Root Traits and Erosion Resistance of Three Endemic Grasses for Estuarine Sand Drift Control

Abstract

In southern Taiwan, rivers sporadically cease to flow and dry up in winter. The exposed dry riverbeds are very vulnerable to wind erosion. The strong northeast monsoon often induces serious estuarine sand drift and fugitive dust, which cause damages to agricultural crops, human health and infrastructures. Giant reed (Arundo formosana), common reed (Phragmite australis) and the wild sugarcane (Saccharum spontaneum) are pioneer grass species in estuary areas. They have great potential to reduce wind erosion and control windblown dust on agricultural lands. Nevertheless, their root traits, biomechanical characteristics and wind erosion resistance have not been investigated. In this research, the root traits were investigated utilizing the hand digging technique and the WinRHIZOPro System. Root pullout resistance and root tensile strength were estimated using vertical pullout and root tensile tests. Wind tunnel tests were executed to evaluate the wind erosion resistance using six-month-old plants. The results demonstrated that the growth performance and root functional traits of S. spontaneum are superior to those of A. formosana and P. australis. Additionally, the root anchorage ability and root tensile strength of S. spontaneum plants are notably greater than those of A. formosana and P. australis plants. Furthermore, the results of the wind tunnel tests showed that the wind erosion resistance of A. formosana is remarkably higher than those of S. spontaneum and P. australis. This study demonstrates that A. formosana and S. spontaneum are superior to P. australis, considering root traits, root anchorage ability, root tensile strength and wind erosion resistance. Taken together, our results suggest that S. spontaneum and P. australis are favorable for riverbed planting, while A. formosana is applicable for riverbank planting in estuary areas. These results, together with data on the acclimation of estuarine grasses in waterlogged soils and brackish waters, provide vital information for designing planting strategies of estuary grasses for the ecological engineering of estuarine sand drift control.

1. Introduction

Taiwan, a Pacific island with mountainous terrain and fragile geologic features, is very susceptible to water erosion and landslides [1]. During the rainy season, rainstorms often cause severe riverbed erosion. The torrential floods transport abundant sand and deposit to downstream areas and estuaries [2,3]. Strong northeastern monsoon winds usually induce serious estuarine sand drift disasters in southwestern Taiwan [4,5]. Estuarine sand drift control has become the major challenge in the management of estuarine ecosystems [6,7,8,9]. In northern Bangladesh, traditional erosion management approaches have been applied to control riverbank erosion and enhance livelihood resilience [10]. The estuarine vegetation is very efficient at trapping drifting sand and reducing aeolian erosion [11], and its use is regarded as an effective ecological engineering method for sand drift control [12,13].

Estuaries, with waterlogged soils and brackish waters, pose great challenges to plants [14]. Previous studies indicated that P. australis and S. spontaneum are salt-tolerant species [15,16]. The salt-adapted ecotype of P. australis is a C₃–C₄ plant [17], while S. spontaneum is a C₄ plant, and both exhibit superior photosynthetic performance under saline and warm temperature conditions. Endemic pioneer estuary plants are known to play a key role in estuarine sand drift control [4]. In Taiwan, Arundo formosana, Phragmite australis and Saccharum spontaneum are endemic pioneer grass species in estuary areas [18].

Root traits and biomechanical characteristics have notable influences on sand erosion [19]. The grass root system is classified as massive (M) [20]. It has been demonstrated that the root anchorage ability is strongly affected by root traits, and grass with a profuse fibrous root system has a great anchorage force that prevents soil erosion [21]. Generally, the plant root tensile strength is highly correlated with the pullout resistance force [22]. Burylo et al. [23] indicated that high root surface and high tensile strength are important for soil stabilization. Hamidifar et al. [24] demonstrated that vetiver grass (Vetiveria zizanioides L.) can increase soil cohesion and the soil internal friction factor and enhance riverbank shear strength. Several studies revealed that plant coverage can notably decrease the wind erosion rate [13,25]. However, the root traits, mechanical properties and wind erosion resistance of A. formosana, P. australis and S. spontaneum have not been explored. Thus, this research is focused on exploring the root traits, biotechnical properties and wind erosion resistance of these estuarine grasses. The comparison of these traits among the examined species can provide a better strategy for the environmental engineering of estuarine sand drift control.

