Aggregate (crushed rock) is one of the most common surfacing materials used on low volume forest roads [1
]. It reduces both wheel load stress to the subgrade material and maintenance costs, and provides better driving comfort than native-surface roads [3
]. It also helps reduce sediment production from forest roads [4
]. However, road aggregate deteriorates due to traffic, weather, and material properties [7
], and, as a consequence, increased sedimentation occurs. Aggregate material can also be deformed and lost from the road surface due to vehicle traffic which then requires additional road maintenance. Therefore, it is important to understand how forest road traffic changes aggregate properties and movement (i.e., traffic-induced processes that change forest road aggregate).
Traffic-induced processes have been speculated on, and occasionally discussed, as part of field-based observations. For example, Reid and Dunne [8
] mentioned (1) “breakdown of the surfacing material” and (2) “forcing upward of fine-grained sediment from the roadbed as traffic pushes the surfacing gravels into the bed”. Many other research studies have observed and discussed these processes, but often in varying terms. For example, the breakdown of road surfacing material (aggregate) was observed by Swift [4
], described as “crushing” [9
], and referred to as “powdering by traffic”, “particle attrition”, or “abrasion” [4
]. It has also been described as “the mechanical degradation of surface aggregate under traffic loading” [13
]. The second process, forcing upward of fine-grained sediment from the roadbed, was also observed by Swift [4
], and has been further described as “piping” (enabling roadbed soil to pipe through the road prism; [6
]), “pressing larger particles down through a matrix of fine sediment” [9
], “churning” [10
], “fine materials move through the pore spaces to the surface” [14
], “pumping of fine particles onto the surface” [12
], and “subgrade mixing” [13
]. Our paper will refer to the breakdown of surfacing material as “crushing” and the forcing upward of fine-grained material as “subgrade mixing”.
In addition to crushing and subgrade mixing, there is another traffic-induced process that alters forest road aggregate. Traffic moves loose aggregate material off the road by tire action [15
], which is called “sweeping”. Traffic-induced sweeping results in the movement of loose aggregate particles to the roadside and shoulder. This is particularly prevalent at sharp corners where road surface aggregate tends to accumulate at the roadside (Figure 1
]. Two extreme cases of sweeping can be considered: sweeping-out (losing large-size particles, such as gravel, near tire tracks) and sweeping-in (receiving large-size particles near the shoulder).
It is difficult to investigate traffic-induced processes directly, because they can be related to many factors, such as the physical properties of road aggregate (e.g., particle size and strength) [13
], tire pressure [20
], road conditions (e.g., moisture content and compaction) [11
], road gradient [6
], road type (mainline or secondary road) [6
], traffic use [6
], and proximity to the root systems of neighboring trees [24
]. However, the particle-size distribution (PSD) of road aggregate can be used as a surrogate to infer the traffic-induced processes of crushing, subgrade mixing, and sweeping.
Particle-size distribution is likely one of the most important characteristics of mineral soil and road aggregates and is routinely used to evaluate if the soil or aggregate is appropriate as an engineering material for road construction [2
]. A PSD consists of a number of particle-size fractions (PSFs) for individual particle-size classes and is usually plotted as a cumulative frequency diagram in a semi-logarithmic scale (i.e., particle size using a logarithmic scale on the x
-axis and percent passing on the y
-axis). When traffic-induced processes such as crushing occur, smaller-size particles (fine sediment) are produced, changing the PSD. Therefore, the PSD can be used to infer the traffic-induced processes.
There have been few studies that have analyzed the soil and aggregate PSDs that are used for forest and gravel roads. In New Brunswick, Canada, surface and subgrade materials from two forest roads were analyzed using PSD [26
], and two resurfacing methods were compared using PSDs from unpaved roads [16
]. In addition, Foltz and Truebe [18
] analyzed various aggregate material from the western USA and used PSD to evaluate aggregate quality. The PSD method has also been used to characterize surface and subgrade (base) materials in South Dakota, USA [27
]. In Lithuania, PSDs were observed to become finer as gravel pavement wears away [28
Understanding forest road PSDs can elucidate traffic-induced changes in aggregate size and distribution, and their processes. However, in previous studies, soil and aggregate materials were collected from limited cross-sectional locations. More importantly, previous studies did not investigate road aggregate PSD changes after traffic. To fully understand traffic effects on forest road aggregate, PSDs should be compared from various vertical and horizontal cross-sectional locations as well as after different traffic uses. We could not find any publications in which this is investigated. Understanding traffic effects on aggregate material can help reduce road maintenance efforts and costs, and help identify the sediment source (i.e., road aggregate or roadbed material) and define adequate mitigation measures to reduce sediment production by traffic from forest roads.
