Sustainable Soil Management in Reservoir Riparian Zones: Impacts of Long-Term Water Level Fluctuations on Aggregate Stability and Land Degradation in Southwestern China

Abstract

Soil structural instability in reservoir riparian zones, induced by water level fluctuations, threatens sustainable land use by accelerating land degradation. This study examined the impact of water-level variations on soil aggregate composition and stability based on key indicators, including water-stable aggregate content (WSAC), mean weight diameter (MWD), and geometric mean diameter (GMD). The Savinov dry sieving, Yoder wet sieving, and Le Bissonnais (LB) methods were employed for analysis. Results indicated that, with decreasing water levels and increasing soil layer, aggregates larger than 5 mm decreased, while aggregates smaller than 0.25 mm increased. Rising water levels and increasing soil layer corresponded to reductions in soil stability indicators (MWD, GMD, and WSAC), highlighting a trend toward soil structural instability. The LB method revealed the lowest aggregate stability under rapid wetting and the highest under slow wetting conditions. Correlation analysis showed that soil organic matter positively correlated with the relative mechanical breakdown index (RMI) (p < 0.05) and negatively correlated with the relative slaking index (RSI), whereas soil pH was negatively correlated with both RMI and RSI (p < 0.05). Comparative analysis of aggregate stability methods demonstrated that results from the dry sieving method closely resembled those from the SW treatment of the LB method, whereas the wet sieving method closely aligned with the FW (Fast Wetting) treatment of the LB method. The Le Bissonnais method not only reflected the outcomes of dry and wet sieving methods but also effectively distinguished the mechanisms of aggregate breakdown. The study concluded that prolonged flooding intensified aggregate dispersion, with mechanical breakdown influenced by water levels and soil layer. Dispersion and mechanical breakdown represent primary mechanisms of soil aggregate instability, further exacerbated by fluctuating water levels. By elucidating degradation mechanisms, this research provides actionable insights for preserving soil health, safeguarding water resources, and promoting sustainable agricultural in ecologically vulnerable reservoir regions of the Yangtze River Basin.

1. Introduction

As a vital buffer for sustainable watershed management, the riparian zone mediates ecological exchanges between terrestrial and aquatic systems. Functioning as a critical transition area, it is naturally shaped by fluctuations in water levels driven by irregular precipitation, seasonal climate shifts, and the presence of rivers, lakes, and reservoirs [1,2]. However, dam operations superimpose pronounced, often rapid, hydrological variations on these natural rhythms [3]. Such engineered fluctuations destabilize soil structure, degrade the soil–water environment [4], and ultimately compromise the integrity of riparian ecosystems. Consequences include landslides, soil deformation [5], accelerated erosion [3], and the loss of native vegetation along reservoir margins [6]. Central to these processes is the deterioration of soil aggregate stability, which is governed by both intrinsic soil properties and external hydrological forcing [7]. Consequently, the evolution of ecological environmental factors in the Yangtze River Basin has garnered significant attention in recent years, with an increasing emphasis on the conservation of this region [1]. Riparian zones adjacent to reservoirs are subject to periodic flooding and exposure, leading to variations in the soil moisture content, which subsequently impacts the decomposition and accumulation of organic matter [8]. Ref. [9] concluded that soil moisture had a significant effect on the stability of aggregates in sandy loam soils, with sandy loam aggregates being the most stable after inundation under the driest conditions and the least stable under the wettest conditions. Ref. [10] noted that the aggregate stability remains unchanged following cycles of rainfall and drying, irrespective of the aggregate texture. Furthermore, ref. [11] reported that the stability of the soil structure under flood conditions varies according to different hydrological circumstances, which may be linked to the specific characteristics of the soil and the stresses it experiences. For example, soils with a higher clay content are more susceptible to disintegration during wetting, and their aggregate stability is closely associated with factors such as swelling clays and electrolytes. In contrast, soils with lower clay contents are more significantly influenced by organic matter [7]. Consequently, comprehending the relationships between aggregate stability and external factors is essential for elucidating the stability of soil aggregates.

Soil aggregates are acknowledged as essential constituents of soil structure [12]. Additionally, soil aggregates provide physical protection to soil organic carbon (OC) by trapping carbon, inhibiting microbial decomposition through coating and isolation effects [13,14]. Distinct fractions of aggregates influence soil nutrient retention and availability, pore structure, hydraulic properties, and biological activity in diverse manners. Consequently, there is a significant correlation between aggregate size distribution and soil quality [15]. The ability of these soil aggregates to endure external forces while maintaining their structural integrity is referred to as aggregate stability [8]. This concept is particularly important concerning ecosystem processes and functions, as it serves as a crucial indicator of ecological stability [16]. The maintenance of soil aggregate stability, which is defined as the inherent resistance of aggregates to mechanical disintegration, dispersion, or dissolution, is vital for mitigating soil erosion, reducing degradation, decelerating the decomposition of organic carbon, and addressing global climate change [17].

The mechanisms underlying aggregate stability in the context of hydrological stress remain inadequately understood. Whereas the Three Gorges Reservoir region has been the subject of extensive and prolonged research concerning soil aggregates in riparian zones [18,19], few studies have examined the effects of long-term, anti-seasonal, extreme dry–wet alternations resulting from periodic fluctuations in water levels within the Southwestern Reservoir regions. This study attempts to fill this knowledge gap by assessing the soil aggregate stability and fragmentation mechanisms in the Southwest Reservoir of China. The southwestern reservoir area in Sichuan Province, located in the lower Yalong River basin, is a sub-basin of the Jinsha River. The Jinsha River, home to a cluster of large cascade reservoirs in the upper Yangtze River, holds important strategic significance in China’s hydropower development strategy. Consequently, this study focuses on the riparian zone of the Ertan Reservoir in the Yalong River Basin. A combination of dry and wet sieving methods is employed to analyze the distribution of water-stable and non-water-stable aggregate fractions, whereas the LB method is utilized to investigate the mechanisms of aggregate breakdown. The primary aim of this study is to evaluate the effects of long-term and anti-seasonal flooding on aggregate composition and stability across different elevations and soil layers. The anticipated outcomes of this research are expected to offer scientific insights for the prevention and management of soil erosion in the riparian zones of Southwestern Reservoir.

