Soil Infiltration Characteristics and Driving Mechanisms of Three Typical Forest Types in Southern Subtropical China

Abstract

Plant roots and soil properties play crucial roles in regulating soil hydrological processes, particularly in determining soil water infiltration capacity. However, the infiltration patterns and underlying mechanisms across different forest types in subtropical regions remain poorly understood. In this study, we measured the infiltration characteristics of three typical stands (pure Phyllostachys edulis forest, mixed Phyllostachys edulis-Cunninghamia lanceolata forest, and pure Cunninghamia lanceolata forest) using a double-ring infiltrometer. Stepwise multiple regression and structural equation modeling (SEM) were employed to analyze the effects of root traits and soil physicochemical properties on soil infiltration capacity. The results revealed the following: (1) The initial infiltration rate (IIR), stable infiltration rate (SIR), and average infiltration rate (AIR) followed the order pure Phyllostachys edulis stand > mixed stand > pure Cunninghamia lanceolata stand. (2) Compared to the pure Cunninghamia lanceolata stand, the IIR, SIR, and AIR in the pure Phyllostachys edulis stand increased by 6.66%, 35.63%, and 28.51%, respectively, while those in the mixed stand increased by 28.79%, 28.82%, and 33.51%. (3) Fine root biomass, root length density, non-capillary porosity, and soil bulk density were identified as key factors influencing soil infiltration capacity. (4) Root biomass and root length density affected infiltration capacity through both direct pathways and indirect pathways mediated by alterations in non-capillary porosity and soil bulk density. These findings provide theoretical insights into soil responses to forest types and inform sustainable water–soil management practices in Phyllostachys edulis plantations.

1. Introduction

Soil water infiltration, one of the most critical processes governing water exchange between soil and the external environment, serves as a pivotal component in forest hydrological cycles [1]. Infiltration capacity not only directly reflects soil water conservation potential but also indirectly influences the utilization efficiency of soil water resources [2] and regulates the growth dynamics of forest communities [3]. Significant variations in root traits and soil physicochemical properties across different forest types may profoundly alter infiltration characteristics and subsequent hydrological processes [4,5,6,7]. Therefore, investigating soil water infiltration patterns and their driving mechanisms under distinct stand types is essential for evaluating regional hydrological cycling.

Soil physicochemical properties are key determinants of infiltration capacity [8,9]. During forest succession, improved soil properties enhance infiltration, whereas compacted soils with high bulk density hinder water exchange [10]. Non-capillary porosity [11]—including macropores formed by soil fauna (e.g., earthworms), decayed root channels, and cracks from shrink–swell clays [12]—and root–soil interactions mediated by soil organic matter critically regulate infiltration and water movement [6,13]. Soil texture, defined by the proportional composition of sand, silt, and clay particles, fundamentally controls soil respiration and pedogenesis, with clay content exerting the strongest inhibitory effect on infiltration [14].

Plant roots further modulate infiltration through dual mechanisms [15,16]. First, roots directly or indirectly reshape soil pore architecture (size, distribution, and connectivity), where coarse roots promote macropore formation while fine roots enhance pore spatial heterogeneity [17]. Second, root exudates alter soil organic carbon, aggregate stability, and microbial activity, thereby modifying soil permeability [18]. Variations in root morphological traits—including biomass, length density, surface area, and volume density—drive spatial heterogeneity in infiltration capacity across plant species [19]. However, debates persist regarding the role of root diameter: fine roots (<1 mm) may enhance infiltration by increasing microporosity [20], whereas both fine (<1 mm) and coarse roots (>2 mm) could suppress permeability through pore occlusion [21]. Current studies predominantly focus on isolated effects of root traits or soil properties on infiltration, leaving coupled mechanisms underexplored—a critical knowledge gap limiting the holistic understanding of infiltration processes.

In this study, we hypothesize that distinct root–soil interactions across forest types drive divergent infiltration patterns. Our objectives are to (1) quantify variations in root traits and soil physicochemical properties among three subtropical stands (pure Phyllostachys edulis forest, mixed Phyllostachys edulis-Cunninghamia lanceolata forest, and pure Cunninghamia lanceolata forest); (2) characterize their soil water infiltration dynamics; and (3) elucidate synergistic mechanisms by which roots and soil properties jointly regulate infiltration capacity.

2. Materials and Methods

2.1. Study Area

The study was conducted in the Soil and Water Conservation Demonstration Park (30°36′ N, 119°35′ E) located in Anji County, Huzhou City, Zhejiang Province, China (Figure 1). This region experiences a typical subtropical maritime monsoon climate characterized by four distinct seasons, with approximately 70% of the annual precipitation (1400 mm) concentrated between May and August. The mean annual temperature is 16.6 °C. The area is recognized as the “Bamboo Town of China” and the birthplace of the ecological development concept “lucid waters and lush mountains are invaluable assets”. The basic soil characteristics of the study area are summarized in

Table 1. Basic information of the plot.

