Effects of Flooding Duration on Plant Root Traits and Soil Erosion Resistance in Water-Level Fluctuation Zones: A Case Study from the Three Gorges Reservoir, China

Abstract

The water-level fluctuation zone (WLFZ) of the Three Gorges Reservoir (TGR) experiences seasonal submergence and exposure, resulting in soil structure degradation and intensified erosion. This study investigated how flooding duration affects root development and the erosion resistance of root–soil complexes in the WLFZ of the TGR. Two representative herbaceous species were chosen for this study: Xanthium sibiricum, an annual with a taproot system, and Cynodon dactylon, a perennial with a fibrous root system. Root traits, soil erodibility K-value, shear strength, and soil texture were measured from plant and soil samples collected at different flooding durations (145–175 m elevations). Our results showed that prolonged flooding significantly suppressed root growth, particularly in the 145–155 m zone, where root length density and root tips were markedly reduced (p < 0.05). Soil erodibility increased with flooding duration, with erodibility K-values ranging from 0.050 ± 0.002 to 0.062 ± 0.001 t·hm²·h/(MJ·mm·hm²), while shear strength declined correspondingly. Textural shifts from silty loam to silt were observed at zones experiencing extended flooding, contributing to aggregate instability and decreased internal friction angles. Notably, Cynodon dactylon demonstrated superior soil reinforcement capacity compared to Xanthium sibiricum, with its root volume and surface area significantly correlated with reduced K-values (p < 0.01) and enhanced shear strength (p < 0.001), enabling it to better prevent bank erosion under flooding conditions. These findings underscore the importance of root morphological traits in maintaining soil stability under hydrological stress and highlight the potential of perennial fibrous-rooted species for vegetation-based erosion control in fine-textured riparian zones. This study provides a theoretical basis and practical reference for ecological restoration in the WLFZ of the TGR and similar environments.

1. Introduction

The water-level fluctuation zone (WLFZ) refers to the area around a reservoir that is periodically exposed to the water surface due to fluctuations in water level [1]. The WLFZ of the reservoir belongs to the wetland category, serving as a transitional zone between land and water. The Three Gorges Reservoir (TGR) is located in central China, spanning Chongqing Municipality and Hubei Province, with a central coordinate at approximately 30°49′ N, 111°00′ E. The WLFZ in the TGR is a unique area formed by the artificial regulation of water levels during the operation of the Three Gorges Dam. Its water level regulation exhibits distinct seasonal reversal characteristics: the water level is lowered to meet flood control needs during the summer flood season, while the water level rises to ensure downstream water supply and hydroelectric power generation in the winter dry season. Consequently, a zone undergoing regular alternations between submergence and exposure has developed between 145 m and 175 m, encompassing an elevation range of 30 m [2]. The ecological environment of the WLFZ belongs to the ecological fragility, transience, and complexity region [3,4]. Ecological problems, including serious water and soil loss, aggravated sediment accumulation, soil pollution from the interaction between land and water, habitat degradation, severe soil erosion, and a reduction in ecosystem diversity, are prevalent [5]. Research on the erosion resistance of the root–soil complexes in the WLFZ plays a crucial role in revealing the soil consolidation and protective effects of roots under fluctuating water levels, improving slope stability, and strengthening soil and water conservation.

Soil erosion is a severe ecological problem that is faced by the WLFZ and has attracted more and more widespread attention. The severity of soil erosion along the shoreline areas of the TGR varies significantly. Studies have found that mainstream shorelines are more susceptible to erosion than tributary shorelines, low-elevation banks are more ecologically vulnerable than high-elevation ones, and soils with finer particles exhibit relatively lower erosion rates [6,7,8]. Native woody species, such as Salix matsudana, and herbaceous species, including Cynodon dactylon and Eclipta prostrata, in the TGR have been demonstrated to exhibit effective growth strategies under submergence stress, indicating their considerable potential for vegetation restoration in the WLFZ [9,10]. Moreover, studies have shown that increasing planting years of tea plantation significantly enhances soil structural stability and shear strength in the TGR, mainly through improvements in organic matter, aggregate stability, and root density. Meanwhile, vegetation restoration—particularly afforestation on medium to steep slopes—effectively mitigates soil erosion, indicating that slope gradient and vegetation recovery jointly regulate the spatial variability of soil shear strength [11,12]. The soil aggregate structure in the WLFZ is severely damaged because of the influence of long-term flooding and drainage [13]. In addition to the inherent strength of the soil itself, water-level fluctuations have become an important factor in the soil erosion process in the WLFZ [14,15]. Vegetation coverage, as a common soil erosion control technique, has also been shown to yield good results in the WLFZ [16].

