Straw Return or No Tillage? Comprehensive Meta-Analysis Based on Soil Organic Carbon Contents, Carbon Emissions, and Crop Yields in China

Abstract

Conservation tillage methods, including straw return (SR) and no tillage (NT), are widely used to improve the soil organic carbon (SOC) content and crop yield. However, applying SR or NT separately has become a common practice for farmers producing different crops or those in different regions. Evaluating the effects of SR or NT on the SOC content, carbon emissions, and crop yield are important for guiding the correct application of conservation tillage and promoting sustainable agricultural development. Therefore, we conducted a meta-analysis based on 1014 sets of data obtained in China to assess the effects of SR and NT on the SOC content, carbon emissions, and crop yield. Compared with no straw return, SR increased the SOC content and crop yield by 10% and 8.6%, respectively, but with no significant impact on carbon emissions. Compared with conventional tillage, NT increased the SOC content by 2.9% and reduced the carbon emissions and crop yield by 18% and 3.9%, respectively. We also found that SR combined with NT had an additive effect, where the combination improved SOC more than applying SR or NT alone. If applying SR or NT alone, the specific climatic conditions, soil characteristics, and field management strategies need to be considered to maximize SOC. In particular, SR should be used in limited hydrothermal conditions (low temperature or low precipitation) and areas where rice–wheat rotation is implemented. NT can be used under any climate conditions, but it can effectively increase the SOC content in continuous wheat cropping areas.

1. Introduction

The soil organic carbon (SOC) pool is the largest carbon pool in terrestrial ecosystems, and it plays a crucial role in the carbon cycle [1]. Small changes in SOC may have significant effects on the coupled carbon cycle–climate system [2]. However, due to problems, such as the unreasonable exploitation of land resources, over-tillage, and soil erosion, the SOC content of farmland in China is generally low and more than 30% lower than the world average [3,4,5,6,7]. Thus, improving the SOC content of arable land is an issue of great concern in agriculture.

Conservation tillage methods, including straw return (SR) and no tillage (NT), are widely recognized as win–win strategies for increasing the SOC content and mitigating the effects of climate change [8,9]. The main effects of SR are increasing the soil organic matter (SOM) content, microbial activity, and nutrient availability to accelerate the formation of macro-aggregates (aggregate sizes > 0.25 mm), thereby increasing the SOC content [10,11]. NT minimizes soil disturbance to reduce the damage to soil aggregates, as well as enhances the capacity for adsorbing organic molecules and the aggregation of soil minerals to promote SOC storage and soil fertility improvement [12,13]. However, in practical applications of conservation tillage, implementing SR or NT separately has become a common practice for farmers producing different crops or those located in different regions [14,15,16,17]. For example, under most NT conditions, farmers remove straw so as not to impede seeding. Conversely, under CT conditions, straw is usually plowed up and mixed with the soil.

The effects of conservation tillage on the SOC content can be influenced by many variables. In particular, the SOC content is strongly influenced by environmental factors (e.g., climate factors, soil bulk density, pH, and soil texture) and field management practices (e.g., cropping system, planting mode, experimental duration, and N application rate) [18,19,20,21]. For example, climate factors, especially the mean annual temperature (MAT) and mean annual precipitation (MAP), are critical factors that constrain improvements in the SOC content in China due to their effects on the microbial biomass and organic matter decomposition rate [22]. Previous studies have shown that the sensitivity of the rate of decomposition of organic matter to temperature and precipitation decreases as the temperature and precipitation increase. Thus, compared with no straw return (NSR), SR can contribute to increasing the SOC content under low-temperature and low-precipitation conditions [23,24]. In addition, the soil texture has significant effects on the SOC content due to the influence of tillage practices. For example, Wang et al. observed that SR significantly increased the SOC content compared with NSR in clay loam soils [25], and the SOC content was significantly increased by implementing NT in sandy soil compared with conventional tillage (CT). Similarly, Berhane et al. suggested that crop rotation could be more beneficial for increasing the SOC content compared with continuous cropping under SR practices [26]. Moreover, maize, rice, and wheat are the three most important cereal crops in China, with a harvested area of 96 million hectares in 2022 according to FAOSTAT data, which comprised 52.04% of the total harvested crop area in China [27]. Applying SR or NT in the production of these three major food crops to increase the SOC content while simultaneously minimizing greenhouse gas emissions is considered the key to sustainable agricultural development in China [28,29], but the appropriate use of SR and NT in the production of major crops in different cropping systems is still controversial. Therefore, it is important to quantitatively assess the effects of environmental factors and field management practices on changes in the SOC content under different tillage practices.

