Abstract

This study investigates the contribution of road deposited sediment (RDS) to clogging and the operational lifecycle of highway filter drains (HFDs). RDS samples were collected from nine Scottish trunk roads and fractionated into grain size classes to determine their particle size distributions (PSDs). Results show that RDS PSDs, and the percentage of each grain size fraction, are highly variable. However, despite being collected from different trunk roads, PSD trends are similar, with individual RDS particles ranging in size from <63 μm to >10,000 μm. Medium sand, coarse sand, fine gravel and medium gravel make up 84.1% of the total particle mass concentration, with particles >1,000 μm mostly mineral or asphalt. The study also reveals that the dynamic nature of a trunk road catchment dictates that grading envelopes are essentially instantaneous values. These findings indicate that large particles from the road surface contribute to clogging and have the potential to reduce the operational lifecycle of HFDs. The study also demonstrated that assuming a single RDS PSD profile for filter drain asset management purposes is unlikely to be representative of a trunk road catchment profile.

INTRODUCTION

Background

Highway filter drains (HFDs) are a widely used road drainage system and it is estimated that surface water runoff from 43% of the strategic trunk road network in Scotland is serviced by HFDs (Transport Scotland 2016). However, despite their popularity, HFDs are not a ‘fit-and-forget’ drainage system. This is because they are prone to clogging, a process that develops over time and is the result of the physical accumulation of road deposited sediment (RDS) in the graded stone, which results in significantly restricted or terminated flow channels (Bruen et al. 2006).

Clogging is exacerbated by the fact that HFDs have no pre-treatment system. As such, there is no mechanism that permits interception, removal, or storage of RDS contained in road runoff. Without any means to intercept RDS, it has the potential to overload the HFDs' graded stone, leading to clogging and premature failure, with examples shown in Figure 1. As a result, the operational lifecycle of an HFD has been estimated to be only 10 years (Stylianides et al. 2015). It has even been suggested that HFDs prone to frequent vehicular over-run will have an operational lifecycle of around 6 years, because compaction of the graded stone reduces the void space and promotes a cake-layer on the surface of the HFD (Bruen et al. 2006).

Figure 1

Clogged HFDs due to: (a) surface cake-layer, (b) pine needles, (c) vehicle over-run, (d) litter accumulation, (e) wear-and-tear of the high friction surface course, (f) vegetation intrusion.

Figure 1

Clogged HFDs due to: (a) surface cake-layer, (b) pine needles, (c) vehicle over-run, (d) litter accumulation, (e) wear-and-tear of the high friction surface course, (f) vegetation intrusion.

RDS PSD grading envelopes

A review of published research evidences variability in the composition of RDS particle size distribution (PSD) grading envelopes. Ball et al. (1998), investigating the build-up of RDS on a suburban road, reported the median particle size values, d50, ranging from 44 to 91 μm. Viklander (1998), in a study exploring RDS PSD grading envelopes, after periods of snowmelt, reported d50 ranging from 1,000 to 4,000 μm. Sansalone et al. (1998), measuring the physical characteristics of solids transported in lateral road runoff, recorded a d50 ranging from 370 to 785 μm, with a mean of 555 μm. More recently, Adachiwe& Tainosho (2005), collecting RDS from a road and street gutter, reported d50 ranging from 740 to 980 μm.

Sansalone & Tribouillard (1999) and Regenmorter et al. (2002) in studies exploring RDS accumulation on road surfaces have identified that individual RDS particles can range in size from 1.0 μm to greater than 10,000 μm. Similarly, Maglionico & Pollicino (2004), in a study exploring the build-up of RDS on urban road surfaces, demonstrated that RDS is well-sorted in terms of size, with individual particle diameters ranging from 53 to 4,000 μm and d50 ranging from 100 to 600 μm. It has also been shown that particles exceeding 1,000 μm can account for a significant percentage of the RDS mass fraction. Ellis & Revitt (1982), collecting RDS from road and street gutters, determined that particles within the range 500 to 2,000 μm dominate, whilst Sartor & Boyd (1972), found that between 74.1% and 92.3% of the particles were within the range 43 to 4,800 μm, the mean being 86.5%. Sansalone et al. (1997) investigated rainfall runoff from a motorway and found that 30% of the RDS mass was between 1,000 and 10,000 μm. Similarly, Sansalone & Tribouillard (1999) determined that particles with a diameter >10,000 μm can exceed 7% of the RDS mass fraction.