2. Materials and Methods

2.1. Plant Sampling and Raising

Three 40 cm × 40 cm sample plots at 3 m intervals across three transects lines of A. formosana, P. australis and S. spontaneum were established at the Pachang river estuary vegetation site situated in Budai Township, Chiayi County, Taiwan (120°7′46′′ E, 23°21′34′′ N) (Figure 1) in January 2021. Two hundred ramets of each species were randomly chosen and carefully collected with a hand shovel. The mean height of the ramets of A. formosana, P. australis and S. spontaneum was 4.1 ± 0.4 cm, 4.2 ± 0.2 cm and 4.3 ± 0.5 cm, respectively. The ramets were cleaned, placed wet in Ziplock bags, transported to the university nursery and transplanted into high wooden boxes (30 × 30 × 100 cm, L × W × H) for investigation of their root traits and biomechanical properties, as well as into small wooden boxes (30 × 30 × 40 cm, L × W × H) for the wind tunnel test. Prior to transplanting, the boxes were filled with soil gathered from the same estuary sampling site. This soil is sandy, with a mean dry weight of 16.2 kN m⁻³ and a water content of 10.2 ± 2.1%. The soil contained 93.1% sand, 5.7% silt and 1.2% gravel. As for the soil’s chemical characteristics, we measured pH 8.5, electrical conductivity of 0.06 ds m⁻¹, total nitrogen content of 0.04%, soluble phosphorus content of 18.0 mg kg⁻¹, soluble potassium content of 35.0 mg kg⁻¹ and organic matter content of 0.26 mg kg⁻¹. For the analysis of growth traits and root biomechanical properties, 30 ramets of each species were planted into 30 tall boxes. For the wind tunnel test, 40 ramets of each species were planted into 120 small boxes, individually. Ten boxes with bare soil collected from the same sampling site served as a control. All boxes were randomly arranged in tree nursery plots under ambient conditions and irrigated every 4 days. The boxes were revolved weekly to reduce shadowing. Jointly, the azimuth of the boxes was consistent.

2.2. Growth Traits

In the estuary areas of southern Taiwan, A. formosana, P. australis and S. spontaneum are endemic pioneer plants adapted to the challenging environment. They are perennial grasses, with hard culm and sharp blades, distributed along riverbeds and riverbanks. (Figure 2).

Pilot experiment revealed the boxes had enough room for plant development. Six months after transplanting, 10 plants of each species were sampled at random for inspecting their root traits. Before root excavation, plant height (H) and stem base diameter (S_bd) were recorded. The roots were extracted and cleaned with water to examine root growth performance. Root system images were recorded to examine root configuration and growth trait. Root traits were measured utilizing the WinRHIZO software [26], although the water displacement method was applied to estimate the root volume [27]. Roots and shoots were dried at 70 °C for 48 h to measure the dry biomass. Then, the root traits were computed [28]. Simultaneously, single root sections were collected for tensile testing.

2.3. Vertical Pullout Test

Ten plants of each species were selected at random for pullout resistance measurements. The soil had an average dry weight of 15.6 kN m⁻³ and a moisture content of 12.3 ± 1.8% estimated at a depth of 20 cm utilizing digital a soil moisture meter [29]. Prior to the pullout test, plant height and stem base diameter were recorded. The stem was severed 15 cm from the base, wrapped with tape, and hooked to the pulling device with iron fasteners. The pullout test was carried out utilizing an uprooting machine [30]. The maximum pullout resistance (F_max, N) and displacement were digitally registered for numerical analysis.

2.4. Root Tensile Test

Single root samples collected from root excavation were classified into different diameter classes. The root samples were processed and stored [31]. Damaged root sections were discarded. Two hundred and ten root sections of each species were examined. All tensile tests were executed within 24 h after sampling. The tensile tests were carried out utilizing a tensile testing machine [32]. For each species, 100 root samples were tested, and only 66 root sections of A. formosana, 81 root sections of P. australis and 76 root sections of S. spontaneum ruptured in the middle section. The root tensile strength (T_si, MPa) was calculated utilizing the subsequent equation [33]:

$$T_{si}=\frac{{4\mathrm{F}}_{max}}{{\pi \mathrm{D}}_{\mathrm{i}}{}^2}$$

(1)

where F_max is the maximal tensile force at breakage (N), and D_i is the mean root section diameter (mm) estimated in the middle point.

Furthermore, the relation between root tensile strength (T_s) and diameter (D) was computed utilizing the subsequent equation [34]:

$$T_s=\alpha ·D^{-\beta }$$

(2)

where $\alpha$ and $\beta$ are experimental constants related to each species.

