Response of Soil Aggregate Stability to Phosphorus, Nitrogen, and Organic Fertilizer Addition: A Meta-Analysis

Abstract

Soil is a highly significant resource for human survival, and agglomerates, as the basic unit of the soil structure, not only enhance the soil fertility and control the biological validity of nutrients but also strengthen the soil’s erosion resistance. The mass application of fertilizers may significantly affect crop growth and the soil structure, and the rational application and dispensing of fertilizers will be an urgent issue to be addressed. Therefore, the effect of fertilizer application on the stability of water-stable soil aggregates needs to be studied under different meteorological and soil conditions to draw more general and feasible conclusions. Our meta-analysis of data from 220 independent observations from 56 published studies found that fertilizer application increased the mean weight diameter (MWD) by an average of 18% compared to the no-fertilizer treatment. Among the nitrogen (N), phosphate (P), and organic (OM) fertilizer treatments, the organic fertilizer treatment had a more significant stimulatory effect on the MWD (26%). Among the different fertilizer levels, a low level of phosphorus (<40 kg·ha⁻¹ yr⁻¹), a high level of N (>120 kg·ha⁻¹·yr⁻¹), and a low level of organic fertilizer (<5000 kg·ha⁻¹·yr⁻¹) increased the MWD by 19%, 14%, and 41%, respectively. Across the soil types and land use types, the response to the MWD was positive for red soils and paddy fields, and the organic fertilizer’s stimulatory effect was more significant than that of the chemical fertilizer. The correlation analysis showed that the response rate of the MWD was negatively correlated with the response rate of the soil pH and bulk density (BD) and positively correlated with the response rate of the soil organic carbon (SOC) and microbial mass carbon (MBC). Meanwhile, the partial least square structural equation model (PLS-SEM) showed that the meteorological factors were the main factors affecting the stability of the soil aggregates, while the secondary factors were the soil’s physical and chemical properties. Therefore, this study found that the long-term use of organic fertilizer instead of partial fertilizer is better than the use of chemical fertilizer alone, while more attention should be paid to the influence of temperature and rainfall on the stability of fertilizer in aggregate soil in the future.

1. Introduction

Due to the influence of economic development and human activities in society, land degradation and soil drought have become increasingly significant issues affecting economic growth, social development, and rural production. Climate change and agricultural activities are likely to be the leading causes of these problems [1]. In the meantime, China’s massive demand for food has created an urgent need to improve land quality, while land degradation is closely related to agricultural activities, among which fertilization is one of the main reasons. For example, common agricultural practices often lead to the misuse of chemical fertilizers, which damage the structure of soil aggregates and soils and waste soil nutrients [2]. Therefore, it is vital to improve soil quality to enhance agricultural production.

Soil organic carbon (SOC) is a crucial index of soil fertility, and soil organic carbon content changes affect soil aggregation. Soil organic carbon is closely related to the formation and stability of soil aggregates [3]. It was reported that soil organic carbon avoids microbial decomposition via adsorption onto the surface of clay minerals and encapsulation in soil aggregates. As an essential component of the soil structure, soil aggregates are the material basis of a good soil structure and not only directly or indirectly affect the soil fertility and crop yield [4] but also increase the stability of the soil aggregates that can improve the soil carbon sink function and reduce greenhouse gas emissions [5]. There are many indicators for evaluating the stability of soil aggregates, among which the mean weight diameter (MWD) has long been used as an indicator of soil aggregate stability, and when the value of the MWD is higher, it indicates that the stability of the soil aggregates is better [6,7]. The stability, particle size, and quantity distribution of soil aggregates are affected by the fertilization type, fertilization level, soil properties, and climatic conditions. Many papers published in China and elsewhere have shown that the decomposing residues of fertilizers after application could stimulate microbial activity, forming mycelium and sugars, and that soil particles would then be cemented by these substances to form soil aggregates [8,9]. The formation of soil aggregates is driven by physical, chemical, and biological factors, with cementing agents being the main formation condition. It had been suggested by Tisdall et al. [3] that large agglomerates were mainly formed by the gelling of mycelium and organic residues, while small agglomerates were formed by the gelling of either polysaccharides or inorganic colloids through cationic bridges.

