Distribution Pattern and Structure of Vascular Plant Communities in Riparian Areas and Their Response to Soil Factors: A Case Study of Baoan Lake, Hubei Province, China

Abstract

The vascular plant community in a riparian area is the main substrate and vehicle of many ecological functions for the lakeshores of grass-type shallow lakes. However, there is still a lack of knowledge regarding the responses of vascular plants to soil factors of the habitat in riparian areas, which restricts the ecological adaptation management for riparian vegetation. In this work, a typical grass-type shallow lake (Baoan Lake) in the Yangtze Basin in Central China was taken as the study area. We describe the plant species distribution and community structure in riparian areas under two habitat types (lake and tributary) and their responses to soil factors. The results showed that (1) the soil chemical factors have a significant effect on the distribution and community structure of vascular plants, even though there was a significant interaction among three group factors of soil habitats; (2) compared with other factors, the total nitrogen (TN) and available phosphorus (AP) have the most significant correlations with the distribution of vascular plants; (3) the rate of soil nutrient sorption determines the distribution of vascular species, closely related to the biological characteristics of plants and the microbial enzymatic activity in soil; and (4) vascular plant diversity and the proportion of perennial plants were generally higher in the lakeshore areas than in the tributaries and showed a low-high-low “hump-shaped” species richness and diversity distribution. The Shannon-Wiener index value increased with the increasing soil-available phosphorus in the surface soil layer. Therefore, this study advanced our knowledge of the species distribution and diversity patterns of lakeshores and tributaries, providing scientific and theoretical guidance for the biodiversity conservation and sustainable ecosystem management of grass-type shallow lakes.

1. Introduction

Natural riparian zones are the most diverse, dynamic, and complex biophysical habitats in the terrestrial regions of Earth [1,2]. Riparian regions across the world are critical biodiversity hotspots due to their endemism, rich biodiversity, and significant spatial and temporal variations compared with other regions [3]. Moreover, vascular plants in the riparian ecosystem play a vital role in the maintenance of multiple ecosystem functions, including the control of riverbank erosion, thermic regulation of the riverbank and the associated water body, and filtration and retention of nutrients and maintenance of water quality [4,5]. However, under the background of global change, the dual interferences of climatic and anthropogenic factors make the soil environment on which vascular plants depend less stable, seriously affecting ecosystems’ structure and function.

Previous studies have demonstrated that the distribution and composition of plants in riparian areas are controlled by factors such as the water flow, soil, topography, groundwater fluctuations, and landscape type. Fu et al. demonstrated the differentiation characteristics of the plant community and their relationships with ecological factors [6]. They revealed that the spatial differentiation characteristics in the typical steppe were chiefly driven by precipitation. In contrast, the influencing factor in the meadow steppe was soil nutrients, followed by temperature and precipitation. Li et al. also investigated the effects of sediment types on the distribution and diversity of plant communities in the Poyang Lake wetlands [7]. Among the different environmental factors, as reported by the reviews of Ledesma et al. [8], Henriques et al. [9], Fagundes [10], and Ding et al. [11], soil conditions and topography underlie habitat heterogeneity, climate, and biology. They control the ecological process of the ecosystem and play a key role in community distribution and characteristics, determining the distribution type, quantity, and quality of riparian vegetation [12]. Overall, the study of environmental factors affecting plant distribution in the riparian area, including details about diversity, has improved. Despite this progress, specific studies on the response of plants to soil factors in riparian areas need to be supplemented and developed. Although some reports mention the effects of soil conditions, in-depth research has yet to be conducted [13]. Furthermore, differences in the influence of soil factors on species composition and community structure between tributaries and lakeshore habitats in the same water system are poorly understood. The lake’s water system is a complex system consisting of tributaries and the lake water body, with hydrological connectivity and material exchange. Soil conditions in the riparian area are the result of hydrology and geomorphology, which can comprehensively reflect the influence of habitat on plant distribution and characteristics. The marked spatial heterogeneity of the soils throughout the area of Baoan Lake makes it an ideal place for field studies of the ecology of riparian communities in different habitats. The influence of habitat factors on plant communities has been the central issue in lake riparian ecosystems. Hence, understanding these issues can represent essential theoretical solutions for lake conservation.

Hubei Province is known as the ‘province of a thousand lakes’ in China. Baoan lake is a substantial water body in the Yangtze River Basin. Grass-type shallow lakes have great potential for fishery resources and are rich in aquatic plant resources worldwide. In this study, a total of 25 sampling sites were selected surrounding the lake. We investigated the effects of soil factors on the species distribution and community structure of vascular plants in lakeshore and tributary riparian areas. We tested the following hypotheses: the plant distribution and diversity will be influenced by soil and topographic factors, with the importance of the influencing factors that may differ in lakeshore and tributary habitats; soil chemistry properties have the greatest impact on plant distribution and species richness in riparian areas; the distribution of particular species in riparian areas may show a dependence or preference for some influencing factors; species diversity will be significantly higher in lakeshore areas than in tributary areas and is consistent with the Intermediate Disturbance Hypothesis. More specifically, the objectives of this study were: (1) to determine the most important soil properties that affect plant distribution and diversity patterns, (2) to survey the distribution pattern of vascular plants and analyze the dominant factors of spatial variation of vascular plants, (3) to identify the impact of topographic and soil factors on the variation of species diversity within the riparian area, and (4) to compare differences in species diversity between lakeshores and tributary riparian areas. This study provides theoretical guidance for the conservation of vascular plants and the ecological restoration of the riparian area in Baoan Lake.

