Peak Soil Erosion Risk in Mixed Forests: A Critical Transition Phase Driven by Moso Bamboo Expansion

Abstract

Driven by climate change and human activities, the expansion of highly invasive moso bamboo (Phyllostachys edulis) into coniferous forests induces a serious ecological imbalance. Its rapidly spreading underground roots significantly alter soil structure, yet the mechanisms by which this expansion affects soil detachment capacity (Dc), a key soil erosion parameter, remain unclear. While bamboo expansion modifies soil physicochemical properties and root characteristics, influencing Dc and, consequently, soil erosion resistance, the underlying mechanisms, particularly stage-specific variations, are not thoroughly understood. In this study, we examined Japanese white pine (Pinus parviflora Siebold & Zucc.) forest (CF), moso bamboo–Japanese white pine mixed forest (MF), and moso bamboo forest (BF) as representative stages of bamboo expansion. By integrating laboratory-controlled measurements of soil physicochemical properties and root traits with field-based flume experiments, we comprehensively investigate the effects of moso bamboo expansion into CF on soil detachment capacity. The results of the study can be summarized as follows: (1) Expansion of moso bamboo significantly changed soil physicochemical properties and root characteristics. Soil bulk density was the highest in the MF (1.13 g·cm⁻³), followed by the CF (1.08 g·cm⁻³) and BF (1.03 g·cm⁻³); non-capillary porosity increased significantly with expansion (CF 0.03% to MF 0.10%); and although the stability of aggregates (MWD) increased by 24.5% from the CF to MF, root mass density (RMD) in the MF (0.0048 g·cm⁻³) was much higher than that in the CF (0.0009 g·cm⁻³). This intense root competition between forest types, combined with increased macroporosity development, compromised overall soil structural integrity. This weakening may lead to a looser soil structure during the transition phase, thereby increasing erosion risk. (2) There were significant stage differences in Dc: it was significantly higher in the MF (0.034 kg·m⁻²·s⁻¹) than in the CF (0.023 kg·m⁻²·s⁻¹) and BF (0.018 kg·m⁻²·s⁻¹), which revealed that the MF was an erosion-sensitive stage. (3) Our Partial Least Squares Structural Equation Modeling (PLS-SEM) results revealed that soil physicochemical properties (soil moisture content and soil total nitrogen) dominated Dc changes through direct effects (total effect −0.547); in comparison, root properties indirectly affected Dc by modulating soil structure (indirect effect: −0.339). The results of this study reveal the dynamics and mechanisms of Dc changes during bamboo expansion, and for the first time, we identify a distinct Dc peak during the mixed forest transition phase. These findings provide a scientific basis for moso bamboo forest management, soil erosion risk assessment, and optimization of soil and water conservation strategies.

1. Introduction

In the context of global climate change and increasingly pervasive human activities, forest ecosystems are undergoing profound structural and functional transformations. The expansion of certain species, characterized by accelerated growth and robust dispersal capabilities [1], has emerged as a significant global ecological concern. Among such species, moso bamboo (Phyllostachys edulis), a clonal plant [2], is invading native forests at unprecedented rates [3,4] across subtropical regions globally [5], particularly in subtropical Asia. This expansion is driven by its extensive rhizome system and environmental adaptability. Crucially, it not only alters species composition and biodiversity [5,6] but it also threatens critical ecosystem services, such as soil erosion resistance [7], root characteristics [8], and hydrological processes [9].

Soil erosion is a primary driver of land degradation and poses a major challenge to ecosystem sustainability. Subtropical regions are particularly vulnerable to this process due to seasonal heavy rainfall and frequent storms and typhoons. Vegetation plays a pivotal role in erosion resistance, with different plant types modulating erosion processes through variations in surface cover, litter input, soil structure, and root characteristics [10]. The replacement of native vegetation by moso bamboo may disrupt soil and water conservation functions [11]. At present, however, the mechanisms by which bamboo expansion impacts soil erosion—particularly in hydrologically sensitive subtropical areas—remain poorly understood.