2.5. Wind Tunnel Test

The wind tunnel test was used to estimate the erosion resistance of these species. A wind tunnel was set up utilizing an iron framework covered with polycarbonate sheets on all sides and on the top, except for its rear end (500 cm × 90 cm × 120 cm, L × W × H). A wind generator was connected to the wind tunnel, and wind-blown sand was collected using an acrylic sand collector [35]. For the wind erosion tests, the testing wind speeds of 6.1 ± 0.2 m s⁻¹ and 10.2 ± 0.5 m s⁻¹ were reached by keeping the distance between the fan and the plant box at 60 cm and 90 cm, respectively. The day temperature in the tunnel was 28 ± 3 °C, with 45–50% relative humidity. Before testing, the average soil moisture (1.2 ± 0.2% in top 10 cm layer of soil) was measured utilizing a digital soil moisture meter, and vegetation cover images of each sample box were estimated using ImageJ system [13]. The plants in the boxes were cautiously thinned to coverages of 20%, 40% and 60%. Boxes with bare soil acted as a control. Then, plant boxes of dissimilar coverages were placed in the wind tunnel individually, and blown for 6 min. For each species, the tests included 4 coverages and 2 speeds, with 10 replicates. The amounts of sand blown were gathered and estimated for the wind erosion resistance analysis. Since this experiment was focused on the wind erosion of a sandy soil of a riverbed in the estuarine area, the land slope was not considered. The riverbed was flat during the dry season.

2.6. Statistical Data Analysis

One-way ANOVA and Tukey’s HSD tests were applied to analyze the significance of the differences in root traits, root biotechnical properties and wind erosion resistance among the three grasses. Regression analysis in SPSS was used to explore the relations between uprooting resistance and root traits. The relations between root tensile resistance, tensile strength and root diameter among the three grasses were investigated utilizing Microsoft Excel Regression analysis (Excel 2013, Microsoft, Redmond, WA, USA).

3. Results

3.1. Growth Traits

Statistical data revealed that the plant growth traits differed greatly among the three grass species, except for the root tip number (

Table 1. Growth performance of the plants of the three grass species.

Growth Parameters	A. formosana	P. australis	S. spontaneum	ANOVA (F)
H (cm)	83.75 ± 4.81 c	177.92 ± 13.14 b	375.83 ± 16.91 a	138.448 ***
Sbd (mm)	6.23 ± 0.26 b	6.84 ± 0.55 b	11.91 ± 0.32 a	61.586 ***
RT	1421.42 ± 240.01 a	1226.33 ± 212.61 a	1813.75 ± 156.68 a	2.108 ns
TRL (m)	31.91 ± 4.54 b	22.34 ± 2.84 c	39.32 ± 3.65 a	5.177 **
Rb (g)	82.58 ± 15.76 b	36.25 ± 6.43 b	217.33 ± 23.09 a	32.250 ***
Sb (g)	72.17 ± 10.73 b	61.58 ± 6.39 b	427.33 ± 40.39 a	72.747 ***

Grass height (H), stem base diameter (S_bd), root tips (RT), total root length (TRL), root biomass (R_b), shoot biomass (S_b). Values with dissimilar letters in the same row indicate remarkable dissimilarity (Tukey’s HSD test) among the species. N = 12. Levels of significance: ns, non-significant, ** p < 0.01, *** p < 0.001.

). On average, S. spontaneum plants (375.83 ± 16.91 cm) were remarkably taller than P. australis (177.92 ± 13.14 cm) and A. formosana (83.75 ± 13.14 cm) plants. The stem base diameter was the largest in S. spontaneum (11.91 ± 0.32 mm) and smaller in P. australis (6.84 ± 0.55 mm) and A. formosana (6.23 ± 0.26 mm). The total root length was notably longer for S. spontaneum (3931.61 ± 365.19 cm) and A. formosana (3190.97 ± 453.93 cm) than for P. australis (2233.74 ± 283.61 cm). The root biomass of S. spontaneum (0.22 ± 0.02 kg) was also remarkably greater than that of A. formosana (0.08 ± 0.16 kg) and P. australis (0.04 ± 0.01 g). The shoot biomass of S. spontaneum (0.43 ± 0.04 kg) was remarkably higher than that of A. formosana (0.07 ± 0.01 kg) and P. australis (0.06 ± 0.01 kg). The root system of A. formosana, P. australis and S. spontaneum was classified as fibrous M- (massive) type according to Yen [20]. S. spontaneum plants grew longer root systems than A. formosana and P. australis plants. In addition, S. spontaneum plants had larger lateral root extensions than A. formosana and P. australis plants (Figure 3). The root area distribution revealed that most roots distributed in the top 20 cm and were sparser below (Figure 4). Altogether, the C₄ S. spontaneum plants exhibited better growth performance than P. australis and A. formosana plants.

