International Journal of Civil Infrastructure (IJCI)

ISSN: 2563-8084

1. Introduction

About 80% of world roads are unpaved [1] and it could be more than 90% in rural areas of the developing countries [2]. Those are both gravel and earth roads. The latter are based on compacted natural soils forming the surface layer which accommodates traffic runs. Gravel roads on the other hand may have additional layer of borrowed granular material which is compacted to form the surface layer. Rural roads help rural social and economic growth. Despite this, they attract little investments in engineering and maintenance. As result, they are usable in dry season though with unwanted dust, and become muddy, slippery, with rills, gullies and potholes in rainy season. Only 37% of people in rural areas of developing countries have access to all-weather road within 2 km compared to 94% in developed countries [1], [3]. Sub-Saharan Africa (SSA) is the most vulnerable region with heavier burden on rural women and children [4]. A systematic investigation on erodibility of soils in earth roads was undertaken, key erosion types (Table 2) and factors affecting erodibility (Figures 4 and 5) discussed. Envelopes for soil erodibility trends due to both raindrop energy and subsequent flow velocity (Figure 7) have been given to shed light on necessary measures for combating erosion in those roads that are vital for developing countries. Moreover, the laboratory testing of sandy soil mixed with its 0%, 5%, 10%, 15% and 20%, and testing of a granular subbase material have been conducted with results confirming the findings of the literature.

2. The global meaning of earth roads

The poor condition of rural roads negatively affect the development due to their impact to the social, economic and development factors which hugely depend on the movement of people and goods. For example, Tanzania (TZ) in 1988 lost 50% of cotton harvest in three regions, 80% of rice paddy in one region and more than 50% of seeds and fertilizers in another region due to poor rural transport [4]. Currently, 10% to 40% of TZ’s agricultural harvest cannot be moved to desired markets and Cook et al. [5] attribute 89% of the problem to poor rural roads condition. Table 1 gives percentages of earth roads in selected countries to highlight their importance. Percentages of unpaved roads in the world, in SSA, and in countries such as Brazil, Canada and USA show that the good condition of those roads is imperative for the wellbeing of several communities.

Table 1: Unpaved roads in the world, SSA and eight selected countries with long road networks.

	[3]				[6]
	1998		2008		2017
	Length (*106 km)	Unpaved (%)	Length (*106 km)	Unpaved (%)	Length (*106 km)	Unpaved (%)
World	29.912	49.2	33.839	42.6	64.285	≥ 70
USA	6.310	41.0	6.494	32.6	6.586	34.6
China	2.210	82.0	3.730	46.5	4.577	11.6
India	3.010	51.0	3.320	48.0	4.699	39.0
Brazil	1.630	90.4	1.633	87.1	1.580	86.5
Japan	1.152	25.1	1.204	20.2	1.218	18.5
Canada	0.902	64.7	1.042	60.1	1.042	60.1
France	0.893	0.0	1.027	0.0	1.028	0.0
Russia	0.948	22.8	0.982	21.0	1.283	27.7
SSA	1.837	86.9	2.296	84.8	2.380	80.6

3. A permanent threat to earth roads

The integrity of unpaved roads depends on factors which together resist natural and functional stresses. Usually earth roads are made of locally selected soils which may have to be stabilized to gain engineering properties for construction. They need to be built and maintained in accordance with suitable standards and procedures. Those roads can fail due to lack of bearing capacity, overloading and surface erosion. The latter is manifested in formation of potholes, rills and gullies which if not addressed can make the road impassable. This study grouped 564 erosion studies into six categories as shown in Figure 1. Only about 15% of these related to unpaved roads. Further, 71 studies on surface erosion relevant to earth roads were detailed (Figure 2). 34% and 33% of those focused on sheet and rill erosion respectively while the remainder covered splash and gully erosions.

3.1. Splash erosion

Known as rainfall erosion [7], [8], rain impact and rain induced erosion [9], splash is the starting point of surface water erosion [10]. Soil particles are detached by the raindrops impact kinetic energy (KE) [11]. This KE is absorbed by deforming, wetting, dislodging and upward reactive forces [12], [13], [14]. The reactive force entrains and moves particles but reduces due to its sensitivity to wind, soil type and soil water functions [12]. Detachment processes differ between bare and vegetated slopes with the latter dictated by plant canopy, leaf interception and raindrop size [15] while roughness, density and humidity are key for bare slopes. Therefore, earth roads are erosion detachment limited but covered slopes are transport limited [16].