In the WLFZ of the TGR, previous studies have mostly focused on either vegetation or soil individually, with limited attention paid to their interaction. To address this gap, this study investigated the root–soil complexes of two dominant species, Xanthium sibiricum and Cynodon dactylon. As an annual tap-rooted herb and a perennial fibrous-rooted herb, they respectively represent contrasting root morphologies and life cycles, making them suitable model species for exploring how root traits influence soil erosion resistance under flooding disturbance. We hypothesize that the contrasting root morphologies (fibrous and taproot) would result in different soil reinforcement capacities under flooding disturbance, with fibrous roots providing more extensive soil binding and lateral stabilization, whereas taproots mainly contribute to vertical anchorage with limited horizontal reinforcement. By analyzing soil erodibility and shear strength, this study aimed to assess the erosion resistance of root–soil complexes under varying flooding durations. Furthermore, this study uniquely integrates root morphological traits with soil mechanical indicators across a flooding gradient, offering a novel perspective on plant–soil interactions under fluctuating hydrological regimes. Although Cynodon dactylon is effective for soil stabilization, its potential invasiveness requires that restoration practices balance erosion-control benefits with ecological risks through species diversity and proper management. The results are expected to reveal how flooding duration influences root adaptation and soil mechanical properties, providing theoretical and empirical support for vegetation-based ecological restoration in the WLFZ of the TGR and similar riparian zones.

2. Methodology

2.1. Study Site

As shown in Figure 1, the study area is located within the WLFZ of the Shibao Town section in Zhongxian County, Chongqing (107°32′–108°14′ E, 30°03′–30°35′ N), which is situated in the central section of the TGR and has been widely recognized as a representative area of the reservoir’s WLFZ due to its typical seasonal water-level fluctuations (up to 30 m). For the upstream and downstream of the TGR, the Zhongxian River section has the comprehensive representative advantages of combining upstream erosion and downstream sedimentation characteristics, obvious terrain undulations, and typical water level fluctuation gradients. The bedrock of this region is primarily composed of purple–red medium-thick layers of silty shale, interspersed with calcareous sandstone and dolomitic limestone layers, all of which exhibit typical purple-red coloration. Influenced by the parent rock, the soil developed on the surface is mainly purple soil, with localized terrain units containing yellow soil (accounting for 15–20%) and paddy soil (accounting for 10–15%). These two soil types show a distinct contrast in profile morphology and physical-chemical properties compared to the purple soil. According to the World Reference Base for Soil Resources (WRB), the soils in the study area, derived from purple sandstone and shale, are weakly developed with indistinct horizon boundaries, and can be classified mainly as Regosols. The region is situated in the humid subtropical southeastern monsoon climate zone, where rainfall and temperature coincide. The region experiences concentrated rainfall between June and August, with yearly precipitation around 1150 mm and an average temperature of 19.2 °C annually. Under prolonged waterlogging stress, the vegetation in the WLFZ is dominated by a typical and monocultural Cynodon dactylon community, accompanied by species such as Xanthium sibiricum, Polygonum hydropiper, and others [17]. Before flooding, the land use in this area mainly consisted of dryland farming and rice cultivation [18].