The root functional traits appeared remarkably dissimilar among the three grasses (

Table 2. Means ± SEs of root functional traits of the plants of the three grass species.

Root Traits	A. formosana	P. australis	S. spontaneum	ANOVA (F)
RD (kg m−3)	1.53 ± 0.29 b	0.67 ± 0.12 b	4.02 ± 0.43 a	32.250 ***
RLD (km m−3)	0.59 ± 0.08 a	0.41 ± 0.05 b	0.73 ± 0.07 a	5.177 *
RSA (cm2)	2973.47 ± 466.77 b	1995.13 ± 241.06 b	5276.63 ± 442.11 a	18.061 ***
RTD (g cm−3)	0.23 ± 0.02 a	0.09 ± 0.01 b	0.20 ± 0.01 a	18.402 ***
RV (cm3)	343.75 ± 46.82 b	400.83 ± 72.38 b	1085.42 ± 83.52 a	35.469 ***
SRL (m g−1)	0.54 ± 0.12 a	0.85 ± 0.16 a	0.21 ± 0.03 b	7.354 **
SRA (m2 g−1)	0.0047 ± 0.001 a	0.0049 ± 0.001 a	0.0027 ± 0.01 b	7.334 **

Root density (RD), root length density (RLD), total root surface area (RSA), root tissue density (RTD), root volume (RV), specific root length (SRL), specific root area (SRA). Dissimilar letters in the same row specify notable dissimilarities (Tukey’s HSD test) among the species. N = 12. Levels of significance: ns, non-significant, * p < 0.05, ** p < 0.01, *** p < 0.001.

). The average root density of S. spontaneum (4.02 ± 0.43 kg m⁻³) was much greater than those of A. formosana (1.53 ± 0.29 kg m⁻³) and P. australis (0.62 ± 0.12 kg m⁻³). The root length density of S. spontaneum (0.73 ± 0.07 km m⁻³) and A. formosana (0.59 ± 0.08 km m⁻³) was greater than those of P. australis (0.41 ± 0.05 km m⁻³). The root surface area of S. spontaneum (5276.63 ± 442.11 cm²) was significantly larger than those of A. formosana (2973.47 ± 466.77 cm²) and P. australis (1995.13 ± 241.06 cm²). The root tissue density of A. formosana (0.23 ± 0.02 g cm⁻³) and S. spontaneum (0.20 ± 0.01 g cm⁻³) was greater than those of P. australis (0.09 ± 0.01 g cm⁻³). The root volume of S. spontaneum (1085.42 ± 83.52 cm³) was also significantly greater than those of P. australis (400.83 ± 72.38 cm³) and A. formosana (343.75 ± 46.82 cm³). However, the specific root lengths of P. australis (0.85 ± 0.16 m g⁻¹) and A. formosana (0.54 ± 0.12 m g⁻¹) were remarkably greater than that of S. spontaneum (0.21 ± 0.03 m g⁻¹). The specific root areas of P. australis (0.0049 ± 0.001 m⁻² g⁻¹) and A. formosana (0.0047 ± 0.001 m⁻² g⁻¹) were notably greater than that of S. spontaneum (0.0027 ± 0.001 m⁻² g⁻¹). Overall, the root functional traits of S. spontaneum and A. formosana were remarkably larger than those of P. australis. Among the functional traits, root density, total root surface area and root volume are the most important for plant growth.

3.2. Pullout Resistance

The plant pullout force raised with a displacement up to the peak and decreased as the roots fractured (Figure 5). The maximal pullout force for S. spontaneum (1.44 ± 0.35 kN) was four times that of P. australis (0.36 ± 0.68 kN) and about five times that of S. spontaneum (0.28 ± 0.07 kN) (

Table 3. Means ± SEs of the maximal pullout force for the three grass species.

Biomechanical Properties	A. formosana	P. australis	S. spontaneum	One-Way ANOVA (F)
Maximal pullout resistance (kN)	0.28 ± 0.07 b	0.36 ± 0.68 b	1.44 ± 0.35 a	63.775 ***

Different letters in the same row specify remarkable dissimilarities (Tukey’s HSD test) among the species. N = 12. Significance level: *** p < 0.001.