Previous studies evaluated the impact of SR or NT on the SOC content of farmland in China, but systematic summaries and comparisons were not conducted, and the best tillage practices for increasing the SOC content were not identified. Due to large differences between the studies, it is difficult to directly compare their results, but meta-analysis is suitable for merging the results obtained from multiple studies of the same research topic to provide comprehensive conclusions [30]. Thus, in the present study, we conducted a meta-analysis in order to achieve the following: (1) to assess the effects of SR and NT on the SOC content, carbon emissions, and crop yield; (2) to identify the factors that might influence how SR and NT affect soil carbon sequestration and yields; and (3) to identify the best cropping patterns for use in different regions or crops.

2. Materials and Methods

2.1. Data Collection

We collected papers published before March 2024 to compare the effects of SR and NT on the SOC contents, carbon emissions, and crop yields in different regions throughout China. The studies were retrieved from the Web of Science and China National Knowledge Infrastructure using the keywords “straw mulching”, “straw return”, “soil organic carbon”, “no tillage”, “carbon emissions”, and “crop yield”. The retrieved studies were screened according to the following criteria: (1) the studies were conducted in China; (2) the selected studies were limited to field experiments that included SR or NT; (3) at least one of the target variables was reported (SOC, carbon emissions, or crop yield); (4) there were at least three replicates of each treatment; (5) when multiple sampling dates were reported in the results, the latest observed values were used; (6) studies were limited to the three major grain crops comprising maize, rice, and wheat; and (7) relevant parameters were reported as means with the standard error (SE) or standard deviation (SD), or calculated from the data. Data were extracted from figures using Get-Data Graph Digitizer (version 2.25). If only the SE was reported, the SD was obtained using the formula: SD = SE × $\sqrt{\mathrm{n}}$ [31]. If the SD or SE was not provided, the SD was calculated as 1/10 of the mean value [32]. In studies that reported SOM, the values were converted into SOC concentrations using a conversion factor of 0.58 [33]. Climate data were sourced from the publications. If the MAT and MAP values were not reported, they were obtained from the WorldClim website according to the location of the experiment [34,35].

This study collected and collated relevant data published by scholars in China and abroad from 2000 to 2024. In total, 188 SOC observations were collected to compare NT and CT, and 118 observations were collected to compare SR and NSR. The numbers of crop yield observations were 365 and 262 to compare NT with CT and SR with NSR, respectively. The numbers of carbon emissions observations were 46 and 35 to compare NT with CT and SR with NSR, respectively. The locations of the experimental sites are illustrated in Figure 1.

2.2. Data Categorization

The terms were defined as follows: SR, all crop residues returned to the field after the previous crop was harvested; NSR, all crop residues removed after the previous crop was harvested; NT, no tillage of the land after the previous crop was harvested; and CT, plowing of the land after the previous crop was harvested. In order to facilitate comparisons of the effects of SR and NT on SOC, carbon emissions, and crop yields in different studies, the relevant experimental information extracted from the studies was grouped according to the conditions (Table S1). Based on the similarity of the climatic characteristics and geographical conditions, China was divided into six regions. Basic information for the samples is shown in Table S2. At least 10 observations per subgroup, or more than 5 observations but derived from at least two independent studies, were included in the analysis for each level of categorization [36]. Meta-analysis was conducted to examine how various factors influenced the effects of SR and NT and to identify sources of heterogeneity.

2.3. Meta-Analysis

In order to systematically compare the differences in SOC, carbon emissions, and crop yields between the experimental and control groups (SR and NT), we quantified the effects by using the logarithm of the response ratio (lnR) as the effect value [37].

$$\mathrm{l}\mathrm{n}\mathrm{R}=\mathrm{l}\mathrm{n}({\displaystyle \frac{{\mathrm{X}}_{\mathrm{t}}}{{\mathrm{X}}_{\mathrm{c}}}})=\mathrm{l}\mathrm{n}{\mathrm{X}}_{\mathrm{t}}-\mathrm{l}\mathrm{n}{\mathrm{X}}_{\mathrm{c}}$$