We hypothesize that all three traffic-induced processes (crushing, subgrade mixing, and sweeping) could alter the physical properties of forest road aggregate. Therefore, the goal of this study was to understand how traffic changes forest road aggregate PSD and, in particular, the magnitude of those changes for various traffic uses, thus contributing to a better understanding of forest roads, traffic, erosion, and their inter-relationships. The specific objectives were to (1) compare forest road aggregate PSDs with respect to vertical and horizontal location within a forest road cross-section and traffic use, and (2) infer how the traffic-induced processes changed the PSDs.
The study results help infer crushing, subgrade mixing, and sweeping-in processes by comparing PSDs from the quarry and forest road aggregate. We inferred that crushing occurred at U-T-L, U-I-L, and all cross-sectional locations on the H road, because the PSDs from these locations and traffic uses were finer than from the other locations. Crushing was only observed near the tire track and road surface (U-T and U-I) on the L road (Table 3
) and was observed at all cross-sectional locations with increasing traffic use (H) (Table 4
), indicating that the crushing occurred near the tire track on the road surface first, than in the deeper aggregate. Subgrade mixing might have occurred at B-S-H, because the PSD here was finer than the rest of the PSDs on H and the location of subgrade mixing (B-S-H) was the bottom layer (Table 4
). Also the subgrade PSD was much finer than any of the road aggregate PSDs (Figure 5
), confirming a strong likelihood of subgrade mixing. Considering that crushing occurred at the other cross-sectional locations on H, we inferred that both crushing and subgrade mixing occurred at B-S-H. We also inferred that sweeping occurred at U-S-L, because the PSD at U-S-L was coarser than at the other locations on the L road and the location of the sweeping-in (U-S-L) was the shoulder on the road surface (Table 3
Our study results partially agree with the results from Toman and Skaugset [13
] in which they observed that subgrade mixing did not occur on three forest roads in northwestern California and Oregon. However, our data indicate subgrade mixing might have occurred at one cross-sectional location on the H road. In addition, Toman and Skaugset [13
] speculated that fine sediment came from either the surface aggregate itself or its breakdown by traffic, which agrees with our data that crushing is the dominant traffic-induced process influencing sediment production from forest roads.
No ruts were observed from the study roads and, therefore, 0.1 m deep aggregate was enough to prevent soil loss from the roadbed. Swift [4
] reported that 0.05 m of aggregate depth was not enough, but 0.15 m prevented ruts and reduced soil loss from forest roads. When designing an aggregate-surfaced road, 0.1 m of minimum aggregate thickness is often chosen [1
] and our study supports these previous findings.
Stress from vehicle loads causes the physical properties of forest road aggregate, such as PSD, to change. Compressive stress on a forest road is concentrated under the loading axis and is highest at or near the road surface [42
]. We found that crushing occurred in the upper vertical layer at the tire track and in-between locations (U-T-L and U-I-L) where light traffic tires passed. Light traffic could pass on tire track and in-between locations, because light traffic vehicles have relatively narrow axle width (1.6–1.8 m) and the forest roads were relatively wide (5 m) leaving a large surface for driving. However, light traffic did not provide enough stress to crush the road aggregate below the upper layer (U). Likewise, with heavy traffic, crushing occurred at all cross-sectional locations, indicating that traffic passed on all horizontal locations including the shoulder. Heavy traffic vehicles have wide axle widths (1.8–2.2 m [44
]) and many often have dual tires. Some heavy vehicles, such as Commercial Motor Vehicles (CMVs), are even wider (up to 2.6 m [45
]). Overloaded vehicles used by the USDA Forest Service included a tracked loader with a width of 3.7 m and an axle load of 41 Mg [44
]. In addition to heavy loads, tracked loaders do not use inflated tires but use steel tracks that likely crush the aggregate surface. Therefore, heavy logging vehicles and equipment can exert more structural damage (e.g., crushing) to a forest road than light traffic. For example, a passenger vehicle weighing 1.8 Mg with a tire pressure of 207 kPa (30 psi) would need to pass 528 times to cause the same structural damage to a forest road as a single pass by a standard 36 Mg log truck with a typical tire pressure of 550 kPa (80 psi [46
]. Therefore, heavy traffic, including logging equipment, can provide enough stress to change the PSDs of all cross-sectional locations of forest road aggregate.