2. Materials and Methods

2.1. Study Area

Baoan Lake (E 114°39′—119°49′, N 30°12′—30°18′) is a typical grass-type shallow lake located in the southeast of Hubei Province, China, which belongs to the Liangzi Lake water system (Figure 1a). Located in the alluvial plain of the middle reaches of the Yangtze River, Baoan Lake has great hydrological connectivity, linking the entire water system. The lake has a surface area of 39.3 km² with a mean depth of 2.27 m and a Secchi depth of 0.5 m. Situated in a subtropical monsoon climate region, Baoan Lake has good heat and moisture conditions: the average annual temperature is 16.8 ℃ and the average annual precipitation is 1283 mm. The water of the lake system flows from south to north, flowing out of the lake from Donggougang Gate (northwest side) into Changgang and then discharging into the Yangtze River. The topography of the lake shore is flat in the west, whereas in the east, there are some low hills or hollows, with decreasing elevation from southeast to northwest. The main land-use types include fishponds, agricultural lands, forests, and construction land. Farming methods are mainly conventional human or cattle farming.

2.2. Vegetation Investigation

Field investigations and sampling were conducted in August 2020. These transects were chosen nonrandomly, mainly based on site accessibility, were isolated from each other and covered dominated plant communities. Based on empirical values for the minimum area of riparian plant communities [14], we determined the length of the study transect to be 60 m. At each of the 25 transects, starting near the water and moving away at 20 m intervals, three zones were selected to capture plant changes within the riparian area (Figure 1d). In each zone, a place of 10 m×10 m was homogeneous in terms of vegetation structure was identified. A 1 m × 1 m representative herbaceous quadrat with 3 replicates within the 10 m × 10 m was randomly selected for plant sampling by avoiding the low-lying, precipitous slopes and areas heavily influenced by human activities, and 225 quadrats in total were obtained. Each quadrat was investigated for species name, numbers, height, and cover of all vascular plants. All vascular plants in herbaceous quadrats (1 m²) were mowed and brought back to the laboratory to measure their biomass (Figure 1c). We converted all species synonyms into accepted names, and all intraspecies records were merged to the species level. To minimize the interference of cultivated species with the sampling results, we removed all the cultivation records (Figure 1b). According to the water level data obtained at Donggougang hydrometric station, the mean tide level (MTL) in the study area was 18 m, the multi-year average low water level (MALWL) was 15.8 m, and the multi-year average high water level (MAHWL) was 18.9 m (Figure 1d).

2.3. Investigation and Measurement of Soil Factors

The main habitat factors obtained for analysis were the topography as well as physical and chemical soil characteristics. During the field sampling, each sample site’s latitude, longitude, and elevation were measured with a handheld GPS and geological compass. Regarding the topographic information, the horizontal distance from the plant quadrat to the water edge was measured with a measuring tape (100 m) as the distance from the water. The average slope of each transect was measured with a CI slope meter purchased from Haglof; the lake depth at the edge of the riparian area was measured with a steel ruler.

Soil samples were collected at the locations of each plant quadrat; three undisturbed samples (using the cutting ring method) and one disturbed sample were collected for replication from the surface soil (0–20 cm) and deep soil (20–40 cm), respectively. We measured the soil saturated hydraulic conductivity (Ks), soil bulk density (SBD), soil initial moisture content (θ₀), soil saturated moisture content (θ_S), particle size distribution, soil pH, available nitrogen (AN), total nitrogen (TN), available phosphorus (AP), and total phosphorus (TP). Mixed soil samples were ground and sieved in the laboratory to 2 mm, and all roots and visible plant remains were removed. Ks was determined using the constant-head method [15]. SBD, θ₀, and θ_S were measured using the oven-drying volumetric ring method after samples were oven-dried at 105 °C for 24 h to a constant mass [16]. The particle size distribution of the samples was determined using laser diffraction (Bettersize2000; Shengke Instrument, Dandong, China). The pH was determined with a calibrated combination pH electrode based on a mixture of soil and ultrapure water at a mass ratio of 1:2.5. The AN was determined using an alkaline hydrolysis–diffusion method [17]. The soil was hydrolyzed with 1 molL⁻¹ NaOH to convert the readily hydrolyzable nitrogen to NH₃, which was absorbed by H₃BO₃ after diffusion. The NH₃ in the H₃BO₃ uptake solution was then titrated with a standard acid, from which the total alkaline nitrogen content of the soil was calculated. TN was determined using the Kjeldahl acid digestion method using a H₂SO₄ mixed catalyst [16]. The AP was extracted from soil samples with 0.03 M NH₄F + 0.025 M HCl [16]. The TP content was determined using the molybdenum blue colorimetry method after digesting the samples with perchloric acid [16].