Soil detachment capacity (Dc) is a core parameter characterizing soil erosion resistance, defined as the mass of soil particles detached from the main soil body per unit time and area under specific hydrodynamic conditions [12,13]. Distinct from macroscopic soil loss measurements, Dc directly reflects the soil’s resistance to particle detachment under runoff shear stress, serving as a critical process parameter for understanding and predicting erosion dynamics [14]. Assessing the soil and water conservation benefits of moso bamboo forests therefore requires a thorough understanding of how bamboo affects Dc. Dc is influenced by multiple factors, encompassing hydraulic properties [15], soil physicochemical characteristics (including bulk density, porosity, organic matter content, aggregate stability, and texture) [16,17], and plant root-system attributes (such as root biomass, spatial distribution patterns, and root–soil cementation effects) [18]. Moso bamboo possesses a well-developed underground rhizome–root system. The characteristics of this root system are considered to have significant potential for improving soil structure and enhancing erosion resistance. The results of a number of studies demonstrate a positive correlation between soil aggregate stability [19], organic matter content [20], and root biomass [21] on the resistance of soil erosion. Such findings suggest that moso bamboo may resist Dc by changing the physical and chemical properties of the soil (such as increasing organic matter and promoting aggregation) and shaping a unique root structure.

The expansion of moso bamboo is a dynamic and evolving process, typically progressing through distinct stages: from native forests (such as coniferous forests) to mixed bamboo–wood forests, ultimately culminating in pure moso bamboo forests [22]. Throughout this developmental continuum, both soil properties and root characteristics exhibit marked transformations. For instance, moso bamboo has a dense, shallowly distributed root network [23]. This may have complex effects on soil structure: On one hand, its roots can enhance soil retention through physical entanglement [3]. On the other hand, its rapid growth and competition may lead to the decline of the root system of the original stand and the destruction of soil aggregates due to its rapid growth and competition. The litterfall and root exudates of moso bamboo may potentially alter the accumulation and turnover of soil organic matter, thereby influencing the formation and stability of soil aggregates [24,25,26]. At present, however, there is a relative paucity of research on the characteristics of Dc dynamics and its dominant mechanisms under different stages of moso bamboo expansion, particularly on how soil properties and root characteristics synergistically or independently affect Dc. This lack of research limits our ability to accurately assess the risk of soil erosion in areas of moso bamboo expansion and develop effective management strategies.

To address these gaps, we focus on a subtropical moso bamboo expansion area in China’s Yili region (Jiangsu Province). We selected three forest types representing expansion stages: Native Japanese white pine forests (Pinus parviflora Siebold & Zucc.), mixed moso bamboo–pine forests, and pure moso bamboo forests. By integrating laboratory analyses of soil physicochemical properties and root characteristics with controlled flume experiments, we aim to achieve the following systematic objectives: (1) characterize the dynamics of soil physicochemical properties, root characteristics, and soil detachment capacity during the expansion of moso bamboo; and (2) quantify the effects of soil physicochemical properties and root characteristics on Dc, determine the dominant driver (soil properties vs. roots) governing Dc, and elucidate its underlying mechanisms. Our findings provide critical insights for optimizing soil conservation strategies in bamboo-invaded subtropical ecosystems.

2. Materials and Methods

2.1. Study Area Description

The study area is located in the Yili mountainous area (119.65° E, 31.20° N) at a mean altitude of 183.3 m above sea level, in Jiangsu Province, China, characterized by a subtropical humid monsoon climate. The region has an average annual temperature of 15.8 °C and receives 1276.5 mm of annual precipitation [27] (Figure 1). The topography primarily comprises low hills with an elevation of approximately 200 m. The native vegetation consists of Japanese white pine (Pinus parviflora Siebold & Zucc.) forests, which have been increasingly invaded by moso bamboo (Phyllostachys edulis), forming a gradient of coniferous forests (CFs), mixed bamboo–pine forests (MFs), and pure moso bamboo forests (BFs).

2.2. Sample Site and Soil Sample Collection

Along the gradient of moso bamboo expansion, three types of typical stands were selected (Figure 1): CF, which is the initial phase of bamboo invasion, characterized by the predominance of Japanese white pine; MF, the midpoint of bamboo invasion, marked by a mixture of moso bamboo and Japanese white pine; and BF, the subsequent phase of bamboo invasion, typified by the dominance of moso bamboo (

Table 1. Sample overview.

Plot Type	Plot Number	Elevation (m)	Dominant Tree Species	Stem Count	DBH (cm)
CF	Plot 1	189.7	Japanese white pine	6	18.62 ± 14.93
	Plot 2	190.7	Japanese white pine	13	49.93 ± 10.42
	Plot 3	186.9	Japanese white pine	12	59.42 ± 11.40
MF	Plot 4	184.5	Moso bamboo, Japanese white pine	6, 7	24 ± 3.46, 58 ± 10.53
	Plot 5	185.0	Moso bamboo, Japanese white pine	10, 8	29 ± 1.41, 63.26 ± 16.26
	Plot 6	187.0	Moso bamboo, Japanese white pine	10, 10	27 ± 5.16, 49.5 ± 10.61
BF	Plot 7	173.6	Moso bamboo	33	37.5 ± 3.73
	Plot 8	177.0	Moso bamboo	22	38.67 ± 2.52
	Plot 9	174.9	Moso bamboo	23	40 ± 3.29

Note: The experimental design comprised three replicated plots for each of the three stand types, totaling nine plots. Values represent mean ± standard deviation (n = 3 plots per type). CF is Japanese white pine forest, MF is mixed Japanese white pine and moso bamboo forest, BF is moso bamboo forest, and DBH is diameter at breast height.