Stand	Altitude	Slope (°)	Aspect of Slope	Plant Density	Mean Tree Height (m)	Mean DBH (cm)	Canopy Density
PPF	73	10°	West	2100	16	12	0.75
MPF	75	10°	South-west	1350	12	10	0.85
PCF	76	10°	South-west	925	10	10	0.6

. The soils are classified as Yellow-Brown Earth (Hapludalfs according to USDA Taxonomy), corresponding to Acrisols in the FAO classification system, with a solum thickness ranging from 40 to 60 cm. The understory is densely covered with shrubs and herbs, dominated by three vegetation types: pure Cunninghamia lanceolata forest, pure Phyllostachys edulis forest, and mixed Cunninghamia lanceolata-Phyllostachys edulis forest.

2.2. Experimental Design

2.2.1. Plot Establishment

Three stand types were investigated: (1) mixed Phyllostachys edulis−Cunninghamia lanceolata forest (MPF), (2) pure Phyllostachys edulis (moso bamboo) forest (PPF), and (3) pure Cunninghamia lanceolata (Chinese fir) forest (PCF). Stand composition criteria were strictly defined:

PPF: >90% bamboo coverage with minimal fir presence;

MPF: Approximately 1:1 bamboo–fir mixture ratio [22];

PCF: <10% bamboo content with dominant fir coverage.

Three typical forest stand types with similar site conditions—pure Phyllostachys edulis (moso bamboo) forest (PPF), mixed Phyllostachys edulis–Cunninghamia lanceolata forest (MPF), and pure Cunninghamia lanceolata (Chinese fir) forest (PCF)—were selected as study objects (see Table 1 for stand characteristics). Within each stand type, two 20 m × 20 m standard plots were randomly established. Five 1 m × 1 m soil infiltration measurement points were systematically positioned at each plot using a stratified random sampling design.

2.2.2. Soil Infiltration Measurements

Field measurements were conducted in May 2024. Prior to formal experiments, gravimetric soil moisture content was determined through soil sampling across forest types, with all measurement points showing no significant differences (p > 0.05) in moisture levels (

Table 2. Soil physical and chemical properties (mean ± SD) of different stand and at a depth of 0–40 cm.

Soil Depth (cm)	Forest Types	W	BD (g∙(100 cm−3))	TP (%)	CP (%)	NCP (%)	SOC (g∙kg−1)	pH	TN (g∙kg−1)	Sandy (%)	Silty (%)	Clay (%)
0–10	PPF	0.23 ± 0.01 a	1.11 ± 0.08 a	49.11 ± 3.47 a	38.64 ± 2.25 a	10.47 ± 2.02 a	29.02 ± 4.05 a	4.91 ± 0.37 a	2.46 ± 0.28 a	69 ± 4.21 a	19.94 ± 2.8 a	11.07 ± 1.85 ab
	MPF	0.22 ± 0.01 a	1.19 ± 0.03 a	42.82 ± 4.71 b	21.32 ± 14.25b	8.72 ± 0.99 ab	22.54 ± 2.03 a	4.44 ± 0.46 ab	1.71 ± 0.31 a	65.03 ± 6.43 a	21.29 ± 6.18 a	13.69 ± 2.86 a
	PCF	0.22 ± 0.01 a	1.2 ± 0.05 a	48.12 ± 3.26 ab	40.78 ± 4.05 a	7.34 ± 0.83 b	28.22 ± 15.21 a	4.28 ± 0.1 b	1.64 ± 0.78 a	69.66 ± 2.69 a	20.47 ± 2.82 a	9.88 ± 0.83 b
10–20	PPF	0.21 ± 0.02 a	1.28 ± 0.06 b	42.43 ± 2.15 a	35 ± 1.25 a	7.43 ± 1.03 a	17.7 ± 3.4 a	4.78 ± 0.1 a	1.32 ± 0.32 a	62.81 ± 11.5 a	23.31 ± 11.8 a	13.88 ± 2.08 a
	MPF	0.2 ± 0.01 a	1.41 ± 0.06 b	44.04 ± 5.54 a	37.66 ± 5.63b	6.38 ± 0.09 b	13.77 ± 3.35 a	4.75 ± 0.18 a	0.91 ± 0.21 a	64.73 ± 2.55 a	23.34 ± 2.15b	11.92 ± 3.54 b
	PCF	0.22 ± 0.01 a	1.46 ± 0.02 a	44.79 ± 2.09 a	38.89 ± 2.03 a	5.9 ± 0.06 c	12.79 ± 0.6 b	4.96 ± 0.23 a	0.84 ± 0.04 b	62.57 ± 3 b	26.41 ± 4.64 ab	11.03 ± 2.04 b
20–30	PPF	0.2 ± 0.02 a	1.27 ± 0.01 b	39.89 ± 4.1 a	31.87 ± 2.97 a	5.82 ± 1.87 a	11.19 ± 1.32 a	5.17 ± 0.27 a	0.9 ± 0.25 a	56.91 ± 8.27 a	31.52 ± 6.78 a	11.57 ± 3.44 a
	MPF	0.19 ± 0 a	1.33 ± 0.02 b	39.63 ± 2.84 a	27.95 ± 11.2 b	7.27 ± 0.8 b	11.08 ± 1.76 a	5.2 ± 0.18 b	0.77 ± 0.19 b	60.81 ± 6.69 b	26.25 ± 8.54 b	12.94 ± 2.12 b
	PCF	0.21 ± 0 a	1.47 ± 0.1 a	44.19 ± 3.27 a	37.52 ± 3.3 ab	6.66 ± 0.11 b	8.42 ± 4.51 a	5.2 ± 0.1 b	0.62 ± 0.21 b	59.62 ± 5.02 b	26.47 ± 5.42 b	13.92 ± 0.83 b
30–40	PPF	0.19 ± 0.02 a	1.37 ± 0.03 a	42.14 ± 3.29 a	37.59 ± 3.36 a	4.56 ± 0.7 a	7.35 ± 1.55 a	5.43 ± 0.34 a	0.53 ± 0.09 a	40.89 ± 7.81 a	49.3 ± 3.52 a	9.81 ± 4.83 a
	MPF	0.18 ± 0.01b	1.39 ± 0.02 b	36.46 ± 2.71 a	31.03 ± 2.42 ab	5.43 ± 0.43 b	6.68 ± 1.12 b	5.24 ± 0.3 a	0.48 ± 0.07 a	44.05 ± 10.68 ab	40.37 ± 10.56 b	15.58 ± 1.2 b
	PCF	0.2 ± 0.01 a	1.45 ± 0.01 a	39.69 ± 2.92 b	34.43 ± 2.83 b	5.26 ± 0.48 c	7.65 ± 0.74 ab	5.49 ± 0.18 a	0.57 ± 0.07 a	50.09 ± 1.06 b	38.04 ± 2 b	11.87 ± 2.1 b