RDS PSD grading envelopes are shaped by both natural phenomena and anthropogenic activities. Natural phenomena are primarily associated with atmospheric deposition (Murphy et al. 2014). Anthropogenic activities are local in nature and are associated with erosion of soil and roadside verges, vegetation detritus from surrounding land use and vehicular motion and related activities. RDS from the latter is derived from vehicle tyre and brake wear, degradation of the road surface course material, corrosion from vehicle and road infrastructure, road maintenance activities, discarded litter etc. (Loganathan et al. 2013; Charters et al. 2016). For example, break wear is predominantly mechanical in nature and produces particles, which vary considerably in size, ranging from ultrafine fraction to coarse fraction (Grigoratos & Martini 2015).

Numerous research studies have demonstrated that dislodgement processes (wind, rain, vehicle over-run, annual-average-daily-traffic (AADT) flow, size of contributing impervious area, level of road maintenance) and hydraulic sorting (wind direction/strength, rain intensity, road gradient and surface roughness) influence RDS PSD grading envelopes. Vaze & Chiew (2002) highlight that due to dislodgement, only about 15% of dry samples collected from a road surface have a particle size <100 μm. Adachi & Tainosho (2005) recorded the median particle size as being between 8% and 10% and Lau & Stenstrom (2001) derived a figure of about 3%. Similarly, Sansalone & Tribouillard (1999) determined that only about 10% of the RDS mass is <50 μm and Ball et al. (1998) determined that the percentage of particles <200 μm ranges from 10% to 30%, with a mean of 16.8%. The aforementioned research does not, however, align with findings by Li et al. (2005) who reported that between 30% and 60% of the particle mass can be found in particles <50 μm or Gunawardana et al. (2014) who determined that more than 70% of the road deposited solids particles at all of the sites they tested are finer than 150 μm. It is therefore possible that differences in catchment characteristics, road dynamics and testing procedures have contributed to the discrepancies in the research noted above.

Differences in the graded stone specification of sustainable drainage systems employing filtration and the PSD grading envelopes of stormwater particulates result in a differing balance between hydraulic profiles and filtration response. Li & Davis (2008) suggest that the critical factor determining the success of straining in filtration systems using a graded stone is the ratio between the diameter of the graded stone and that of the stormwater particulates. Knowledge of RDS PSD grading envelopes is therefore critical to the understanding of clogging and operational lifecycle of HFDs.

However, the mechanisms that govern RDS generation and PSD grading envelopes across strategic trunk road networks are lacking and poorly understood. The current lack of knowledge is a consequence of the difficulty in safely accessing strategic trunk roads for research purposes and highlights a gap in the research and performance data pertaining to this field of study. Research that has been published tends to have been derived from geographical areas outwith the UK and, apart from a few studies, is the result of research on local roads. Moreover, RDS tends to have been collected from roads in urban areas, on one road, at one location. The research studies that are commonly referenced therefore do not have the range of AADT flows that are representative of Scottish trunk road network flows. Similarly, RDS samples have been collected from roads where international standards for the specification of the road surface course differ from those typically used in Scotland. To complicate matters further, differences in sample collection, processing and analysis have resulted in a wide range of concentrations and distributions being reported. Taken together, these inconsistencies suggest that the RDS PSD research commonly referenced may potentially not be representative of current Scottish trunk road grading envelopes, which suggests there is a substantial research information gap.