2.2. Experimental Design

2.2.1. Sampling Plots and Sample Collection

Due to the varying flooding durations and differences in species richness across different elevation zones of the WLFZ, the sampling plots were classified based on elevation. According to the water level data of the TGR, the soil at elevations of 165–175 m was exposed in March, the soil at elevations of 155–165 m was exposed in mid-April, and the water level dropped to its lowest point at 145 m in June, exposing the soil at elevations of 145–155 m. In August, the water level rose back to 160 m. Based on the periodic exposure of approximately 10 m soil profiles at monthly intervals (March–June) and the distinct vertical stratification observed in the study area, sampling zones were systematically established at 10 m elevation intervals.

Root samples from two plant species and corresponding soil samples at different elevations in the WLFZ were collected for analysis. The plants selected were Xanthium sibiricum and Cynodon dactylon, which are widely distributed in the WLFZ. The elevations were classified into a low-water level zone (L, 145–155 m), a mid-water level zone (M, 155–165 m), a high-water level zone (H, 165–175 m), and an unflooded control group (CK, 175–185 m). Three sampling points were set at each elevation. Meanwhile, plant root and soil samples from the WLFZ were collected in mid-April, mid-June, and mid-August 2024. These three sampling periods were designed to serve as temporal replicates, ensuring consistency and comparability of observations across time while accounting for potential seasonal variation. Samples collected in different seasons were treated as temporal replicates and analyzed together to increase statistical robustness. Although seasonal variation may influence plant growth and soil properties, a detailed analysis of seasonal effects was beyond the scope of this study. To improve sampling reliability, two plots were set at each elevation to account for variations in plant cover (dense vs. sparse), with soil samples taken from multiple depths at each site.

The root sampling methodology was as follows: representative plots within the plant growth area were selected, and undisturbed soil cores were collected vertically using a 25 cm × 25 cm soil sampler to preserve the structural integrity of both the soil and root systems. The root–soil blocks were then transported to the laboratory, where roots were gradually separated from the complex by rinsing with distilled water in a bottom-up manner. The roots were isolated and collected layer by layer. After rinsing, the roots were sieved to remove any remaining sediment and then placed in Ziploc bags for storage and subsequent analysis of root traits. A minimum of three replicate sampling points were established at each elevation to ensure representativeness and data reliability.

During the sampling of the root–soil complexes, surface vegetation was first removed. Subsequently, soil samples were collected from the surface down to a depth of 0–40 cm without disturbing the original soil structure. Sampling was performed in four layers, with a 10 cm interval between each layer. Four independent samples were collected with a 60 cm³ soil ring sampler to measure the shear strength of the root–soil complexes. In addition, bulk soil samples were collected and transported to the laboratory for particle size distribution and organic carbon content analysis; each weighed approximately 1 kg.

2.2.2. Sample Processing and Testing

The root samples were rinsed and separated using deionized water and then placed in a glass Petri dish with distilled water. Microscope precision forceps (with 0.1 mm accuracy) were used to remove any adhering impurities. High-resolution root images were obtained using an Epson Perfection V 700 Pro digital scanner, ensuring that the roots were fully extended in the distilled water during scanning. The root length density, total root volume, root surface area, and root tips were quantitatively assessed using the WinRHIZO Pro root image analysis system and standardized per cubic meter of soil volume.

Shear strength testing was conducted using the ZJ strain-controlled direct shear apparatus (Nanjing Soil Instrument Factory). During the experiment, vertical pressures of 50, 100, 150, and 200 kPa were sequentially applied to four samples from the same batch to obtain the corresponding shear strength values. The shear strength–vertical pressure relationship line was then plotted based on Mohr–Coulomb theory [19]. The equation is as follows:

τ = stgφ + C

(1)

In the equation, τ represents the shear strength of the soil (kPa); s is the vertical pressure applied to the shear plane (kPa); φ is the internal friction characteristic between soil particles, also known as the internal friction angle (°); and C reflects the inherent bonding ability of the soil, also known as cohesion (kPa).