Notes: W, natural water content; BD, soil bulk density; TP, soil total porosity; CP, soil capillary porosity; NCP, soil non capillary porosity; SOC, soil organic carbon; TN, soil nitrogen content; Sandy, soil sand content; Silty, soil silt content; clay, soil clay content—the same as below. Different lowercase letters indicate significant differences among the forest types within the same soil depth at a significant level of p < 0.05.

). At selected sites, surface debris including litter and extraneous materials was removed. A double-ring infiltrometer (inner ring: 30 cm diameter; outer ring: 60 cm diameter) [23] was vertically installed 3 cm below the soil surface following standard protocols. To establish vertical unidirectional flow, 4 g L⁻¹ Brilliant Blue FCF solution (inner ring) and deionized water (outer ring) were simultaneously maintained at 4 cm constant head levels, effectively minimizing lateral seepage artifacts [24]. The time required for each 0.5 cm hydraulic head decline in the inner ring was recorded. When the head dropped to 1 cm, the solution was replenished to the 4 cm baseline. The test continued until achieving steady-state infiltration, defined by (a) cumulative infiltration depth ≥ 4 cm and (b) invariant time intervals between consecutive 0.5 cm head drops during at least three successive measurements.

2.2.3. Soil Sampling and Root Analysis

Soil properties including initial water content (W), bulk density (BD), total porosity (TCP), and non-capillary porosity (NCP) were determined using the core cutter method. Soil particle size distribution was analyzed following the Chinese national standard (GB/T 19077-2016).

A vertical soil profile was excavated along the diameter of the inner infiltration ring. Undisturbed soil samples were collected from six depth layers (0–10 cm, 10–20 cm, 20–30 cm, 30–40 cm, 40–50 cm, and 50–60 cm) at an adjacent profile unaffected by infiltration. Each layer includes cutting ring samples (100 cm³ in volume) and bulk soil samples (intact soil of 3534.3 cm³ in volume).

Roots were manually separated and trimmed from bulk soil samples and categorized into three diameter classes: 0–1 mm, 1–2 mm, and >2 mm. Subsequently, the morphological parameters (length, surface area, and volume) of roots from different diameter classes were quantified using the WinRHIZO Pro root scanning system. Following scanning, all root samples were air-dried at ambient temperature and then weighed to determine biomass.