). Detailed elevation characteristics of each sampling location are provided in Table 1. In each forest stand, three parallel sample strips (10 m × 50 m) were established, and within each of these strips, three 10 m × 10 m sample plots were delineated (resulting in a total of nine sample plots). The number and diameter at breast height of each dominant tree species in each sample plot were investigated. Furthermore, nine samples of primary soil measuring 0–10 cm in depth were collected from the surface layer for the flume experiment. These samples were obtained using a 200 cm³ ring knife, with the samples collected based on a 3 m × 3 m grid method. In addition, three samples of primary soil were collected using the diagonal method with a smaller 100 cm³ ring knife. This method was followed in order to determine the soil bulk density. Concurrently, three samples of disturbed soil were collected to ascertain the soil’s physicochemical properties.

2.3. Determination of Soil Detachment Capacity

Dc was determined under simulated natural runoff conditions using a customized flume-based scouring system (Figure 2). The scouring experimental setup primarily consisted of a water supply reservoir, a flow stabilization tank, and a scouring flume. The scouring flume was rectangular in shape (length × width × height: 400 cm × 20 cm × 20 cm). A layer of soil was uniformly adhered to the interior of the flume using adhesive to simulate the roughness of a natural soil surface. Prior to sample collection, a pre-marked line was drawn 2 cm from the inner edge of the hollow ring cutter using a permanent marker, followed by subsequent sampling and separation procedures. The flume experiment was conducted under three slope gradients (5°, 10°, and 15°) combined with three flow rate conditions (0.24 L·s⁻¹, 0.36 L·s⁻¹, and 0.48 L·s⁻¹). The surface velocity of the water flow was measured nine times, using the fluorescent dye method (KMnO₄), and the average value was multiplied by a discount factor of 0.8 to obtain the average flow rate [14]. The slope of the flume and the flow rate were adjusted to the set values, and once the water flow was smooth, the measurement was performed [21]. Two cross-sections were established at 2 m and 1 m downstream from the flume outlet. At each cross-section, three measurement points were positioned: two at 1 cm from both sidewalls and one along the centerline. Water depth measurements were performed using calipers at each designated point, with the average value of the three measurements representing the cross-sectional water depth. The mean water depth was subsequently determined by averaging values from all cross-sections. Flow shear stress and stream power were calculated using the following equations:

$$\tau =\rho ghS$$

(1)

$$\omega =\tau \upsilon$$

(2)

where $\tau$ is the shear stress of the water flow (Pa), $\rho$ is the density of the water flow (kg·m⁻³), $g$ is the acceleration of gravity (m·s⁻²), $h$ is the water depth (m), $S$ is the slope sinusoidal (%), $\omega$ is the power of the water flow (kg·s⁻³), and $\upsilon$ is the average flow velocity (m·s⁻¹).

Following the determination of basic hydraulic parameters, timing was initiated simultaneously with the lifting of the baffle. The test was terminated when the sample was eroded to the 2 cm mark. However, considering the potential resistance of certain soils to erosion, the test was also stopped if the elapsed time reached 5 min, even if the 2 cm erosion depth was not achieved. After completing the flume experiment, we carefully removed the ring knife and collected all of the soil samples that had been washed away and transferred them to a sealed bag. Thereafter, we placed the washed soil in a 105 °C oven to dry and performed subsequent quality measurements. The soil detachment capacity was subsequently determined using the following equation [15]:

$$Dc={\displaystyle \frac{W_a{-W}_b}{tA}}$$

(3)

where Dc is the soil detachment capacity (kg·m⁻²·s⁻¹), $W_b$ is the dry weight of the soil sample before the test (kg), $W_a$ is the dry weight of the soil sample after the test (kg), $t$ is the scouring time (s), and $A$ is the cross-sectional area of the scoured soil sample (m²).