OBJECTIVES OF STUDY

Research mapping the mechanisms that govern RDS generation and PSD grading envelopes across strategic trunk road networks is limited in number and scope. This is a consequence of the difficulty in safely accessing high-speed trunk road networks for research purposes and highlights a gap in research and performance data pertaining to this field of study.

This paper reports on a field study comparing RDS PSD grading envelopes from across the strategic Scottish trunk road network to determine the variables that govern the generation, spatial variability and PSD grading envelopes of RDS within different trunk road catchments, with the aim of establishing the impact RDS has on clogging and the operational lifecycle of HFDs.

MATERIAL AND METHODS

Site description

In this study, a total of 23 RDS samples were collected from nine trunk roads. These being the A7, A68, A90, A912, M8, M9, M80, M90 and M876 (Figure 2).

Figure 2

Roads where RDS samples were collected.

Figure 2

Roads where RDS samples were collected.

A full spectrum of trunk road layouts, including slip-roads, roundabouts, traffic lights, junctions and straight sections of road, was included. Roads were also selected on the basis that they should be representative of typical surface course design mixtures. All roads were considered 100% impervious, had a cross-fall that diverts runoff directly to the edge of the road and had a clearly defined drainage catchment area comprising only the road surface.

Table 1 summarizes the site information for the 23 trunk road network RDS sampling sites.

Table 1

Site information

Site ID & road classification (A = A road, M = motorway) Eastings & northings Land usea Soil type Speed limit (mph) AADT Road surface materialb Number of running lanes Road surface texture depth (mm) 
A7(1) 349,345, 614,255 U-C alluvial 30 1,686 0.5–0.7 
A7(2) 348,845, 613,720 alluvial 60 1,686 0.7–0.9 
A7(3) 348,845, 613,720 alluvial 60 1,686 0.7–0.9 
A7(4) 349,495, 614,380 alluvial 60 1,686 0.9–1.1 
A68(1) 365,087, 620,115 brown earth 60 11,843 0.7–0.9 
A68(2) 359,300, 629,995 mineral gleys 60 3,294 0.3–0.5 
A90(1) 341,640, 734,298 U-C brown earth 70 31,081 >1.1 
A90(2) 324,355, 726,130 mineral gleys 70 28,545 0.7–0.9 
A90(3) 328,910, 729,550 brown earth 70 26,846 0.7–0.9 
A90(4) 360,750, 763,100 brown earth 70 16,725 0.3–0.5 
A90(5) 360,595, 762,925 brown earth 70 16,725 0.3–0.5 
A912(1) 312,330, 720,584 brown earth 70 8,334 0.5–0.7 
M8(1) 300,453, 668,303 brown earth 70 28,617 >1.1 
M8(2) 302,383, 669,658 U-C-I mineral gleys 70 55,094 0.9–1.1 
M8(3) 296,588, 665,678 brown earth 70 23,243 >1.1 
M9(1) 281,565, 688,235 mineral gleys 70 21,796 0.7–0.9 
M9(2) 294,955, 678,565 brown earth 70 41,741 0.9–1.1 
M9(3) 293,185, 679,550 mineral gleys 70 32,314 0.7–0.9 
M9(4) 290,610, 682,370 U-C-I mineral gleys 70 32,314 0.7–0.9 
M9(5) 281,594, 688,211 mineral gleys 70 32,314 >1.1 
M80(1) 280,229, 680,261 brown earth 70 68,194 0.7–0.9 
M90(1) 283,920, 682,751 humus-iron podzols 70 30,993 0.3–0.5 
M876(1) 312,745, 698,510 alluvial 70 33,705 >1.1 
Site ID & road classification (A = A road, M = motorway) Eastings & northings Land usea Soil type Speed limit (mph) AADT Road surface materialb Number of running lanes Road surface texture depth (mm) 
A7(1) 349,345, 614,255 U-C alluvial 30 1,686 0.5–0.7 
A7(2) 348,845, 613,720 alluvial 60 1,686 0.7–0.9 
A7(3) 348,845, 613,720 alluvial 60 1,686 0.7–0.9 
A7(4) 349,495, 614,380 alluvial 60 1,686 0.9–1.1 
A68(1) 365,087, 620,115 brown earth 60 11,843 0.7–0.9 
A68(2) 359,300, 629,995 mineral gleys 60 3,294 0.3–0.5 
A90(1) 341,640, 734,298 U-C brown earth 70 31,081 >1.1 
A90(2) 324,355, 726,130 mineral gleys 70 28,545 0.7–0.9 
A90(3) 328,910, 729,550 brown earth 70 26,846 0.7–0.9 
A90(4) 360,750, 763,100 brown earth 70 16,725 0.3–0.5 
A90(5) 360,595, 762,925 brown earth 70 16,725 0.3–0.5 
A912(1) 312,330, 720,584 brown earth 70 8,334 0.5–0.7 
M8(1) 300,453, 668,303 brown earth 70 28,617 >1.1 
M8(2) 302,383, 669,658 U-C-I mineral gleys 70 55,094 0.9–1.1 
M8(3) 296,588, 665,678 brown earth 70 23,243 >1.1 
M9(1) 281,565, 688,235 mineral gleys 70 21,796 0.7–0.9 
M9(2) 294,955, 678,565 brown earth 70 41,741 0.9–1.1 
M9(3) 293,185, 679,550 mineral gleys 70 32,314 0.7–0.9 
M9(4) 290,610, 682,370 U-C-I mineral gleys 70 32,314 0.7–0.9 
M9(5) 281,594, 688,211 mineral gleys 70 32,314 >1.1 
M80(1) 280,229, 680,261 brown earth 70 68,194 0.7–0.9 
M90(1) 283,920, 682,751 humus-iron podzols 70 30,993 0.3–0.5 
M876(1) 312,745, 698,510 alluvial 70 33,705 >1.1 

