Comparative Assessment of Woody Species for Runoff and Soil Erosion Control on Forest Road Slopes in Harvested Sites of the Hyrcanian Forests, Northern Iran

Abstract

Soil erosion and surface runoff on forest road slopes are major environmental concerns, especially in harvested areas, making effective mitigation strategies essential for sustainable forest management. The study compared the effectiveness of three selected woody species on forest road slopes as a possible mitigating action for runoff and soil erosion in harvested sites. Plots measuring 2 m × 3 m were set up with three species—alder (Alnus glutinosa (L.) Gaertn.), medlar (Mespilus germanica L.) and hawthorn (Crataegus monogyna Jacq.)—on the slopes of forest roads. Within each plot, root abundance, root density, canopy percentage, canopy height, herbaceous cover percentage, and selected soil characteristics were measured and analyzed. Root frequency and Root Area Ratio (the ratio between the area occupied by roots in a unit area of soil) measurements were conducted by excavating 50 × 50 cm soil profiles at a 10-cm distance from the base of each plant in the four cardinal directions. The highest root abundance and RAR values were found in hawthorn, followed by alder and medlar in both cases. The same order of magnitude was evidenced in runoff (255.42 mL m⁻² in hawthorn followed by 176.81 mL m⁻² in alder and 67.36 mL m⁻² in medlar) and the reverse order in terms of soil erosion (8.23 g m⁻² in hawthorn compared to 22.5 g m⁻² in alder and 50.24 g m⁻² in medlar). The results of the study confirm that using plant species with dense and deep roots, especially hawthorn, significantly reduces runoff and erosion, offering a nature-based solution for sustainable forest road management. These results highlight the need for further research under diverse ecological and soil conditions to optimize species selection and improve erosion mitigation strategies.

1. Introduction

The alteration of natural forest landscapes through human interventions, including logging operations and road development, often leads to increased runoff and soil erosion, posing significant challenges for sustainable forest management [1]. Among the key contributors to these issues, forest operations constitute an aggressive form of human interference in natural ecosystems [2]. If forest operations are carried out without sufficient knowledge of the interactions among all environmental factors, ecosystem stability can be disrupted and irreparable damage can be caused [3]. Forest roads constitute a critical infrastructure element in sustainable forest resource management, enabling efficient access for timber harvesting, ecological monitoring, and conservation practices, while simultaneously posing significant environmental challenges to forest ecosystem integrity. The construction of forest roads often requires the removal of large portions of natural cover. Road slopes alter surface and subsurface water flow, whereas excavations and embankments lead to the creation of steep, uncovered walls that increases the volume of surface runoff and erosion [4,5,6,7]. Additionally, these areas are affected by harvesting equipment, which sometimes operates not only on the road surface but also on adjacent slopes, further disturbing soil structure and increasing the risk of erosion and sediment displacement [8,9].

Root characteristics, particularly root surface area ratios, are crucial factors influencing runoff and erosion. Plants intercept rainfall through their canopies, reducing its impact, and facilitate water infiltration into the soil via root systems that enhance soil porosity and stability [10].

In natural forest ecosystems, vegetation plays a critical role in soil protection and stabilization. Comprising diverse native tree and shrub species, this vegetation shields the soil from direct raindrop impact and regulates soil moisture through enhanced evaporation and transpiration [11,12]. Additionally, it improves water infiltration or promotes water accumulation at the plant base, thereby mitigating erosion. Root systems form a mechanically resistant network that prevents soil displacement by counteracting compressive forces from runoff and soil weight through tensile root strength [12,13,14].

The effectiveness of these mechanisms varies depending on root traits such as depth, density, and spatial distribution, which influence water infiltration and sediment retention [15,16]. Furthermore, soil properties including texture, organic matter content, and moisture significantly affect runoff generation and sediment transport [17].

Using tree and shrub species is considered as one of the most effective and common methods to reduce soil erosion on forest road embankments [18]. Shinohara et al. [19] investigated the effect of vegetation, litter, and roots on sediment and runoff in bamboo forests and broad-leaved evergreen forests using 1 m × 2 m plots. The results revealed significant differences between the bamboo plots and the control plots, indicating that the bamboo species was highly effective in reducing sediment and runoff due to its spreading root system as well in enhancing water retention. Genet et al. [20], observed that species diversity reduces sedimentation. According to the same authors, as species diversity increases in forested slopes, the combined root volume of different plant species within a given soil area enhances soil strength and permeability. However, comparative studies on the effects of shrub and tree species on soil erosion remain limited [21,22]. Implementing methods to reduce runoff and sediment on exposed slopes is a key strategy for combating erosion [23,24], especially in harvested sites where extra pressure occurs during forest operations.