2.4. Determination of Soil Properties and Root Traits

(1)

Physical properties

Soil bulk density (BD), moisture content (SMC), and porosity were determined using the cutting ring method with the following equations [28]:

$$BD=\left(W_4{-W}_0\right)/V$$

(4)

$$SMC=\left(W_1{-W}_4\right)/\left(W_4-W_0\right)$$

(5)

$$TP=\left(W_2{-W}_4\right)/V\times 100\%$$

(6)

$$CP=\left(W_3{-W}_4\right)/V\times 100\%$$

(7)

$$NCP=TP-SCP$$

(8)

where $BD$ is the bulk density (g·cm⁻³); $SMC$ is the soil moisture content (%); $TP$ is the total porosity (%); $CP$ is the capillary porosity (%); $NCP$ is the non-capillary porosity (%); $W_0$ is the weight of the core ring (g); $W_1$ is the weight of the moist soil sample plus the core ring (g); $W_2$ is the weight of the saturated soil sample plus the core ring (g); $W_3$ is the weight of the soil sample after free drainage plus the core ring (g); $W_4$ is the weight of the oven-dried soil sample plus the core ring (g); and $V$ is the volume of the core ring (100 cm³).

The mechanical composition of the soil was analyzed with a Malvern 2000 laser particle size analyzer (Malvern Panalytical Ltd., Malvern, UK) [20]. Soil aggregates were measured using dry and wet sieving methods [23], and their stability was quantified using the mean weight diameter (MWD), calculated using the following equation:

$$MWD={\sum }_1^{n+1}{\displaystyle \frac{r_{i+1}+r_i}{2}}\times m_i$$

(9)

where $r_i$ denotes the aperture of the ith mesh sieve (mm), $m_i$ denotes the weight percentage (%) of soil aggregates on the ith mesh sieve, and n denotes the number of mesh sieves.

(2)

Soil chemical properties

Soil pH was determined using the potentiometric method, with a pH meter (Shanghai Yidian Scientific Instrument Co., Ltd., Shanghai, China) [29]. Soil organic carbon and total nitrogen were measured using an elemental analyzer (Elementar Analysensysteme GmbH, Langenfeld, Germany) [30]. Total phosphorus and available phosphorus content in the soil were determined using UV-Vis spectrophotometry (Shimadzu Corporation, Kyoto, Japan) [31]. Total potassium and available potassium were measured using atomic absorption spectrophotometry (PerkinElmer, Inc., Waltham, MA, USA) [32].

(3)

Root system characterization

Soil from each core was submerged in water. Thereafter, the roots were carefully separated and thoroughly rinsed using a sieve. The cleaned roots were first oven-dried at 105 °C for 15 min and subsequently dried at 60 °C until a constant weight was achieved. The root mass density (RMD) was calculated using the following formula [14]:

$$RMD={\displaystyle \frac{m}{V}}$$

(10)

where $RMD$ is the root mass density (g·cm⁻³), $m$ is the dry mass of the roots, and $V$ is the volume of the core ring (100 cm³).

2.5. Statistical Analysis

Statistical analyses were performed using SPSS 26.0, Origin 2023, and SmartPLS 4.0. After confirming data normality via Q-Q plots and the Shapiro–Wilk test (p > 0.05) and assessing variance homogeneity using Levene’s test (a = 0.05), one-way ANOVA was employed to compare the variability in soil physicochemical properties and soil detachment capacity (Dc) across different bamboo forest expansion stages, with post hoc test (LSD for equal variances, Tamhane’s T2 for unequal variances) selected based on homogeneity results. Pearson correlation analysis was conducted to quantify the linear relationships between Dc and both soil physicochemical properties and root system characteristics. Data visualizations, including bar plots, box plots, and correlation matrices, were generated using Origin 2023. To elucidate the complex influencing pathways of bamboo forest expansion on Dc, Partial Least Squares Structural Equation Modeling (PLS-SEM) was implemented in SmartPLS 4.0. Based on theoretical expectations, soil physicochemical properties and root system characteristics were specified as exogenous latent variables, with Dc as the endogenous latent variable [14]. A systematic assessment of the measurement model and the structural model was rigorously performed to ensure model quality. Non-significant paths (p > 0.05) were iteratively removed to achieve a parsimonious and statistically sound final structural equation model depicting the key drivers influencing Dc.