). Regression analysis revealed a positive relation between the maximal pullout resistance and some functional traits, i.e., total root length and root length density. Linear regressions of pullout resistance (P_r) and total root length (TRL) for A. formosana, P. australis and S. spontaneum plants were P_r = 0.009TRL − 0.691 (R² = 0.329, p = 0.051), P_r = 0.008TRL + 18.284 (R² = 0.02, p = 0.295) and P_r = 0.019TRL + 68.499 (R² = 0.43, p = 0.012), respectively. Linear regressions of pullout resistance (U_r) and root length density (RLD) for A. formosana, P. australis and S. spontaneum plants were U_r = 48.297RLD − 0.706 (R² = 0.33, p = 0.051), U_r = 12.64RLD + 18.285 (R² = 0.02, p = 0.295) and U_r = 104.089RLD + 68.473 (R² = 0.45, p = 0.010), respectively (

Table 4. Relations between maximal pullout resistance force and root traits for the three grasses.

Morphological Traits	Species	Regression Equation	R2	p
TRL (cm)	A. formosana	Pr = 0.009TRL − 0.691	0.329	0.051
	P. australis	Pr = 0.008TRL + 18.284	0.02	0.295
	S. spontaneum	Pr = 0.019TRL + 68.499	0.43 *	0.012
RLD (km/m3)	A. formosana	Pr = 48.297RLD − 0.706	0.33	0.051
	P. australis	Pr = 42.64RLD + 18.285	0.02	0.295
	S. spontaneum	Pr = 104.089RLD + 68.473	0.45 *	0.010

Maximal pullout resistance force (P_r), total root length density (TRL), root length density (RLD). Significance level: * p < 0.05.

). In conclusion, with greater root length and root length density, the pullout resistance of S. spontaneum was remarkably greater than those of A. formosana and P. australis.

3.3. Root Tensile Strength

The ANOVA results showed that root diameter, tensile resistance and tensile strength differed notably among the three grasses. The mean root diameter for P. australis (2.08 ± 0.11 mm) was remarkably larger than those of S. spontaneum (0.92 ± 0.02 mm) and A. formosana (0.81 ± 0.02 mm). The mean root tensile resistance force of P. australis (32.3342.69 ± 2.90 N) was notably greater than those of S. spontaneum (17.09 ± 0.39 N) and A. formosana (11.70 ± 0.38 N). Additionally, the mean root tensile strength was the topmost for S. spontaneum (32.51 ± 1.08 MPa), lower for A. formosana (26.02 ± 0.09 MPa), and the lowest for P. australis (12.19 ± 0.61 MPa) (

Table 5. Means ± SEs of root diameter, tensile resistance and tensile strength for the three grass species.

Root Parameters	A. formosana	P. australis	S. spontaneum	One-Way ANOVA (F)
Root diameter (mm)	0.81 ± 0.02 c	2.08 ± 0.11 a	0.92 ± 0.02 b	126.01 ***
Tensile resistance force (N)	11.70 ± 0.38 c	32.33 ± 2.90 a	17.09 ± 0.39 b	36.15 ***
Tensile strength (MPa)	26.02 ± 0.99 b	12.19 ± 0.61 c	32.51 ± 1.08 a	117.74 ***

Dissimilar letters in the same row indicate remarkable dissimilarities (Tukey’s HSD test) among the species. Level of significance: *** p < 0.001.

). Furthermore, the root tensile resistance decreased with the decrease of the root diameter (Figure 6).

However, the root tensile strength increased with a decreasing root diameter (Figure 7). Overall, the root tensile strength of S. spontaneum was notably higher than those of P. australis and A. formosana.

3.4. Wind Erosion Resistance

Statistical analysis demonstrated the leeward soil surface wind speed differed among the three grasses. The average leeward wind speed varied remarkably among the species at windward wind speeds of 6.1 ± 0.2 and 10.2 ± 0.5 m s⁻¹. At a windward wind speed of 6.1 ± 0.2 m s⁻¹ and a vegetation cover of 20%, the average leeward wind speed of A. formosana (2.17 ± 0.08 m s⁻¹) was about 10% higher than those of S. spontaneum (1.97 ± 0.03 m s⁻¹) and P. australis (1.90 ± 0.02 m s⁻¹), while at the vegetation cover of 60%, the leeward wind speeds of S. spontaneum (1.21 ± 0.08 m s⁻¹) and P. australis (1.02 ± 0.02 m s⁻¹) were about 250% higher than that of A. formosana (0.29 ± 0.03 m s⁻¹) (

Table 6. Means ± SEs of leeward wind speeds at a windward wind speed of 6.1 ± 0.2 m s⁻¹ and different vegetation covers for the plants of the three grasses.