When there is enough stress, aggregate material may be crushed or moved down into the finer-textured subgrade material, but its movement is dependent on material strength, particle size of the aggregate, and road conditions such as water content and compaction. The material strength of cohesive soils and rocks can be approximated by using uniaxial compression strength [47
]. Forest roads are usually well compacted with a uniaxial compression strength of 0.6–1 MPa (very stiff soil; very difficult to move with hand pick, pneumatic tool needed for excavation), ranging up to over 200 MPa (very strong rock; quartzite, dolerite, gabbro, and basalt). The normal and shear stress from traffic might not be enough to crush forest road aggregate particles, especially below the upper layer. However, the contact stress between tire and aggregate particle or between aggregate particles, might be enough to cause the crushing [48
] because aggregate particles have polyhedron shapes with many flat faces and sharp edges [49
], resulting in high contact stresses produced when the sharp edges are pressed. Farmani et al. [50
] found that higher contact stress was distributed across the areas between large-size aggregate particles, indicating that this type of stress is more likely to break large-size particles. A material often loses its strength over time when experiencing stresses lower than the level that would cause instantaneous failure. This “fatigue” ultimately leads to fracturing if the stresses continue [51
]. In nature, fatigue is a major factor in the physical weathering processes of rocks [52
]. Erarslan and Williams [53
] noted that static and cyclic loading due to vehicle-induced vibrations and traffic often caused rocks to fail at a lower stress. Vehicles used in forest operations, especially log trucks, have multiple axles and can exert repeated, cyclic loadings on the road aggregate. Therefore, traffic can cause the aggregate particles to break by fatigue failure. The roads we used had basalt as the aggregate material because this material is locally available and commonly used to surface forest roads throughout the Pacific Northwest [19
]. There are a variety of aggregate tests developed by ASTM and AASHTO [18
], including Los Angeles Abrasion (AASHTO T96 standard [54
]) to measure aggregate’s resistance to crushing (mechanical breakdown) from traffic; however, they are beyond the scope of this study. Even if the study roads had a stronger aggregate material, heavy traffic and logging equipment could change the PSDs of some or all of the cross-sectional locations to a 0.1 m depth.
The heavy traffic road (H) was subjected to application of calcium chloride for dust abatement. Calcium chloride is used to help hold fine particles on the road surface together, thus reducing dust, surface raveling (loose aggregate), washboarding (corrugations), and maintenance costs [55
]. Therefore, the effects of calcium chloride on subsequent PSDs was likely limited to the road surface. Since our work indicates no significant differences in PSDs from the upper and middle layers on the H road, we could not detect the effect of calcium chloride on PSDs using an aggregate sampling depth of 33.3 mm. A different sampling method might be needed to investigate the effect of dust abatement on road aggregate PSDs.
Particle size is an important factor for aggregate crushing. Soil and rock (aggregate) strength increases with decreasing particle size [58
]. Forest road aggregate usually consists of various-sized particles to meet the Forest Service’s specifications for surface course aggregate [2
]. Stress from traffic likely crushes larger-size particles into smaller ones within the road aggregate. Our data shows the changed PSDs did not get finer than a certain PSD limit, close to the particle-size distribution curves at U-I-H or B-I-H, except the particle-size distribution at B-S-H where subgrade mixing likely occurred in addition to crushing (Figure 5
). This curve is the limit to which the stress from traffic can crush road aggregate. Traffic crushes and compacts road aggregate, and subsequently makes aggregate PSDs finer. Therefore, we can consider this PSD limit as the optimum compaction by crushing for a given road condition and aggregate material. A similar concept was noted by Fuller and Thompson [61
] who described an idealized grading that represented the densest state of packing particles (Fuller packing) [62
]. Once the aggregate reaches this limit, stress from traffic is delivered, without further changing the PSD, to the aggregate below or to the subgrade where more crushing or subgrade mixing may occur.