3. Results

3.1. Impact of the Expansion of Bamboo Forests to Coniferous Forests on Soil Physical Properties

The expansion of moso bamboo significantly altered soil physical properties (Figure 3). Notably, bulk density increased in the MF (1.13 ± 0.05 g·cm⁻³) compared to the CF (1.08 ± 0.03 g·cm⁻³) and BF (1.03 ± 0.02 g·cm⁻³), suggesting a trend towards intensified soil compaction during the transitional phase. Non-capillary porosity increased significantly with expansion (MF: 0.10% > BF: 0.09% > CF: 0.03%); in comparison, capillary porosity decreased. Silt content in soil mechanical composition increased with expansion (BF: 84% > MF: 79% > CF: 64%), and sand content decreased (CF: 33% > MF: 19% > BF: 13%). However, soil moisture content dropped sharply to 9.41% in the MF (CF: 23.71% ± 2.1%), and water-holding capacity decreased by 60.3% (p < 0.001).

The stability analysis of aggregates (Figure 4) revealed that during the expansion of moso bamboo into coniferous forests, the proportion of water-stable macroaggregates increased significantly. Notably, the BF exhibited the highest percentage of aggregates > 2 mm (37%), substantially exceeding those in the other two forest types (MF: 26% and CF: 22%). The proportion of small-sized soil aggregates (<0.25 mm) in the BF was exceptionally low (8%). The MF still exhibited a relatively low percentage of small-sized aggregates (9%); in comparison, the CF demonstrated a significantly higher proportion of such aggregates among the three forest types (13%). Moreover, both water-stable and non-water-stable aggregates in coniferous forests showed a marked decrease in mean weight diameter (p < 0.05).

3.2. Impact of the Expansion of Bamboo Forests to Coniferous Forests on Soil Chemical Properties and Root Characteristics

The results of the elemental composition analysis of soil during the process of moso bamboo expansion into coniferous forests exhibited some discrepancies (

Table 2. Impact of the expansion of bamboo forest to coniferous forest on soil chemical properties.

Indexes	CF	MF	BF
pH	4.17 ± 0.10 c	5.16 ± 0.08 b	5.41 ± 0.06 a
SOC (g·kg−1)	3.46 ± 0.85 a	2.75 ± 0.64 b	2.59 ± 0.47 b
N (g·kg−1)	0.17 ± 0.03 a	0.13 ± 0.02 b	0.19 ± 0.04 a
P (g·kg−1)	18.00 ± 5.24 a	14.08 ± 2.74 b	12.42 ± 1.25 b
K (g·kg−1)	22.16 ± 8.17 a	23.66 ± 8.78 a	18.42 ± 9.22 a
NH4+-N (mg·kg−1)	7.65 ± 3.56 b	8.13 ± 1.79 b	14.24 ± 6.72 a
AP (mg·kg−1)	7.91 ± 1.13 a	7.67 ± 1.31 a	8.03 ± 1.23 a
AK (mg·kg−1)	104.5 ± 8.15 a	103.6 ± 11.49 a	105.12 ± 13.66 a
C/N	19.86 ± 3.71 a	20.89 ± 3.01 a	14.22 ± 3.37 b

Note: CF is Japanese white pine forest, MF is mixed Japanese white pine and moso bamboo forest, BF is moso bamboo forest, pH is soil pH, SOC is soil organic carbon, N is soil total nitrogen, P is soil total phosphorus, K is soil total potassium, NH₄⁺-N is ammonium nitrogen, AP is soil available phosphorus, AK is soil quick-acting potassium, and C/N is soil carbon/nitrogen ratio. Lowercase letters indicate significant (p < 0.05) differences in soil chemical properties between different bamboo forest expansion stages.

). Soil pH significantly increased with moso bamboo expansion (BF: 5.41 > MF: 5.12 > CF: 4.71); in comparison, organic carbon (SOC) content decreased (CF: 3.46 g·kg⁻¹ > MF: 2.75 g·kg⁻¹ > BF: 2.59 g·kg⁻¹). Total nitrogen (N) and ammonium nitrogen (NH₄⁺−N) contents were highest in the BF (N: 0.19 g·kg⁻¹ and NH₄⁺−N: 14.24 mg·kg⁻¹). Total phosphorus (P) content was significantly higher in the CF (18.00 g·kg⁻¹) than in the MF (14.08 g·kg⁻¹) and BF (12.42 g·kg⁻¹). In contrast, total potassium (K) and available potassium (AK) exhibited no significant differences among the three forest types. The carbon-to-nitrogen ratio (C/N) was highest in the MF (20.89), exceeding that in the CF (19.86) and BF (14.22).

The root mass density of the BF (0.0184 g·cm⁻³) was significantly higher than that of the other sites, followed by the MF (0.0048 g·cm⁻³), and the lowest was noted in the CF (0.0009 g·cm⁻³; Figure 5).