3.2. Sheet (inter-rill) erosion

Inter-rill erosion occurs if there is enough rain to create surface flow. It is often combined with splash [17] because both are rain detachment affected [18]. However, there is a time gap between splash detachment [19] and overland flow start since thresholds for flow must be met. Sheet erosion affects topmost surface soils [17] with flow stresses detaching loose soils and moving these downslopes. Sheet detachment depends more on rain KE [20] and less on water flow stresses [21].

3.3. Rill erosion

Rill erosion results from sheet flow concentrating into small streams. Rills are narrow and shallow [22], [23] and increase in size as both traffic and rains increase. Rill and sheet erosions differ by detachment and transport processes. The rill erodibility depends on concentrated shear stresses [12] that become greater than critical stresses to detach soil particles. Most sediment on bare slopes is generated by rill erosion [18]. Entrainment and deposition refer to the mass of detached soils and disposed respectively. The net detachment is the activity of excess hydraulic stresses to critical stresses [24]. Also, net sediment deposition is the difference between deposition and entrainment when the former is greater, and the opposite way gives net erosion. Although unlikely, erosion equilibrium happens when entrainment and deposition equate [25].

3.4. Gully erosion

It is an advanced concentrated erosion. Gullies are wide and deep usually with tension cracks and cliffs [22]. These may form if rills are not treated [26] and can destroy the road [27]. [28] describe extreme gullies in Nigeria measuring up to 150m in depth, 0.4m to 5.6km wide and up to 2.5km long. [29] subdivide erosion types on slopes such as earth roads (Figure 3).

Table 2. Water erosion and erosive forces.

Type	Sub-type	Key erosive forces
Water Erosion	Splash erosion	Rainfall kinetic energy [9]
	Sheet erosion	Rainfall kinetic energy and flow shear stresses [30]
	Rill erosion	Concentrated flow shear stresses [24]
	Gully erosion	Concentrated flow shear stresses [17]

4. Interactive factors affecting erosion in earth road

A study underway at the University of Birmingham has identified more than 2200 studies of soil erosion. These were systematically screened and narrowed to 564 good studies from which 99873 data were analysed in detail. The screening process based on the meaningful titles and abstracts at first; then on the methodology and results of studies. Inclusion and exclusion criteria that allowed to retain only studies that dealt with laboratory and field-based investigations on erosion processes and measurements were used. In this way, 219 studies were deliberately put aside for further analysis that helped to identify key factors affecting erodibility of soils in earth roads. Those can be grouped into environment and climate; geology and geotechnical; and road and traffic factors as it is shown in the Figure 4.

4.1. Geology and Geotechnical factors

Figure 5 shows that geology and soil lead the understanding of erosion. 82%, 32% and 31% of studies respectively relate particle size distribution, clay percent and index properties to erodibility. 26%, 24%, 23%, 22% and 21% of studies argued that shear strength, bulk density, organic content and compaction respectively influence erodibility in earth roads. Also, infiltration, permeability, particle stability, salts content, mineralogy and consolidation were reported to influence erosion.

Tests on loamy sand, silt loam and clay loam [31], loess [20], mixes of kaolin and sand, and kaolin, silt and sand [32] and consolidated sandy loam [33] showed that erodibility decreases as clay content and plasticity index (PI) [34] increase. [12] states that there is no single soil property that either does not impact soil erosion or that can alone be used to predict erodibility. [35], [36], [37], [38] argue that silt and fine sand erode more than gravel and clay soils due to weight and cohesion respectively. The soil shear strength that resists erosion stress decreases with increasing moisture [39], leading to formation of ruts and rills [40]. Figure 6 relates erosion rate (𝐸𝑟) and shear stresses (τ) to PI as recorded during erosion tests on 11 soils. It shows that 𝐸𝑟 decreases with increase in PI whilst τ increase with PI, though relations are a bit tenuous due to limited data.