Species	Leeward Soil Surface Wind Speed (m s−1)
	Vegetation Cover
	0%	20%	40%	60%
A. formosana	2.99 ± 0.09 a	2.17 ± 0.08 a	1.61 ± 0.07 b	0.29 ± 0.03 c
P. australis	2.99 ± 0.09 a	1.90 ± 0.02 b	1.63 ± 0.02 b	1.02 ± 0.02 b
S. spontaneum	2.99 ± 0.09 a	1.97 ± 0.03 b	1.80 ± 0.04 a	1.21 ± 0.08 a

Dissimilar letters in the same column specify notable dissimilarities (Tukey’s HSD test) among the species. N = 10. Level of significance: p < 0.05.

). Additionally, at a windward wind speed of 10.2 ± 0.5 m s⁻¹ and a vegetation cover of 20%, the mean leeward wind speed of A. formosana (3.32 ± 0.10 m s⁻¹) was about 8% higher than those of S. spontaneum (2.39 ± 0.04 m s⁻¹) and P. australis (3.05 ± 0.04 m s⁻¹), while at a vegetation cover of 60%, the leeward wind speeds of P. australis (1.96 ± 0.03 m s⁻¹) and S. spontaneum (1.45 ± 0.06 m s⁻¹) were about 50% higher than that of A. formosana (0.96 ± 0.06 m s⁻¹) (

Table 7. Means ± SEs of leeward wind speeds at a windward wind speed of 10.2 ± 0.5 m s⁻¹ and different vegetation covers for the three grasses.

Species	Leeward Soil Surface Wind Speed (m s−1)
	Vegetation Cover
	0%	20%	40%	60%
A. formosana	5.62 ± 0.07 a	3.32 ± 0.10 a	2.73 ± 0.06 a	0.97 ± 0.06 c
P. australis	5.62 ± 0.07 a	3.05 ± 0.04 b	2.45 ± 0.03 b	1.96 ± 0.03 a
S. spontaneum	5.62 ± 0.07 a	2.39 ± 0.04 c	2.07 ± 0.06 c	1.45 ± 0.06 b

Dissimilar letters in the same column specify remarkable dissimilarities (Tukey’s HSD test) among the species. N = 10. Level of significance: p < 0.05.

). Moreover, the wind erosion rates were notably dissimilar among these grasses. At a windward wind speed of 6.1 ± 0.2 m s⁻¹ and a vegetation cover of 20%, the average wind erosion rates of P. australis (7.43 ± 0.05 g m⁻² s⁻¹) and S. spontaneum (6.41 ± 0.08 g m⁻² s⁻¹) were about 40% higher than that of A. formosana (4.63 ± 0.1 g m⁻² s⁻¹), while at a vegetation cover of 60%, the wind erosion rates of S. spontaneum (0.53 ± 0.02 g m⁻² s⁻¹) and P. australis (0.46 ± 0.01 g m⁻² s⁻¹) were about 90% higher than that of A. formosana (0.11 ± 0.01 g m⁻² s⁻¹) (

Table 8. Means ± SEs of wind erosion rates at a windward wind speed of 6.1 ± 0.2 m s⁻¹ and different vegetation covers for the three grasses.

Species	Wind Erosion Rate (g m−2 s−1)
	Vegetation Cover
	0%	20%	40%	60%
A. formosana	17.87 ± 0.26 a	4.63 ± 0.1 c	1.02 ± 0.04 b	0.24 ± 0.01 c
P. australis	17.87 ± 0.26 a	7.43 ± 0.05 a	3.29 ± 0.08 a	0.46 ± 0.02 b
S. spontaneum	17.87 ± 0.26 a	6.41 ± 0.08 b	0.72 ± 0.03 c	0.53 ± 0.02 a

Different letters in the same column indicate notable dissimilarities (Tukey’s HSD test) among species. N = 10. Level of Significance: p < 0.05.

). At a windward wind speed of 10.2 ± 0.5 m s⁻¹ and a vegetation cover of 20%, the average wind erosion rates of P. australis (22.22 ± 0.25 g m⁻² s⁻¹) and A. formosana (21.73 ± 0.35 g m⁻² s⁻¹) were about 20% higher than that of S. spontaneum (17.51 ± 0.30 g m⁻² s⁻¹), while at a vegetation cover of 60%, the wind erosion rates of P. australis (12.49 ± 0.47 g m⁻² s⁻¹) and S. spontaneum (4.23 ± 0.16 g m⁻² s⁻¹) were about 300% higher than that of A. formosana (1.3 ± 0.03 g m⁻² s⁻¹) (

Table 9. Variation in wind erosion rates under a windward wind speed of 10.2 ± 0.5 m s⁻¹ and different vegetation covers for the three grasses.