The bulk soil samples were air-dried and passed through a 2 mm sieve to remove plant roots and surface debris. Soil particle size distribution was determined using a Mastersizer 2000 laser particle size analyzer (Malvern Instruments Ltd., Malve, UK) equipped with a Hydro 2000MU module. Soil organic carbon (SOC) content was measured using the potassium dichromate oxidation-external heating method. The soil erodibility K-value was then calculated based on the Universal Soil Loss Equation (USLE) model [20], using the following equation:

$$\begin{array}{c} K=0.1317\times \left\{0.2 + 0.3\mathit{exp}\left[-0.0256S_{\mathit{an}}\left(1.0 - {\displaystyle \frac{S_{\mathit{il}}}{100}}\right)\right]\right\}\times {\left({\displaystyle \frac{S_{\mathit{il}}}{C_{\mathit{la}} + S_{\mathit{il}}}}\right)}^{0.3}\\ \times \left\{1.0 - {\displaystyle \frac{0.25C}{\left[C + \mathit{exp}\left(3.72 - 2.95C\right)\right]}}\right\}\times \left\{1.0 - {\displaystyle \frac{{0.7S}_{\mathit{Nl}}}{S_{\mathit{Nl}} + \mathit{exp}\left(-5.51 + 22.9S_{\mathit{Nl}}\right)}}\right\}\end{array}$$

(2)

In the equation, S_an represents the sand content (%); S_il is the silt content (%); C_la refers to the clay content (%); C is the organic carbon content (%); and S_Nl is the proportion of non-sand particles, where S_Nl = 1 − S_an/100.

2.3. Experimental Data Analysis and Statistics

Statistical analyses, including t-tests and analysis of variance (ANOVA), were performed using IBM SPSS Statistics 25, with significant differences determined using the least significant difference (LSD) method. Data fitting and graphical representations were conducted using Origin 2021.

3. Results and Analysis

3.1. Soil Texture Characteristics of the Water-Level Fluctuation Zone

According to the soil textural taxonomy defined by the USDA Natural Resources Conservation Service, the soil particle size distribution in the WLFZ of the TGR area was predominantly composed of silt particles (Figure 2) [21], accounting for 49.61% to 96.77%, followed by sand particles ranging from 1.26% to 49.00%, while clay content was the lowest, merely constituting 0.52% to 4.47%. Soils subjected to different flooding durations were predominantly classified as silty loam and silt soils. Specifically, soil samples from the 155–165 m and 145–155 m zones were identified as silt soils, whereas those from the control group were characterized as silty loam. Notably, within the 165–175 m elevation range, 83.3% of samples exhibited silty loam properties, with the remaining 16.7% being silt soils. Moreover, it was observed that prolonged flooding duration correlated with a progressive increase in silt content, accompanied by a textural transition from silty loam to silt soil.

3.2. Root Characteristics of Xanthium sibiricum and Cynodon dactylon Under Different Flooding Durations in the Water-Level Fluctuation Zone

The root characteristics of Xanthium sibiricum and Cynodon dactylon under different flooding durations in the riparian zone are shown in Figure 3. Since the area at an elevation of 145 m was not fully exposed during sampling, root samples were collected from elevations of 175 m, 165 m, and 155 m, respectively corresponding to the elevation ranges of 165–175 m, 145–155 m, and 155–165 m. For Xanthium sibiricum, root length density, total root volume, root surface area, and root tips all decreased with increasing flooding duration, with the highest values observed at 175 m. Significant reductions were observed at 165 m and 155 m (p < 0.05), with no obviously significant difference between the 165 m and 155 m elevations. The root characteristics of Cynodon dactylon were similar to those of Xanthium sibiricum, which decreased with the increasing flooding duration. For the same flooding duration, Cynodon dactylon exhibited lower root length density, total root volume, root surface area, and root tips compared to Xanthium sibiricum. At 155 m elevation, the root length density, root surface area, and root tips of Cynodon dactylon were significantly lower than those at 175 m and 165 m (p < 0.05), while the total root volume at 155 m was only significantly lower than at 175 m (p < 0.05). Due to the unique characteristics of the Three Gorges WLFZ, water levels began to drop at the end of March, with 175 m exposed first, followed by 165 m, and finally 155 m. The exposure duration was the shortest at 155 m, resulting in much shorter plant growth time, and thus, the root system was in the early stages of growth. Xanthium sibiricum is an annual herb, and a single seed can only germinate once per year. Therefore, with longer exposure time, Xanthium sibiricum tends to reach maturity. In contrast, Cynodon dactylon is a perennial herb that can continue growing from roots that were established before flooding. It has a short growth cycle and a fast growth rate, allowing it to quickly sprout roots and grow well. Thus, there was little difference between the Cynodon dactylon at 175 m and 165 m elevations. However, at 155 m, the plant had not yet reached its peak growth period due to insufficient growth time.

