Introduction

Soil erosion is one of the top-most environmental and ecological concerns in Peninsular Malaysia especially during the northeast monsoon. It causes loss of nutrient-rich topsoil and increases sedimentation, turbidity and level of pollutants in adjacent water bodies (Goodwin and Sin 2014; Martínez-Casasnovas et al. 2016; Vietz et al. 2017). Construction activities and road systems account for the soil erosion and cause impact to the streams adjoining highways (Gernaey et al. 2004; Maier et al. 2004). About 19% of the total land area is eroded due to the construction of public roadway system (Rezaeianzadeh et al. 2014). Some researchers reported that most of the erosion occurred during the rainy season after a disturbance, and half of the erosions were caused by logging operation in road construction. Even during low rainfall events, stream sedimentation is caused by a large volume of overland flow over linearly connected road systems through hydraulic erosion processes. Construction of roads with concrete and bituminous materials changes the dynamics of plant and animal populations. It also alters the flow of materials in the landscape, due to the involvement of exotic materials and changes levels of available resources such as water, light and nutrients (Caraman et al. 2007). Hence, the disturbance of soil during construction activity is considered as a major nonpoint source (NPS) of water pollution by sedimentation (Othman et al. 2015; Van der Helm and Rietveld 2002; Yu et al. 2000). Thus, construction activity is ultimately liable for the exclusion of topsoil, destruction of vegetation, surface runoff and soil erosion (Islam et al. 2018; Lai et al. 2008; Ozsoy et al. 2012; Wilson et al. 2018; Zeng et al. 2003). In comparison with natural areas, the construction sites increases the erosion by about 500 times in USA (Azamathulla and Ghani 2010).

However, since erosion as a natural process cannot be eliminated completely, control and managing of sediment loading through best management practices (BMPs) might reduce its impact (Azamathulla and Ghani 2010; Maerker et al. 2017). Numerous BMPs like porous pavement materials, vegetated buffers and mulches, retention or detention basins and ponds, silt fences, hydro seeding and the placement of natural fiber mats have been proposed and implemented along highways (Dastorani et al. 2010; Kim et al. 2013; Kumar et al. 2012). Bare soil covering by grass or mulch was proposed by Zhang and Stanley (1999) to control soil erosion. The intensity of surface erosion is related to the land use and soil cover (Boix-Fayos et al. 2009; Firat and Güngör 2007; Gautam and Holz 2001; Hamilton et al. 2017; Zhang et al. 2016). The soil coverage reduces the erosive power of raindrops and increases the volume of water infiltration to the soil surface (Panagos et al. 2015; Piacentini et al. 2018; Talei et al. 2010).

An abundant use of BMPs has been proposed worldwide through different research locations; however, the efficiency of some BMPs is still lacking detailed investigation (Akrami et al. 2014; Zhang and Stanley 1999). Considerable literature exists related to erosion control in a humid environment, but only a few studies evaluated the effect of planted and natural vegetation on sediment loss rate in active and established construction areas of hilly topography nearby highway. An evaluation of various vegetation densities that consist of planted and natural vegetation in relation to surface runoff and soil loss based on plot-scale sites on roadside slope adjacent to a highway is presented in this study. The significance of vegetation cover and the mutual interaction of vegetation cover and slope steepness on soil loss are also investigated.

Study area

Hillslope adjacent of the Guthrie Corridor Expressway (GCE), Malaysia, is chosen as a study area. The annual average precipitation is about 1570 mm with an average minimum and maximum temperatures of 26 °C and 36 °C, respectively. A location map of study area along with five study plots is shown in Fig. 1a, b, and characteristics about each plot area is shown in Table 1. A total five numbers of plots were established at two different locations with vegetation densities varies between 5 and 95%. Out of the five plots, four were of size 8 m × 8 m and one of size 5 m × 5 m. The plots are situated in an elevation range between 45 and 75 m, whereas the slope percentage ranges from 50 to 80%. On the basis of the slope values, the study area can be characterized as prone to soil erosion (Dunne 1977). The soil group classified under the group B implies water infiltration rate is ranging from 3.81 to 7.62 mm h−1. As for the soil texture, it falls under the categories of 2 and 3, i.e., fine and coarse granular.