The stability of soil aggregates is not merely a pedological concern but a cornerstone of sustainable development. Degraded soils in riparian zones exacerbate land degradation, directly threatening food security and water resource sustainability. Healthy soil aggregates act as ecological filters: they reduce sediment pollution in reservoirs, enhance groundwater recharge, and maintain water quality for irrigation and aquaculture. As a critical transition between terrestrial and aquatic ecosystems, reservoir riparian zones provide indispensable ecosystem services, including erosion control and biodiversity conservation. Preserving their soil structural integrity is thus vital for achieving Sustainable Development Goals. This study combines soil science with water conservancy engineering to propose the causes of aggregate instability, which is a prerequisite for the sustainable operation of reservoirs and basin management in the southwest region. We hypothesize that long-term anti-seasonal water level fluctuations reduce deep soil organic matter (SOM) by (1) prolonging the flooding time and (2) frequently alternating dry and wet conditions inducing the dispersion of clay particles, jointly weakening the stability of aggregates. Moreover, the Le Bissonnais method can quantify such damage mechanisms more accurately.

2. Materials and Methods

2.1. Study Area Overview

The Ertan Reservoir of China is situated at geographical coordinates of 101°43′36″ E and 26°48′39″ N. The topography of the region is characterized by a gradual incline, with a slope measuring 13°,which is primarily directed towards the north. The overall elevation of the reservoir region is recorded at 1150 m. Additionally, the dominant soil composition in the vicinity of the reservoir is classified as red loam. The vegetation is dominated by Cynodon dactylon, Abutilon theophrasti and Persicaria lapathifolia, with a coverage rate of 83%. Climatically, the region falls within a subtropical monsoon zone where frequent summer rainstorms intensify the physical stress exerted by water-level fluctuations on soil aggregates.

2.2. Sample Collection and Pretreatment

Sampling was carried out in early July 2022 along the riparian zone of the Ertan Reservoir, which is located at coordinates of 26°48′39″ N and 101°43′36″ E on the Yalong River (refer to Figure 1). The riparian zone of the Ertan Reservoir is characterized by anti-seasonal hydrological conditions. The selection of three sampling intervals at different elevations was guided by historical water level data obtained from the China Yangtze River Hydrological Network (http://www.cjh.com.cn/swyb_syqbg.html, accessed on 3 August 2025). Specifically, samples collected from low elevations (L) were submerged for a duration exceeding seven months, those from medium elevations (M) were submerged for a period ranging from five to six months, and samples from high elevations (H) were submerged for a duration of three to four months (

Table 1. Rooted soil sampling. The elevation gradient data are referenced based on the Yellow Sea elevation. At the L elevation, the flooding duration of rooted soil exceeds 7 months whereas, at the M elevation, the flooding duration ranges from 5 to 6 months. At the H elevation, the flooding duration of rooted soil is between 3 and 4 months.

Location	Geographical Position	Elevation Gradient	Elevation Gradient Range	Number of Samples
Ertan	N 26°48′39″, E 101°43′36″	L	1150.579~1155.963 m	131
		M	1155.963~1159.279 m	144
		H	1159.279~1163.424 m	134

).

In this investigation, the research area at each elevation was stratified into three segments, with three sampling points established within each segment. Soil samples, each measuring 60 cm³, were collected utilizing a ring knife (Φ 61.8 mm × 20 mm), alongside undisturbed soil blocks measuring 10 × 10 × 10 cm, which were obtained for the purpose of assessing aggregate composition. The collected soil samples were subsequently transported to the laboratory for further analysis. The undisturbed soil blocks were fragmented into smaller pieces, with debris being meticulously removed, and the samples were allowed to air-dry naturally. Both dry and wet sieving methods were subsequently employed to evaluate the aggregate composition. The LB method was specifically utilized to assess the stability of the aggregates within the 3–5 mm size range. Additionally, the soil particle size distribution, pH, and organic matter content were determined using a Malvern Mastersizer 2000 laser diffraction instrument (Malvern Instruments Ltd., Malvern, UK), a Rex Electric Chemical pH–3E precision pH meter (Shanghai Precision Scientific Instruments Co., Ltd., Shanghai, China), and a potassium dichromate oxidation method with external heating [20], respectively.

2.3. Measurement Indicators and Methods

In this research, the Savino dry sieving method (D) and the Yoder wet sieving method (W) were employed to assess the aggregate composition of undisturbed soil blocks sourced from the riparian zone of a reservoir. The LB method was utilized with three distinct treatments to elucidate the mechanisms underlying aggregate breakdown: FW, slow wetting (SW), and prewetting disturbance (WS) [21]. The procedural steps for the Le Bissonnais method are delineated as follows:

(1) FW treatment: A sample of 5 g of aggregate is rapidly submerged in 50 g of deionized water for 10 min, after which the excess water is removed using a pipette.

(2) SW treatment: A sample of 5 g of aggregate is placed on filter paper and subjected to a tension of −0.3 kPa. The aggregates are allowed to saturate for 30 to 40 min to ensure complete wetting.

(3) WS treatment: A sample of 5 g of aggregate is immersed in ethanol to facilitate air displacement. Following a 10-min soaking period, the ethanol is removed, and the aggregates are transferred to a 50 mL conical flask filled with deionized water. Additional deionized water is added until the total volume reaches 200 mL. The flasks are then sealed and shaken 20 times, after which they are allowed to stand for 30 min to allow the coarse dispersed fraction to settle. The water in the graduated cylinder is subsequently drained using a pipette.

After treatment, the aggregates were thoroughly dried in an oven at 40 °C and then subjected to sieving through a series of sieves with aperture sizes of 3 mm, 2 mm, 1 mm, 0.5 mm, 0.25 mm, 0.1 mm, and 0.05 mm. The mass of aggregates retained on each sieve was recorded, with each sieving method replicated three times throughout the experiment.