2.4. Data Analysis

We detected the biomass and then used it to calculate each species’ importance value (IV) for identifying characteristic species and comparing the community differences among plots. Plant samples (aboveground biomass and roots) were collected and placed in envelopes for further drying at 65 °C for 48 h to a constant mass [18]. The importance value is a measure proposed by Curtis and Mcintosh (1950) to indicate the dominance of a species in a community.

The importance value was calculated using the following formulae [19]:

For the trees:

$$I{V}_{tree}=\left(relative density+relative dominance+relative frequency\right)∕3$$

(1)

For shrubs or herbaceous plants:

$$I{V}_{shrubs and herbs}=\left(relative height+relative cover\right)∕2$$

(2)

Riparian plant community diversity was evaluated using the Species Richness Index (R), Shannon-Wiener index (H′), Pielou Evenness index (E), and Simpson Index (D) [20].

Species Richness Index (R):

$$R=S$$

(3)

Shannon-Wiener Index (H′):

$$H^{\prime }=-{\displaystyle \sum }_{i=1}^s{P}_i\ln P_i$$

(4)

Pielou Evenness Index (E):

$$E=\frac{H^{\prime }}{\ln \left(S\right)}$$

(5)

Simpson Index (D):

$$D=1-{\displaystyle \sum }_{i=1}^s{P}_i{}^2$$

(6)

where S is the number of species recorded in quadrats, and P_i is the relative abundance of each species.

As basic presumptions for the parametric tests, we used the Shapiro–Wilk test and the Levene’s test to assess for the normality of the data and homogeneity (or equality) of variances, respectively [21,22]. Statistical analyses of the data included normality tests, correlation analysis, and variance analysis. The Hill numbers of the riparian plant community were calculated using Excel 2019; one-way ANOVA was conducted on each soil factor for both lake and tributary plant communities using SPSS (version 26.0, IBM SPSS Statistics, Chicago, IL, USA). The mantel test was utilized to analyze the relationships between soil variables and plant communities in different habitat types.

The maximum gradient lengths of the first four axes of ordination were calculated using the “decorana” function in R and found to be less than 3.0, indicating that most species exhibit a linear response to potential soil changes. The linear model was determined to be superior; thus, redundancy analysis (RDA) was selected to analyze the effect of soil factors on the differentiation characteristics of the plant community, and the statistical significance of each explanatory variable was validated using the Monte Carlo method with 999 permutations [23]. We used hierarchical partitioning to estimate the importance of each explanatory variable and group of variables (topography and physical and chemical properties). This technique calculates the importance of the variables from all subset models, resulting in more accurate results through an unordered importance assessment [24]. In order to analyze the effects of different habitats and zonation zones on preponderant families and lifeforms of vascular plants, we used a generalized linear mixed model (GLMM), where transect was treated as a random effect, and habitat types and different zones were the fixed variable. We used the R Environment (version 3.3.2) with the lme4 package for all statistical analyses.

The variance inflation factors (VIF) of soil variables were calculated. The VIF values were all less than 10, and there was no collinearity. All the analyses were performed with the R packages “vegan” and “rdacca.hp” [24]. Pearson correlation was used to assess the associations between different soil factors, and the results were visualized using the R package “corrplot” [25].

3. Results

3.1. Soil Habitat Characteristics of Lakeshore and Tributary

We used a hierarchical partitioning to estimate the importance of each explanatory variable and group of variables (chemical, physical, and topographic). The importance of the joint effect of physical and chemical factor groups of soil was relatively obvious, reaching 7.79%. The importance of the joint effect by all three groups together reached 14.43% (Figure 2c). The soil chemistry factors had the highest unique explanatory power, contributing 58.94% of the total explanation. TP, AN, and pH in tributary riparian areas showed stronger potentially important shared effects in explaining variation in spatial plant distribution than in lakeshores. However, a similar trend was observed in both habitat types: soil chemistry was the most important factor influencing plant community composition, followed by the physical factors and, finally, topographic factors (Figure 2). In lakeshore habitats, the variable of highest importance was TN (14.52%), followed by AP (13.56%) and TP (12.84%). Correspondingly, in tributary habitats, the variable of highest importance was TN (21.26%), followed by pH (16.6%) and TP (10.30%).

Most soil chemical properties showed different levels between the two habitats (

Table A2. Mean and ranges of soil factors for lakeshore and tributary riparian areas.