3.3. Characteristics and Influencing Factors of Soil Detachment Capacity

The MF exhibited the highest Dc (0.034 kg·m⁻²·s⁻¹), significantly surpassing both the CF (0.023 kg·m⁻²·s⁻¹) and BF (0.018 kg·m⁻²·s⁻¹) (Figure 6a). Our correlation analysis results (Figure 6b) revealed the key relationships affecting Dc, with them demonstrating that Dc was significantly negatively correlated with soil moisture content (r = −0.80, p < 0.001) and significantly positively correlated with soil carbon-to-nitrogen ratio (r = 0.59, p < 0.01). The influencing factors of Dc exhibited significant correlations, with some demonstrating highly significant relationships (p < 0.01). This finding indicates the presence of collinearity among factors affecting Dc. Our PLS-SEM results (Figure 6c) revealed distinct pathways: soil physicochemical properties, particularly soil moisture content (SMC) and soil total nitrogen (N), exerted a strong direct negative effect on Dc (total effect −0.547). In contrast, root system characteristics primarily influenced Dc indirectly (indirect effect −0.339), achieved by modifying soil physicochemical properties, which subsequently exerted their direct negative effect on Dc.

4. Discussion

4.1. Effects of Soil Properties and Root Traits on Soil Detachment Capacity

The expansion of moso bamboo profoundly alters physicochemical properties and root system architecture through cascading effects triggered by root competition and litter input, with these dynamic responses exhibiting distinct stage-specific characteristics. Our results demonstrated that the MF, as an expansion transition stage, exhibits an abrupt 233% increase in soil non-capillary porosity (CF: 0.03% to MF: 0.10%) combined with a sharp 62.3% decrease in soil moisture content (CF: 23.7% to MF: 9.4%). These findings indicate that during the expansion of moso bamboo into the CF, the soil structure becomes more porous, potentially creating a critical window for soil erosion vulnerability [4]. This phenomenon may be jointly driven by decreased soil water-holding capacity and intensified tree water competition. Moso bamboo roots spread rapidly, disrupting the original soil structure and reducing water retention. Simultaneously, during mixed forest stages, rapidly growing new bamboo shoots require substantial amounts of water. Parent bamboos transfer large volumes of water to these shoots [33], thus escalating water competition with conifer roots [24]. Notably, despite the significant increase in soil non-capillary porosity in the MF stage, its capillary porosity decreased by 16.3% (CF: 0.43% to MF: 0.36%). This reduction in effective water-holding capacity further exacerbates the risk of water stress during this stage [13]. A reduction in capillary pores is often a sign of soil structure degradation, increased bulk density, and decreased total porosity. This can slow down water infiltration rates, reduce total infiltration, and prevent more water from draining away promptly; the excess water becomes surface runoff and exacerbates erosion. Concurrently, a decline in soil water-holding capacity may induce water stress, which affects growth by inhibiting net photosynthetic rate and transpiration in seedlings [34]. Furthermore, bamboo, possessing deep roots, sustains water uptake from deeper soil layers. In contrast, shallow-rooted native seedlings often perish due to insufficient surface soil moisture and disruption of beneficial rhizosphere microbial communities, such as mycorrhizal fungi.

Regarding soil chemical properties, the expansion of moso bamboo resulted in a 28.3% decrease in the soil carbon-to-nitrogen ratio (MF: 20.89 to BF: 14.22), indicating accelerated organic matter mineralization during the BF stage [35]. Mixed forests exhibited the lowest soil total nitrogen (N) content. Beyond potential plant competition, shifts in microbial community function are key: microbes with high demand for organic nitrogen mineralization acquire N sources [25], with the activity of potential free-living nitrogen-fixing functional groups (such as nematodes) being suppressed [36]. These processes collectively result in the N trough. In addition, the soil pH of the MF was significantly higher than that of the CF (4.71 to 5.41), likely associated with organic acids secreted by bamboo roots displacing H⁺ ions [24]. This pH elevation typically alters microbial community composition and activity significantly, potentially favoring bacteria over fungi [37]. This shift may partially explain the observed C/N ratio decline. Concurrently, changes in soil structure and accelerated organic matter mineralization synergistically exacerbate base cation leaching [2].