Ultimately, several indices were computed to characterize soil aggregate stability, including the MWD, GMD, WSAC, RSI, and RMI. The specific formulas utilized for the calculation of these indices are as follows:

$$MWD={\sum }_{i=1}^n{X}_i\times w_i$$

(1)

In the formulas, $X_i$ and $w_i$ represent the average diameter (in mm) and the mass percentage content (%) of the $i$, $i+1$th grain size, respectively, and $n$ denotes the number of sieves used.

$$GWD=exp[{\displaystyle \frac{\sum _{i=1}^n{m}_{i}ln{X}_i}{m}}]$$

(2)

$$WSAC={\displaystyle \frac{M_{r>0.25}}{M_t}}\times 100\mathrm{\%}$$

(3)

In the formulas, $M_{r>0.25}$ represents the mass of soil aggregates larger than 0.25 mm, whereas $M_t$ refers to the total mass of the aggregates (in grams).

To distinguish the effects of different wetting treatments in the Le Bissonnais method, the RSI and the RMI are defined as follows:

$$RSI={\displaystyle \frac{{MWD}_{SW}-{MWD}_{FW}}{{MWD}_{SW}}}\times 100\%$$

(4)

$$RMI={\displaystyle \frac{{MWD}_{SW}-{MWD}_{WS}}{{MWD}_{SW}}}\times 100\%$$

(5)

In the formulas, ${MWD}_{FW}$, ${MWD}_{SW}$ and ${MWD}_{WS}$ represent the MWD for the FW, SW, and WS treatments, respectively. The larger the values of RSI or RMI, the more easily the aggregates undergo dispersion or mechanical breakdown [22].

2.4. Data Analysis

Data processing and analysis in this experiment were performed using Excel 2019 and SPSS 25.0 software, whereas Origin 2021 was used for graph creation. One-way analysis of variance (ANOVA) combined with LSD (Least Significant Difference) test was employed to compare the significance of differences between groups (with a significance level of p < 0.05). The Pearson correlation method was applied to analyze the correlations between the data.

3. Results

3.1. Soil Physicochemical Properties

Table 2. Distribution along the elevation of soil physical and chemical properties along the riparian zone. The differences in letters indicate significant variations in altitude. In the table, pH and Soil organic matter show no significant differences with variations in soil layer.

Elevation Gradient	Depth/cm	Sand (%)	Silt (%)	Clay (%)	Soil Organic Matter (g/kg)	pH	Texture Class
L	0~10	4.83	62.70	32.47	15.24 ± 1.02 B	7.62 ± 0.26 B	Silty clay loam
	10~20	4.86	61.17	33.98	14.14 ± 2.26 B	7.79 ± 0.14 B	Silty clay loam
	20~30	4.76	59.34	35.90	13.5 ± 1.03 B	7.76 ± 0.19 B	Silty clay loam
M	0–10	3.84	51.47	44.69	16.2 ± 0.93 B	6.96 ± 2.04 C	Silty clay
	10~20	4.43	55.96	39.61	15.36 ± 1.39 B	6.85 ± 2 C	Silty clay loam
	20~30	4.32	57.09	38.59	13.22 ± 2.93 B	6.76 ± 1.67 C	Silty clay loam
H	0–10	2.24	41.38	56.38	16.02 ± 2.23 A	8.2 ± 0.25 A	Silty clay
	10~20	3.37	48.98	47.45	15.83 ± 1.14 A	8.16 ± 0.33 A	Silty clay
	20~30	3.38	50.026	46.60	15.26 ± 2.21 A	7.94 ± 0.26 A	Silty clay

presents the physicochemical properties of the soil analysed in this study. The data indicate that the SOM content tends to increase with both elevation and soil layer. Additionally, the pH initially decreases with increasing elevation before subsequently increasing, whereas a general decrease in pH is observed with increasing soil layer. Notably, for soils at equivalent depths but varying elevations, there was a statistically significant increase in organic matter and pH with elevation (p < 0.05). Conversely, at a constant elevation, an increase in soil layer correlates with an increase in organic matter content, except at H, where an inverse trend is noted. The pH levels demonstrate a distinct pattern, initially decreasing and then increasing with elevation, whereas an increase in soil layer is associated with a decrease in pH, again except for H, where the opposite trend is observed. This study revealed significant variations in organic matter content and pH in relation to changes in soil layer, with organic matter and pH decreasing as soil layer increased. The soil composition is predominantly silty, with clay being the most abundant component, followed by sand, which is the least prevalent. These findings indicate that, as elevation increases, the sand and silt contents generally decrease, accompanied by an increase in the clay content. At L, an increase in soil layer corresponds with a decrease in the sand and silt contents, whereas the clay content increases. At M, the clay content follows the trend of 0–10 cm > 20–30 cm > 10–20 cm with increasing soil layer. At H, the trend is as follows: 0–10 cm > 10–20 cm > 20–30 cm.

3.2. Distributed Along the Elevationof Soil Aggregate Particle Size Using Dry and Wet Sieving Method

The compositions of the non-water-stable aggregates are illustrated in Figure 2a. The data presented indicate that the proportions of aggregates exceeding 5 mm in size are the most substantial, ranging from 46.61% to 83.55%. In contrast, the distributions of aggregates within the other particle size categories are relatively uniform, varying between 1.68% and 11%. Notably, within the same soil layer, an increase in elevation is correlated with a significant increase in the content of aggregates larger than 5 mm (p < 0.05), whereas the proportions of aggregates in the smaller size classes markedly decrease (p < 0.05). Specifically, when comparing the low elevations (L) to the medium (M) and high (H) elevations, there are increases in the number of aggregates greater than 5 mm by 22.90% and 43.70%, respectively. Concurrently, the contents of aggregates in the 5–3 mm, 3–2 mm, 2–1 mm, 1–0.5 mm, 0.5–0.25 mm, and <0.25 mm size classes decrease by 42.33% and 57.94%, 35.60% and 49.59%, 23.49% and 61.75%, 26.70% and 59.87%, 21.40% and 71.78%, and 34.83% and 69.05%, respectively. At a consistent elevation, no significant differences in aggregate content are detected across various soil layers (p < 0.05). As the soil layer increases, the proportions of aggregates larger than 5 mm decrease, ranging from 46.60% to 58.14%, whereas the contents of aggregates in the smaller size classes increase. When the 0–10 cm soil layer was compared with the 10–20 cm and 20–30 cm layers, the contents of aggregates exceeding 5 mm decreased by 16.64% and 19.84%, respectively. Moreover, the contents of aggregates in the 5–3 mm, 3–2 mm, 2–1 mm, 1–0.5 mm, 0.5–0.25 mm, and <0.25 mm size classes increased by 11.94% and −0.64%, 24.01% and 24.82%, 25.79% and 32.12%, 33.61% and 38.93%, 21.56% and 36.31%, and 28.68% and 54.52%, respectively.

