25 mm) and micro-aggregates (WSA < 0.25 mm), and the higher the content of macro-aggregates, the better the agglomeration and stability of soil aggregates [46,47]. The present study showed that PT/ST had higher soil water-stable aggregates of 0.25–5 mm than the other three treatments in the 0–50 cm soil layers, and 3ST + PT increased soil water-stable aggregates of 1–5 mm in the 0–20 cm soil layers. This is mainly because three-year ST was less sensitive to aggregate breakage in 0–20 cm soil layers than 20–50 cm soil layers, and formed a relatively stable soil profile in 0–20 cm soil layers [48]. And the fourth-year PT disrupted soil stability and hindered formation of water-stable large aggregates below the 20 cm soil layers [48,49]. However, a rotation tillage of PT/ST overturned and mixed the upper and lower soil layers under annual tillage practices in 0–50 cm soil layers; the soil profile had more stable soil aggregates [18,50]. In addition, soil aggregates stability can be expressed by MWD and GMD, which are correlated with the amount of WSA > 0.25 mm [51]. In the 20–50 cm soil layers, MWD and GMD under PT/ST were higher than those under other treatments. The result was in agreement with the observation of Wang et al. that rotation tillage could improve soil aggregates [52]. On the contrary, a study showed that no tillage improved the characteristics of GMD and MWD comparing with rotation and conventional tillage [53]. In the present study, NT did not disturb soil, but it significantly increased soil ρ_b. With crop residues and other organic matter gathering in the topsoil under NT, soil microbial activities could not be promoted, while soil aggregates stability was reduced instead [54]. The results were consistent with the soil aggregates distribution that we observed above. Overall, rotation tillage of PT/ST is an effective tillage practice to improve soil aggregate stability in the Loess Plateau.

Soil chemical properties are related to soil physical characteristics and altered by land tillage practices. In the present study, RDA results showed the same conclusion, that θ_g, ρ_b, P_c, P_t, and soil aggregate stability were significantly correlated with soil nutrients and enzyme activities. Nitrogen is an important element participating in producing chlorophyll pigment and photosynthesis [55]. With the decrease in soil water content, soil nitrate nitrogen was significantly increased instead of leaching with soil depth [56]. However, the present study showed that soil nitrate nitrogen and θ_g decreased simultaneously with soil depth, and soil nitrate nitrogen content under PT/ST was the highest. It may be due to plant nitrogen utilization that the adoption of PT/ST was conductive to reducing soil nitrate nitrogen leaching and promoting nitrate nitrogen absorption and utilization by plants. In addition, the content of SOC under PT/ST was the highest in the 0–50 cm soil layers. However, it is well known that NT could increase SOC storage in the topsoil [57,58]. Here, with increased ρ_b and decreased θ_g under NT, topsoil SOC was washed away instead of penetrating deep into the soil [59,60]. Soil enzyme activities, as comprehensive indicators reflecting the state of soil fertility, also play an important role in soil nutrients cycling [61]. Adoption of PT/ST significantly increased soil invertase activity in the 0–30 cm soil layers to promote SOC accumulation.

A previous study elucidated that NT contributes to mitigating disadvantageous effects caused by conventional tillage and increased crop yield [5]. However, continuous no tillage could cause soil compaction in the topsoil and subsoil, and the long-term influence of no tillage is dependent on soil type and climatic conditions [62]. In the Loess Plateau region of this experiment, NT decreased soil aggregate stability and increased ρ_b, resulting in uneven distribution of soil water and nutrient. Decreased SOC and soil nitrogen could reduce root growth, resulting in poor yield [63]. Soil quality reflects the integration of soil physical, chemical, and biological characteristics. Here, the grain yield was positively correlated with SQI values in the 0–10 cm and 40–50 cm soil layers, and PT/ST had the highest value of SQI, followed by treatments of 3ST + PT, 3PT + ST, and NT (Figure 4, Table 5). These observations agreed with the result that grain yield under PT/ST was higher than other treatments, especially NT. In arid or semiarid fields, precipitation is the dominant environmental factor affecting wheat growth and production. Precipitation during summer fallow (P_F) accounts for approximately 60% of total precipitation, and it was positively correlated with grain yield in the Loess Plateau [63]. Table 1 shows that the P_F was extremely low in the fourth year; this is why the yield was reduced in the fourth year (Table 6). In addition, no tillage in the dry season could prevent rainfall from infiltrating and remaining in the deep soil, and it is incapable of protecting soil quality [64]. Without sufficient precipitation, NT amplified the hardship whereby winter wheat had high yield. These results indicate that the adoption of PT/ST could be a potential approach to improve winter wheat yield based on progressive soil quality in the Loess Plateau. However, the study maintained the endless discussion regarding whether good soil might be influenced by land tillage practices.

5. Conclusions

This study showed that different tillage practice during summer fallow influenced soil characteristics. Long-term no tillage exacerbated soil compaction and accumulated fertilizers and straw on the topsoil, resulting in uneven distribution of organic matter, which was not conductive to the improvement of soil chemical quality for a dryland wheat field. However, the adoption of rotation tillage loosened and mixed the soil evenly, allowing the effective allocation of soil organic matter and moisture. In addition, the results of RDA and PCA analysis indicated that the indicators of ρ_b, θ_g, soil urease activity, P_t, soil particle size of 1–5 mm, P_c, soil water-stable aggregates of 1–5 mm, soil water-stable aggregates of 0.25–1 mm, soil particle size of 0.25–1 mm, soil particle size of >5 mm, AN, soil catalase activity, and GMD were suitable to calculate SQI in 0–50 cm soil layers in the Loess Plateau. And rotation tillage significantly increased the SQI and grain yield. Thus, these results elucidate that rotation tillage is a reliable cultivation pattern in dryland wheat fields.