The variation in root system characteristics further elucidates the ecological mechanisms underlying expansion (Figure 5). The root mass density of the BF increased by 58.6% compared to the CF (0.19 to 0.30 kg·m⁻³). The dense, shallow horizontal root network of bamboo [23] not only enhances soil shear resistance [14] but also suppresses vertical extension of coniferous root systems through allelopathic exudates [38]. Deep-rooted conifers provide long-term deep fixation for steep slopes. Such slopes are at risk of deep-seated sliding. In contrast, moso bamboo has a powerful horizontal root system, making it able to provide long-term shallow stability reinforcement for shallow soil slopes. It can also provide stability for slopes affected by rainfall-induced shallow landslides. However, the heterogeneous distribution of root systems during the MF stage (characterized by the coexistence of horizontal moso bamboo roots and vertical Japanese white pine roots) may trigger a “root interference effect”, leading to localized fragmentation of soil structure [22]. This finding explains why the stability of soil aggregates in the MF remains lower than that in the BF despite a 24.5% (1.14 to 1.42 mm) increase in the mean weight diameter (MWD) of soil aggregates compared to the CF.

4.2. Effects of the Invasion of Bamboo Forests on Soil Detachment Capacity

The results of this study demonstrate that the Dc of the MF (0.034 kg·m⁻²·s⁻¹) was significantly higher than that of the CF (0.023 kg·m⁻²·s⁻¹) and BF (0.018 kg·m⁻²·s⁻¹), a phenomenon that may be driven by a synergistic multifactorial effect [39]. First, the dramatic changes in soil physicochemical properties during the MF stage serve as the primary driver: the elevated soil bulk density coupled with increased non-capillary porosity collectively enhances soil aeration. However, this process reduces capillary water-holding capacity, leading to soil desiccation and weakened interparticle cohesion, thereby rendering the soil more susceptible to erosion by surface runoff [10]. Secondly, the MF stage exhibited a significantly higher carbon-to-nitrogen ratio (20.89%) compared to the moso bamboo forest (14.22%), accompanied by weaker soil ammonification [26]. This finding indicates a slower organic matter decomposition rate and insufficient humus accumulation, which collectively undermine aggregate stability and further exacerbated soil detachment. In addition, there are notable differences in root system characteristics between the BF and MF: the MF exhibits lower root mass density (intermediate between CF and BF), with an underdeveloped root network that provides weaker soil stabilization. In contrast, the dense root system of the BF significantly enhances erosion resistance through physical entanglement [3] and the secretion of organic matter [19,40].

Our correlation analysis results demonstrated that soil detachment capacity was positively correlated with bulk weight and carbon-to-nitrogen ratio and negatively correlated with pH, soil moisture content, total nitrogen, and root mass density. This result is consistent with the results of existing studies: high-bulk-weight soils are prone to form surface crusts due to their compact structure [41]; however, they are more likely to disintegrate under the action of water flow shear. Under low pH conditions (acidic soils), positively charged iron–aluminum oxides combine with negatively charged clay particles and organic matter to form stable aggregates [42]. However, as pH increases following moso bamboo expansion, electrostatic adsorption weakens, resulting in an elevated soil-detachment risk. In addition, increased root mass density directly suppresses Dc by promoting aggregate formation [43] and enhancing root–soil cohesion [19], while also indirectly influencing the detachment process through the modulation of soil porosity and moisture content.

Notably, the lowest Dc in the BF may be related to the combined effect of its high root mass density (significantly higher than that of the other stands) and non-capillary porosity: the dense root network enhances the soil shear strength, whereas the moderate pore structure balances water permeability and water-holding capacity, and it reduces the erosion potential of runoff. The high root mass density and aggregate stability in the BF reduced Dc. In contrast, MF, as a transitional stage, has not yet reached a steady state of soil structure and root distribution, leading to a peak in its Dc. The vulnerability of mixed forests may be due to a combination of factors, including structural instability caused by the transition period, low soil moisture content, and a root network that is not yet fully developed and is highly heterogeneous.

In this study, the driving mechanism was further elucidated through PLS-SEM: soil physicochemical properties (soil moisture content and soil total nitrogen) had a dominant, direct effect (−0.547) on Dc; in comparison, root properties had an indirect effect (−0.339) by modulating soil properties. This finding substantiates the “vegetation–soil–erosion” cascade theory, demonstrating that bamboo forest expansion reshapes the dynamics of soil anti-erodibility through modifications in stand structure and subterranean ecological processes. The authors of future studies should integrate long-term positioning observations to quantify the interactive effects between root morphological characteristics (e.g., root diameter distribution and spatial configuration) and soil microbial communities on detachment capacity across different expansion stages, thereby improving the ecohydrological modeling for bamboo forest expansion.