Minimizing the environmental impacts of forest road construction is a key concern in forest management, especially in cases of high ecological importance such as the Hyrcanian forests in northern Iran. These forests are the richest forests in the country and have been extensively exploited over time. They play a vital role in sustaining practically most of the watersheds in the region, directly affecting the lives of downstream communities in various ways. Runoff generated in the upstream areas has caused numerous floods, especially in recent years, with sediment entering the water sources and reducing the quality of the water fed in downstream communities [22,25]. To mitigate these adverse effects, seeds are sown, or plants are planted along the forest road slopes after their construction [26]. However, the success and potential of these treatments hasn’t been extensively examined. The aim of this study was to assess the success of this soil protection strategy by comparing different plant species, and possibly connecting their root characteristics, such as root abundance and root area to soil area index, to their ability to reduce runoff and sediment yield under field conditions. We hypothesized that differences both in terms of root structure and soil conservation would be evidenced among the examined species. Such findings can support sustainable forest management strategies and inform policymakers on best practices for soil and water conservation in the Hyrcanian forests and similar ecosystems.

2. Materials and Methods

2.1. Study Area

The field studies were conducted on forest road slopes located in Shenrood forest, in Guilan province, northern Iran, between 37°0′ N to 37°8′ N latitude and 49°38′ to 49°52′ E longitude (Figure 1) that were seeded with the species under investigation after the construction of a forest road. There is a large network of forest roads in the area due to timber harvesting activity. The forest was composed predominantly of Oriental beech (Fagus orientalis Lipsky) growing on clay-loamy soil. The mean tree height was 21 m and stand density was 180 trees per hectare. The elevation of the study area ranges between 1300–1600 m. The mean annual temperature is 15 °C with the lowest values in February. There are more than 150 rainy days per year in the study area resulting in very notable soil erosion. The mean annual precipitation recorded at the proximate climatology station is 1200 mm. The maximum mean monthly precipitation of 120 mm usually occurs in December, while the minimum monthly rainfall of 25 mm occurs in August. This study was conducted between May 2023 and May 2024. The amount of precipitation during this period was measured at 965 mm [27].

2.2. Experimental Plot Characteristics

The measurements were conducted in 18 runoff and erosion test plots dimensioned 2 m × 3 m and equipped with waterproof walls, a flow guide, and a runoff collection tank with a capacity of 20 L (Figure 2). Three species were investigated in this study: Alnus glutinosa (L.) Gaertn., Mespilus germanica L. and Crataegus monogyna Jacq. (hereinafter referred to as alder, medlar and hawthorn, respectively) with six replications (plots) established for each species at the study site. The specific species were selected because they are native and naturally grow abundantly in the forest areas of northern Iran. Alder is a tree species, whereas medlar and hawthorn are shrub species. All plots were designated on sites having a loamy soil texture, as well as similar hydrological conditions and physiographical characteristics. The average slope of the plots in this section was 30%. In each plot, plants with an average canopy diameter of 2 m and a collar diameter of 5 cm were chosen. The runoff collection tanks were installed at the lower end of each plot. At the end of each month, the runoff and sediment trapped in the tank were drained and transported to the laboratory. The runoff volume was measured using graduated containers. Sediment particles were separated from the runoff by passing it through filter paper, followed by oven-drying and weighing.

2.3. Soil Measurements

Soil texture was determined using the Bouyoucos hydrometer method [28] in the laboratory. Soil samples from the depth interval of 0–10 cm were collected using a soil hammer and a cylinder dimensioned 5 cm in diameter and 10 cm in height. The collected samples were promptly weighed and were later oven-dried initially at 105 °C for 24 h and then at 65 °C for an additional 48 h [29]. Bulk density was calculated based on the apparent specific gravity of aggregates, using the dry weight of the sample and the volume of water displaced in a water-filled pycnometer. Finally, the apparent specific gravity was calculated by dividing the dry weight of the sample by the saturated volume. Moreover, initial soil moisture content was calculated as Equation (1):

$$\mathrm{SM} (\%)={\displaystyle \frac{{\mathrm{W}}_{\mathrm{w}}-{\mathrm{W}}_{\mathrm{d}}}{{\mathrm{W}}_{\mathrm{d}}}}\times 100$$

(1)

where: SM is the soil moisture content (%), W_w is the wet soil mass (g), and W_d is the dry soil mass (g).