Figure 2b shows the distributions of water-stable aggregates, revealing that the highest concentrations are found in aggregates larger than 5 mm and smaller than 0.25 mm, followed by those in the 2–1 mm, 1–0.5 mm, 0.5–0.25 mm, and 5–3 mm size categories. Conversely, aggregates smaller than 5 mm presented the lowest concentrations. Notably, a significant elevation-dependent increase in the aggregate content greater than 5 mm in size was observed (p < 0.05), alongside a marked decrease in the aggregate contents that were smaller than 0.25 mm in size (p < 0.05). The aggregate contents in the 1–0.5 mm range also displayed elevation-dependent variations, with the highest concentration recorded at medium elevations (M), followed by lower elevations (L), and the lowest concentration at high elevations (H). Within a single soil layer, the aggregate contents greater than 5 mm increased by 246.27% and 371.97% at M and L, respectively, whereas the content of aggregates in the 5–3 mm, 3–2 mm, 2–1 mm, 0.5–0.25 mm, and <0.25 mm categories decreased by 36.70%. The 1–0.5 mm grain size initially increased, followed by a subsequent decline, with increases of 14.65%, 30.30%, 50.13%, and 58.36% and a decrease of 20.76%. At the same elevation across different soil layers, the aggregate contents did not significantly vary (p < 0.05). However, as the soil layer increased, the contents of aggregates greater than 5 mm, 5–3 mm, and 3–2 mm decreased, whereas the contents of aggregates less than 0.25 mm increased. The variability in the content of 2–1 mm aggregates with soil layer indicates that the 20–30 cm layer presented with the highest concentration, followed by the 0–10 cm layer, with the 10–20 cm layer showing the lowest concentration. At the same elevation, when comparing the 0–10 cm layer to the 10–20 cm and 20–30 cm layers, there were decreases in the contents of aggregates greater than 5 mm and 5–3 mm, with reductions of 19.89% and 52.98%, 35.34% and 85.24%, and 45.71% and 51.33%, respectively. Conversely, increases were observed in the 1–0.5 mm, 0.5–0.25 mm, and <0.25 mm aggregate contents, with increases of 9.11% and 14.20%, 15.72% and 15.95%, and 25.77% and 39.29%, respectively. The 2–1 mm grain size range exhibited an oscillatory trend, characterized by a decrease of 5.91%, followed by an increase of 2.07%.

3.3. Determination of the Elevation Distribution of Aggregate Grain Size via the Le Bissonnais Method

The distribution of aggregate grain size following the application of FW treatment is depicted in Figure 3a–c. The analysis of the aggregate contents revealed that aggregates with sizes of 5–3 mm and <0.05 mm were present in the highest proportions, followed by those with sizes of 2–1 mm, 3–2 mm, 1–0.5 mm, and 0.5–0.25 mm. Conversely, the aggregates with sizes of 0.25–0.1 mm and 0.1–0.05 mm presented the lowest levels. Statistically significant differences (p < 0.05) were noted with respect to elevation; specifically, the contents of the 5–3 mm and 3–2 mm aggregates increased, whereas the content of the <0.05 mm aggregates decreased. Additionally, the 0.5–0.25 mm and 0.1–0.05 mm aggregates reached their highest levels at medium elevations (M), followed by those at low elevations (L), with the lowest levels observed at high elevations (H). No significant differences (p < 0.05) were detected across the different soil layers. The aggregate contents within the 0–10 cm soil layer ranged from 6.73% to 26.33%. When the aggregate contents at the L, M, and H elevations were compared, a notable increase of 58.11% was detected for the 5–3 mm, 3–2 mm, and 0.5–0.25 mm aggregates. In contrast, the contents of the 2–1 mm, 1–0.5 mm, 0.1–0.05 mm, and <0.05 mm aggregates decreased by 26.27% and 38.26%, respectively. The changes in aggregate composition with elevation in the 10–20 cm and 20–30 cm soil layers mirrored those observed in the 0–10 cm layer, with aggregate contents ranging from 5.60% to 27.07% and 7.20% to 32.80%, respectively. Compared with those in the 0–10 cm layer, the contents of the 5–3 mm, 3–2 mm, and 2–1 mm aggregates in the 10–20 cm and 20–30 cm layer generally decreased by 5.46% and 14.18%, 32.88% and 52.61%, and 9.55% and 26.78%, respectively, Increases in aggregate content of varying percentages were observed across different size categories, indicating that complex interactions are influenced by elevation.

The particle size distributions of the aggregates following WS treatment are depicted in Figure 3d–f. The distribution pattern of these particles was similar to that of the FW distribution, with no statistically significant differences in aggregate content observed across various elevations (p < 0.05). In the 0–10 cm soil layer, the aggregate contents varied between 6.53% and 28.13%. Compared with those at the L, M, and H elevations, the aggregate contents in the 5–3 mm, 3–2 mm, and 0.5–0.25 mm size groups increased by 13.81%, 11.94%, and 22.17%, 14.39%, and 0.72%, respectively. Conversely, the aggregate contents of the 2–1 mm, 1–0.5 mm, 0.25–0.1 mm, 0.1–0.05 mm, and <0.05 mm particles decreased by 8.82% and 13.24%, 12.38% and 8.42%, 13.28% and 11.42%, 10.09% and 14.25%, and 9.78% and 22.30%, respectively. The changes in aggregate content with elevation in the 10–20 cm and 20–30 cm soil layers were consistent with those observed in the 0–10 cm layer, with aggregate contents ranging from 6.73% to 24.80% and from 7.33% to 22.33%, respectively. Compared with the 0–10 cm layer, the 10–20 cm and 20–30 cm layers presented reductions in the contents of 5–3 mm, 2–1 mm, and 0.1–0.05 mm aggregates, with decreases of 4.15% and 9.44%, 7.87% and 18.60%, and 5.04% and 5.80%, respectively. Moreover, the contents of the 3–2 mm, 0.25–0.1 mm, and <0.05 mm aggregates increased by 4.60% and 1.63%, respectively. The aggregate contents in the 1–0.5 mm and 0.5–0.25 mm size ranges tended to decrease, followed by increases of 1.68% and 2.06%, respectively. Additionally, the contents of the 0.05–0.25 mm and 0.25–1 mm aggregates increased by 8.42% and 4.83%, respectively.