Parameters	Layer	Lakeshore		Tributary
		Mean (SE)	Min–Max	Mean (SE)	Min–Max
Ks(cm min−1)	0–20 cm	1.94 ± 0.36	0–6.38	1.10 ± 0.25	0–5.13
	20–40 cm	1.25 ± 0.31	0–5.89	0.74 ± 0.12	0–1.96
SBD (g cm−3)	0–20 cm	1.35 ± 0.03	1.14–1.60	1.40 ± 0.02	1.20–1.57
	20–40 cm	1.39 ± 0.02	1.22–1.60	1.45 ± 0.01	1.28–1.54
θ0 (g g−1)	0–20 cm	0.29 ± 0.01	0.15–0.41	0.26 ± 0.01	0.21–0.34
	20–40 cm	0.30 ± 0.02	0.15–0.45	0.25 ± 0.01	0.18–0.35
θS (g g−1)	0–20 cm	0.33 ± 0.01	0.23–0.43	0.31 ± 0.01	0.24–0.38
	20–40 cm	0.33 ± 0.01	0.24–0.45	0.30 ± 0.01	0.24–0.37
Cclay	0–20 cm	13.9 ± 0.43	9.49–17.03	13.28 ± 0.43	10.12–19.51
	0–20 cm	14.25 ± 0.26	11.18–15.61	13.62 ± 0.33	10.37–17.01
Csilt	0–20 cm	66.36 ± 0.39	61.66–68.77	66.20 ± 0.85	56.30–72.62
	20–40 cm	67.46 ± 0.39	63.30–70.42	67.37 ± 0.63	59.30–74.33
Csand	20–40 cm	19.27 ± 0.78	15.78–28.85	20.49 ± 1.14	7.87–33.16
	20–40 cm	18.39 ± 0.62	14.56–25.52	19.00 ± 0.75	13.04–29.27
pH	0–20 cm	6.47 ± 0.15	5.51–7.63	6.95 ± 0.09	5.51–7.78
	20–40 cm	6.62 ± 0.11	5.58–7.76	6.94 ± 0.09	5.51–7.78
AN (mg L−1)	0–20 cm	113.97 ± 2.94	90.81–135.81	114.31 ± 2.39	81.88–135.28
	20–40 cm	62.32 ± 0.94	56.35–72.90	61.84 ± 0.87	55.01–69.91
AP (mg L−1)	0–20 cm	14.22 ± 0.97	8.32–21.97	14.19 ± 1.82	1.16–35.42
	20–40 cm	8.86 ± 0.73	6.03–24.70	5.58 ± 0.35	0.58–8.37
TP (mg L−1)	0–20 cm	0.46 ± 0.01	0.36–0.56	0.57 ± 0.04	0.35–1.25
	20–40 cm	0.22 ± 0.01	0.13–0.38	0.37 ± 0.03	0.21–0.75
TN (mg L−1)	0–20 cm	1.57 ± 0.07	1.13–2.22	1.05 ± 0.03	0.75–1.33
Distance (m)	/	2.29 ± 0.54	−0.50–11.00	2.71 ± 0.58	−0.70–13.50
Slope (◦)	/	0.45 ± 0.05	0.01–0.80	0.45 ± 0.08	0.01–1.50

). In the lakeshore’s soil, the TN (lakeshore: 1.57 ± 0.07 mg/L, tributary: 1.05 ± 0.03 mg/L) and AP (lakeshore: 8.86 ± 0.73 mg/L, tributary: 5.58 ± 0.35 mg/L) obtained the higher values, exhibiting a high soil nutrient content. Another difference was that the range of slope variation in the tributaries was higher, with the slope of the riparian areas in the tributary ranging from 0.01 to 1.5, averaged at 0.45. Moreover, K_S was significantly lower in the tributaries than in the lakeshore area. Soil physical and chemical properties, including TN and K_S, exhibited significant and positive correlations (Spearman’s test (r > 0.4, p < 0.001)). A saturated moisture content (θ_s) was strongly correlated with AN in the surface layer (r = 0.68, p < 0.001), and AN was negatively correlated with SBD (r = −0.76, p < 0.001). (Figure 3).

3.2. Distribution and Composition of Vascular Plant Communities

In total, 87 species of plants in the study area of Baoan Lake were identified, belonging to 24 orders and 36 families and 73 genera. Only one species of gymnosperms was Pinus elliottii, and the rest were angiosperms. The first two axes together accounted for 44.76% and 53.79% of the plant distribution in the riparian areas of lakeshores and tributaries, respectively. Blue arrows indicate selected habitat factors, and red points indicate species (Figure 4). The Monte Carlo permutation test indicated that the eigenvalues for all canonical axes were significant (p < 0.01). As shown in Figure 4a, the first RDA axis was significantly related to the TN, TP, and soil bulk density; the TN of the soil exhibited a negative correlation, and the TP of the soil had a positive correlation. This indicated that the first RDA axis mainly reflected changes in soil nutrients. The second RDA axis was significantly correlated with pH and soil sand content, which indicates that the second RDA axis mainly reflects changes in soil acidity and mechanical composition. Accordingly, the distribution of vascular plants in tributaries was shown in Figure 4b. The first RDA axis was significantly related to the TN, TP, and AN, thus mainly reflecting changes in soil nutrients. The second axis reflected the physical properties of the riparian area and was significantly correlated with soil particle size composition. In terms of species distribution, the two types of the riparian area had some similarities. The perennial gramineous plants represented by Paspalum distichum, Cynodon dactylon, and Heteropogon contortus exhibited a strong positive correlation with TN of the soil. Humulus scandens and Phaenosperma globose exhibited a strong positive correlation with the pH of the soil. Eichhornia crassipes and Setaria viridis exhibited a strong positive correlation with physical properties such as the C_sand of the soil. In addition, some species showed habitat-type selectivities, such as Bidens pilosa and Daucus carota. Daucus carota showed a positive correlation between its distribution and TP, although only in lakeshore areas.