3.3. The Erodibility K-Value Characteristics of Root–Soil Complexes Under Different Flooding Durations in the Water-Level Fluctuation Zone

The erodibility K-value ranged from 0.050 ± 0.002 t·hm²·h/(MJ·mm·hm²) to 0.062 ± 0.001 t·hm²·h/(MJ·mm·hm²) in this study. This index was calculated from the particle size distribution and organic carbon data. As shown in Figure 4, the variation pattern of the soil erodibility K-value at the same soil depth is as follows: 145–155 m > 155–165 m > 165–175 m > control group (CK), indicating that the erodibility K-value decreases with the reduction in flooding duration. In the 0–10 cm, 10–20 cm, and 20–30 cm soil layers, the erodibility K-values of the 145–155 m soils are significantly higher than those of the control group (CK) (p < 0.05). The erodibility K-values of the 145–155 m and 155–165 m soils are significantly higher than those of the 165–175 m and control group (CK) (p < 0.05) in the 30–40 cm soil layer. There was no significant difference in the erodibility K-value of the WLFZ soils as the soil depth increased. The K-values showed a trend of first increasing and then decreasing, due to the shallow soils being more susceptible to structural damage from the floods.

3.4. Shear Strength Characteristics of Root–Soil Complexes Under Different Flooding Durations in the Water-Level Fluctuation Zone

The shear strength–vertical pressure relationship lines of root–soil complexes in the WLFZ are shown in Figure 5. Based on the relative positions of the lines, it could be observed that the shear strength of root–soil complexes decreases with increasing flooding duration, following the order: control group (CK) > 165–175 m > 155–165 m > 145–155 m. Moreover, the shear strength of root–soil complexes showed a significant linear increase with increasing vertical pressure (p < 0.05).

Combined with the shear strength data and Mohr–Coulomb theory, the cohesion and internal friction angle of root–soil complexes under different altitudes (which can also be regarded as flooding durations) in the WLFZ were plotted as shown in Figure 6 and Figure 7. The internal friction angle gradually decreases with increasing flooding duration in the 0–40 cm soil depth, following the order (R² > 0.8, p < 0.05): control group > 165–175 m > 155–165 m > 145–155 m. In contrast, cohesion shows no significant linear correlation with flood duration (R² < 0.7, p > 0.05).

3.5. Correlation Analysis Between the Erodibility K-Values, Shear Strength of Root–Soil Complexes, and Root Traits Under Different Flooding Durations

The results of the correlation analysis among the erodibility K-values, shear strength (under 150 kPa vertical pressure), and the root characteristics indices of Xanthium sibiricum are shown in Figure 8. The results indicated that there existed a significant negative correlation between the erodibility K-value and soil shear strength at 150 kPa vertical pressure (p < 0.01). However, both the K-value and shear strength showed weak correlations with root length density, total root volume, root surface area, and root tips of Xanthium sibiricum. Among the root trait indicators of Xanthium sibiricum, positive correlations were observed among root length density, total root volume, root surface area, and root tips, with correlation coefficients ranging from 0.56 to 0.97. Notably, a significant positive correlation (p < 0.01) was found between the total root volume and the root tips, while the remaining root characteristics indices exhibited obviously significant positive correlations (p < 0.001).