4.2. Environment and Climate factors

Environment and climate erosion factors include rain, rain intensity and duration, raindrop size, raindrop shape, rain falling height, weathering, flow stresses, water chemistry, freeze-thaw, humidity, wet - dry cycles and slope properties. Rain, its intensity and duration are with more effects on soil erodibility with 56%, 53% and 46% of studies respectively (Figure 5). Flow shear stresses and stream power were reported in 24% of studies each while 23% and 12% of studies argued impact of road position within a slope to erosion. Heavy rains increased erosion on silt loam and clay loam soils [41], [7] and on sandy loam and loam soils [41]. More erosion was reported due to high rain intensity [42], [43] and duration [44], [45], [46]. Usually, rain detaching energy depends on drops size, shape and velocity, wind and drop surface striking angle [47]. Salles et al. [19] studied erosivity in terms of raindrops size and velocity, and particle detachment energy on fine sand and silt loam. The study shows that energy decreases from clay to silts and then increases with particle size. Figure 7 gives an envelope for KE thresholds for splash detachment with respect to soil mean particle size (D50) and another one for sheet erosion critical velocity versus soil particle size. Also, splash, sheet and rill erosions increase with stream and unit stream powers [35], [48], [49], [50], [51].

Table 3. Factors and erosion trends in unsurfaced roads.

4.3. Road and Traffic factors

Traffic wheels disturb the surface and avail loose material for entrainment before and during the rainfall [48], [52], [53]. Other road factors include cuts and fills which can divert both surface flow and groundwater regimes; traffic volume, type and frequency which logically define the level of surface disturbance by traffic wheels; road geometry and size which can affect both amount and velocity of water flow at the road surface, and the surface roughness which can decelerate the flow and promote surface infiltration. Road length, geometry, drainage and maintenance activities are argued by 26%, 23%, 13%, 13% and 12% of studies respectively as influential to earth roads erosion (Figure 5). The ruts influence on erosion was reported in 7% of studies. Table 3 gives trends of erosion (E) due to some factors (F) in terms of increase (I) and decrease (D) generally.

5. Laboratory experiments

Erodibility tests were conducted in the civil engineering laboratory of the University of Birmingham, UK. These were performed on six various soils, namely a sand (S) soil mixed with its 0%, 5%, 10%, 15% and 20% china clay (CC). The sixth soil was a subbase material. Soils were compacted at maximum dry densities to simulate suitable densities in earth road construction. Moreover, unsoaked CBR tests were carried out immediately after compaction, and after one, five, ten and fifteen days to investigate the impact of clay content on the load bearing capacity. For sand - china clay mixes, CBR increased with increasing China clay content and with increase in time after compaction as in Figure 8. Subbase CBR values were 21.1%, 36.5%, 62.3%, 74.2 and 77% respectively after compaction and after one, five, ten and fifteen days. CBR immediately after compaction and after fifteen days are shown in Table 4.

Two slopes, 0% and 6% were chosen according to Petts et al. [68] who recommends slopes not greater than 6% for earth road. A soil testing box of dimensions 0.6m x 0.3m x 0.17m length, width and height respectively was deemed suitable [69]. Erodibility tests were carried out using a 30 mm/hr simulated rainfall for 30 minutes. Ngezahayo et al. [70] provides details on both the design of the rainfall simulator and the soil testing box.

Table 3. Factors and erosion trends in unsurfaced roads.

Properties	S+0%CC	S+5%CC	S+10%CC	S+15%CC	S+20%CC	Subbase
Clay (%)	0	5	10	15	20	2
Silt (%)	3	3	3	3	3	5
Sand (%)	87	84	80	74	72	15
Gravel (%)	10	8	7	6	5	78
PI (%)	0	4.2	6.8	9.4	12.2	0
D50 (mm)	0.54	0.42	0.46	0.48	0.49	6
OMC (%)	10.5	10.3	9.6	8.9	9.4	8.8
MDD (Mg/m3)	1.82	1.94	1.95	2.03	2.1	2.19
CBR after compaction (%)	0.47	3.1	4.3	6.1	6.5	21.1
CBR after 15 days (%)	1.1	9.8	17.6	19.8	22.5	77
OC (%)	0.4	0.9	1.3	1.5	2.3	3.5
k (cm/s)	2.62*10-2	2.1*10-3	3.06*10-4	2.29*10-4	1.55*10-4	6.5*10-2
PI = plasticity index, D50 = mean particle size, OMC = optimum moisture content, MDD = maximum dry density, CBR = California Bearing Ratio, OC = organic content, k = coefficient of permeability