3.3. Structural Characteristics of the Vascular Plant Community

We characterized the structure of vascular plant communities in terms of the proportions of species (belonging to different dominant families and lifeforms) and alpha diversity indices under different environmental conditions.

As shown in Figure 5, the Shannon-Wiener and Species Richness index values in lakeshores showed a humped distribution pattern from zone A to zone C, i.e., the highest value occurred in zone B of moderate disturbance. In contrast, zone B of the tributaries had lower average Shannon-Wiener and Species Richness index values than the same zones (zone B) in the lakeshore (

Table 1. The test of significance (mean ± SD) of Shannon diversity difference in different zones of two types of riparian habitat.

Zone	Lakeshore		Tributary		t	Sig.(2-Tailed)
	Mean (SD)	Min–Max	Mean (SD)	Min–Max
A	1.11 ± 0.40	0.29–1.74	1.44 ± 0.35	0.74–1.90	−2.178	0.040
B	1.59 ± 0.34	1.14–2.18	1.14 ± 0.48	0.44–1.80	2.524	0.022
C	1.11 ± 0.55	0.56–1.85	1.23 ± 0.55	0.54–1.89	−0.385	0.708

). There was little difference in the Shannon-Wiener index and Simpson index values across the three zones in tributary habitats. The Pielou evenness index was observed to decrease with the increase in the distance from the water, while the Species Richness index showed an increasing trend from zone A to zone C.

Different diversity indices showed different correlations with the soil and topographic factors, such as the Shannon-Wiener index and Simpson index, which were significantly positively related to the available p (p < 0.05) and TP in the surface soil layer, but the Species Richness index was significant negatively correlated to the slope (p < 0.05) (

Table 2. Correlation analysis of plant community structure indicators with major soil factors.

Factors	Diversity Indices				Community Species Characteristics
	R	H′	E	D	Biomass	Tree	Density	Perennial	Annual
Csand.sl	−0.19	−0.17	−0.13	−0.21	0.12	0.48 **	−0.05	0.01	−0.09
Csand.dl	0.18	0.17	0.06	0.21	0.31	0.39	−0.03	−0.13	0.15
SBD	0.08	−0.04	−0.07	0.23	0.23	0.04	0.23	−0.01	0.46 *
Form	0.13	−0.21	−0.31	−0.23	0.37 *	−0.15	0.14	0.01	0.25
Slope	−0.48 *	0.01	0.21	0.07	−0.44 *	0.54 *	−0.08	−0.17	0.13
AN.sl	−0.25	0.11	0.26	0.10	−0.31	0.27	−0.34	−0.22	−0.29
AN.dl	−0.03	0.12	0.1	0.12	0.08	−0.2	−0.11	0.05	−0.28
AP.sl	−0.17	0.37 *	0.45 **	0.38 *	−0.38	0.15	−0.39	−0.23	−0.35
AP.dl	−0.3	0.15	0.38	0.19	−0.36	−0.14	−0.27	−0.02	−0.45 *
TP.sl	−0.07	0.35 *	0.33	0.35 *	−0.29	0.2	−0.37	−0.32	−0.21
TP.dl	−0.15	0.26	0.34	0.32	−0.33	0.34	−0.31	−0.42 *	0.06
TN.sl	−0.15	0.22	0.32	0.25	−0.24	−0.18	−0.21	0.08	−0.48 *

*, significant at the 0.05 level (bilateral); **, significant at the 0.01 level (bilateral); R, Species Richness index; H′, Shannon-Wiener index; E, Pielou Evenness index; D, Simpson Index; SBD, soil bulk density; AP, available phosphorus; TN, total nitrogen; TP, total phosphorus; C_sand, sand content. Variables ending with “.sl” and “.dl” are variables representing the surface layer soil variation (0–20 cm) and deep layer soil variation (20–40 cm), respectively.

). It can be seen from Table 2 that the number of woody plants showed a significantly positively related to the soil sand content and the slope (p < 0.05). In contrast, the proportion of Annual herbs were significantly positively related to SBD (p < 0.05). Among the five community species characteristics, except the density of distribution of vascular plants, the other four characteristics all showed significant correlation to some soil or topographic factors.

The GLMM (

Table 3. Effect of different habitats and zonation zones on preponderant families and lifeforms of vascular plants assessed by the generalized linear mixed model.