4.3. Limitations and Significance of This Study

This study delivers critical insights for soil erosion risk assessment in subtropical bamboo expansion zones. During mixed forest transition stages, it is recommended to prioritize the implementation of soil and water conservation measures such as vegetative mulching and root soil stabilization measures (including planting deep-rooted Chrysopogon zizanioides). More significantly, we reveal a universal pattern whereby plant invasions induce predictable ecosystem vulnerability peaks. This factor is exemplified by Eichhornia crassipes—a globally invasive aquatic species—which exhibits maximal biotic suppression during initial colonization [44], mirroring the erosion susceptibility peak identified in our bamboo system. Such stage-locked disruption maxima transcend ecosystem boundaries, providing a mechanistic framework to forecast critical vulnerability windows in habitats invaded by aggressive clonal plants. Notably, our results revealed a pronounced Dc peak during the MF stage, though its persistence during other stages remains undetermined. Despite successfully quantifying the stage-specific impacts of Phyllostachys edulis encroachment on Dc using a space-for-time substitution approach, this study is inherently constrained by single-timepoint sampling, which limits the assessment of intra-stage temporal dynamics. In addition, microbial community structure, biodiversity indices, and enzyme activities—key mediators of soil aggregation—were not measured. Integrating these biological indicators with physicochemical datasets will aid in elucidating microbial–erosion linkages across invasion stages. To address these limitations, the authors of future studies should employ long-term monitoring or process-based ecosystem modeling to determine whether this transitional peak represents a transient state or sustained phenomenon in stable mixed forests. Specifically, effective long-term monitoring requires establishing multi-site, multi-year fixed observation networks in key areas (e.g., representing different stages of moso bamboo expansion) to capture long-term dynamics, combined with the integration of technologies like remote sensing and GIS for efficient, large-scale, high-frequency monitoring of key indicators. Furthermore, building cross-institutional data-sharing and collaborative platforms is crucial to ensure the continuity and standardization of such monitoring efforts. Building upon this monitoring foundation, mechanistic modeling integrating multi-scenario simulations and sensitivity analyses will be required in order to quantify parameter contributions to stability thresholds and identify context-dependent critical values. Implementing these comprehensive long-term monitoring and modeling strategies will significantly advance our understanding of moso bamboo expansion dynamics, thereby providing a solid scientific foundation for developing more proactive and adaptive sustainable management measures. Despite these limitations, the significance of this work lies in revealing the dynamics and mechanisms of Dc changes during bamboo expansion and identifying a distinct Dc peak during the mixed forest transition phase. This study and its findings advance our comprehension of bamboo forest expansion’s ecological impact and provide a scientific basis for moso bamboo forest management, soil erosion risk assessment, and optimization of soil and water conservation strategies.

5. Conclusions

In this study, we systematically investigated three forest stands representing distinct stages of moso bamboo expansion into coniferous forests: CF, MF, and BF. Through comprehensive analysis of soil physicochemical properties, root system characteristics, and dynamic changes in Dc, we elucidated the mechanisms underlying the impact of moso bamboo expansion on soil erosion resistance. The principal conclusions are as follows:

(1)

Moso bamboo expansion significantly altered soil structure and nutrient stoichiometric characteristics: The MF stage exhibited the highest soil bulk density (1.13 g·cm⁻³), with non-capillary porosity increasing markedly with bamboo expansion (reaching 0.10% in MF). The mean weight diameter (MWD) of soil aggregates demonstrated a 12.1% enhancement in the MF compared to the CF, though the MF still demonstrated lower structural stability than the BF. Soil nutrients exhibited stage-dependent variations during bamboo expansion: Total phosphorus (P) content decreased during the MF stage (14.08 g·kg⁻¹); in comparison ammonium nitrogen (NH₄⁺-N) content increased significantly in the BF stage (14.24 mg·kg⁻¹). The carbon-to-nitrogen ratio (C/N) exhibited a 28.3% reduction during the transition from the MF to BF. Furthermore, root mass density (RMD) demonstrated spatial differentiation, with the BF displaying the highest values (3.12 kg·m⁻³), significantly exceeding those in the MF (2.05 kg·m⁻³) and CF (1.78 kg·m⁻³) during the expansion process.

(2)

During the expansion of moso bamboo into this CF, Dc demonstrated a significant nonlinear response, reaching its peak (0.034 kg·m⁻²·s⁻¹) during the MF stage. This value was notably higher than that observed in the CF (0.023 kg·m⁻²·s⁻¹) and BF (0.018 kg·m⁻²·s⁻¹).

(3)

Our PLS-SEM analysis results confirmed that soil moisture content and soil total nitrogen exert primary direct effects on Dc. In comparison, root characteristics exerted indirect effects on Dc by modulating soil properties.

The results of this study will contribute to understanding the impact of moso bamboo expansion and the drivers of soil detachment capacity, which are important considerations for effective bamboo forest management and soil erosion control.