(1)

$$\mathrm{V}(\mathrm{l}\mathrm{n}\mathrm{R})={\displaystyle \frac{{\mathrm{S}\mathrm{D}}_{\mathrm{t}}^2}{{\mathrm{N}}_{\mathrm{t}}{\mathrm{X}}_{\mathrm{t}}}}+{\displaystyle \frac{{\mathrm{S}\mathrm{D}}_{\mathrm{c}}^2}{{\mathrm{N}}_{\mathrm{c}}{\mathrm{X}}_{\mathrm{c}}}}$$

(2)

In the formula, lnR is the effect value, V(lnR) is its variance, where X_t is the average of the treatment (SR or NT), X_c is the average of the control (NSR or CT), N_t and N_c are the numbers of corresponding replicates, respectively, and SD_t and SD_c are the corresponding standard deviations.

Meta-analysis involves weighted calculation of the effect values for each independent study to obtain the overall average effect value, lnRR. In the calculation, it is necessary to determine the weighting factor W_i for each independent study, and the specific formulae are (3) and (4):

W_i = 1/V_i

(3)

$$\mathrm{l}\mathrm{n}\mathrm{R}\mathrm{R}={\displaystyle \frac{\sum \mathrm{l}\mathrm{n}{\mathrm{R}}_{\mathrm{i}}\times {\mathrm{W}}_{\mathrm{i}}}{\sum {\mathrm{W}}_{\mathrm{i}}}}$$

(4)

The standard error of lnRR (SD) and 95% confidence interval were calculated used the following formulae:

$$95\%\mathrm{CI}=\mathrm{lnRR}\pm 1.96\mathrm{SD}$$

(5)

$$\mathrm{SD}=\sqrt{1/\sum {\mathrm{W}}_{\mathrm{i}}}$$

(6)

If the 95% confidence interval overlapped with 0, SR or NT had no significant effect on the variable. If the 95% confidence interval exceeded 0, SR or NT had a significant promoting effect on the variable. If the 95% confidence interval was less than 0, SR or NT had a significant inhibitory effect on the variable.

A response to SR or NT with significant effects on SOC, carbon emissions, and crop yields was calculated as a percentage change (E) by using:

$$\mathrm{E}=({\mathrm{e}}^{\mathrm{l}\mathrm{n}\mathrm{R}\mathrm{R}}-1)\times 100\%$$

(7)

All analyses were performed using the R package “metafor” (version 4.3.2). We obtained mean effect sizes and 95% confidence intervals by using a random effects model. A funnel plot was used to test for publication bias, and the R package “regtest” was applied to detect asymmetry in the funnel plot, where p > 0.05 indicates no publication bias in the data. The publication bias test results indicated publication bias (Figure S1). The R package “glmulti” was used to analyze the combinations of important variables that affected the relative changes in the SOC content under the two tillage practices. A threshold value of 0.8 was set to identify the most important variables.

3. Results

3.1. Overall Effects of Tillage Practices

The average SOC content, yield, and carbon emissions in China were 17 g·kg⁻¹, 7281 kg·ha⁻¹, and 4094 kg·ha⁻¹ in maize, rice, and wheat under NSR, respectively (Table S3), and 19 g·kg⁻¹, 7822 kg·ha⁻¹, and 3678 kg·ha⁻¹ under SR. However, the average SOC content, yield and carbon emissions were 18 g·kg⁻¹, 8103 kg·ha⁻¹, and 6571 kg·ha⁻¹ in maize, rice, and wheat under CT, respectively, and 19 g·kg⁻¹, 7866 kg·ha⁻¹, and 5441 kg·ha⁻¹ under NT. In addition, we used a random effects model to calculate the combined effect sizes for the SOC content, carbon emissions, and yields in agricultural fields under SR and NT (Table S4). Overall, compared with the control, SR significantly increased the SOC content and yield by 10% and 8.6%, respectively, but had no significant effect on carbon emissions. NT increased SOC by 2.9% compared with the control and reduced the carbon emissions and yield by 18% and 3.9%, respectively (Figure 2).

3.2. Differences in SOC Contents among Regions and Field Management Conditions

Under SR, the maximum increase in the SOC content was found in the southwest region of China (29%), followed by the east region (15%), northeast region (13%), and northwest region (12%) (Figure 3a). SR had no effects in the north and central regions. Under NT, clear increases in the SOC contents were found in the east region (14%) and northwest region (12%), and although the relative changes in the SOC contents were positive in the southwest (11%), northeast (2.7%), and north (1.7%) regions and negative in the central regions (–0.13%), these changes were not significant (Figure 3d).