The use of chemical fertilizers plays a significant role in agricultural output. Adding organic matter to agrarian fertilizer can help improve the soil structure, decrease the bulk density, and affect the soil aggregates’ stability [10]. Fertilizer application has different effects on soil aggregates. Some studies [11,12] showed that long-term fertilizer application can increase the number of macro-aggregates, although others [13,14] reported no significant effect of fertilizer application on the number of macro-aggregates. Different types of fertilizer applications also have different effects on soil agglomerates. Tian et al. [15] found that organic fertilizer alone replenished the soil nutrients and increased soil the macro-agglomerate content and water stability compared to organic–inorganic mixed treatments. Ma et al. [16] showed that the soil water-stable agglomerates’ mean weight and geometric mean diameter were significantly higher in organic fertilizer treatments than in organic-inorganic fertilizer blends and inorganic fertilizer treatments. Li et al. [17] showed that applying organic fertilizer promoted the formation of large agglomerates, while using chemical fertilizer and organic fertilizer increased the agglomerates’ organic carbon and total nitrogen content in the whole soil and at all the grain levels. Additionally, Řezáčová et al. [18] showed that the application of organic fertilizer alone for four and eight consecutive years was beneficial in improving the stability of soil aggregates. Overall, the fertilizer treatments increased the soil’s mean weight diameter compared to the treatments without fertilizer. Therefore, it is essential to explore the role of proper fertilizer application in relation to the soil quality and stability of soil agglomerates. In recent years, most studies on fertilizer application on soil aggregates have focused on the effect of fertilizer application on the nutrient content of soil aggregates and its dynamics and soil aggregate stability [19]. However, there needs to be a detailed description of the mechanism of the effect of a single fertilizer application on soil aggregate stability. Therefore, the research literature on changes in soil agglomerate stability under long-term fertilizer application trials in China was systematically collected in this report. A meta-analysis was used to quantitatively estimate the magnitude of the increase or decrease in the soil agglomerate stability changes by a single fertilizer application to analyze and explore the variability between the extent of the increase or decrease under different conditions. The study aims to elucidate the role of single fertilizer application on the formation of water-stable agglomerates and to provide a scientific basis for the rational cultivation of the soil, improvement of the soil structure, and reduction of global warming and carbon cycling.

2. Materials and Methods

2.1. Data Compilation

Peer-reviewed journal articles were searched using the Web of Science (WOS), Wanfang, and China National Knowledge Infrastructure (CNKI) databases. The following search term combinations were used to select the studies: (fertilization and aggregates/soil macro-aggregate or aggregate/nitrogen fertilizer or phosphate fertilizer or organic fertilizer) and (fertilizer and aggregates/fertilizer/aggregates/nitrogen fertilizer and aggregates/phosphorus fertilizer and aggregates/organic fertilizer and aggregates) and (nitrogen fertilizer/phosphate fertilizer/organic fertilizer/organic fertilizer/organic fertilizer). The literature on the “single fertilization and aggregate soil stability” themes was selected and collected. To improve the data quality, the literature was screened according to the following criteria: (i) the experiment must be a long-term positioning experiment or a field experiment; (ii) the test must include a control (such as no fertilization) and a treatment (such as a single application of phosphate fertilizer, nitrogen fertilizer, organic fertilizer, etc.), and any other test conditions must be consistent with the control and treatment; (iii) the mean value, standard deviation (SD) or standard error (SE) and sample size (n) of the variables in the papers, tables, and digital charts must be able to be directly extracted; (iv) the paper must have the stability index of the soil aggregates: mean weight diameter (MWD) data, and (v) the test site must be located in China.

The dataset included 220 independent observations from 56 published studies (Figure 1). Here, we considered data concerning different fertilization types, fertilization levels, soil types, and land use types in the same experiment as independent observations (

Table 1. Data groups used in the meta-analysis.