aU – urban, R – rural, I – industrial, C – commercial.

bA – asphaltic, C – concrete.

The field study relied on the availability of traffic management to facilitate safe access to the nine trunk roads; e.g. samples were taken from locations on trunk roads where road maintenance was active with traffic management. A predefined sample number could therefore not be set as there was no guarantee that traffic management would be available within the study timeframe. As a result, the decision was taken to dispense with setting a predefined sample quota and instead the aim was to collect the maximum number of samples possible, as and when traffic management permitted.

RDS sample collection

Mechanical sweeping is not a regular occurrence on any of the roads selected for this study. Therefore, the data produced only permits analysis of the spatial variability of RDS (at a single point in time) but does not allow for an assessment of temporal variability. This approach aligns with research by Sartor & Boyd (1972) and Bian & Zhu (2009), who similarly made no attempt to collect repeat samples to identify how RDS accumulates with time.

RDS was collected from the road surface using a dustpan and brush, similar to Li et al. (2015). Samples were collected 3 days after a rain event as it was found that the accumulated RDS is still generally moist; therefore smaller particles are aggregated and this prevents them from being re-suspended during sample collection. Previous studies identify that almost all RDS accumulates within 1.0 m of the kerb-line, with around 88% within 0.3 m of the kerb (Charlesworth et al. 2003). Consequently, each sample was collected up to a maximum of 1.0 m perpendicular to the edge of the road, over a kerb length of 1.0 m. The mass of RDS available for collection varied from 0.5 kg to >1.0 kg across the 23 study sites, with the variability being related to the omission of regular mechanical sweeping on any of the roads selected for this study. RDS is also heterogeneous in composition and local catchment variables and factors influencing hydraulic sorting dictated the rate, magnitude and distribution of particles dispersed across the trunk roads.

At each site, a detailed survey was undertaken to collect data pertaining to catchment characteristics and the road was assessed using a road surface condition index (RSCI) rating system. The RSCI is a delineation type method based upon inspection and observation, which is then mapped to characteristic descriptions, as shown in Table 2.