2.4. Root Measurements

To measure root abundance and the Root Area Ratio (the ratio between the area occupied by roots in a unit area of soil in mm² m⁻²), subplots measuring 50 cm × 50 cm × 30 cm were dug at a distance of 10 cm from the tree or shrub base in four cardinal directions (north, east south and west) [26] (Figure 3).

According to our preliminary measurements the root development in depths deeper than 30 cm was limited and unlikely to contribute significantly to water or nutrient uptake in the studied environment. In each subplot, the root number, root length and mean diameter of all roots were measured with a caliper with an accuracy of 0.01 mm. The ratio of the root area to the soil area was calculated using Equation (2). Assuming all roots were circular in cross-section, root diameter was measured at the midpoint of each root.

$$\mathrm{R}\mathrm{A}\mathrm{R}={\displaystyle \frac{{\mathrm{A}}_{\mathrm{r}}}{{\mathrm{A}}_{\mathrm{w}}}}\times 100={\displaystyle \frac{\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{a}}_{\mathrm{i}}}{{\mathrm{A}}_{\mathrm{w}}}}\times 100$$

(2)

where: A_r is the total root area (cm²), A_w is soil subplot area (cm²), n is the number of roots in the sample, a_i is the cross-sectional area of the ith root in the sample.

2.5. Additional Measurements

The location of each plot was recorded using a GPS device. Furthermore, the depth of the humus layer, and the height of the tree canopy in each of the plots were measured using meters. The canopy cover percentage was measured as the proportion of the ground covered by the vertical proportion of vegetation crowns [30]. Soil organic carbon (SOC) was determined by means of wet oxidation [31,32].

2.6. Statistical Analysis

The Kolmogorov–Smirnov test was used to test data for normality. Analysis of variance (ANOVA) and Bonferroni post-hoc tests were used to identify statistically significant differences among variable means. Moreover, SPSS version 23 (IBM Corp., Armonk, NY, USA) was used to analyze and model the relations of the factors with runoff and soil erosion rates. First, data were normalized and then randomized.

The General Linear Model (GLM) approach was applied to detect the effects of experimental plot attributes on (i) water runoff and (ii) sediment yield. The nominal variable Species was treated as fixed factor, whereas the continuous variables Clay, Silt and Sand percentages, Canopy height, Forest floor, Bulk density, SOC, Soil moisture, Root density and RAR as covariates. The general form of the GLM was:

$$\gamma = \mu + {{\alpha }_{\mathrm{i}}}^{\text{(Species)}} + {\sum }_{\mathrm{k}}{\mathrm{b}}_{\mathrm{k}}{\mathrm{x}}_{\mathrm{k}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(3)

where μ = overall mean, α = fixed effect, β₁, β₂ … βκ, are constants x₁, x₂ … x_k examined variables and ε = random error.

First, the data were screened for outliers. Full factorial models were constructed to assess potential interaction effects on the dependent variables. To evaluate the regression models’ goodness-of-fit, F tests were performed, while t-tests were applied to determine the significance of model coefficients. Then, insignificant factors were removed in order to create reduced models for predictive purposes of runoff and sediment yield, respectively. Validation of models’ normality and homoscedasticity was obtained graphically through Q–Q plots and residuals vs. fitted values plots.

3. Results

3.1. Soil Texture

The texture of all studied samples was heavy clay soil, with clay constituting, in most cases, more than 58% of the sample mass (range: 58.89%–67.89%) and sand more than 20% of it (range: 21.61%–23.39%) (

Table 1. Particle size distribution (%) of soil samples, showing the mean ± standard error for clay (<2 μm), silt (2–50 μm), and sand (50–2000 μm) fractions across different treatments in the research area.