Soil Type	Type of Land Use	Type of Fertilization	P Addition Level (kg·ha−1·yr−1)	N Addition Level (kg·ha−1·yr−1)	OM Addition Level (kg·ha−1·yr−1)
Black soil	Cultivated land	P	>120 (high level)	>120 (high level)	>15,000 (high level)
Red soil	Paddy field	N	40–80 (low level)	80–120 (medium level)	10,000–15,000 (medium level)
Yellow soil	Garden	OM	<40 (lowest level)	40–80 (low level)	5000–10,000 (low level)
				<40 (lowest level)	<5000 (lowest level)

). The GetData (version 2.20) software was used to extract the data from the digitized graph in the paper. Overall, our data set covers different fertilization types, such as phosphate, nitrogen, and organic fertilizers; different soil types, such as black soil, red soil, and yellow soil; and different land use types, such as arable land, farmland, and garden land. The P addition levels varied from 6.0 to 1444 kg·ha⁻¹·yr⁻¹, the N addition levels ranged from 4.5 to 1558 kg·ha⁻¹ yr⁻¹, and the OM addition levels varied from 13.0 to 225,000 kg·ha⁻¹·yr⁻¹. Therefore, according to the data, the horizontal gradients of a single application of nitrogen and phosphate fertilizer were set as >120, 80–120, 40–80, and <40 kg·ha⁻¹·yr⁻¹, while the horizontal angles of the organic fertilizer were set as >15,000, 10,000–15,000, 5000–10,000, and <5000 kg·ha⁻¹·yr⁻¹. Moreover, the rainfall ranged from 120 to 1795 mm. The annual temperature ranged from −1.11 to 19.2 °C.

2.2. Meta-Analysis

We followed the methods that Zhou et al. [20] used to evaluate the responses of the MWD to P, N, and OM addition. In brief, it involved the use of MetaWin 2.1 software for the purpose of calculation. The response ratio (RR, natural log of the ratio of the mean value of a concerned variable in the P, N, and OM addiction treatment to that in the control) was used here as an index of the magnitude of the P, N, and OM addition effect. The RR was calculated as follows:

$$\mathrm{RR}=\frac{{\mathrm{lnX}}_{\mathrm{t}}}{{\mathrm{lnX}}_{\mathrm{c}}}={\mathrm{lnX}}_{\mathrm{t}}-{\mathrm{lnX}}_{\mathrm{c}}$$

(1)

where X_t and X_c are the means of a particular variable in the P, N, or OM addition and control treatments, respectively. If an RR > 0 indicates that the single fertilization has a positive response to the response variable, the single fertilization will improve the MWD. The following equation estimated its variance (v):

$$\mathrm{v}(\mathrm{RR})=\frac{{\mathrm{SD}}_{\mathrm{t}}}{{\mathrm{N}}_{\mathrm{t}}{\mathrm{X}}_{\mathrm{t}}^2}+\frac{{\mathrm{SD}}_{\mathrm{c}}}{{\mathrm{N}}_{\mathrm{c}}{\mathrm{X}}_{\mathrm{c}}^2}$$

(2)

where N_t and N_c are the numbers of fertilized and unfertilized samples, SDc and SDt are the variances of the control and treatment groups, respectively, and ($\mathrm{SD}=\mathrm{SE}\sqrt{\mathrm{N}}$).

MetaWin 2.1 software was used to calculate the response ratio, and then the random effects model was used to calculate the average weighted response ratio (RR₊₊):

$${\mathrm{RR}}_{++}=\frac{{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{m}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{k}}{\mathsf{\omega }}_{\mathrm{ij}}{\mathrm{RR}}_{\mathrm{ij}}}{{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{m}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{k}}{\mathsf{\omega }}_{\mathrm{ij}}}$$

(3)

The calculation formula of the weighted standard error (S) is as follows:

$$\mathrm{S}\left({\mathrm{RR}}_{++}\right)=\sqrt{\frac{1}{{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{m}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{k}}{\mathsf{\omega }}_{\mathrm{ij}}}}$$

(4)

The 95% confidence interval (95%CI) can be calculated as follows:

$$95\%\mathrm{CI}={\mathrm{RR}}_{++}\pm 1.96S\left({\mathrm{RR}}_{++}\right)$$

(5)

where i = 1, 2, 3, …, m; j = 1, 2, 3, …, k; m is the number of groups, and k is the number of comparisons in group i. If the RR₊₊ is positive, it is a positive response; otherwise, it is a negative response.

When the confidence interval included 0, the P, N, or OM addition of the MWD was insignificant (p > 0.05). When all the confidence intervals were more significant than 0, adding P, N, or OM significantly increased the MWD (p < 0.05). In contrast, adding P, N, and OM significantly decreased the MWD (p < 0.05).