Variables	Habitats Difference		Zoning Difference
	Lakeshore (Reference: Tributary)		Zone A (Reference: Zone B)		Zone C (Reference: Zone B)		Zone A (Reference: Zone C)
	CE (%)	p	CE (%)	p	CE (%)	p	CE (%)	p
Preponderant families of vascular plant species
Asteraceae	0.205	0.962	−10.925 *	0.003	0.854	0.849	−11.779 *	0.010
Poaceae	7.767	0.258	−28.206 *	0.001	−8.751	0.233	28.614 *	0.007
Amaranthaceae	−5.409	0.498	11.445	0.176	−17.169	0.104	28.614 *	0.007
Apiaceae	0.498	0.684	−2.906 *	0.035	−2.681	0.114	−0.224	0.890
Polygonaceae	16.168 *	0.009	12.877	0.058	0.499	0.952	12.378	0.128
Lifeforms of vascular plant species
Woody plants	0.364	0.244	−1.797 *	0.001	1.348 *	0.001	−3.146 *	0.001
Annual herbaceous	−2.739	0.760	−12.458 *	0.049	−0.265	0.973	−12.193	0.123
Perennial herbaceous	29.472 *	0.015	5.493	0.655	−15.551	0.321	21.044	0.173

CE, Contrast estimate, with positive and negative values indicating increases and decreases in the proportion of species under different conditions, respectively; * indicates p < 0.05.

) indicated that the proportions of perennial herbaceous and species belong to Polygonaceae in the community were significantly correlated with habitat type (p < 0.05). The distribution of many dominant species showed a zonation change (from zone A to zone C), reflecting variations in community structure at the transect level. Species belonging to Amaranthaceae and Polygonaceae had a higher proportion of distribution in zone A of the lakeshore (p < 0.05), with representative species such as Alternanthera philoxeroides and Polygonum hydropiper showing better submergence tolerance. In contrast, the Gramineae species had a higher proportion of distribution in zone B (p < 0.05), with species such as Setaria viridis and Paspalum distichum. The species belonging to Asteraceae were significantly more abundant in zone B and zone C than in zone A (p < 0.05), with the main species such as Polygonum perfoliatum and Lactuca sibirica.

Lifeforms reflect different species’ life cycle and ability to utilize resources. The GLMM showed a significantly higher proportion of annual herbaceous in zone B than in zone A (CE:12.46%, p < 0.05). Perennial herbs decreased from zone A to zone C. However, the proportion of woody plants significantly increased from zone A to zone C (p < 0.05). Zones B and C had the most affluent compositions of vegetation lifeforms. The communities developed a spatially stratified structure of tree–grass or tree–shrub–grass.

4. Discussion

4.1. The Importance of Habitat Factors

It has been proven that riparian areas are intensive places of water–land material and energy exchanges, with complex and diverse habitat variations [8,26]. In general, the most critical soil habitat factors were TN and TP. This result tied well with previous studies by Liu et al., who studied the factors driving the relationships between vegetation and soil properties in the Yellow River Delta. They demonstrated that NO₃⁻, soil organic matter, and TP were strongly related to vegetation properties at the subregion scale [27], indicating that soil factors significantly affected plant growth at medium scale (0–400 km²). Building on this, we went further and found that the critical soil factors affecting the distribution of vascular plants were different in the lakeshore and tributaries of Baoan Lake. In tributary riparian areas, pH also played an important role in the distribution of plants [28]. Soil pH was considered the fundamental variable among soil properties for vegetation growth [29]. We believed this difference was inextricably linked to the terrain and human disturbance. The discharge of domestic water and stronger surface runoff leaching impacted soil pH levels for plants in tributary riparian areas [30,31]. Some of the tributary sections were found to be surrounded by villages, which were more vulnerable than the lake water body to unreasonable human fertilizer management practices and domestic pollution, such as the indiscriminate discharge of crop and livestock farming wastewater. For example, the application of excessive organic fertilizer tends to increase soil pH [30].

In the lakeshore riparian area, specifically due to the regular inundation–emergence of water levels, the shared effect of soil nutrients and the physical properties on plant distribution was significant (Proportion: 22.22%, p < 0.05). Existing studies have confirmed the interaction between physical characteristics and chemical factors. Adhikari et al. studied the influence of topography on soil nutrient changes in a silvopasture system. They found that topography habitat factors can significantly impact the retention and transfer of soil nutrients [32]. In addition, previous studies have shown that soil particle fractions can influence soil nutrient contents. Similarly, Jiao et al. studied the variations in soil nutrients and particle sizes under different vegetation types in the Yellow River Delta and found that the distribution and transformation of soil nutrients can affect the soil texture [33]. Therefore, a combination of topography and soil physical and chemical habitat factors can better explain the composition characteristics of plant communities in riparian areas than focusing on only one of these groups of factors.