2.3. Calculation Methods

(1)

Root Morphological Parameters

Root morphological indices included total root length (RL, cm), root surface area (RSA, cm²), and root volume (RV, cm³). Three density-based parameters were calculated as follows:

$$R_{LD}={\displaystyle \frac{R_L}{V}}$$

(1)

$$R_{SAD}={\displaystyle \frac{R_{SA}}{V}}$$

(2)

$$R_{VD}={\displaystyle \frac{R_V}{V}}$$

(3)

Here, $R_{LD}$, $R_{SAD}$ and $R_{VD}$ represent root length density (cm per 100 cm³), root surface area density (cm² per 100 cm³), and root volume density (cm³ per 100 cm³), respectively; $R_L$, $R_{SA}$, and $R_V$ denote total root length (cm), root surface area (cm²), and root volume (cm³) within the soil volume; and $V$ is the total semi-cylindrical soil volume (3534.3 cm³ in this study).

(2)

Soil Infiltration Rate Calculation

The cumulative infiltration amount (CIA, mm) was defined as the total water depth infiltrated before reaching steady-state conditions. The soil infiltration rate was calculated as follows [25]:

$$i={\displaystyle \frac{∆I}{∆t}}\times 600$$

(4)

Here, $i$ is the infiltration rate (mm h⁻¹); $∆I$ is the water head decline (cm) in the inner ring during time interval $∆t$(minute); and 600 is the dimensionless conversion factor.

2.4. Data Processing and Analysis

Data were statistically analyzed using Excel 2016 (Microsoft Corp., Redmond, WA, USA) and SPSS 18.0 (IBM Corp., Armonk, NY, USA). One-way analysis of variance (ANOVA) and Duncan’s multiple comparison test (α = 0.05) were performed to assess differences among forest types. Pearson correlation analysis was conducted to examine relationships between root morphological characteristics and soil physicochemical properties. Principal component analysis was employed to identify key factors influencing soil infiltration capacity. Structural equation modeling (SEM) was implemented in R Studio 4.3.1 using the “lavaan” package (v0.6-16) to quantify direct and indirect effects of root traits and soil properties on infiltration, with model visualization generated by the “semPlot” package (v1.1.6). All data presented in figures and tables are expressed as means ± standard deviation (n = 5 measurement points per stand type).

3. Results

3.1. Soil Infiltration Characteristics Across Three Stand Types

Soil water infiltration characteristics of the three stand types are presented in Figure 2. The mean initial infiltration rate (IIR) ranged from 947.7 to 1220.5 mm·h⁻¹, with the MPF showing significantly higher values than both pure stands (PPF, PCF) (Figure 2a).

For the stable infiltration rate (SIR), mean values varied between 272.0 and 369.0 mm·h⁻¹, following the order PPF > MPF > PCF (Figure 2b). The average infiltration rate (AIR) exhibited a range of 373.1–498.2 mm·h⁻¹, where both PPF and MPF demonstrated significantly higher rates than PCF (p < 0.05; Figure 2c). Cumulative infiltration amounts (CIAs) differed significantly among stands (p < 0.05), with means of 386–423 mm ranked as PPF > MPF > PCF (Figure 2d).

3.2. Soil Physicochemical Properties and Root Morphological Traits Across Stand Types

3.2.1. Soil Physicochemical Characteristics

As shown in Table 2, soil bulk density (BD), pH, clay, and silt contents generally increased with soil depth across all stands, while other properties decreased progressively.

The PPF exhibited significantly higher non-capillary porosity (NCP) and total porosity (TCP) in the 0–20 cm layer compared to other stands (p < 0.05). PPF also showed elevated soil total nitrogen (TN) and organic carbon (SOC) concentrations at 10–30 cm depths. Conversely, PPF demonstrated lower BD values in the 10–40 cm layer and reduced clay content throughout 0–40 cm profiles (p < 0.05). Sand and silt contents in PPF were marginally but significantly lower (p < 0.05) below 10 cm depth, while natural water content (W) showed minimal variation among stands.

3.2.2. Root Traits Across Stand Types

Fine roots (<1 mm) dominated all soil layers across stands (Figure 3. For roots of 0–1 mm and 1–2 mm diameters, four morphological parameters (RBD, RLD, RSAD, and RVD) consistently followed the order pure Phyllostachys edulis stand (PPF) > mixed stand (MPF) > pure Cunninghamia lanceolata stand (PCF) throughout the profile (Figure 3a–h). In contrast, roots > 2 mm exhibited a distinct pattern: MPF > PPF > PCF (Figure 3i–l).

Significant inter-stand differences were primarily observed in fine-to-medium roots (<2 mm). Within the 0–20 cm layer, PPF demonstrated higher RBD, RLD, RSAD, and RVD of <2 mm roots compared to PCF, with MPF showing intermediate values (p < 0.05, Figure 3a–h). Below 20 cm depth, these parameters showed no statistical divergence among stands (p > 0.05). For coarse roots (>2 mm), morphological parameters across stand types exhibited the order MPF > PPF > PCF, though without statistical significance (p > 0.05, Figure 3i–l).