Species	Particle Size Distribution
	Clay (%)	Silt (%)	Sand (%)
Medlar	67.89 ± 1.31	10.50 ± 0.59	21.61 ± 1.31
Hawthorn	60.89 ± 1.71	15.83 ± 1.12	23.28 ± 1.69
Alder	58.89 ± 2.18	17.61 ± 0.91	23.39 ± 2.17

).

3.2. Root Abundance and Root Area Ratio

Hawthorn demonstrates the highest average root number at 34.92 ± 5.46, per 50 × 50 × 30 cm excavated soil subplot significantly surpassing both alder (24.83 ± 7.15 roots) and medlar (19.33 ± 3.69 roots) (F = 47.611, df = 2, p < 0.001). The root abundance ranges further illustrate these differences: hawthorn roots span from 28 to 44, alder roots from 16 to 41, and medlar roots from 13 to 28 per excavated soil subplot (

Table 2. Descriptive statistics for the number of roots per excavated 50 × 50 × 30 cm soil subplots among the species investigated.

Species	N	Mean	SD	S.E.	Range
Medlar	24	19.33 a	3.69	1.12	13–28
Alder	24	24.80 b	7.14	2.12	16–41
Hawthorn	24	34.92 c	5.47	1.64	28–44

Note: N = number of observations; SD = standard deviation; S.E. = standard error of the mean. Different letters denote statistically significant differences among group means.

).

Statistical differences were also found among the three species in terms of their root density (F = 14.827, df = 2, p < 0.001). Hawthorn had the highest density (0.009 roots cm⁻²) followed by medlar (0.006 roots cm⁻²) and alder (0.005 roots cm⁻²) (

Table 3. Descriptives statistics of root density among the investigated species.

Species	N	Mean	SD	S.E.	Range
Medlar	24	0.015 a	0.006	0.001	0.008–0.030
Alder	24	0.021 b	0.005	0.001	0.011–0.031
Hawthorn	24	0.027 c	0.009	0.002	0.014–0.050

Note: N = number of observations; SD = standard deviation; S.E. = standard error of the mean. Different letters denote statistically significant differences among group means.

).

3.3. Runoff and Sediment Yield

The highest runoff was measured in medlar 255.42 ± 245.9 mL m⁻² followed by 176.81 ± 76.6 mL m⁻² in alder and 67.36 ± 21.41 mL m⁻² in hawthorn (Figure 4). The results of the one-way ANOVA test indicated that there are significant differences in the mean runoff among the species studied (F = 193.04, df = 2, p < 0.001). Furthermore, Bonferroni’s post-hoc test revealed that the mean runoff in medlar was significantly higher than in both alder and hawthorn. Among the three species, hawthorn exhibited the lowest mean runoff.

The highest mean sediment yield was observed in medlar (50.24 ± 0.66 g L⁻¹), followed by alder (22.5 ± 0.03 g L⁻¹) and hawthorn (8.23 ± 0.25 g L⁻¹). These differences were statistically significant (F = 5.449, df = 2, p = 0.007) and post-hoc comparisons revealed that the mean runoff in medlar was significantly higher than in both alder and hawthorn. Additionally, hawthorn exhibited the lowest mean runoff compared to the other two species (Figure 5).

3.4. Regression Models

The statistical characteristics of the two developed models are presented in

Table 4. Analysis of the variance table for the GLM examining runoff (Model 1) and sediment yield (Model 2).

Model	Dependent Variable	Source	SS	df	η2p	F	p-Value
1	Runoff	Corrected Model	1,003,310	5	0.659	18.527	<0.001
		Intercept	24,332	1	0.045	2.247	0.140
		Bulk density	185,139	1	0.263	17.094	<0.001
		SOC	129,158	1	0.199	11.925	0.001
		Surface cover	92,919	1	0.152	8.579	0.005
		Species	83,987	2	0.139	3.877	0.027
		Error	519,869	48
2	Sediment yield	Corrected Model	5.580	4	0.529	13.780	<0.001
		Intercept	0.237	1	0.046	2.342	0.132
		Bulk density	1.077	1	0.178	10.639	0.002
		SOC	1.906	1	0.278	18.826	<0.001
		Species	1.280	2	0.205	6.321	<0.001
		Error	4.407	50

Note: SOC = Soil organic content.

and the parameters of their estimates in

Table 5. Parameter estimates for the GLM models.