We mainly used a partial least square structural equation model to analyze the causal relationship between the meteorological and soil physical and chemical indexes and the MWD. The partial least squares structural equation model (PLS-SEM) is a partial least squares-based analysis of variance method. An iterative estimation combining a principal component analysis with a multiple regression is also a causal modeling method. The principle is to assume causality in a set of latent variables that can be represented separately by a group of observed variables. It usually includes some basic linear regression models and many observed variables. Technically, verifying the covariance between the observed variables makes it possible to estimate the basic linear regression model’s coefficient values (NFI, SRMR, and χ²), thus statistically testing whether the assumed model is suitable. If appropriate, our relationship model can be considered reasonable [21].

3. Results

3.1. Response of MWD to P, N, and OM Addition

The weighted RR of the MWD across all 220 pairs of comparisons was 0.1836 (18.45, p < 0.05) (Figure 2a,b). Application of phosphorus, nitrogen, and organic fertilizer significantly (p < 0.05) increased the MWD by 11%, 9%, and 26%, respectively. Among them, the effect of the organic fertilizer on the MWD was more significant (Figure 2b). The P addition increased the MWD by 19% and 9% (p < 0.05) at the lowest and low P treatments (Figure 3a). The MWD increased by 14% and 14% (p < 0.05) at the medium and high N treatments (Figure 3b). Meanwhile, the OM addition increased the MWD by 41%, 25%, 11%, and 13% (p < 0.05) at the lower, low, medium and high OM treatments (Figure 3c).

3.2. Differences between Soil Aggregate Stability under Different Soil Types and Land Use Types

The addition of P, N, or OM did not significantly (p > 0.05) increase or decrease the MWD in black soil (Figure 4a). On the contrary, the addition of P, N, or OM had a significant (p < 0.05) positive effect on the increase in the MWD in red soil (Figure 4b). Additionally, adding organic fertilizer increased the MWD more than phosphorus and nitrogen, which were 44%, 17%, and 14%, respectively. Furthermore, only the addition of organic fertilizer significantly (p < 0.05) increased the MWD by 30% in yellow soil (Figure 4c). In general, adding P, N, or OM in red soil increased the MWD more than in black and yellow soil.

In terms of the land use types, the MWD response in the paddy field (weighted RR: 0.27, p < 0.05) was significantly greater than those in the garden (weighted RR: 0.18, p < 0.05) and cultivated land (weighted RR: 0.16, p < 0.05) (Figure 4d–f). The addition of organic fertilizer significantly (p < 0.05) increased the MWD more than phosphorus and nitrogen, which were 27%, 12%, and 9% in cultivated land (Figure 4d), respectively. The addition of P, N, or OM had a significant (p < 0.05) positive effect on the increase in the MWD in the paddy field (Figure 4e). Furthermore, the addition of organic fertilizer significantly (p < 0.05) increased the MWD by 17%, while the addition of N mainly (p < 0.05) decreased the MWD by 241% in the garden (Figure 4f).

3.3. The Correlations of the Responses of the MWD with Meteorological and Soil Factors

A linear relationship was found between the RR of the MWD and the mean annual precipitation (MAP) (r = 0.2133, p = 0.0044) and mean annual temperature (MAT) (r = 0.1535, p = 0.0484). When the MAP and MAT were close to 434 mm and 3.7 ℃, respectively, P, N, or OM addition had the most potent stimulating effect on the RR of the MWD (Figure 5a,b). The RR of the MWD exhibited significant negative relationships with the RR of the pH (Figure 6a, r = −0.5717, p = 0.0068) and bulk density (BD, Figure 6b, r = −0.2962, p = 0.0005), and significant positive relationships with the RR of the soil organic carbon (SOC, Figure 6c, r = 0.2054, p = 0.0364) and microbial biomass carbon (MBC, Figure 6d, r = 0.5397, p = 0.0208). Moreover, the RR of the MWD was positively (r = 0.4671, p = 0.0438) correlated with the soil pH (Figure 6e) and showed a negative linear relationship (r = −0.3648, p = 0.0162) with the BD (Figure 6f). In addition, the RR of the MWD exhibited significant positive associations with the SOC (Figure 6g, r = 0.2143, p = 0.0323), and a negative linear relationship was found between the RR of the MWD and the MBC (r = −0.5187, p = 0.0476), exhibiting a downward trend with an increasing MBC (Figure 6h).