3.2.3. Relationships Between Soil Properties and Root Traits

Pearson correlation analysis demonstrated depth-specific relationships between soil properties and root traits (

Table 3. Correlation between soil physicochemical properties and root traits.

Soil Depth (cm)	Root Traits	Soil Physical Index					Soil Chemical Index			Soil Texture
		W	BD (g∙(100 cm−3))	TP (%)	CP (%)	NCP (%)	SOC (g∙kg−1)	pH	TN (g∙kg−1)	Sandy (%)	Silty (%)	Clay (%)
0–10 cm	RBD (<2 mm)	0.249	−0.713 **	0.072	0.104	0.626 *	0.028	0.568 *	0.398	−0.135	0.109	0.087
	RLD (<2 mm)	0.196	−0.841 **	0.185	0.284	0.671 **	0.074	0.601 *	0.392	−0.083	0.101	−0.003
	RSAD (<2 mm)	0.144	−0.734 **	0.197	0.219	0.734 **	0.065	0.468	0.37	−0.144	0.134	0.064
	RVD (<2 mm)	0.178	−0.818 **	0.002	0.189	0.616 *	−0.259	0.413	0.016	0.065	0.019	−0.162
	RBD (>2 mm)	0.101	−0.067	−0.275	−0.017	0.204	−0.117	−0.049	−0.037	−0.022	−0.178	0.339
	RLD (>2 mm)	0.116	−0.358	−0.297	−0.19	0.475	−0.161	0.114	−0.003	−0.083	−0.08	0.299
	RSAD (>2 mm)	0.115	−0.182	−0.261	−0.023	0.296	−0.122	0.003	−0.026	−0.029	−0.147	0.301
	RVD (>2 mm)	0.108	−0.04	−0.247	0.044	0.149	−0.104	−0.053	−0.032	−0.008	−0.182	0.318
10–20 cm	RBD (<2 mm)	−0.209	−0.610 *	0.11	0.068	0.075	0.656 **	0.052	0.547 *	−0.329	0.288	0.247
	RLD (<2 mm)	0.035	−0.563 *	0.136	0.173	−0.191	0.646 **	0.168	0.556 *	−0.457	0.412	0.309
	RSAD (<2 mm)	−0.22	−0.496	0.082	0.04	0.091	0.650 **	−0.088	0.5	−0.174	0.093	0.305
	RVD (<2 mm)	−0.099	−0.136	0.527 *	0.517 *	−0.242	0.248	0.301	0.22	−0.336	0.296	0.248
	RBD (>2 mm)	−0.301	−0.283	−0.232	−0.404	0.669 **	0.611 *	−0.203	0.3	0.29	−0.263	−0.193
	RLD (>2 mm)	−0.084	−0.05	−0.124	−0.272	0.536 *	0.527 *	−0.207	0.131	0.25	−0.2	−0.242
	RSAD (>2 mm)	−0.259	−0.364	−0.101	−0.237	0.486	0.653 **	−0.129	0.385	0.072	−0.067	−0.045
	RVD (>2 mm)	−0.246	−0.279	−0.123	−0.269	0.530 *	0.642 **	−0.156	0.335	0.154	−0.14	−0.101
20–30 cm	RBD (<2 mm)	−0.273	−0.075	−0.113	−0.234	0.222	0.127	−0.019	0.123	0.509	−0.455	0.205
	RLD (<2 mm)	−0.052	−0.158	0.395	0.131	0.318	0.229	−0.218	0.344	0.345	−0.236	−0.067
	RSAD (<2 mm)	−0.146	−0.197	−0.136	−0.034	−0.09	0.107	−0.188	0.145	0.156	−0.145	0.077
	RVD (<2 mm)	0.08	−0.351	0.194	0.338	−0.306	0.262	−0.408	0.442	0.035	0.113	−0.391
	RBD (>2 mm)	−0.244	0.067	0.004	−0.184	0.439	−0.368	0.441	−0.394	0.495	−0.577 *	0.578 *
	RLD (>2 mm)	−0.307	−0.078	−0.073	−0.204	0.267	−0.058	0.135	−0.046	0.484	−0.479	0.327
	RSAD (>2 mm)	−0.3	−0.01	−0.05	−0.208	0.356	−0.209	0.295	−0.22	0.505	−0.544 *	0.464
	RVD (>2 mm)	−0.23	0.092	0.022	−0.151	0.426	−0.373	0.454	−0.411	0.482	−0.571 *	0.589 *

Notes: * t-p < 0.05; ** t-p < 0.01.