Fig. 1
figure 1

Location (a) and aerial view (b) of the study area

Full size image
Table 1 Characteristic of the plots
Full size table

Methods

Plot characterization and plant transplantation

In this study, some of the vegetation cover is seeded with microbes. The seeds of Melastoma malabathricum L and Lantana camara L are germinated with a surrounding temperature of 30 °C. About three weeks later, they are placed in a PVC pipe filled with sandy loam soil and placed in a greenhouse at the Institute of Biological Sciences, Faculty of Science, University of Malaya. When the seedlings reached about 1.0 m height, they are then transplanted to the site and a 50 g of commercialized bio-fertilizer is applied a week after transplantation. Bio-fertilizer is buried at 10 cm of soil depth adjacent to the transplanted plants in the microbe treated planted plot (PDM). A plot that received the transplantation but did not use the bio-fertilizer is named as the non-microbe planted plot (PDNM). At some plots, although the transplantation of Melastoma malabathricum L is not applied, this species is still available naturally. By applying bio-fertilizer to the naturally grown plants, the plot is characterized as natural plants with microbes (NDM). A plot that has a few naturally grown plants of similar species but without bio-fertilizer is known as a natural bare non-microbe plot (NBNM). Similarly, a plot with few naturally grown plants and with bio-fertilizer is known as a natural bare microbe plot (NBM). The name assigned on each plot (i.e., PDM, PDNM, NDM, NBNM, NBM) depending on either natural or planted vegetation cover, density of vegetation and also the inclusion of bio-fertilizer which indicates the presence of microbes or not.

Field settings

A tipping bucket rain gauge is installed near to the experimental plots for measuring precipitation accurately. Therefore, a site reconnaissance survey is the first thing to do to identify the most suitable location for the rain gauge installation. As the experimental plots are located on hillslope, the crown of the small territory was considered as a suitable position for the rain gauge installation (Shih 1982). A data logger is attached to the rain gauge to facilitate collection of the precipitation data. For isolating the runoff from different plots, these are enclosed with 25-cm-high wooden barriers. The sketch diagram of the barrier setup and the layout plan of experimental plot along with sediment collector, pipes, barrier, etc., are depicted (Figs. 2, 3, 4). The size designed of sediment collectors to be used for all plots is same. Normally, the volume of sediment collector depends on how much water would be accumulated to it as runoff. And the total volume of runoff water depends upon characteristics of the areas which receive the rainfalls and generate runoff for the given rainfall events. In this research, the volume of sediment collector is taken as 2.75 m3. It is considered to be sufficient in accumulating the runoff and eroded soil for the semi-monthly retention period because the effective areas receiving rainfall events are actually less than the actual plot area since all study plots are situated in slope landscape. In addition, in the study area the frequency of intensive rainfalls that are capable of generating runoff is found very low (Fig. 5).

Fig. 2
figure 2

Sketch of the barrier

Full size image
Fig. 3
figure 3

Layout plan of the experimental plot a location 1, b location 2

Full size image
Fig. 4
figure 4

A sketch of sediment collector

Full size image
Fig. 5
figure 5

Frequency distribution of rainfall events, duration and intensity

Full size image

Rainfall, surface runoff and sediment

As the experiment is performed during the monsoon season, the data for rainfall, runoff and sediment were collected on fortnightly basis. In each observation, the accumulated runoff is measured by volume and the sediment yield is determined by weight. Other than the settled sediments, suspended load in the sedimentation tank is also evaluated using the TSS analyzer for achieving accurate results. The wet eroded soil deposited at the bottom of the tank (after removing the water) is taken out and weighed in the laboratory after drying it at 105 ± 5 °C for 24 h. Rainfall is observed for the monsoon period, which is started from November 2014 to March 2015. The observed rainfall events with intensities that exceed 2.75 mm h−1 are considered as erosive and therefore are used for runoff and erosion computation. In this study, the observed rainfall and soil loss are measured at plots, whereas for the runoff, the SCS method is used to derive this parameter. The SCS runoff curve number method had been used to predict the potential runoff in Alor Gajah and the Jasin basin (Adham et al. 2014). The collected rainfall and sediment data at plot areas are collected every 15th day of the month.

Vegetation cover and topography

A digital camera is used to capture the actual density of vegetation. Using a portable camera, a ground image of the plot is captured vertically from 2.0 m above the ground to examine the distribution of vegetation cover based on the plot basis. Once the image is captured, it is then processed using the maximum likelihood algorithm in the ArcGIS10.1. In each image, a few representative pixels of the two classes namely bare soil and vegetation cover are selected as a reference for the algorithm classification. However, as the different images contain different pixel counts, a weightage area for the two classes is applied in the computation. The percentage of vegetation ground cover is computed by taking a ratio between the area of vegetation cover and the total area (i.e., vegetation cover and bare surface). Despite the fact that photograph images might not have been captured with a particular scale, the results obtained are normalized as all images are acquired in the same manner. Slope steepness is measured based on the digital elevation model (DEM) of the study plot.