3.4. Path Analysis of Influencing Factors of the MWD Response

This study used the PLS-SEM model to analyze the soil aggregate stability response influencing factors to adding P, N, and organic fertilizer. The fitting parameters of the conceptual model were all within the target range, which proved that the model data were reliable. The results showed that the soil pH, soil organic carbon (SOC), mean annual temperature (MAT), mean annual rainfall (MAP), bulk density (BD), and microbial biomass carbon (MBC) all had specific effects on the stability of the soil aggregates (MWD) (standardized path coefficient (SPC), SPC = 0.334, 0.010, 0.352, −0.319, 0.064, 0.177). In addition to the MAP’s adverse effects on the MWD, BD, and SOC, other paths had positive results. In short, the meteorological factors had a more significant impact on the stability of the soil aggregates, and the soil factors were secondary factors (Figure 7).

4. Discussion

4.1. Response of MWD to P, N, and OM Addition

Based on 220 independent observations, the soil and climate responses to P, N or OM addition were examined. The potential mechanisms by which soil and climate factors regulate P-, N-, or OM-induced changes in the MWD were discussed. The results of the meta-analysis showed that the addition of P, N or OM had a significant effect on the MWD, indicating that the addition of P, N or OM affected the soil structure and soil quality. This outcome is consistent with the results of previous studies [22,23]. This is because the input of nitrogen fertilizer, phosphate fertilizer, and organic fertilizer provides many available carbon sources and nutrient elements for microorganisms in the soil, improving the effectiveness of the substrates. In addition, the soil microenvironment and physical and chemical properties will be changed, which may affect the soil aggregates’ stability.

The stability of soil aggregates is highly dependent on the type and extent of the fertilizer application. For the different fertilizer types, our study showed that P, N, or OM fertilizer treatments significantly increased the MWD, with the OM greatly expediting the increase in the MWD, which is consistent with the results of Hati et al. [24]. The reason for the difference may be that organic fertilizer provides sufficient carbon and nutrients for the growth and reproduction of microorganisms and produces a solid positive incentive effect, thus affecting the stability of soil aggregates. Wang et al. [25] found that long-term treatment with manure significantly enhanced the tendency of water-stable micro-agglomerates to aggregate into water-stable macro-aggregates. Zhang et al. [26] and Karami et al. [27] also found that combining fertilizer and organic fertilizer significantly increased the MWD value of soil aggregates. Moreover, some studies reported that the stability of >0.25 mm soil aggregates under organic fertilizer treatment was greater than that of nitrogen fertilizer alone. This may be because organic manures increase the role of soil organic cement in the agglomeration process and boost soil aggregate formation [28].

In addition, P and N fertilizers alone also contributed significantly to the stability of the soil aggregates, probably owing to the enhancement of soil microbial activity and enhanced breakdown of organic matter by applying phosphorus or nitrogen fertilizers, respectively. It has been reported that the complex salt-based effect of phosphate fertilizer application improves colloidal material [29]. Wang et al. [30] also found that nitrogen fertilizer application could promote macro-agglomerate formation by alleviating the limitation of soil microorganisms, such as Actinomycetes, by nitrogen.

Under different fertilization levels, the MWD was increased by 9% and 19%, respectively, under the ultra-low and low phosphorus treatments. This was consistent with the finding of Zhang et al. [23] that agglomerate stability decreases with increasing phosphorus fertilizer. Blanco-Canqui et al. [31] also found that the quality of soil aggregates decreases with reduced rates of over-application of phosphorus fertilizer. Therefore, long-term mono-application of phosphorus fertilizer tends to lead to nitrogen deficit and land degradation. The MWD increased by 14% in both the high and medium nitrogen fertilizer treatments. This phenomenon may be related to the addition of nitrogen-inducing cementing substances. It has been found that high N fertilizer application caused smaller particles in the soil to gel into large water-stable soil aggregates, which were associated with substances such as root systems and secretions [32]. In the case of both high and low organic fertilizer treatments, the MWD was increased, with the ultra-low organic fertilizer treatment showing the most significant increase of 41%. Contrary to the results of existing studies, Gao et al. [33] and Zhao et al. [34] concluded that the MWD of soil aggregates was higher in high-volume organic fertilizer than in low- and medium-volume organic fertilizer. This may be because high organic fertilizer application reduces the soil bulk, increases the soil porosity, leads to increased microbial activity, accelerates organic carbon decomposition within soil aggregates, reduces the cementitious material, and thus affects soil aggregate formation and stability. Oorts et al. [35] suggested that the soil aggregate content and stability were significantly and positively correlated with the soil fertility levels. However, it was also believed that the application of chemical fertilizer alone would accelerate the mineralization of soil organic matter by soil microorganisms and cause soil slumping and destruction of the soil aggregate structure, which would not be conducive to the survival of soil aggregate stability [26]. For example, the long-term application of chemical fertilizers led to soil slumping in paddy soils, while organic fertilizer significantly improved the soil aggregate stability [36]. This shows that changes in fertilizer application patterns affect the process of soil agglomeration formation and alter the soil agglomeration structure and stability.