Model	Adjusted R2	Parameter Estimate	B	S.E.	t		95% CI
						p-Value	Lower Bound	Upper Bound
1	0.623	Intercept	−575,726	419,476	−1.372	0.176	−1,419,139	267,687
		Bulk density	772,288	186,791	4.134	0.000	396,719	1,147,857
		SOC	−119,134	34,499	−3.453	0.001	−188,498	−49,770
		Surface cover	−4084	1394	−2.929	0.005	−6888	−1281
		[species = alder]	−47,736	42,157	−1.132	0.263	−132,499	37,027
		[species = hawthorn]	−114,574	41,858	−2.737	0.009	−198,735	−30,414
		[species = medlar]	0.000
2	0.491	Intercept	−1.697	1.243	−1.365	0.178	−4.194	0.801
		Bulk density	1.849	0.567	3.262	0.002	0.710	2.988
		SOC	−0.429	0.099	−4.339	<0.001	−0.628	−0.230
		[species = alder]	−0.184	0.125	−1.470	0.148	−0.434	0.067
		[species = hawthorn]	−0.406	0.115	−3.526	0.001	−0.638	−0.175
		[species = medlar]	0

Note: SOC = Soil organic content.

. According to Model 1, the main factors affecting water runoff are Species, Bulk density, SOC and Surface cover percentage. In the case of sediment yield (Model 2), its magnitude is largely determined by Species, Bulk density and SOC. In both models, the parameter estimates for alder were lower than those for medlar but were not statistically significant, as indicated by respective confidence intervals and the significance values (Model 1: p = 0.263; Model 2: p = 0.148).

4. Discussion

As initially hypothesized, the three examined species exhibited statistically significant differences in terms of root abundance and RAR. The highest values in both sets of measurements were found in hawthorn, followed by alder and finally medlar. Runoff and sediment yield were observed in the reverse order. This pattern suggests that higher root abundance and RAR may contribute to reducing runoff and sediment yield. A study in four sites located in the UK by Webb et al. [33] showed that trees with dense root systems play an important role in reducing surface runoff and sediment, which is consistent with our results. Also, species with extensive roots were found to be more efficient in controlling surface water and preventing soil erosion, according to previous studies in Italy [34] and Iran [35] which is consistent with our findings for hawthorn, alder, and medlar.

Previous erosion studies have demonstrated the positive effect of increased root density in the forest topsoil layer on soil erosion [36,37,38,39] as well as on forest road slopes [38,40,41]. Roots increase the soil shear strength and the components of soil resistance [38,42]. Furthermore, an increasing density of roots in the soil also increases soil permeability [43,44] by modifying the grain size of the soil and the soil structure. The pores become larger and during rainfall events, facilitating a larger amount of rain to penetrate the soil and runoff decrease [43,45,46,47].

However, the fact that statistically significant differences were found among the three treatments cannot be solely attributed to the different root system architectures. In fact, both root abundance and RAR are not components of the GLM models, possibly suggesting that the magnitude of their effects was lower than those of other factors, despite our initial efforts to create comparable conditions among the three treatments (similar slope, canopy diameter and tree/shrub root collar dimensions). Thus, our findings may suggest that in similar multifactorial settings a larger number of factors should be examined.

Bulk density was identified in both models and positively correlated to runoff and sediment yield. Generally, compacted soils are characterized by higher bulk density and lower porosity [48,49], that alter fluid transport and a range of other ecological functions of the soil [50]. High bulk density values lead to a higher possibility of runoff and erosion formation, especially during heavy rainfall incidents due to saturated hydraulic conductivity [51]. Forest machinery traffic [49,52] has been widely documented as a major source of soil compaction. During various phases of forest operations, forest road slopes are often affected, especially when alternative skid trails are used for wood extraction. Soil organic carbon was also correlated to lower runoff and sediment yield in our study. This is an expected result, as soil organic carbon influences the soil physical properties in several ways [53]. Increased soil organic carbon facilitates faunal activity, thus enhancing soil porosity and soil fertility. Furthermore, fresh organic matter stimulates the activity of macrofauna such as earthworms that create borrows and increase water infiltration [54]. In addition, soil organic carbon is responsible for the stabilization of soil aggregates that has beneficial effects in terms of soil erosion [39,55,56,57].