Species	Wind Erosion Rate (g m−2 s−1)
	Vegetation Cover
	0%	20%	40%	60%
A. formosana	38.79 ± 0.23 a	21.73 ± 0.35 a	6.95 ± 0.09 c	1.30 ± 0.03 c
P. australis	38.79 ± 0.23 a	22.22 ± 0.25 a	19.89 ± 0.30 a	12.49 ± 0.47 a
S. spontaneum	38.79 ± 0.23 a	17.51 ± 0.30 b	10.21 ± 0.25 b	4.23 ± 0.16 b

Different superscripts in the same column specify notable dissimilarities (Tukey’s HSD test) among the species. N = 10. Level of significance: p < 0.05.

). Taken together, the results clearly show that the wind erosion resistance was the highest for A. formosana, lower for S. spontaneum and the lowest for P. australis for decreasing wind speed and erosion rate.

4. Discussion

4.1. Growth Traits

The statistical data showed that the plant growth performance differed remarkably among the three grass species. Most growth traits, i.e., plant height, stem base diameter, total root length, root biomass and shoot biomass, were notably larger for S. spontaneum plants than for A. formosana and P. australis plants. The sandy soil from the estuary area considered in this study has a low nutrient content. Thus, grass growth and vegetation development are restricted. Plant growth of estuary plants is very crucial for sand stabilization in estuary areas. In general, vegetation plays an important role in the conservation of estuarine ecosystems [14]. Past studies showed that P. australis and S. spontaneum can adapt to flooding and salinity [15,16,17]. Hsu [18] also indicated that A. formosana is beneficial for riverbank and slope stabilization. Earlier studies demonstrated that plants with better root traits and greater biomass have better survival and growth than smaller plants [36,37]. Overall, our results show that C₄ S. spontaneum plants with better growth traits than A. formosana and P. australis plants are more favorable for sand drift control and ecosystem restoration in estuaries. In general, salt-adapted C₄ plants, such as S. spontaneum and Miscanthus floridulus, are expected to have higher photosynthetic capacity and higher nutrient use efficiency and thus greater growth advantage over C₃ species along saline coast lines [38]. In future studies, M. floridulus, a salt-adapted C₄ plant found in Green Island and Taiwan, may also be useful for sand drift control.

The root systems of A. formosana, P. australis and S. spontaneum plants are categorized as fibrous M- (massive) type, consistent with Yen’s study [20]. Earlier research demonstrated that grasses with fibrous M-type roots are favorable for erosion control and slope stabilization [39,40]. Furthermore, S. spontaneum and A. formosana grew deep roots up to 130 cm and 100 cm, respectively. In contrast, P. australis developed its roots only to a depth of 70 cm. Our results also showed that S. spontaneum possesses greater total root length and root biomass than A. formosana and P. australis, implying better competitiveness in nutrient acquisition and water uptake in estuary sand soils [41,42]. On the other hand, S. spontaneum and P. australis are salt-tolerant and can grow well in riverbeds in estuary areas [4]. Consequently, S. spontaneum is recommended for priority estuary riverbed planting, while P. australis, with its smaller root system, is appropriate for second-choice plantings. Furthermore, A. formosana is suggested for riverbank planting.

In addition, our data revealed that all root functional traits, except for specific root length and specific root area, grew remarkably larger for S. spontaneum plants than for A. formosana and P. australis plants. Earlier studies demonstrated that nutrient acquisition is closely related to root biomass, root density, root tissue density and total root surface area [43,44,45]. In the coastal estuary area in Taiwan, A. formosana, P. australis and S. spontaneum can form a thick vegetation cover on estuary sandy soils [18]. In general, endemic estuarine plants can adapt to the harsh environments in estuary areas and are beneficial for sand stabilization. Overall, our results indicate that salt-adapted S. spontaneum has better root functional traits and adaptability than A. formosana and P. australis.

4.2. Pullout Resistance

Data from the pullout tests demonstrated that the pullout force of S. spontaneum plants was remarkably higher than those of A. formosana and P. australis plants. The regression analysis of pullout resistance and root traits exhibited positive relations with total root length and root length density. Additionally, the lateral root extension of S. spontaneum plants was larger than those of A. formosana and P. australis plants. These findings are in agreement with earlier research [21,46,47]. Overall, S. spontaneum demonstrated the greatest anchorage ability among the three grasses and appears to be beneficial for fixing sands in estuary areas.