The particle size distributions of the aggregates following the SW treatment are depicted in Figure 3g–i, revealing a pattern that is largely comparable to that observed in the FW and WS distributions. Statistical analysis revealed no significant difference in aggregate content with increasing elevation (p < 0.05). Within the 0–10 cm soil layer, the aggregate contents vary from 5.27% to 36.27%. When comparing the aggregate sizes of 5–3 mm, 3–2 mm, 1–0.5 mm, and 0.5–0.25 mm at L,M,H elevations, increases of 20.11% and 28.62%, 9.49% and 25.86%, and 60.87% and 14.99%, respectively, were observed. In contrast, the aggregate sizes of the 0–1 mm, 0.25–0.1 mm, 0.1–0.05 mm, and <0.05 mm aggregates decreased by 24.80%, 13.36%, and 18.98%, respectively, compared with those of the L aggregates, with reductions of 12.25% and 18.38% and 46.40% and 66.51%, respectively. The trends in aggregate variation with elevation in the 10–20 cm and 20–30 cm soil layers were consistent with those in the 0–10 cm layer, with aggregate contents ranging from 4.53% to 34.4%. Notably, the overall decreases in the 10–20 cm and 20–30 cm layers for the 5–3 mm and 3–2 mm aggregates were 10.89% and 20.82%, 15.60% and 39.17%, and 9.19% and 3.98%, respectively. Conversely, increases in the 1–0.5 mm, 0.5–0.25 mm, and <0.05 mm aggregate sizes were noted, with increases of 6.90% and 39.88%, 69.45% and 144.21%, and 15.00% to 21.19%, respectively.

3.4. Determination of Aggregate Stability via Dry and Wet Sieving Methods

As illustrated in Figure 4a, the dry sieving method revealed no statistically significant differences (p < 0.05) in either the MWD or the GMD with respect to increasing elevation. Additionally, no statistically significant differences were detected between various elevations within the same soil layer and the MWD values at L. Specifically, the MWD values for M and H decreased from 4.16% to 10.73% and 12.73% to 15.08%, respectively, whereas the GMD values for M and H decreased between 8.31% and 17.39% and 24.24% and 25.87%, respectively, compared with those of L. In contrast, significant increases in both the MWD and GMD were noted as the soil layer increased (p < 0.05). The MWD increased by 11.67% to 16.00% in the 0–10 cm soil layer and by 26.79% to 36.12% in the 10–20 cm and 20–30 cm layers, respectively, relative to L. Compared with those of L, the GMD values for the 10–20 cm and 20–30 cm soil layers increased by 21.75% and 15.08%, respectively. Conversely, the GMD values for M and H decreased by 8.31% and 17.39%, respectively, in comparison with those at L, whereas the increases in the 10–20 cm and 20–30 cm layers were 21.75% and 27.84%, respectively.

As illustrated in Figure 4b, the wet sieving method did not significantly affect the MWD or GMD values with increasing elevations (p < 0.05). The trends observed in the MWD and GMD values that were derived from the wet sieving method were similar to those obtained via the dry sieving method. A comparative analysis of the MWD values from the wet sieving method against those from the dry sieving method indicated reductions ranging from 31.27% to 60.77%. Similarly, a decline in GMD values was noted, with a range of 18.31% to 36.78%. Importantly, the GMD values for the same soil layer at different elevations, compared with those at L elevations, exhibited decreases in the MWD values at M and H elevations of 9.32% to 22.88% and 16.59% to 39.15%, respectively. The GMD values at the M and H elevations, in relation to the L elevation, decreased by 16.54% to 27.69% and 29.07% to 42.90%, respectively. Furthermore, for equivalent elevations across distinct soil layers, both the MWD and GMD values significantly increased with increasing soil layer (p < 0.05). Specifically, in comparison with those in the 0–10 cm layer, the MWD values in the 10–20 cm and 20–30 cm layers increased by 70.49% to 88.14% and 112.88% to 191.80%, respectively. Similarly, the GMD values in the 10–20 cm and 20–30 cm layers, compared with those in the 0–10 cm layer, increased by 77.38% to 109.30% and 159.94% to 203.24%, respectively.

As illustrated in Figure 4c, the WSAC index did not significantly change with increasing elevation (p < 0.05). When the dry sieving method was used, comparisons for the same soil layer at varying elevations revealed that, relative to L, M and H resulted in WSAC reductions of 0.65–1.64% and 1.68–3.12%, respectively. Conversely, when the same elevation was used across different soil layers, the WSAC index significantly increased with increasing soil layer (p < 0.05). Specifically, when comparing the 0–10 cm soil layer, the WSAC values at the 10–20 cm and 20–30 cm depths increased by 2.00%–2.35% and 3.95%–5.50%, respectively. Notably, the WSAC values that were obtained through the wet sieving method were consistently lower than those derived from the dry sieving method, with reductions ranging from 11.10% to 36.83%. The most pronounced decrease was observed in the 20–30 cm soil layer at L. Overall, the WSAC values increased with both elevation and soil layer, peaking in the 0–10 cm layer at H. In the context of the wet sieving method, for the same soil layer at different elevations, M and H presented decreases in WSAC of 0.16–0.82% and 0.10–2.96%, respectively, compared with L. Additionally, for a given elevation and varying soil layers, the WSAC index displayed a significant positive correlation with soil layer (p < 0.05). Compared with the 0–10 cm layer, the 10–20 cm and 20–30 cm layers presented increases of 26.06–30.07% and 30.07–34.85%, respectively, across L, M, and H.