SR increased the SOC content regardless of the crop rotation system (Figure 3b). Crop rotation had no significant impact on the SOC content under NT, but continuous cropping led to a significant increase in SOC (10%) (Figure 3e). The planting modes had different effects on the SOC contents. Under SR, the maize, wheat–maize, and rice–wheat cropping modes increased the SOC contents by 17%, 9.1%, and 23%, respectively (Figure 3c). Under NT, the wheat and rice–wheat cropping modes increased the SOC contents by 44% and 5.9%, respectively (Figure 3f). However, there were no significant differences in the effects of other cropping patterns on the SOC contents under SR and NT.

Compared with the control, the SOC contents under SR were 18%, 8.2%, and 9.9% higher under low (<120 kg·hm⁻²), medium (120–240 kg·hm⁻²), and high N application rates (>240 kg·hm⁻²), respectively (Figure 4a). Medium N application rates increased SOC under NT (14%). Compared with the control, NT had no significant effects on the SOC contents at low and high application N rates (Figure 4g). These results suggest that SR should be combined with low N application rates and NT with medium N rates.

The effects of SR and NT on SOC varied with the experimental duration and experimental start time. In particular, the SOC contents increased with the experimental duration under SR, where the SOC contents increased by 8.6%, 12%, and 15% after experimental durations of <3, 3–7, and >7 years, respectively (Figure 4b). The SOC contents increased gradually under NT after experimental durations of <3 years (3.6%) and 3–7 years (7.7%) compared with the corresponding values under CT, but the SOC contents tended to decrease after >7 years (–0.42%) (Figure 4h). In addition, we found that under SR, the SOC contents increased by 6.0% and 12% when the experiments were conducted in 2000–2010 and after 2010, respectively (Figure 4c). The SOC contents increased by 8.1% under NT after 2010, but there were no significant effects on the SOC contents before 2000 and during 2000–2010 (Figure 4i).

Different forms of SR also affected the SOC content compared with the control. The overall increase in the SOC content with covering was 12%, which was higher than that with digging (4.8%) (Figure 4d). In addition, the increase in the SOC content was greater when SR was combined with rotary tillage (RT) or NT compared with the control, i.e., 14% and 13% higher, respectively, followed by CT, with a 4.6% increase in the SOC content; we also found that the combination of SR and NT resulted in 8.4% higher SOC content compared to the combination of SR and CT (Figure 4e). The relative rate of change in the SOC content was higher under NT combined with SR (4.8%) (Figure 4j) than NT without SR.

3.3. Responses of SOC Contents to Environmental Factors

The SOC contents decreased as the MAT and MAP increased, where the SOC contents decreased by up to 0.84% for every 1 °C increase in the MAT (Figure 5a, Table S5). The MAT explained 12.37% of this variation. The SOC contents decreased by 0.83% for every 100 mm increase in the MAP (Figure 5b, Table S5). The MAP explained 6.25% of the variation in the SOC contents under SR. However, the variations in the SOC contents under NT were not explained by differences in the MAT and MAP (p > 0.05) (Figure 5c,d).

SOC, SOM, and the pH also influenced the SOC contents under SR. The proportions of the variance among studies (R²) explained by SOC, SOM, and the pH were 3.72%, 4.93%, and 61.71%, respectively. For every 1 g·kg⁻¹ increase in the initial SOC, the SOC content decreased by up to 0.42% after SR, and for every 1 g·kg⁻¹ increase in the initial SOM, the SOC content increased by 0.27% after SR. However, the SOC content increased by 1.10% for every 1 unit increase in the pH (

Table 1. Meta-regression results for SOC content under soil conditions. SR, straw return; NT, no tillage; SOC, soil organic carbon; SOM, soil organic matter; BD, soil bulk density. p value indicates significance at p lower than 0.05. * represents significance at p lower than 0.05; ** represents significance at p lower than 0.01; R², heterogeneity explained by explanatory variables; r, correlation coefficients.