4.3. Root Tensile Strength

The root tensile strength plays a vital role in root reinforcement and slope stabilization [28,48,49]. Our results revealed that root tensile resistance force and tensile strength varied notably among the three species included in this study. The root tensile resistance force was the highest for P. australis, lower for S. spontaneum and the lowest for A. formosana, whereas the root tensile strength was the highest for S. spontaneum, lower for A. formosana and the lowest for P. australis. The root tensile force increased as the root diameter increased, whereas the root tensile strength decreased with an increasing root diameter, consistent with previous studies [34,50,51]. These relations were attributed to a decrease in root cellulose content with an increasing root diameter [52] and to an increase in cellulose content as well as a decrease in lignin content [53]. Further study is required to examine the root chemical components of these estuary plants and elucidate the effects of chemical components on root tensile strength. On the other hand, vetiver grass, with its high tensile strength, is a good alternative for riverbank protection. The reason for using these three species instead of others such as vetiver grass is because the exotic vetiver grass may have some impact on the local estuarine ecosystem.

4.4. Wind Erosion Resistance

In general, estuary plants can reduce wind speed and block blown sand. Estuary plants, such as A. formosana, P. australis and S. spontaneum, can adapt to harsh environments. Estuarine vegetation can decrease wind erosion by decreasing wind speed and acts as a stabilizing agent in estuarine ecosystems [19,54]. This research demonstrated the remarkable effect of estuarine vegetation on curtailing wind speed and erosion rate. The wind speed reduction for A. formosana plants was remarkably higher than that of P. australis and S. spontaneum plants. Moreover, the wind erosion rate for A. formosana plants was notably lower than those of S. spontaneum and P. australis plants. Previous research indicated that the vegetation cover plays a vital role in decreasing wind speed and wind erosion [11,13,55]. Altogether, our findings clearly indicated that the wind erosion resistance of A. formosana was the highest, followed, in order, by those of S. spontaneum and P. australis. However, A. formosana cannot adapt to waterlogged soils [20]. Thus, we recommend that S. spontaneum is the priority species, and P. australis is the second species for estuarine riverbed planting. However, A. formosana is more suitable for estuarine riverbank planting.

In Taiwan, strong northeast monsoon winds and typhoons frequently cause serious wind erosion and sand drift hazards in estuaries. Sand drift control has become an important issue in estuary management. Estuarine plants play key roles in wind erosion control and vegetation restoration as well as in the sustainability of estuarine ecosystems [56]. A. formosana, P. australis and S. spontaneum are endemic estuarine grasses beneficial for sand drift control. However, there are challenges to their practical application, such as those posed by serious droughts, freezing temperature and heavy pollution, which may hinder the survival and growth of these estuarine grasses.

5. Conclusions

This research showed that the root systems of A. formosana, P. australis and S. spontaneum belong to the fibrous M-type. S. spontaneum plants showed notably better root traits, e.g., total root length and root biomass, than A. formosana and P. australis plants. Root functional traits, e.g., root density, total root surface area and root volume of S. spontaneum plants appeared also remarkably greater than those of A. formosana and P. australis plants. Moreover, the pullout resistance of S. spontaneum plants was notably higher than those of P. australis and A. formosana plants. In addition, S. spontaneum plants demonstrated greater root tensile strength than A. formosana and P. australis plants, which will contribute to soil reinforcement. However, A. formosana plants exhibited higher wind erosion resistance than S. spontaneum and P. australis plants. Taken together, our findings clearly indicate that the C₄ species S. spontaneum is the first-choice plant, while P. australis is the second-choice species for estuarine riverbed planting; in contrast, A. formosana is suitable for estuarine riverbank planting. This study also demonstrated that root traits, root anchorage ability, root tensile strength and wind erosion resistance are important factors for selecting estuarine plants for sand drift control. These findings contribute to our understanding of how estuarine plants are valuable for estuarine ecological engineering. Furthermore, we recommend that specific companion planting methods should be developed, such as mixing with other endemic estuarine plants, e.g., Spanish needle (Bidens pilosa), Peacock-plume grass (Chloris barbata), Blady grass (Imperata cylindrical), Bayhops (Ipomoea pescaprae), salt-adapted C₄ Pacific Island silvergrass (M. floridulus) and Beach vitex (Vitex rotundifolia) in order to advance the biodiversity and sand drift control as well as the sustainability of estuarine ecosystems.