3.5. Determination of Soil Aggregate Stability via the Le Bissonnais Method

As illustrated in Figure 5a, the FW treatment did not result in any statistically significant differences (p < 0.05) in the MWD across various elevations. Nevertheless, the MWD values tended to decrease with increasing elevation across the different soil layers, with reductions ranging from 3.86% to 10.87% for M and from 9.23% to 18.30% for H, in contrast to the MWD values at L. Additionally, the GMD indices for M and H also decreased, with reductions ranging from 5.17% to 17.54% and 20.03% to 29.46%, respectively. Conversely, increases in both the MWD and GMD were noted in deeper soil layers at the same elevation (p < 0.05). Specifically, the MWD for the 10–20 cm and 20–30 cm layers increased by 20.46% and 35.84%, respectively, compared with the MWD values of the 0–10 cm layer. The GMD indices for the 10–20 cm and 20–30 cm layers rose by 16.17% to 18.30% and 16.17% to 18.30%, respectively, relative to those at L. In contrast, the GMD indices for M and H decreased by 5.17% to 17.54% and 20.03% to 29.46%, respectively, compared with L. The increases in GMD were 16.17% to 31.00% and 42.92% to 62.02%, respectively.

As shown in Figure 5b, the WS treatment did not significantly affect (p < 0.05) the MWD with increasing elevation. Within a single soil layer, the reductions at M and H at varying elevations ranged from 2.60% to 7.48% and from 6.92% to 14.62%, respectively, compared with the MWD at L. The decreases in the GMD for M and H were observed to be between 5.30% and 13.07% and between 13.95% and 21.55%, respectively, in comparison with L. Furthermore, no significant differences (p < 0.05) were noted in the MWD and GMD across different soil layers at the same elevation, although both metrics increased with increasing depth of the soil layer. A comparative analysis of the MWD values for the 0–10 cm soil layer against those for the 10–20 cm and 20–30 cm layers revealed increases of approximately 0.36% and 0.92%, respectively. Additionally, the GMD values for the 10–20 cm and 20–30 cm layers increased by 0.64% to 15.57% and by 19.54% to 31.11%, respectively, compared with those of the 0–10 cm layer.

As illustrated in Figure 5c, the results indicated that there were no statistically significant differences (p < 0.05) in the MWD with respect to increasing elevation under the SW treatment. Additionally, when different elevations within the same soil layer were compared, the reductions for M and H were observed to be between 8.24% and 10.66% and between 15.10% and 18.76%, respectively, relative to the MWD at L. The GMD values for M and H, in comparison with those for L, decreased from 11.19% to 23.47% and from 25.70% to 37.50%, respectively. In the different soil layers at the same elevation, the MWD significantly increased (p < 0.05) with increasing depth. Specifically, when the MWD indices of the 0–10 cm soil layer were compared, the increases in the 10–20 cm and 20–30 cm layers were recorded at 10.42% and 22.14%, respectively. Furthermore, the increases for the 10–20 cm and 20–30 cm layers were 23.51% and 41%, respectively, compared with those of the 0–10 cm layer. The 10–20 cm and 20–30 cm layers presented increases ranging from 23.51% to 41.34% and from 44.69% to 72.02%, respectively.

As shown in Figure 6a, there was no statistically significant difference (p < 0.05) in the WSAC with increasing elevation across the three LB methods. Conversely, reductions in WSAC were noted with increasing elevation in various soil layers under the FW treatment, with decreases ranging from 3.26% to 6.93% at M elevations and from 7.75% to 13.63% at H elevations, in comparison to L elevations. Furthermore, for the same elevation, a significant increase in the WSAC was observed as the soil layer deepened (p < 0.05). Specifically, for L, M, and H, the increases in WSAC were recorded as 0.16% to 0.82% and 0.10% to 2.96% for the 10–20 cm and 20–30 cm soil layers, respectively, compared with the 0–10 cm soil layer. Additionally, the reductions in WSAC under the WS and SW treatments ranged from 1.83% to 45.41% and from 5.81% to 64.81%, respectively, compared with those under the FW treatment.

As illustrated in Figure 6b, the sensitivity of soil aggregates to dissipation and mechanical fragmentation can be assessed through the RSI and the RMI. The distributions of the RSI and RMI for aggregates at varying water table elevations and soil layers are presented in Figure 6b. The data indicate that within the riparian zone of the reservoir, under conditions characterized by alternating wet and dry environments, the RSI consistently exceeds the RMI. This observation suggests that soil aggregates are more prone to dissipation than to mechanical fragmentation. The RSI values ranged from 0.24 to 0.45, whereas the RMI values were between 0.05 and 0.22. Notably, as the elevation increased, the RSI tended to increase, whereas the RMI tended to decrease. Following a seven-month inundation period, the mid-soil layer aggregates in the reservoir riparian zone demonstrated increased sensitivity to dissipation. Throughout the inundation period from May to June, the sensitivity of the aggregates across the three soil layers to dissipation remained stable. However, during the inundation period from March to April, the sensitivity of the reservoir riparian zone aggregates to dissipation decreased in the order of 0–10 cm > 10–20 cm > 20–30 cm. Additionally, the sensitivity of the reservoir riparian zone aggregates, as indicated by the RMI, in the 0–10 cm and 10–20 cm layers increased with increasing water elevation, whereas the 20–30 cm layer initially increased but then decreased, reaching a maximum value of 0.22 at H elevations in the 0–10 cm layer and a minimum value of 0.05 at L elevations in the 20–30 cm layer. At both L and H elevations, the RMI followed the order of 0–10 cm > 10–20 cm > 20–30 cm, whereas at M elevations, the order was 0–10 cm > 20–30 cm > 10–20 cm.

3.6. Correlation Analysis Between the Stability of the Soil Aggregates and Their Contents and Soil Physicochemical Properties Determined via the Five Methods

Figure 7 illustrates the relationships between the indicators of soil aggregate fragmentation mechanisms (e.g., RSI and RMI) and various physicochemical properties of soil, primarily pH and SOM. The analysis revealed a positive correlation between SOM and both pH and the RMI, whereas a negative correlation existed between SOM and the RSI (p < 0.05). Additionally, the pH has a negative correlation with both the RSI and RMI (p < 0.05).