). In the 0–10 cm layer, fine root (<2 mm) morphological parameters—biomass density (RBD), root length density (RLD), surface area density (RSAD), and volume density (RVD)—showed significant correlations (p < 0.05) with soil bulk density (BD) and non-capillary porosity (NCP), while coarse root (>2 mm) traits exhibited no significant associations with soil properties. Within the 10–20 cm layer, BD and NCP were significantly correlated with RBD and RLD of fine roots (<2 mm) (p < 0.05), whereas NCP displayed significant positive correlations with RBD, RLD, and RVD of coarse roots (>2 mm) (p < 0.05). Notably, soil organic carbon (SOC) consistently correlated with all root metrics across both fine and coarse roots (p < 0.05).

3.3. Analysis of Influencing Factors on Soil Infiltration Rates

3.3.1. Principal Component Analysis

The principal component analysis (PCA) of soil properties yielded a loading matrix as presented in

Table 4. Principal component analysis of soil properties and root traits of the forest types.

Principal Component	PC-1	PC-2	PC-3	PC-4	PC-5
Eigenvalue	9.254	4.764	2.764	1.72	1.602
Variance	40.235	20.714	12.019	7.477	6.967
Cumulative variance	40.235	60.949	72.968	80.446	87.412
Eigenvectors
W	−0.22	0.168	0.511	−0.42	0.047
BD	−0.818	−0.109	0.341	−0.194	0.013
TP	−0.042	0.843	0.093	0.371	−0.13
CP	−0.065	0.858	0.324	0.172	0.032
NCP	0.688	−0.434	−0.121	0.489	−0.177
SOC	0.269	0.276	0.04	0.494	0.741
pH	0.422	0.675	−0.101	−0.134	−0.363
TN	0.572	0.518	−0.302	0.21	0.479
Sand	−0.271	−0.75	0.206	0.496	−0.202
Silt	0.121	0.819	−0.008	−0.465	0.072
Clay	0.41	0.08	−0.494	−0.221	0.345
RBD1	0.955	0.14	0.159	−0.015	−0.058
RLD1	0.904	0.28	0.232	0.06	−0.083
RSAD1	0.898	0.07	0.301	0.189	−0.039
RVD1	0.656	0.241	0.37	0.076	−0.461
RBD2	0.759	−0.455	0.332	−0.2	0.151
RLD2	0.783	−0.376	0.449	0.014	0.11
RSAD2	0.823	−0.331	0.367	−0.198	0.134
RVD2	0.702	−0.326	0.431	−0.299	0.257
IIR	0.256	−0.481	−0.646	−0.192	0.018
SIR	0.869	0.057	−0.378	−0.015	−0.15
AIR	0.761	−0.234	−0.499	−0.17	−0.071
CIA	0.705	0.189	−0.348	−0.043	−0.325

Notes: The SIR, IIR, and AIR represent infiltration rates in steady, initial, and average infiltration, respectively; CIA represents cumulative infiltration amounts.; RBD1 represent fine root (<2 mm) biomass density; RLD1 represents fine root (<2 mm) length density; RSAD1 represent fine root (<2 mm) surface area density; RVD1 represent fine root (<2 mm) volume density; RBD2 represent coarse root (>2 mm) biomass density; RLD2 represents coarse root (>2 mm) length density; RSAD2 represent coarse root (>2 mm) surface area density; RVD2 represent coarse root (>2 mm) volume density.

. Five principal components (PCs) with eigenvalues > 1 collectively accounted for 87.412% of the cumulative variance. In PC-1, high-loading parameters included BD and NCP, root traits of fine roots (RBD, RLD, and RSAD), and soil infiltration parameters (SIR, AIR, and CIA). BD exhibited negative correlations with all these factors, while positive correlations were observed among other parameters. PC-2 was predominantly characterized by soil TP, CP, silt content, and NCP, along with the IIR. Notably, The IIR showed a positive correlation with NCP but negative correlations with silt, TP, and CP. Similarly, PCs 3–5 primarily featured high loadings from soil clay content, SOC, W, and NCP, coupled with infiltration parameters IIR and AIR. Infiltration rates demonstrated positive correlations with these factors except for W. Collectively, the five PCs explained 87.412% of the total variance, revealing systematic relationships: soil BD, TP, CP, W, and silt content were negatively correlated with infiltration capacity, whereas root traits (RBD and RLD) and soil clay, NCP, and SOC showed positive correlations with infiltration.

3.3.2. Structural Equation Model of the Infiltration Process

Considering the combined influence of soil properties and root traits on infiltration, we performed a path analysis of the infiltration process. The optimal structural equation model and its path coefficients are presented in Figure 4.