Results

Rainfall and runoff patterns

This section presents observation result of recorded rainfall at those five plots. Based on the data collected starting from November 2014 until March 2015, the recorded observed mean monthly rainfall was 310, 273, 220, 77 and 240 mm, respectively. In addition to this, magnitude, duration and intensities of rainfall pattern are also analyzed of which their values are found to vary between 3.6 to 52.6 mm, 15 to 396 min and 1.03 to 49.37 mm h−1, respectively. The rainfall frequency, duration and intensity graphs are depicted (Fig. 5) whereby it can be seen that in most rainfall events the highest recorded rainfall depth was about 5 mm with 20 min duration of rainfall and intensity of 5 mm h−1. The effect of vegetation cover on the surface runoff of hillslope area is investigated as the vegetation cover, and slope steepness have contributed to the amount of runoff directly. In this study, the amount of runoff yielded in relation to the number of rainfall events for all plots is depicted (Fig. 6). Out of five plots, the NBNM plot has yielded the largest amount of runoff followed by the NBM, PDM and PDNM. The amount of surface runoff in plots NBM, PDM and PDNM is almost similar. For the surface runoff at the NBNM plot, the amount is about equal to the event rainfall. This might be due to minimal vegetal cover and therefore less infiltration process into the subsurface to occur. Conversely, only a few events of runoff are observed at the NDM plot. Correlation between rainfall and surface runoff is also shown (Fig. 7), and result shows that most of the R2 values are well correlated with values are nearly 0.99.

Fig. 6
figure 6

Surface runoff and rainfall events of all plots

Full size image
Fig. 7
figure 7

Relationships between rainfall and surface runoff

Full size image

Factors affecting surface runoff and soil loss

This section investigates on the slope steepness and vegetation cover that affect the surface runoff and soil loss. First, in terms of slope steepness, it is found that the slope values of the study plots are varied within considerable range as they are located in hillslope (i.e., 77%, 78%, 68%, 56% and 76% for NBNM, PDM, PDNM, NBM and NDM plots, respectively). Second, the amount of surface runoff for the corresponding sequence plots of slope steepness is found to be 1058.77, 502.42, 438.92, 580.19 and 16.02 mm, respectively. The surface runoff of the respective plots is significantly varied due to difference in their slope steepness and infiltration rate. Infiltration rate is increased simultaneously with the increase in vegetation cover in a particular plot. That’s why the runoff is resulted from the NDM plot which occupied by the vegetation cover of 96% is minimum. Whereas the maximum runoff is obtained from the NBNM plot whose vegetation cover is of only 5% (Table 1). Here, the maximum runoff that resulted from the NBNM plot is found about 9 times higher than that of the NDM plot despite the slope steepness of those plot is nearly equal. Hence, the role of the vegetation cover in reducing runoff is clearly depicted. The rest of plots have demonstrated a positive relationship between runoff and slope steepness regardless of the vegetation cover type. The relationship between runoff and slope steepness with the presence of a vegetation fraction is almost consistent among all plots.

Usually, plots with steeper slope had greater soil loss than those with the lower values of slope steepness. The mean soil loss obtained was about 1833.4, 1173.32, 367.08, 192.42 and 11.92 g m−2 for the NBNM, PDM, PDNM, NBM and NDM, respectively, as depicted (Fig. 8). Erosion occurred at the NBNM was about 1.5, 5 and 9 times more than that of the PDM, PDNM and NBM, respectively. On the contrary, the erosion rate was found to be lowest for the NDM plot owing to high dense developed shrubs. The shrubs make the top soil stable enough for protection against heavy rainfall and any detachment hazards (Greacen and Sands 1980; Khaki et al. 2015). The rate of soil loss increases with slope steepness and vice versa with the existence of vegetal coverage (Chang et al. 2014). Hence, the erosion rate has individually demonstrated a negative and positive relationship with vegetation cover and slope steepness, respectively (Wang et al. 2014; Zhang et al. 2011). In the steepest slope topography, the availability of vegetation cover serves as the best criterion in controlling soil erosion.

Fig. 8
figure 8

Mean soil loss

Full size image

Conclusions

In this study, rainfall-induced soil erosion is assessed based on the observation of soil loss and rainfall data from study plots located on hillslope adjacent to the GCE with the main focus on examining the effect of various vegetation density on the five experimental plots with either natural or planted vegetation and either with or without microbes. As the experiment incorporates natural and planted vegetation and microbes at those plots, the effect is somehow complicated. However, the concern of this study is to investigate various density of vegetation at those plots in relation to the soil loss that occurs during the monsoon season. In terms of vegetation, we can conclude that the naturally developed vegetation types especially fern and sod grass that grows on hillslope can contribute well in reducing the potential erosion. Based on the counted rainfall events data, the general pattern observed about the depth and duration is in most events, and the frequent rainfall depth occurs is about 5 mm with 20 min rainfall duration. As for plots, plot without vegetation cover (i.e., NBNM) yields greater soil loss and with maximum runoff whereas the least is exhibited by the plot having natural dense microbe (NDM). Meanwhile, the planted vegetation is found to be somehow effective in reducing the rate of soil loss. A positive relationship is found between runoff and slope steepness regardless of the vegetation cover type.