The correlation analysis among the erodibility K-values, shear strength (under 150 kPa vertical pressure), and the root traits of Cynodon dactylon is presented in Figure 9. The soil erodibility K-value was significantly negatively correlated with total root volume (r = –0.55, p < 0.01) and root surface area (r = –0.43, p < 0.05), while it showed weak correlations with root length density and root tips. Soil shear strength (under 150 kPa vertical pressure) was significantly positively correlated with root length density (p < 0.05) and highly significantly correlated with total root volume and root surface area (p < 0.001). Compared to Xanthium sibiricum, all root traits of Cynodon dactylon exhibited significant correlations with one another (p < 0.001).

4. Discussion

4.1. Effect of Flood Duration on Root Traits of Typical Plants in the Water-Level Fluctuation Zone

With prolonged flooding duration, the root growth of plants declined, which may be attributed to the reduction in oxygen availability within soil pores. Under flooded conditions, oxygen content in soil pores significantly decreased, affecting root respiration and metabolism, inhibiting cell elongation and division, and ultimately leading to reductions in root length density and root tips. Extended flooding resulted in root decay or growth stagnation [22]. Flooding also induces nutrient leaching and oxygen deficiency in the soil [23], which explains the overall lowest root trait values observed at 155 m elevation for both species. Moreover, the duration of flooding may influence the interactive effects of soil texture on root morphology and development. Soils subjected to longer submergence showed a significant increase in silt content, while those with shorter flooding durations contained higher proportions of clay and maintained a silty loam texture (Figure 2). Firstly, changes in soil particle size distribution can cause alterations in soil physicochemical properties and enzyme activity, thereby regulating root structure, growth kinetics, and developmental processes [24]. Compared to soils dominated by silt, balanced textures such as loam contribute to better aeration, nutrient retention, and root growth. These areas have finer texture and higher clay content, which may improve soil structure and nutrient retention, further supporting root growth and stability. Additionally, root-derived exudates facilitate the formation of soil aggregates to varying extents, thereby enhancing the mechanical properties of the soil matrix [25,26].

However, even at the same elevation, Xanthium sibiricum and Cynodon dactylon exhibited distinct root growth strategies. As an annual xerophytic herb, Xanthium sibiricum was more sensitive to flooding duration, while the perennial nature of Cynodon dactylon enabled it to adopt dormancy strategies, granting better tolerance to long-term submergence and mitigating hypoxia stress in roots. As a result, the root trait values of Cynodon dactylon showed no significant differences between 165 m and 175 m, with a significant decrease only at 155 m [27,28]. Cynodon dactylon can also adapt to inundation by increasing root diameter and modifying aerenchyma structure to maintain root function [29]. Regarding the functional traits and stoichiometric characteristics of herbaceous plants, Xanthium sibiricum exhibited markedly different inter-trait relationships compared to Cynodon dactylon and Polygonum hydropiper, with most of its functional traits being positively correlated [30]. The conclusions align with the trends identified in the root functional traits of Xanthium sibiricum and Cynodon dactylon in this study.

4.2. The Erodibility K-Value and Shear Strength Characteristics of Root–Soil Complexes Under Different Flooding Durations

Compared with the control group, the soil erodibility K-values in the WLFZ were significantly higher, and the K-values increased with longer flooding duration, indicating greater soil instability and reduced erosion resistance due to flooding. Vertically, the erodibility of flooded soils in the WLFZ generally exhibited a pattern of rising first and subsequently declining with increasing depth, suggesting that deeper soil layers, characterized by a more complex structure, tend to exhibit stronger resistance to erosion. As flooding duration decreased, vegetation cover became denser and root systems more developed. Well-developed root systems contributed to increased organic matter and improved soil structure [31,32]. Meanwhile, according to the K-value calculation formula, higher organic matter content significantly enhances soil resistance to erosion. Plant roots can also enhance the stability of soil aggregates through the secretion of organic compounds and mechanical entanglement, thus reducing soil erodibility [33]. This study further confirms these findings and demonstrates that soils with high root content at elevations of 165–175 m exhibited significantly higher organic matter content than soils with low root content, suggesting that plant roots play a critical role in organic matter accumulation. Additionally, this study found that soil erodibility increased with depth, aligning with previous studies that reported deeper soil structures are more prone to damage and exhibit higher erodibility [34,35].