Aggregate is made by crushing rocks in a quarry and is delivered to a road construction site by vehicles such as dump trucks. It is then dumped out of the vehicle and spread on subgrade material. During this process, particle-size segregation might occur; as observed in some road aggregates [17
]. While road aggregate is being delivered, vehicles vibrate and shake aggregate particles, which might potentially cause particle-size segregation. This particle-size segregation is often called the “Brazil-nut” effect which originated from the observation that shaking mixed nuts made the Brazil nuts (the biggest) move upward and end up on top of the mixed nuts [64
]. This phenomenon occurs for granular materials having a wide size distribution, even at the molecular level, making large particles rise to the top of a mixture, and fine particles move downwards [64
]. However, the shapes of the aggregate particles are polyhedrons [49
], thus there can be interlocking of the aggregate particles which prohibits movement. Therefore, it is not known if particle-size segregation occurs during aggregate transport. Once aggregate arrives at the road construction site it is dumped out of the vehicle. When aggregate particles are dumped, large particles tend to roll down to the edge of the pile causing particle-size segregation. This is similar to a scree slope with rockfall sorting in which larger particles are at the bottom edge of the slope [67
] and has been reported when granular materials with different particle properties (e.g., size, density, shape, resilience, angle of repose, and cohesiveness) are deposited [65
]. Aggregate dumped and spread on subgrade likely has the same particle-size segregation occurring at road edges (i.e., shoulder). Interestingly, this study points out some inconsistencies. On the N road, PSDs were uniform (Table 2
). However, at U-T-N, U-S-N, and M-S-N they were coarser than the PSD from the quarry (Table 5
), indicating that particle-size segregation may occur from the quarry to the newly constructed road.
Root and soil interactions may also be another cause of changes in road aggregate, because vegetation root systems can loosen compacted road surface [69
]. Also, roots from nearby trees can cause upheaval or displacement of the road surface [24
]. This road surface change has particularly been noted on paved roads, but could also occur on unpaved roads, such as aggregate or native-soil-surfaced roads. However, we did not test for this on our road surfaces because of the relatively short timeframe that our study sites had been in use.
Knowledge of traffic-induced changes to forest road aggregate PSD is important for reducing road maintenance efforts and sediment production from forest roads. This study provides new data to help understand traffic-induced changes in PSD. Increased traffic changed the PSDs by all three processes we examined: crushing, subgrade mixing, and sweeping. We inferred these traffic-induced processes by comparing forest road aggregate PSDs from vertical and horizontal locations, and traffic uses. No differences in PSD were found on the no traffic road, indicating no processes occurred. Light traffic made the PSDs at U-T and U-I finer, indicating crushing occurred, and made the PSD at U-S coarser, indicating sweeping occurred. Heavy traffic and logging equipment with wide axles made all the PSDs finer, which indicates crushing occurred at all cross-sectional locations. In addition, the PSD at B-S was finer than the other locations and can be explained by subgrade mixing in the bottom layer at the shoulder.
Crushing appears to have been the dominant process on our study roads based on PSD comparisons. Crushing changed the PSDs up to a certain limit, but no finer. Having road aggregate close to the PSD limit would reduce aggregate crushing and fine sediment production. Crushing and subsequent compaction can cause the aggregate volume to change, resulting in permanent deformation of road surface with ruts, washboards, and potholes; requiring road maintenance. Therefore, having aggregate close to the PSD limit (i.e., the optimum compaction) will reduce road maintenance. Further, particle-size segregation of road aggregate may occur during transport, dumping, and surfacing the subgrade, but additional work is needed to investigate this segregation and to produce aggregate with more desirable PSDs.
Subgrade mixing can also be a dominant process in other geographical locations with different road conditions, aggregate, and subgrade properties (e.g., soft, weak subgrade, and wet road conditions). Understanding the physical processes on different road conditions will help mitigate sediment production from forest roads and reduce road maintenance efforts by providing information for best management practices. For example, strengthening the surface material (e.g., surface stabilization) is recommended if the dominant process is crushing; strengthening the subgrade (e.g., geotextile reinforcement on subgrade) if subgrade mixing is the dominant process; and collecting and recycling large aggregate particles on the shoulder and roadside for road resurfacing if sweeping occurs excessively. Future study is recommended to investigate traffic-induced processes in other locations where subgrade mixing or sweeping is the dominant process, and the effects of road treatments and management practices on the traffic-induced processes, for better road management.