Variable	SR			NT
	p Value	R2 (%)	r	p Value	R2 (%)	r
SOC	<0.05 *	3.72	−0.19	0.63	0.00	−0.04
SOM	<0.05 *	4.93	−0.22	0.43	0.00	−0.05
pH	<0.01 **	61.71	0.79	<0.01 **	59.68	0.77
BD	0.06	24.92	0.50	<0.01 **	48.73	0.70
Soil texture	<0.01 **	32.41		<0.01 **	11.31

and Table S5). The pH and bulk density affected the SOC contents under NT. The SOC contents increased by up to 12.73% for every 1 unit increase in the pH and increased by 2.55% for every 1 unit (g·cm⁻³) increase in the bulk density (Table 1 and Table S5). The proportions of the variance among studies explained by the pH and bulk density were 59.68% and 48.73%, respectively. Compared with NSR, the increases in the SOC contents under SR in loamy, sandy, and clay soils were 24.47%, 7.66%, and 8.72%, respectively (Figure 4f). However, NT only increased the SOC content by 5.89% in loamy soil conditions (Figure 4k).

3.4. Relative Importance of Effects on SOC Contents under SR and NT

Model selection analysis confirmed that the effects of SR on the SOC contents were mainly determined by the MAT, region, experimental duration, cropping system, and planting mode (Figure 6a). The relative changes in the SOC contents were also influenced by the tillage method, experimental start time, SR method, initial SOC, initial SOM, pH, and soil texture. Moreover, the effects of NT on the SOC contents were mainly determined by the soil texture, planting mode, and MAP (Figure 6b). The relative changes in the SOC contents were also influenced by the N application rate, cropping system, region, experimental start time, SR, pH, bulk density, and soil texture.

4. Discussion

4.1. Overall Impacts on Farming Practices

In the present study, we investigated the effects of SR and NT on the SOC contents of agricultural fields in China using meta-analysis. SR increased the SOC content by 10.00% (Figure 2a). A previous meta-analysis by Zhao et al. found that SR significantly increased the SOC content of fields in China by 14.50%, while Xia et al. showed that SR increased the SOC content by 14.90% at a global scale [38,39]. SR also increased the crop yield by 8.60% (Figure 2b), and a similar result was reported by Liu et al. [40,41]. As a natural fertilizer, straw is rich in a variety of nutrients (N, P, K, etc.) and organic matter, which can change the quantity and composition of the soil microbial community, thereby improving the soil quality and crop yields [42,43,44,45]. We found no significant differences in the carbon emissions under SR in China (Figure 2c), but a previous global study indicated that SR can increase CO₂ emissions by 2.18% [46].

Compared with CT, NT increased the SOC contents by 2.92% and decreased the crop yield by 3.89% (Figure 2a,b). Similar results were obtained in a previous meta-analysis based on China [47] and a global meta-analysis [48], which showed that NT increased the SOC contents by 3.80% and reduced the yield by 5.00%, respectively. NT affects the SOC content by closing some pathways for water and air exchange between the surface and atmosphere [49]. The reduced crop yields under NT are mainly due to soil compaction, stunted seedling growth, and increased frequencies of weeds and pests [50,51]. In addition, we found that NT reduced carbon emissions by 18.13% (Figure 2c), which is consistent with previous findings [52].

Our results also showed that combining SR with NT had an additive effect, where the combination of SR and NT increased the SOC contents more than implementing SR or NT alone (Figure 4e,j). A meta-analysis by Islam et al. indicated that the combination of SR and NT was highly beneficial, where it increased the surface soil SOC contents by 14.30% and 7.38% compared with SR or NT alone, respectively [53]. Meanwhile, Luo et al. demonstrated that SR combined with NT is a more effective management measure to increase SOC content through 69 experimental data sets worldwide [54]. In an 11-year in situ field trial in the North China Plain, Xu et al. found that tillage treatments combining SR and NT resulted in the greatest increase in SOC content compared to SR or NT alone [55]. SR combined with NT increases the storage and sequestration potential of SOC, and this additive effect is mainly due to the combination of SR with NT increasing the SOM, microbial diversity, and soil cation-exchange capacity, thereby increasing the soil fertility and SOC content [56]. It has also been suggested that the increase in SOC is due to the fact that the combination of SR and NT protects the soil structure from disruption and promotes agglomerate incorporation of SOC while reducing the rate of organic carbon mineralization [57].