Sediments from erodibility tests were collected every five minutes and oven dried before being weighed up and sieved for particle size distribution analysis. Major findings are that erodibility reduced with clay content increase in (S+CC) mix soils and with slope reduction. The first 5 minutes of the rainfall were characterised by particle detachment, followed by surface saturation and runoff formation. The maximum sediment was collected between 5 and 15 minutes of the rainfall, suggesting that most particles (silt and fine sand) are detached and washed within the first 15 minutes of the rainfall in earth roads. The first 5 minutes were of higher infiltration rate for soils with less clay content (S+0%CC and S+5%CC), therefore less erosion was observed. The bigger particle size resisted detachment from rain drops, thus the subbase soil was less erodible when compared to (S+CC) soils.

Among the findings, the particle size distribution curves [70] showed that the soil particle size in the sediment gradually increased with increase in the rainfall duration. The sediment collected after the first five minutes of rainfall contained more fine soils (clay and silt) than the sediment collected after 10 minutes of rainfall. This rule of erosion was respected until the last sediment collected after 30 minutes of rainfall which was mainly of granular soil particles (mainly medium and coarse sand). It was also obvious from the reduction in the cloudiness of the runoff as the rainfall duration increased to indicate that generally fine materials eroded before more granular soil particles.

Results are plotted in the Figures 9, 10 and 11 for the tests done on slope = 6%, and in the Figures 12, 13 and 14 for the tests done on the slope = 0%. Respectively, results are for sediment quantities collected at every five minutes intervals; cumulative sediment quantity and cumulative erosion rate for the 30 minutes erodibility tests under simulated rainfall intensity of 30 mm/hr. The minimum quantities of sediment were collected after the first 5 minutes of rainfall due to raindrops having to detach particles, infiltrate the surface, saturate the surface and then create surface flow. The next 10 to 15 minutes were of higher runoff and most detached particles would be transported during this period after which sediment transport decreased gradually.

6. Implication for earth roads

Even though a detailed analysis of published data is reported elsewhere, the findings in this paper show that soils used at the surface of earth roads need very careful consideration as they are most affected by both the climate (temperatures, rainfall) and traffic loadings. The latter impose shear stresses that can dislodge surface particles, which when exposed to dry conditions may be eroded by natural wind or that generated by moving vehicles. If the dislodged soils are exposed to rainwater, they may be transported due to splashing, or surface water flow. Also, inundated surface soils may be dislodged due to traffic wheel-soil interaction. Therefore, in addition to the ability of soils to support traffic loadings and traffic speed as key design factors, designs need also to consider soil-wheel interactions. Since it is not possible to control rain intensity and duration, it is possible to engineer road design to minimise the loss of soils due to erosion. In addition, these designs need to pay particular attention to drainage, surface slopes and degree of compaction. In some cases, regardless of the level of compactions soils will remain highly susceptible to erosion. In such instances, particle size of soil may need to be modified to ensure that the road is constructed using well graded interlocking granular soils that achieve high compaction levels. Also, the amount of clay content in the uppermost surface layer must be controlled to have the plasticity index of about 10% [67], [71] as such soil will resist detachment due to rainfall by both improved surface cohesion and high level of compaction. Where such soils cannot be available, some form of soil stabilisation techniques may need to be used.

7. Conclusion

Bulk of the 71 studies found that interrill and rill erosion were the most prominent types of erosion that cause most of earth road failures. Judging by both the number of studies and the laboratory experiments results, the five most important factors that affect erosion in those roads were particle size distribution of the soil, slope, amount of rainfall, rainfall intensity and duration of rainfall. Further, the literature showed that soils with higher plasticity needed higher critical shear stresses to initiate erosion in those soils. This agrees with laboratory results which showed less erosion for soils with more clay content, thus higher plasticity. Those soils resisted both the raindrops kinetic energy which detaches particles and the subsequent flow shear stresses better than soils with less clay content and lower plasticity. Low plasticity soils, comprising silts and fine sands, are likely to erode. Dislodgement of particles due to rain drops shows that silts and fine sands are likely to be easily dislodged by both drops kinetic energy and surface flow stresses. Larger and finer soils particles need higher energy to be dislodged and it could be due to the robustness of particles and cohesion forces respectively.

Acknowledgements

The Commonwealth Scholarship Commission in the UK are acknowledged for financial support. The authors would like to thank the reviewers for their thoughtful comments and efforts towards improving the manuscript.