For the IIR, soil clay content and NCP exerted significant positive effects, with standardized path coefficients of 0.76 and 0.42, respectively. RBD and RLD showed no direct effects, yielding total effects of 0.362 and −0.327, respectively (Figure 5). Regarding the SIR, only soil NCP had a significant positive impact (standardized coefficient = 0.60). RBD again had no direct effect, whereas RLD exhibited a direct effect of 0.60; the total effects of RBD and RLD on the SIR were 0.13 and 0.168, respectively (Figure 5). For the AIR, BD negatively influenced infiltration (standardized coefficient = −0.59), while clay content was positively related (standardized coefficient = 0.39). RBD showed a direct positive effect of 0.45, and RLD had no direct effect; the resulting total effects were 0.577 for RBD and 0.698 for clay content (Figure 5). Finally, for the CIA, soil BD again had a negative effect (−0.29) and NCP a positive effect (0.45). Both RBD and RLD contributed direct effects of 0.18 and 0.25, respectively (Figure 5).

4. Discussion

4.1. Effects of Forest Types on Root Traits and Soil Properties

Our study revealed substantial differences in root characteristics (0–30 cm depth) among the three stand types, particularly in the PPF stand, which exhibited significantly higher fine root (<2 mm) RBD and RLD in the 0–20 cm layer compared to MPF and PCF stands. Root biomass serves as an indicator of root quantity, and root length density reflects the spatial distribution and entanglement of roots in the soil [3]. Because moso bamboo typically invades other stands through vigorous rhizome expansion [26], stands with a higher proportion of moso bamboo likely have a more abundant root system in terms of both number and spatial distribution. Research has reported that during the expansion of moso bamboo into evergreen broadleaf forests, it employs a high fine root turnover strategy, with fine root biomass and length being highest in PPF, intermediate in MPF, and lowest in evergreen broadleaf forests [27]. Fine roots (<2 mm) showed greater stand-type sensitivity than coarse roots (>2 mm), indicating stronger stand-type dependency of fine roots. These fine roots likely contribute substantially to SOC accumulation and aggregate formation via root exudates and microbial interactions [28,29]. For instance, Cai et al. [30] reported 7.03- and 1.57-fold increases in fine root biomass and RLD during bamboo expansion into fir forests, aligning with our findings on PPF’s root modification capacity.

The PPF stand demonstrated superior surface soil (0–20 cm) hydraulic properties, with 34.2% higher NCP and 8.4% lower BD than PCF (Figure 2), consistent with Chen et al.’s findings on soil BD and NCP changes during bamboo invasion into artificial oak forests [31]. The reduced BD in PPF surface soils can be attributed to the vigorous expansion of bamboo rhizomes, which enhance soil porosity and pore connectivity, creating larger and more numerous soil voids [32]. This mechanism aligns with our observation that PPF surface soils had significantly higher NCP than MPF and PCF (p < 0.05).

Regarding soil nutrients, PPF surface soils showed higher SOC and TN content, following the trend PPF > MPF > PCF. This contrasts with Bai et al.’s findings [33], potentially due to moso bamboo’s ability to increase surface root biomass, whose exudates stimulate microbial activity and SOC accumulation. In terms of soil texture, PPF surface soils had significantly higher clay content than PCF (p < 0.05), though clay remained less abundant than sand and silt fractions, consistent with Xie et al.’s observations in southern moso bamboo forests [34]. However, clay content decreased sharply in deeper PPF soil layers, differing significantly (p < 0.05) from the gradual vertical changes in MPF and PCF. This may result from the strong penetrating force of bamboo rhizomes, which disperse soil particles. Additionally, PPF’s understory species richness significantly increases [35], and the abundant herbaceous vegetation combined with bamboo’s rapid regeneration enhance litter input, thereby boosting surface carbon and nitrogen accumulation. Alternatively, rhizome turnover may replenish soil carbon and nitrogen pools, elevating total SOC and TN content [36].

Notably, fine root traits—including RBD and RLD—exhibit strong associations with SOC, BD, and NCP. The influence of root systems on soil texture and physicochemical properties has been extensively documented [37,38]. Specifically, the denser distribution of fine roots enhances soil aggregation efficiency [39]. During root growth, physical displacement of soil particles occurs, directly shaping the spatial configuration of NCP [40]. Mechanistically, fine roots reduce soil BD by increasing total porosity, as their lower density occupies soil voids and decreases overall soil compactness. Furthermore, root decomposition contributes substantial organic carbon inputs to soils [41], while the remnants of decayed roots create persistent macropores [42]. Ecologically, the development of fine root length correlates with stand site quality. The spatial distribution and morphological status of plant fine roots critically determine water/nutrient uptake efficiency, thereby exerting cascading effects on topsoil physicochemical characteristics [43].