For soils subjected to different flooding durations, shear strength increased significantly with applied vertical stress at all depth profiles, while increasing flooding duration led to a decrease in shear strength. This trend may result from several interacting factors. Prolonged flooding may reduce shear strength through multiple mechanisms: longer inundation increases soil moisture content and pore water pressure, leading to a reduction in effective stress and, consequently, shear strength, consistent with the findings of Turrakheil K. S. [36]. On the soil side, continuous flooding caused the breakdown of water-stable aggregates, resulting in looser soil structure and rearrangement of particles, which reduced internal friction and thus shear strength [37,38]. Furthermore, the internal friction angle of the root–soil complexes decreased progressively with longer flooding duration, consistent with previous findings [39], while cohesion showed no significant correlation with flooding duration, likely due to differences in the initial moisture content of root–soil complexes caused by submergence. In terms of plant influence, flooding duration significantly affects root shear strength, which decreases as flooding duration increases. Longer submergence leads to root mortality and the loss of associated benefits such as organic matter secretion and physical entanglement, thereby reducing soil shear strength [40]. Organic matter can directly influence key shear strength parameters, including cohesion and internal friction angle. In root–soil complexes composed of both soil and plant roots, improved coupling between roots and soil enhances shear strength. Root characteristics such as geometric structure and water content influence soil stabilization capacity, while root density, diameter, and distribution depth significantly impact soil shear strength [37,41,42]. In summary, extended periods of submergence alter soil mechanical properties by modifying both soil composition and root functionality.

Beyond biological effects, variations in soil texture also account for the observed changes in erodibility and shear strength. In zones subjected to prolonged inundation, a textural shift from silty loam to silt was noted, which may reflect a weakening of soil aggregation and water retention capacity. Soils with a high silt content exhibit lower cohesive force, rendering them more susceptible to particle detachment [43]. In contrast, areas experiencing shorter flooding durations retained a higher clay fraction [44,45]. This higher clay content may improve soil aggregation and cohesion, providing better structural stability. Vegetation restoration and related biological activities further enhance these improvements, which support the erosion resistance of these areas.

4.3. Effects of Root Structural Differences on the Anti-Erosion Capacity of Root–Soil Complexes in the Reservoir Water-Level Fluctuation Zone

The correlations between root traits of the two plant species and soil shear strength and the erodibility K-value differed significantly, primarily due to disparities in their root structures and functional characteristics. Cynodon dactylon possesses a well-developed fibrous root system, with greater root volume and surface area, which enhances the cohesion among soil particles through stronger physical entanglement and reinforcement effects, thereby significantly increasing shear strength and reducing soil erodibility [46]. Moreover, this dense root structure contributes to the stabilization of soil aggregates, mitigates the risk of raindrop impact and runoff-induced erosion, and plays a crucial role in preserving soil nutrients [47,48]. In contrast, Xanthium sibiricum is a taproot-type plant with relatively shallow and concentrated root distribution. Although its total root volume may be relatively large in some plots, the lack of effective lateral expansion limits its reinforcement capacity on the overall soil structure, resulting in a weaker contribution to the improvement of soil erodibility [49,50]. The widespread distribution and higher density of fibrous roots provide stronger physical entanglement and surface area, enhancing shear strength; although the taproot can penetrate vertically, it lacks horizontal expansion. This observation aligns with the viewpoint of Stokes, who emphasized that fibrous root systems, owing to their higher root density and more extensive spatial distribution, are more beneficial for enhancing soil structural stability and erosion resistance [51]. Therefore, morphological differences in plant root systems are key factors underlying the contrasting performance of the two species in terms of soil mechanical properties observed in this study. Among the root traits of Cynodon dactylon, total root volume and root surface area demonstrated stronger correlations with both soil shear strength and K-value. This could be explained by the contribution of increased root volume to a more cohesive and intact soil structure, thereby enhancing shear strength and erosion resistance. Additionally, root surface area represents the interface between roots and soil, which is closely related to shear strength parameters, including the internal friction angle and cohesion. A larger surface area can facilitate greater interparticle cohesion and aggregate stability. For ecological restoration of the WLFZ, Xanthium sibiricum or mixed sowing suits high elevations (165–175 m), Cynodon dactylon and other perennials suit low to mid elevations (145–165 m), and mixing with natives like Polygonum hydropiper helps prevent single-species dominance.