As illustrated in Figure 8, a positive correlation exists between the bulk stability of soil aggregates, primarily measured by the MWD, and the presence of various grain sizes. Specifically, the MWD_D value was significantly positively correlated with the 5 mm grain size (p < 0.001) but not significantly correlated with the 3 mm (p > 0.05) and 2 mm (p > 0.05) sizes. Conversely, a negative correlation was observed with the 1 mm grain size (p < 0.05). In terms of the MWD_W, a positive correlation was noted with the 5 mm grain size (p < 0.001), whereas the correlations with the 3 mm (p > 0.05) and 2 mm (p > 0.05) grain sizes were nonsignificant. Negative correlations were identified with the 1 mm (p < 0.05), 0.5 mm (p > 0.05), 0.25 mm (p < 0.05), and <0.25 mm (p < 0.001) grain sizes. The MWD_FW similarly demonstrated a positive correlation with the 5 mm grain size (p < 0.001) and nonsignificant correlations with the 3 mm (p > 0.05) and 2 mm (p > 0.05) grain sizes but negative correlations with the 1 mm (p > 0.05), 0.5 mm (p > 0.05), 0.25 mm (p < 0.05), and <0.25 mm (p < 0.001) grain sizes. The MWD_WS was positively associated with the 5 mm grain size (p < 0.01) and not significantly correlated with the 3 mm (p > 0.05) and 2 mm (p > 0.05) grain sizes, whereas it was negatively correlated with the 1 mm (p > 0.05), 0.5 mm (p < 0.05), 0.25 mm (p < 0.01), and <0.25 mm (p < 0.05) grain sizes. Finally, the MWD_SW was positively correlated with the 5 mm grain size (p < 0.01), whereas the correlations with the 3 mm (p > 0.05) and 2 mm (p > 0.05) grain sizes were nonsignificant, and negative correlations were found with the 1 mm (p > 0.05), 0.5 mm (p > 0.05), 0.25 mm (p < 0.05), and <0.25 mm (p < 0.01) grain sizes.

The results of the correlation analysis among the various methodologies are presented in Figure 9. The results indicate that MWD_D is positively correlated with MWD_W, MWD_FW, and MWD_SW (p < 0.001), as well as with MWD_WS (p < 0.01). Moreover, MWD_W is positively correlated with MWD_FW (p < 0.001) and with MWD_WS and MWD_SW (p < 0.01). Additionally, the MWD_FW shows a positive correlation with MWD_WS, while MWD_SW also exhibits a positive correlation (p < 0.001). Furthermore, the MWD_WS is positively correlated with the MWD_SW (p < 0.001).

4. Discussion

4.1. Response of Soil Aggregate Composition to Changes in Water Level

The periodic variations in water level within the reservoir riparian zone of the Southwest Reservoir area result in different flooding durations, depths, frequencies, and amplitudes of wetting–drying cycles across various elevations. These fluctuations have a significant effect on the compositions of both the water-stable and non-water-stable soil aggregates. The findings of this study indicate that with increasing elevation, the proportions of non-water-stable aggregates (>5 mm) and water-stable aggregates notably increase, whereas the amounts of water-stable aggregates and non-water-stable aggregates (<0.25 mm) significantly decrease. Additionally, the contents of non-water-stable aggregates in the other size categories markedly decreased, whereas the contents of water-stable aggregates significantly increased. No statistically significant differences in the proportions of non-water-stable and water-stable aggregates were detected with respect to soil layer (p < 0.05). As the soil layer increased, the quantity of non-water-stable aggregates (>5 mm) decreased, whereas the contents of aggregates in other size categories tended to increase. Specifically, the number of water-stable aggregates greater than 2 mm in size decreases, whereas the content of water-stable aggregates less than 2 mm in size generally increases. The observed response of water-stable aggregates to variations in soil layer is closely associated with the organic matter content. As the soil layer increases, the organic matter content correspondingly increases, which facilitates an increased quantity of water-stable aggregates, corroborating the findings of [23].

The variations in aggregate particle size distribution with elevation are different from the findings in the Three Gorges Reservoir area [8]. This observation may be due to the higher elevation of the Southwestern Reservoir area, where greater periodic fluctuations in the water level within the reservoir area result in the fragmentation of larger aggregates into smaller ones. Furthermore, as the water level decreases, the diversity in aggregate sizes increases, potentially due to the collapse of aggregates resulting from rapid water level changes in the riparian zone. Ref. [24] proposed that the primary cause of soil aggregate destruction is the dissipative effect that is produced by the rupture of air bubbles within soil pores. Ref. [25] argued that variations in flooding duration, depth, and frequency across different elevations in a reservoir’s riparian zone significantly influence the soil structural stability. The cyclical nature of water level fluctuations modifies the pore structures within and between aggregates, leading to either shrinkage or swelling of the aggregates and altering the distributions of aggregates across various size fractions. Notably, non-water-stable aggregates, particularly those exceeding 5 mm in size, are significantly more prevalent than their water-stable counterparts. This phenomenon can be attributed to the greater resistance of soil to mechanical breakdown from wind erosion than to hydraulic disintegration [26]. During the wet sieving process, larger non-water-stable aggregates are more susceptible to disintegration into smaller aggregates. This finding indicates that the wet sieving method demonstrates superior reproducibility compared with dry sieving, thereby providing a more accurate representation of the composition and stability of soil aggregates [27]. The Le Bissonnais method effectively differentiates various mechanisms of aggregate destruction and simulates the actual conditions of soil aggregate degradation, establishing itself as the standard for assessing aggregate stability [28]. The FW treatment emulates extreme weather conditions and has the most significant destructive impact on aggregates, whereas the WS treatment replicates the mechanical disintegration effects of runoff, and the SW treatment simulates the wetting effects of light rainfall. The aggregate size distributions resulting from these three treatments were predominantly concentrated within the ranges of 5–3 mm and <0.05 mm, with the remaining fractions being evenly distributed. The aggregates subjected to the FW treatment exhibited a more uniform size distribution, characterized by a reduced presence of larger aggregates. This observation implies that the FW treatment induces the most substantial breakdown of aggregates, indicating that extreme weather events or out-of-season fluctuations in water levels are the primary contributors to soil aggregate destruction in the studied area.