4.2. Effects of Field Management Measures on SOC Contents

The effects of SR and NT on the SOC contents are usually determined by the interactions between multiple parameters. We found that the SOC contents continued to increase under SR, where the SOC contents increased by 14.91% after SR for more than 7 consecutive years, which indicates that the impact of SR on SOC is a long-term process (Figure 4b). The soil fertility is enhanced by long-term SR, thereby avoiding nutrient insufficiencies for crops in fields, and the SOC stocks are highly stable and stress resistant [58,59]. The SOC contents also increased more under SR after 2010 (11.72%) (Figure 4c).

Reasonable cultivation measures can improve the soil structure, physical and chemical properties, water storage, and water utilization efficiency, as well as increase the richness of soil microorganisms to provide a suitable soil environment for promoting seed germination, root growth, and crop development [60,61]. By analyzing the differences in SOC under four tillage measures comprising CT, NT, deep tillage (DT), and rotary tillage (RT) combined with SR, we found that the SR with different cultivation methods significantly increased the SOC contents in the 0–20 cm soil layer, except with the DT cultivation mode. Under SR, RT and NT were most favorable for increasing the SOC contents (Figure 4e). The tillage layer is generally shallow under RT (15–20 cm), and the combination of RT and SR leads to a large amount of straw collecting on the soil surface to form an oxygen-deficient environment, which results in a slower rate of SOM decomposition and a longer residence time for C in the soil [62].

The effects of SR on the SOC contents differed under various cropping modes. The results obtained using meta-analysis showed that the rice–wheat rotation cropping pattern with SR obtained the greatest increase in SOC (by 23.44%) in China (Figure 3b,c), whereas SR had no significant effects under continuous wheat or rice cropping systems, mainly due to long periods of continuous cropping causing problems such as a shallow soil cultivation layer and decreased soil fertility. However, an appropriate rotation tillage mode can effectively address this problem because diverse crops absorb different amounts of soil nutrients, and crop rotation can help to balance the nutrients in the soil and prevent the excessive consumption of certain nutrients [63,64]. The roots of different crops in rotation also provide various substrates to support diverse microbial communities and regulate the soil environment, thereby improving the soil microenvironment [65,66]. The effect of SR on the SOC contents was mainly determined by the experimental duration, tillage system, and cropping pattern (Figure 6a).

Under NT, continuous tillage and wheat growing increased the SOC contents (by 44.09%) compared with rotation tillage (Figure 3e,f). Growing two crops in rotation leads to competition for nutrients, with faster decomposition and the consumption of SOC and nutrients, whereas a single wheat or corn planting system can provide sufficient fallow time and play an important role in improving soil health and fertility, and thus avoid the negative impacts of NT. The effect of NT on the SOC contents was mainly determined by the cropping pattern. Combining NT with the application of N fertilizer can promote the sequestration of soil carbon and bacterial activity levels in the rhizosphere, increase the diversity of soil carbon-sequestering bacterial communities, and further promote the formation of SOC [67]. Our meta-analysis showed that the SOC contents increased at rates of 5.25% to 14.05% under N fertilizer application rates from 0–120 kg·ha⁻¹ to 120–240 kg·ha⁻¹, but the rate decreased to 2.26% when N fertilizer was applied at over 240 kg·ha⁻¹, which is similar to the results obtained in a previous study [68] (Figure 4g). Appropriate N fertilizer application rates can promote the positive impacts of NT on the SOC contents. Similar to SR, the greatest increases in the SOC contents under NT (8.06%) occurred when the trial period started after 2010 (Figure 4i), possibly due to the development and utilization of arable land resources, improved agricultural mechanization levels, and the provision of high-quality germplasm resources through scientific breeding, contributing to higher SOC contents.

4.3. Effects of Climate Conditions on SOC Contents

The differences in the SOC contents among regions were significantly affected by local variables, such as the MAT, MAP, soil texture, and other factors, and thus, the responses of SOC to SR practices varied significantly among regions [69,70]. The highest increases in the SOC contents under SR were observed at lower temperatures (Figure 5a), and similar results were obtained in the meta-analysis conducted by Tian et al. [71]. SR with low precipitation was most favorable for increasing the SOC contents compared with other precipitation amounts (Figure 5b). Low-temperature and low-precipitation conditions inhibit microbial growth and reduce the decomposition of organic matter to increase the SOC content [72,73,74]. Under NT, no significant heterogeneity was found within both MAT and MAP subgroups (p > 0.05) (Figure 5c,d). NT increased SOC under different climatic conditions, but the increases were not statistically significant. Similarly, Hashimi et al. [75] analyzed 64 peer-reviewed articles and found that the differences in effect sizes between different MAT and MAP levels were not significant. Consequently, NT might be suitable for all climatic conditions.