5. Conclusions

This study integrated plant root functional traits and soil mechanical parameters to evaluate the erosion resistance of root–soil complexes under different flooding durations in the water-level fluctuation zone of the Three Gorges Reservoir. By selecting and comparing two representative herbaceous species with contrasting root types, the findings provide new insights into plant–soil interactions and root-mediated stabilization processes in periodically inundated riparian zones. The main conclusions are as follows:

(1)

Flooding suppressed root growth to varying degrees in different plants, especially in the long-term flooding areas at an altitude of 145–155 m, where both root length density and root tips were significantly reduced (p < 0.05). Xanthium sibiricum exhibited greater sensitivity to flooding at middle and high-water level zones, whereas Cynodon dactylon was more affected at middle and low-water level zones, indicating its stronger flood tolerance, when compared with that of Xanthium sibiricum.

(2)

Prolonged flooding notably increased the erosion susceptibility and reduced the shear strength of root–soil complexes in the water-level fluctuation zone. As flooding duration increased, the soil erodibility K-value rose progressively with elevation in the order of CK < 165–175 m < 155–165 m < 145–155 m, while shear strength showed a corresponding decline. These variations were accompanied by a textural transition from silty loam to silt under extended submergence, particularly in low-elevation zones. The higher silt content and reduced clay proportion contributed to aggregate breakdown and weakened particle cohesion, leading to lower internal friction angles and increased vulnerability of the shallow soil layers to structural damage.

(3)

The fibrous root system of Cynodon dactylon significantly reinforced soil structure more effectively than the taproot system of Xanthium sibiricum. The soil erodibility K-value was significantly negatively correlated with the total root volume (p < 0.01) and root surface area (p < 0.05) of Cynodon dactylon, while the soil shear strength was not only significantly positively correlated with the root length density of Cynodon dactylon (p < 0.05), but also extremely significantly positively correlated with the total root volume and root surface area of Cynodon dactylon (p < 0.001). The morphological structure of the root system is thus a key determinant of the erosion resistance of root–soil complexes.

This research provides not only theoretical insights but also practical guidance for ecological restoration. The results highlight that perennial fibrous-rooted species such as Cynodon dactylon can be prioritized in vegetation restoration programs to enhance soil stability and mitigate erosion in the water-level fluctuation zone of the Three Gorges Reservoir. Specifically, low and mid-elevation areas (145–165 m) with longer flooding durations can benefit from the rapid regrowth and strong soil-reinforcing capacity of Cynodon dactylon, whereas high-elevation zones (165–175 m) are more suitable for Xanthium sibiricum or mixed sowing. To prevent dominance by a single species, combining these with native plants such as Polygonum hydropiper is recommended. Furthermore, the experimental design and standardized methodologies employed, such as root trait quantification, soil erodibility assessment, and shear strength testing, are replicable and applicable to other reservoir WLFZs and similar hydrological environments. These aspects enhance the repeatability of the study and provide a framework that can inform ecological restoration and soil conservation strategies beyond the study area. However, this study only focused on two species at a single location in the middle reaches of the Three Gorges Reservoir, future work should expand to diverse plant types, multi-site comparisons, and consider root–microbe interactions to better inform soil stabilization strategies.