4.2. Mechanisms of Soil Infiltration Regulation by Root–Soil Interactions

The decisive influence of root traits on soil infiltration has been widely acknowledged [44,45]. Root systems create preferential flow channels within the soil matrix, significantly enhancing water transport capacity [46]. Our results revealed significant correlations between fine root characteristics (<2 mm) and infiltration performance, consistent with Hao et al.’s findings that fine root biomass (<1 mm) enhances soil infiltration [20], but contrasting with Lu et al.’s conclusion that both fine (<1 mm) and coarse roots (>2 mm) inhibit infiltration. This discrepancy suggests that moso bamboo roots do not obstruct pore networks during water percolation. Specifically, fine root (<2 mm) RBD and RLD exerted significant positive effects on infiltration rates at all stages except the initial phase, diverging from Wang et al.’s observations in loess hilly-gully regions where coarse roots dominated infiltration enhancement [47]. In moso bamboo systems, coarse roots are subordinate in the root network, while fine roots substantially increase pore spatial heterogeneity and modify soil properties [18].

Although Zhang et al. found root surface area to be critical for SIR improvement in vegetation-restored loess soils [48], our study emphasized the importance of RBD and RLD. This distinction likely arises from the horizontally distributed, intricate fine root networks of moso bamboo, which enhance macropore formation and lateral pore connectivity—key drivers of preferential lateral flow [49].

Soil infiltration is a process transitioning from unsaturated to saturated conditions. In the unsaturated state, soil pores contain both water and air (where pores are not completely water-filled), and water movement is governed by both capillary and gravitational forces. The IIR represents average infiltration during the early stage of water entry and is a critical determinant of runoff generation at the onset of rainfall. As the soil progressively wets from dry to saturated conditions, soil moisture content and macroporosity emerge as key influencing factors [50,51]. In this study, the natural moisture content across sampling sites was consistent, with no significant differences (p < 0.05). Principal component analysis (PCA) confirmed the negative influence of soil moisture on the IIR, reinforcing its critical role. Notably, MPF exhibited the highest IIR, along with the lowest capillary porosity and the highest clay content (Table 2). Clay content had a stronger influence, suggesting it primarily drives the enhanced IIR in MPF. This aligns with previous findings that clay promotes the formation of large soil aggregates [52], enhancing their stability and thereby increasing infiltration rates [52,53]. Clay’s explanatory power for infiltration variation was significantly higher than that of sandy soils [54]. NCP was another key factor, serving as the primary pathway for rapid gravitational infiltration during early rainfall by overcoming capillary adsorption forces. The AIR reflects the overall infiltration process, including both initial infiltration and the transition to saturation. Results indicate that BD and clay content are its dominant controls. During soil saturation, compacted soils (high BD) impede water exchange [10]. The SIR represents the late-stage infiltration rate when the soil is fully saturated (pores are completely water-filled, as in groundwater layers). At this stage, water movement is primarily gravity-driven, occurring mainly through macropores, making NCP the decisive factor—almost the sole pathway for post-saturation percolation [14]. Furthermore, the CIA was predominantly governed by NCP, which played a pivotal role throughout the measured infiltration stages. NCP directly connects surface and subsurface soil layers, a well-established mechanism in infiltration dynamics [55]. Soil infiltration is interactively regulated by root characteristics and soil properties [56]. A substantial body of research attributes infiltration modification to root-mediated alterations in soil structure [57,58]. Our principal component analysis revealed that root traits (RBD and RLD), soil texture (particularly clay content), bulk density, and porosity exert critical control on infiltration processes. Structural equation modeling revealed that soil clay content and NCP directly enhanced infiltration, while BD directly suppressed it; RBD and RLD affected infiltration through both direct pathways and indirect mediation of soil properties, consistent with Cui et al.’s findings [59]. This aligns with reported mechanisms whereby RBD increases NCP in surface soils to enhance infiltration [60].

In summary, both PPF and MPF develop through the continuous expansion of their robust root systems. This represents an ongoing process of increasing both root quantity and quality, which consequently modifies the physical properties of surface soil, improves soil infiltration capacity, and ultimately contributes to more stable community development.

5. Conclusions

This study investigated soil infiltration characteristics and regulatory mechanisms in three typical subtropical forest stands. The results demonstrated the following: (1) Compared to PCF, PPF increased the IIR, SIR, and AIR by 6.66%, 35.63%, and 28.51%, respectively, whereas MPF enhanced these rates by 28.79%, 28.82%, and 33.51%. (2) Root–soil systems differed significantly among stand types. PPF and MPF developed higher fine root (<2 mm) RBD and RLD, along with greater soil porosity and lower bulk density than PCF. (3) Fine root (<2 mm) RBD and RLD regulated infiltration capacity through dual pathways: direct hydraulic effects and indirect modulation of soil properties (non-capillary porosity and bulk density). Structural equation modeling identified four dominant factors: fine root trait (RBD, RLD), clay content, non-capillary porosity, and bulk density.