4.4. Effects of Soil Properties on SOC Contents

The initial SOC and SOM contents affected the SOC contents, thereby indicating that the initial fertility contributes to determining the soil C dynamics under SR. The highest SOC content under SR was found when the initial SOC content was at a low level, and the same result was obtained in a previous meta-analysis [76], probably because lower initial SOC values can readily increase to the same threshold level. The MAT, MAP, and other factors can affect the balance between SOC generation and decomposition, but the initial soil properties determine the level of C sequestration [77]. The soil pH is a major factor that influences the stability of biochar and the soil microbial community, and adding biochar to soil is considered a promising method for improving carbon sequestration [78,79]. Under SR, biochar will increase the soil pH, and the SOC content will also increase (Table 1). Similarly, we found that the SOC contents increased under NT as the pH increased (Table 1). Previous studies have widely demonstrated that soil acidification can have deleterious effects on microorganisms, thereby leading to losses of nutrients and the release of toxic heavy metals. Moreover, bulk density is a fundamental physical soil property that affects the supply of soil nutrients and moisture, and it also reflects the productivity [80]. In addition, the soil bulk density is an important parameter for estimating SOC stocks [81]. The relative changes in the SOC contents increased as the soil bulk density increased (Table 1), possibly because the increase in the soil bulk density under NT was accompanied by an increase in the SOM content [82].

The combination of mineral particles of different sizes in soil determines the soil texture. The presence of pores, solidity, percolation, water holding capacity, rate of breakdown of organic matter, and available nutrients are influenced by the soil texture. Thus, the soil texture greatly affects the soil fertility [83]. Our results showed that the SOC contents under SR and NT were higher in loamy soil than sandy and clay soils (Figure 4f,k). Wang and Kong et al. also found that the greatest increases in the SOC contents under SR and NT occurred in loamy soils compared with other soil textures [84,85], possibly because loamy soil has a greater capacity to retain water and fertilizer, and it is rich in minerals with a stable organic matter content [86]. Analysis of the relative importance of different variables demonstrated that the effect of NT on the SOC content was mainly determined by the soil texture (Figure 6b).

4.5. Limitations of Meta-Analysis

In this meta-analysis, we quantitatively analyzed the effects of various factors on the SOC contents under SR and NT in order to understand the mechanisms that might allow these two cultivation measures to improve the SOC content of farmland. In addition, we assessed the effects of SR and NT on crop yields and carbon emissions. However, our study had several limitations. In particular, we cannot guarantee the accuracy of the results because insufficient data were available regarding soil carbon emissions. The effects of SR and NT on the SOC contents may also be influenced by other factors, such as differences in crop varieties, basic soil fertility, and irrigation amounts. Limited data are available on these factors, and some are difficult to extract from previous studies, or more field experiments are required. Consequently, we did not consider these factors in this study. More relevant data should be collected regarding the effects of various factors on the SOC contents under SR and NT to reduce uncertainty.

5. Conclusions

In this study, we used meta-analysis to assess the impacts of SR and NT on the SOC contents, crop yields, and carbon emissions when growing the three major grain crops in China. The results showed that SR increased the SOC contents and crop yields compared with NSR, but it had no significant effects on carbon emissions. Compared with CT, NT increased the SOC contents but decreased the crop yields and carbon emissions. We also found that combining SR with NT had an additive effect, where this combination increased the SOC contents more than implementing SR or NT alone. Therefore, the combination of SR and NT is the best tillage method for increasing the SOC contents. If SR or NT is used alone, the specific climatic conditions, soil characteristics, and field management strategies need to be considered to maximize the SOC contents. In conclusion, SR should be applied under a limited range of hydrothermal conditions (low temperature or low precipitation) and in areas under rice–wheat rotation. NT can be used under any climate conditions, but it is more effective at obtaining higher SOC contents in areas under continuous wheat cropping. Finally, in this meta-analysis, we quantified the effects of SR and NT on the soil SOC contents, carbon emissions, and crop yields based on the available scientific evidence to provide theoretical guidance for the future management of conservation tillage measures and promote sustainable agricultural development. However, it is necessary to obtain more information to explore the underlying mechanisms associated with the changes in the SOC contents under different measures.