4.2. Comparison of Test Methods for Stability of Soil Aggregates

This research employed dry sieving, wet sieving, and Le Bissonnais methodologies to assess the effects of water level variations on the size distributions and stability of soil aggregates. These findings indicate a significant correlation among the various treatment methods. The ranking of the MWD across the different techniques is as follows: FW < WS < SW < wet sieving < dry sieving. Notably, the results obtained from dry sieving are more similar to those from the SW treatment, whereas the outcomes from wet sieving align more closely with those from the FW treatment. This discrepancy contrasts with the findings reported by [26], which may be attributed to variations in climatic conditions and out-of-season fluctuations in water levels within their study area. The dry sieving technique quantifies the proportion of non-water-stable aggregates, thereby illustrating the distribution characteristics of soil aggregates under natural wind erosion conditions. Given that the disintegration of aggregates across various size fractions is not fully realized, a significant quantity of aggregates remains large [29]. Conversely, the wet sieving method assesses the proportion of water-stable aggregates, which reflects the soil’s resilience to water erosion. Although the wet sieving technique encompasses all the mechanisms of aggregate disintegration, it is unable to differentiate the responses of aggregates to the various mechanisms involved [29,30].

The Le Bissonnais method employs various treatments to assess the impact of extreme weather on soil aggregates. The FW treatment effectively simulates the detrimental effects of severe weather events, such as heavy rainfall, resulting in the most pronounced aggregate disintegration. Conversely, the SW treatment mimics the effects of light precipitation or drip irrigation, wherein aggregates are gradually saturated with water, leading to minimal damage due to the limited shrink-swell capacity of the soil. The WS treatment involves mechanical disintegration processes, including wind erosion and tillage [31]. The findings of this study indicate that the FW treatment induces the most substantial breakdown of aggregates, followed by mechanical breakdown (WS), whereas chemical dispersion (SW) has the least impact on aggregate stability. These results are consistent with previous research employing similar experimental methodologies [32,33,34] The observed degradation of the soil aggregates in the study area is likely attributable to internal pressures within the aggregates that exceed the cohesive forces among soil particles, with external factors such as raindrop splash erosion exacerbating the disintegration process [26]. This phenomenon is associated with the frequent summer rainstorms in the region, where periodic fluctuations in water levels result in aggregates being made wet by both rivers and rainwater, ultimately contributing to their disintegration.

In conclusion, the Le Bissonnais method not only captures the outcomes of both dry and wet sieving processes but also offers a more thorough evaluation of aggregate stability by investigating the mechanisms involved in aggregate disintegration. Consequently, this method is more appropriate for assessing the soil aggregate stability in the study area, particularly under fluctuating water level conditions.

4.3. Response of Aggregate Fragmentation Indicators to Changes in the Water Table and Soil Stratigraphy

As shown in Table 2, the results of the present study revealed that soil SOM was unevenly distributed among the different soil particles. The clay particles had higher SOM contents, whereas both the sand and flour-sized particles had lower SOM contents. Similar to the conclusions reached by [35], this difference can be attributed to the different specific surface areas and charge densities of different-sized soil particles, which affect the adsorption of SOM. There was a positive correlation between the organic matter content and pH with increasing elevation. OM serves as a binding agent that facilitates the formation of soil aggregates and enhances their stability [36,37,38]. The varying stabilization properties of organic matter within aggregates contribute to distinct stabilization mechanisms during erosion processes [39,40]. Under conditions of elevated pH, the adsorption of organic matter by Fe³⁺ and other oxides is reduced, thereby allowing a greater proportion of organic matter to engage in the aggregation process [41]. The findings of this study indicate that, with increasing elevation, the soil organic matter content and pH increase, the resistance to shear index (RSI) decreases, and the resistance to mechanical index (RMI) increases. This suggests that prolonged flooding diminishes the impact of dissipative effects on soil aggregates while amplifying the effects of mechanical breakdown. Overall, aggregates are predominantly influenced by dissipative effects. Conversely, mechanical breakdown has a more significant destructive effect on both surface and subsurface soils, whereas its effect on deeper soils is relatively minimal. This discrepancy may be attributed to the fact that mechanical breakdown primarily targets surface soils, with a lesser influence on deeper layers. Additionally, the structural characteristics of aggregates may vary at different soil layers, affecting their sensitivity to mechanical breakdown. To preserve soil structural stability, it is essential to consider factors such as flooding duration and mechanical breakdown and to implement suitable soil conservation measures. Based on the distribution of soil characteristics in the drainage zone of Ertan Reservoir (Table 2) and the analysis of stability indicators, this study reveals that the destruction mechanism of soil aggregates under hydrological fluctuations is dominated by the interaction of multiple factors such as water level–depth–clay particles–SOM–pH: Long-term waterlogging (at low altitudes) promotes the enrichment of clay particles in deep soil through an anaerobic environment, but accelerates the decomposition of SOM and weakens the resistance to mechanical crushing (RMI↓). However, in the high-altitude areas with short-term flooding, a high pH environment is formed due to frequent dry-wet alternations, which inhibit the SOM–clay particle binding and intensifies the dissipation destruction during rapid wetting (RSI↑). In the mid-altitude area, due to the moderate duration of flooding, a relatively low pH, and a balanced ratio of sticky particles to SOM, the optimal stability (with the lowest RMI and RSI) is formed. This interaction indicates that for areas rich in clay particles, organic matter management and pH regulation should be combined to alleviate the dissipation sensitivity while, in areas with long-term waterlogging, the focus should be on enhancing the SOM retention capacity of deep soil.

5. Conclusions

This investigation focuses on the influence of water level variations on soil aggregates within the Ertan Reservoir Area in the Southwestern region, and has unveiled the following critical conclusions: The concentrations of non-water-stable aggregates across various size fractions, as well as water-stable aggregates larger than 5 mm and smaller than 1 mm, are significantly influenced by fluctuations in water level (p < 0.05), while no significant difference in soil layer is observed between non-water-stabilized and water-stabilized aggregates. A negative correlation exists between decreased water level and increased soil layer, both of which are associated with reduced soil aggregate stability. Under different pretreatment conditions, the stability hierarchy is as follows: fast wetting < pre-wet shaking < slow wetting. This suggests that both dissipative and mechanical breakdown mechanisms are primarily responsible for aggregate disintegration. Prolonged flooding exacerbates dissipative breakdown, while mechanical breakdown is modulated by water levels and soil layer, with lower water levels potentially intensifying this process. Among the diverse assessment techniques employed, the Le Bissonnais method, which can capture the outcomes of both dry and wet sieving and effectively differentiate the breakdown mechanisms of soil aggregates, is deemed more appropriate for evaluating soil aggregate stability within the study area. These conclusions can be extended to other reservoir areas with similar hydrological regulation characteristics (such as anti-season regulating reservoirs).