We also found that increases in surface cover decreased runoff quantities in the experimental plots. Similar finding has been reported by previous research (e.g., Du et al. [55]) even during large rainfall events [58]. Surface cover, physically prevents soil crusting from direct raindrop impact that detaches particles and soil aggregates and modifies the soil surface structure causing compaction [59] thereby enhancing rainwater infiltration and reducing runoff [60]. Thus, it preserves soil structure and increases water infiltration. On the contrary, soil texture differences were not found to be related to runoff and soil erosion. This might be due to the relatively limited differences among the three treatments in terms of soil particle distribution, as all soil samples originated from the same area. Nevertheless, soil classification has generally been shown to have limited effect on infiltration capacity [33]. Instead, interactions between soil fauna (e.g., earthworms), roots, plant species richness and soil structure have been reported to play a more important role of greater importance [61,62]. Opinions among forest managers vary on the choice between tree and shrub species for the protection of forest road slopes. Some experts favor shrub species [34,35,63] based on their higher soil conservation efficiency compared with tree species [42]. On the contrary other studies [64,65] reported comparable runoff and soil erosion values between shrub and tree species. In our study, a shrub species outperformed the other two, while medlar—another shrub—showed the poorest performance. Alder, a tree species was positioned between them. Therefore, we suggest that plant type alone is not the determining factor, rather the special characteristics that a species has and its interaction with the site conditions [66,67,68].

Limitations of the Study

The research provides valuable insights, however, there are certain limitations that need to be acknowledged. One of the primary limitations of our study is the absence of control plots with barren soil, which would allow more a detailed analysis and comparisons between the treated and control plots. This was not possible due to the fact that the slopes of the examined forest road had been planted and the removal of the existing vegetation by our research team would inevitably alter soil structure. Therefore, our findings should be interpreted with this limitation in mind, as the effectiveness of the treatments examined might not be fully captured. Future studies should prioritize obtaining rainfall interaction and soil conductivity measurements to gain a more thorough understanding of similar bioengineering practices on runoff and soil erosion.

Another limitation of our study is the limited generalizability of the findings. Although our research offers valuable insights into the use of shrub and tree species for the protection of forest road slopes from runoff and soil erosion, the results may not necessarily apply to other regions. Various factors, such as soil texture, soil organic carbon, climate, and land management practices, can cause significant variation in soil responses to vegetation. Therefore, additional research is needed to explore soil recovery across different environments and soil types to improve the broader applicability of the findings.

5. Conclusions

The results of this study demonstrate that the type and characteristics of the root systems of different plant species have a significant impact on runoff and soil erosion in forest road slopes. The findings reveal that species with denser root systems and higher RAR, such as hawthorn, were more effective in reducing both runoff and sediment yield. Specifically, hawthorn, with the highest root abundance and RAR, exhibited the lowest runoff and sediment yield compared to medlar and alder. Other factors, such as bulk density and soil organic carbon, also played a crucial role in reducing runoff and sediment yield. These results are consistent with previous studies that highlight the impact of compacted soils with lower organic carbon on increased runoff and erosion. Additionally, surface cover acted as a protective factor, reducing runoff by preventing soil crust formation and enhancing water infiltration, thus improving soil structure. However, the study also indicated that the physical differences in soil texture among the three species had little impact on runoff and erosion, possibly due to the minimal variation in soil particle composition across the samples.

In conclusion, this study emphasizes that the selection of plant species for controlling soil erosion and runoff should not be based solely on plant type (tree vs. shrub) but rather on the specific characteristics of each species, such as root system structure and its interaction with environmental conditions. Overall, the findings highlight the importance of a more comprehensive understanding of the factors influencing soil erosion in forested areas and the role of plant species in managing soil and water resources. A major limitation of this study was the absence of control plots with bare soil, which would have allowed for a more precise comparison of the effectiveness of vegetative treatments versus bare soil. Furthermore, the results of this study may not be directly applicable to other regions with different soil types or climatic conditions. Therefore, future research should focus on investigating the role of other factors, such as canopy interception, plant species diversity, and soil fauna, to gain a more thorough understanding of vegetation’s role in soil conservation. Future studies should also consider exploring the recovery of soil properties in different environments to broaden the applicability of these findings.