4.2. Differences between Soil Aggregate Stability under Different Soil Types and Land Use Types

Different fertilizer applications’ effect on soil aggregates’ stability varies between soil types. Our results found that the N, P, and organic fertilizer additions only showed a significant increase in response to the MWD in red soils and a largely insignificant increase in black and yellow soils. This suggests that black and yellow soils treated with N or P fertilizers alone are less susceptible to the formation of large agglomerates and have lower soil stability, which may lead to soil structural degradation. Organic fertilizer, but not chemical fertilizers [37], mainly influenced soil aggregates. For example, Li et al. [38] found that organic fertilizer significantly increased the content of >2 mm aggregates and the MWD value in red soil. Di et al. [39] also found that organic fertilizer significantly increased the content of stable large aggregates (>2 mm) and larger aggregates (0.25–2 mm) in red soil and paddy soil, while chemical fertilizer application had little effect on soil stability and even reduced the soil aggregate stability. This may be explained by the fact that red loamy rice soils with clay minerals are mainly dominated by iron and aluminum oxides, which are more compact and less water-stable [40]. Furthermore, contrary to the results of this study, Zhang et al. [41] found that organic fertilizer significantly reduced the proportion of soil aggregates with a particle size of >2 mm and the MWD values in black soils, while chemical fertilizer application had no significant effect on the proportion of soil aggregates distributed by particle size. Zhang et al. [42] found that chemical and organic fertilizer applications alone increased the soil aggregate content in a black soil study. Yuan et al. [43] also found that organic fertilizer accelerated the turnover of macro-agglomerates in black soils and that the macro-agglomerate turnover increased with increasing organic fertilizer application. This suggests that different soil types and fertility differences, and others, have a great influence on the soil agglomerate distribution, which may be explained by the fact that different types of soil aggregates form different cementing substances and quantities, resulting in an inconsistent response of soil aggregate stability to fertilizer application under different soil types.

Our study found that the N, P, and OM treatments significantly increased the MWD in cultivated and paddy land, while the N fertilizer treatments decreased the MWD in gardens. This phenomenon is inconsistent with the results of existing studies. For example, Qi et al. [44] found that the structural characteristics of the soil stability aggregates in farmland were not significantly different from those in gardens and orchards, while in this study, it was found that the addition of nitrogen resulted in a 241% reduction in the MWD, which may be because nitrogen fertilizer was applied in the garden alone and only N was ingested. In addition, the soil permeability of the garden is good. Still, the drought is barren, the soil erosion is severe, and the water and fertilizer retention ability could be better, thus reducing the formation and stability of aggregates. However, Shao et al. [45] found that the combination of organic and inorganic fertilizers in cultivated land facilitated the soil organic carbon accumulation and significantly increased the soil fertility. Fan et al. [46] reported that different fertilizer treatments on paddy soils greatly affected the MWD of water-stable aggregates. This may be caused by damage to the soil structure due to anthropogenic disturbances such as cropping systems and tillage. Moreover, Luo et al. [47] found that by studying the effects of different land use practices on the content of water-stable macro-agglomerates, the trend in the soil agglomerate content was as follows: wasteland > forest land > garden land > cultivated land. These findings suggest that land use for human activities may directly damage macro-magnets on the one hand and alter the soil environment on the other, thus affecting the plant growth and microbial activity and inhibiting the formation and stabilization of micro-magnets [48].