4.2. Distribution Pattern of Vascular Plant Communities in Response to Soil Factors

It has long been recognized that the relationship between plant communities and soil characteristics is a key issue for the ecological restoration of riparian areas. The distribution of many characteristic species showed a particular preference or dependence on certain environmental factors. We found that Heteropogon contortus, Cynodon dactylon, and Paspalum distichum exhibited high positive correlations with TN. Plant species have different efficiencies in the uptake and use of soil nitrogen, which enables them to adapt and survive in habitats with different soil nutrient levels. Thus, soil nutrients exert a filtering effect on plant distribution by affecting plants with different botanical characteristics [34]. Bai et al. found that the N supply, which has strong effects on Gramineae plant functional traits, is an important factor affecting the growth of marsh plant species [35]. The increase in TN in the soil supply increased Gramineae plant height, leaf area, and aboveground biomass. The aboveground biomass was significantly promoted by the N supply [36]. Moreover, previous studies have reported that microbial communities and enzyme activity are themselves influenced by soil nutrient levels and soil moisture in the riparian zone [37]. The greater increase in the respective nutrient (N and P) contents had a positive effect on the activities of enzymes such as nitrogen (protease) and phosphorus (phosphatase) in the rhizosphere soil of plants [38]. Perennial herbaceous plants, such as Heteropogon contortus, Cynodon dactylon, and Paspalum distichum, benefit from having a higher root density in a small area of soil; we speculate that they may be more sensitive to microbial enzyme activity compared with other plants. At the macroscopic level, this was manifested by the close dependence of their distribution on the TN and P contents of soil (Figure 3). In the future, combining more measurements of soil bacterial and fungal communities to explain the species distribution in the riparian area will bring more accurate results [38].

On the transect scale, the distribution of vascular plants has an obvious zonation phenomenon. The three zones were divided according to multi-year average high water lines and mean tide levels. Middle-distance zone (zone B) and near-water zone (zone A) plant habitats were subject to high-frequency and long-term periodic changes in water levels, and the harsh environmental conditions were not conducive to the growth of tall woody plants with significantly sandy soils and low nutrients [39]. Low-growing herbaceous plants had strong environmental adaptability and a wider ecological amplitude, making them the dominant species in the riparian area’s middle-distance and near-water zones. The representative species distributed in this zone included Polygonum hydropiper, Alternanthera philoxeroides, Artemisia argyi, and Setaria viridis, changing from hygrophytes to mesophytes (

Table A3. Relative abundance of dominant families and species in different habitats of the subzones.

Zones	Dominance	Lakeshore			Tributary
		Dominant Families	Dominant Species	Biomass	Dominant Families	Dominant Species	Biomass
A	First	Amaranthaceae	Altphi	349.98	Polygonaceae	Polhyd	212.08
		33.67%	33.50%		48.04%	14.29%
	Second	Gramineae	Setvir		Amaranthaceae	Altphi
		26.73%	13.71%		42.64%	48.97%
B	First	Gramineae	Eleind	352.61	Polygonaceae	Polper	295.03
		62.63%	13.91%		51.13%	8.84%
	Second	Amaranthaceae	Setvir		Gramineae	Artarg
		24.24%	23.48%		45.54%	8.84%
C	First	Gramineae	Eleind	302.01	Gramineae	Polper	265.95
		41.68%	61.54%		44.69%	8.78%
	Second	Compositae	Bidtri		Polygonaceae	Elydah
		20.22%	9.35%		44.01%	2.74%

). The distant water zones (zone C) were less affected by the changing water level; therefore, the clay content of the soil increased, and the nutrient content was also higher, meaning that most plants were suitable for growth. The proportion of woody plants was highest in zone C (compared to zone A: CE: 3.15, p < 0.05, Table 3), while the lowest proportion of annual herbs was found in zone C compared to the other two zones, as mentioned in Section 3.3. After the settlement of the subshrubs and woody plants, their ability to compete for resources such as sunlight and soil nutrients was significantly higher than that of low-growing herbs due to their larger biomass [40]. This explained the development of mesophytic trees and shrubs, such as Melochia corchorifolia and Morus alba (Table 3, Figure 4). Companion species included shrubs and perennial herbs such as Rosa sertata and Elymus dahuricus. However, we did not consider the nested effect in the RDA (transects nested within tributaries) because of the high intra tributary variability in the response variables considered, as well as our limited sample size.

4.3. Community Structure of Vascular Plants in Response to Soil Factors

We found that the herb-dominant communities were mainly distributed in zone A, shrub–grass communities were mainly distributed in zone B, and tree–shrub–grass communities were mainly distributed in zone C of the riparian area (Table 3). The overall community complexity of Baoan Lake is consistent with competition theory and environmental stability theory [41,42]. The statistical analysis of the diversity indices showed that the lakeshore and tributary plant communities exhibited different patterns of variation from zone A to zone C. Based on correlation analysis and field surveys, this might be due to the steeper slope in zone B of the tributaries compared to the lakeshores (Table 2). Steeper slopes promoted the selection effects of the soil environment on plants, with species such as Artemisia caruifolia and Artemisia annua being more sensitive to changes in slope. This effect also induced the typical “R strategy” of plants to grow in large numbers [43]. These plants relied on their strong ability to propagate to form a large number of seedlings, thus gaining a population growth advantage [44,45]. Their flourishing led to a more homogenous species composition in the riparian area, resulting in a decrease in Shannon-Wiener and Pielou evenness index values.