4.3. Correlation and Path Analysis of the MWD Responses with Soil and Meteorological Factors

Our study found that the response ratio of the MWD was greatest for rainfall of 434 mm and a temperature of 3.7 °C, and the average annual temperature and average annual rainfall are the main factors influencing soil aggregate stability, respectively. This indicates that stronger rainfall increases the striking force of raindrops on soil and increases the degree of dispersion between soil particles. In addition, excessive rainfall also generates an anaerobic environment, which leads to the possible limitation of soil microbial activity and possible dry–wet alternation aggravating the damage to the soil aggregate stability [49]. Therefore, meteorological factors, as external conditions, have a greater influence on the stability of soil aggregates than soil factors. Long-term fertilizer treatments also significantly affected the soil pH, with a negative relationship between the soil pH and WMD response ratio and a significant increase in the MWD in alkaline soils compared to acidic soils. Consistent with the results of existing studies, the addition of organic fertilizer may increase the soil pH compared to chemical fertilizer treatments [50]. He et al. [51] also found that organic fertilizer and the pH affect iron and aluminum oxides, which in turn affect soil aggregate stability. This is also possible because the soil pH affects the structure of soil aggregates by influencing other clay minerals [52,53]. Many studies have found that fertilizer application reduces the soil bulk density and increases soil porosity, and it is believed that organic and inorganic fertilizers are more effective when applied in combination [54,55], which is consistent with the results of this study.

This study showed a positive correlation between the organic matter content and the number and stability of soil aggregates, which was consistent with the results of existing studies. For example, Yao et al. [56] found a linear correlation between the soil organic matter and the MWD of water-stable soil aggregates. Geng et al. [57] found that long-term application of inorganic fertilizer increased the organic carbon content of aggregates, while organic fertilizer application accelerated the agglomeration of small aggregates into large aggregates. Organic matter increased the hydrophobicity of soil aggregates and reduced the damage to soil aggregates caused by air filling in soil particles [58]. However, there are contrary findings. For example, some studies found that the relationship between the soil organic carbon and the MWD was not significant or negatively correlated [59]. This is because the formation, stability, and turnover of aggregates are affected not only by the soil organic matter but also by many factors, such as the gelation of root secretions and the activities of microorganisms between the roots. Microbial biomass promotes soil aggregation through the production of polysaccharides and mycelium [60]. In addition, long-term fertilizer application may inhibit microbial biomass, leading to a significant reduction in microbial mass carbon (MBC) [61], although as the MBC decreases, the stability of soil aggregates increases, which is consistent with the results of this study. The responses of different fertilizer types to the MWD under the influence of the MBC varied. The N fertilizer alone caused a decrease in the soil microbial mass carbon, while long-term organic fertilizer treatment led to an increase in the microbial mass carbon relative to the chemical fertilizer and no fertilizer treatments, resulting in different levels of cementing material affecting the soil aggregate aggregation [62,63].

5. Conclusions

This meta-analysis showed that the organic fertilizer treatments stimulated the MWD more than the phosphorus and nitrogen fertilizers and that organic fertilizer addition promoted the role of soil organic cement in the agglomeration process, which in turn influenced the soil aggregate formation. The lowest amounts of organic fertilizer (<5000 kg·ha⁻¹·yr⁻¹), high N (>120 kg·ha⁻¹·yr⁻¹) and the lowest P (<40 kg·ha⁻¹·yr⁻¹) treatments stimulated the MWD more than different levels of fertilizer application. High N fertilization can cement smaller particles in the soil into large water-stable soil aggregates, while low levels of organic fertilization may lead to increased microbial activity, accelerating the decomposition of organic carbon within the soil aggregates and reducing the cemented material, which in turn affects the soil aggregation and stability. In addition, the stimulating effect of red soils on the MWD was significantly higher than that of yellow soils and black soils. The stimulation of the MWD was also higher in paddy fields than in gardens and cultivated fields, indicating that anthropogenic disturbance was more sensitive to the phosphorus, nitrogen, and organic fertilizer addition in response to the MWD. Meanwhile, meteorological factors were the main factors affecting the stability of the soil aggregates, and the secondary factors were the physical and chemical properties of the soil. Overall, the organic fertilizer additions had an excellent stimulatory effect on the MWD compared to the chemical fertilizers, as they provided a good source of carbon and nutrients for microorganisms. Therefore, organic fertilizer substitution for some chemical fertilizers is more effective than chemical fertilizer application alone in terms of the fertilization and better stabilization of soil aggregates, and more attention should be paid to the effects of temperature and rainfall on the stability of soil aggregates in response to fertilization.