The field survey results indicated stable hydrological conditions on the lakeshores, with less variability in lakeshore water levels and lower flood frequency than in the tributaries. Therefore, we hypothesized that the zones near the water’s edge in lakeshores have higher biodiversity than the tributaries. However, the diversity of zone A of the riparian of tributary beaches was significantly higher than that of the lakeshores (p < 0.05), which exceeded our expectations (Table 1, Figure 5). Lakeshore areas near the water’s edge had a higher proportion of long floating seed-propagating species. These plants are more suitable for low-flowing lake waters because the seeds often have hairs, plumes, or wing-like structures that increase the surface area of the seed, allowing them to provide sufficient surface tension as they float on water. The low proportion of short-floating species on lakeshores may be due to the slower flow velocity, which results in most of the short-floating propagules sinking before reaching the shoreline [46]. Consequently, ensuring a suitable flow velocity in the tributaries will enable some plants to propagate.

In the transect, changes in plant community composition from the hydrosere to the xerosere were evident within the riparian area (Figure 4). The greatest diversity of tree, shrub, and grass communities was found in middle-distance zones. The Shannon-Wiener and Species Richness index values tended to decrease towards the extremes of wet and dry zones (zone A and zone C, respectively) (Figure 5). The relatively low diversity in the near-water zone was mainly due to the complexity and variability of the habitat conditions, such as the regular fluctuations of water levels and the farming practices of fishermen, which could result in harsher habitat conditions. In addition, soil nutrients were more likely to be lost due to the high sand content. Only a few pioneer species with strong resource utilization abilities could survive, such as Alternanthera philoxeroides and Eichhornia crassipes (Table 3). Eichhornia crassipes has a well-developed fibrous root, a number of polygonal columnar cells in the inner part of the stem and long, thin stolons to utilize resources [47]. Alternanthera philoxeroides is a perennial herbaceous plant of the genus Amaranthus, native to Brazil, which has tubular stems with a creeping base and can grow by clonal integration, being more competitive for soil nutrients and light resources. [48,49]. This led to poor growth in communities where the positive species were disadvantaged in the competition for resources, resulting in communities with low richness and diversity. Shrub–grass or tree–shrub–grass communities in middle-distance zones exhibited the highest biodiversity. In these zones, the proportion of sand content and soil porosity decreased, the water-holding capacity increased, and the SBD was maximized [50]. This favored the development of tree and shrub species such as Rosa multiflora and Triadica sebifera, which were accompanied by perennial herbs, such as Elymus dahuricus and Lactuca sibirica. The richness of these zones was the highest, which supports the Intermediate Disturbance Hypothesis [51]. With the increase of offshore distance, the lower Shannon-Wiener index and Pielou evenness index values in the distant water zones were mainly related to the distribution of tall trees (Table 3). In situations where the overall nutrient content was not too high (compared to ecosystems such as forests), competition for resources between species was more intense. We know that the ecological amplitude of trees, shrubs, and herbs overlapped and differed significantly in their ability to capture resources [52]. Although each lifeform contains a range of strategies associated with the tolerance-competition trade-off, in general, trees have a higher competitive capacity and are less stress-tolerant than herbaceous plants. However, in distant water areas with relatively low environmental pressures, the presence of trees can crowd out the resources of herbaceous plants [53]. Dwarf and shallow-rooted herbaceous plants were reduced in number or even disappeared because they did not get sufficient sunlight and soil nutrients in the competition. Due to this, the biomass of nonwoody plants and their distribution density was low, although the community trophic structure in the distant water zone was complex.

5. Conclusions

This study investigated vascular plants in the riparian areas of lakes and tributaries in Baoan Lake. It explored the influence of soil factors on the distribution pattern and community structure of vascular plant communities in different habitats. It can help researchers better understand the role of habitat filtering and soil-driving mechanisms in the community of vascular plant composition. The main findings are as follows.

The soil properties were the most critical factors affecting the distribution and composition of vascular plants in riparian areas. In addition, the shared effects of the different factor groups were equally significant. The most important soil factors affecting the distribution of vascular plants in lakeshore riparian areas are the TN, AP, and TP. In contrast, the critical factors in tributary habitats are the TN, pH, and TP. Soil TN, as the main factor driving the distribution of vascular plants in riparian areas, was mainly related to plants’ biological characteristics. Additionally, biodiversity and the proportion of perennial plants were generally higher in lakeshore riparian areas than in tributaries. The Shannon-Wiener and Species Richness index values of the lakeshore habitats showed a hump-shaped distribution pattern from near to distant water zones. The dominant species of Asteraceae, Gramineae, and woody plants showed a zonation distribution.

Overall, this study provides a foundation for understanding the effect of soil physicochemical and topographical factors on plant communities in shallow lake riparian areas. Ensuring riparian areas’ ecological health and stability is the basis of maintaining lake biodiversity. The results could provide a helpful framework and a scientific basis for researchers to facilitate the analysis of riparian ecosystems.