Table 2

Road surface condition index

Failure mechanism Service condition Rating 
Cracking
The extent/severity of cracking is determined by the percentage of road surface that is subject to transverse, longitudinal, centreline, road edge or alligator cracking. 
Surface is reasonably new and there is very little evidence of cracking. no defects Excellent 
<5% Very good 
5% to 10% Good 
Surface showing early signs of edge, joint, slippage, longitudinal, transverse, alligator, block etc. cracking. 10% to 15% Fair 
Surface showing large areas of edge, joint, slippage, longitudinal, transverse, alligator, block etc. cracking. There is also evidence of localized loss of material. 15% to 20% Poor 
Surface is worn out, lots of wear and tear, typically the entire segment has pockets of fatigue saturation and loss of material. >20% Very poor 
Potholes and patches
A pothole is a hole in a road resulting from the loss of pavement material under traffic. A patch is a pothole or other surface defect that has been repaired. The rating is determined by the percentage of road surface with potholes or patches. 
Surface is reasonably new and there is very little evidence of potholes and patches. no defects Excellent 
<5% Very good 
5% to 10% Good 
Evidence of pothole and patch repairs to road surface. 10% to 15% Fair 
Surface showing large areas of pothole and patch repairs to road surface. There is also evidence of localized loss of material. 15% to 20% Poor 
Surface is worn out. The entire segment has pockets of pothole and patch repairs to road surface and loss of material. >20% Very poor 
Ravelling, loss of surface aggregate or polished aggregate
Ravelling, bleeding, loss or polished aggregate is the progressive disintegration of a pavement surface through loss of both binder and aggregate. The rating is determined by the percentage of road surface impacted. 
Surface is reasonably new and there is very little evidence of ravelling, bleeding, loss or polished aggregate. no defects Excellent 
<5% Very good 
5% to 10% Good 
Surface showing early signs of ravelling, bleeding, loss or polished aggregate. 10% to 15% Fair 
Surface showing large areas of ravelling, bleeding, loss or polished aggregate. 15% to 20% Poor 
Surface is worn out, lots of wear and tear, typically the entire segment has pockets of ravelling, bleeding, loss or polished aggregate. >20% Very poor 
Failure mechanism Service condition Rating 
Cracking
The extent/severity of cracking is determined by the percentage of road surface that is subject to transverse, longitudinal, centreline, road edge or alligator cracking. 
Surface is reasonably new and there is very little evidence of cracking. no defects Excellent 
<5% Very good 
5% to 10% Good 
Surface showing early signs of edge, joint, slippage, longitudinal, transverse, alligator, block etc. cracking. 10% to 15% Fair 
Surface showing large areas of edge, joint, slippage, longitudinal, transverse, alligator, block etc. cracking. There is also evidence of localized loss of material. 15% to 20% Poor 
Surface is worn out, lots of wear and tear, typically the entire segment has pockets of fatigue saturation and loss of material. >20% Very poor 
Potholes and patches
A pothole is a hole in a road resulting from the loss of pavement material under traffic. A patch is a pothole or other surface defect that has been repaired. The rating is determined by the percentage of road surface with potholes or patches. 
Surface is reasonably new and there is very little evidence of potholes and patches. no defects Excellent 
<5% Very good 
5% to 10% Good 
Evidence of pothole and patch repairs to road surface. 10% to 15% Fair 
Surface showing large areas of pothole and patch repairs to road surface. There is also evidence of localized loss of material. 15% to 20% Poor 
Surface is worn out. The entire segment has pockets of pothole and patch repairs to road surface and loss of material. >20% Very poor 
Ravelling, loss of surface aggregate or polished aggregate
Ravelling, bleeding, loss or polished aggregate is the progressive disintegration of a pavement surface through loss of both binder and aggregate. The rating is determined by the percentage of road surface impacted. 
Surface is reasonably new and there is very little evidence of ravelling, bleeding, loss or polished aggregate. no defects Excellent 
<5% Very good 
5% to 10% Good 
Surface showing early signs of ravelling, bleeding, loss or polished aggregate. 10% to 15% Fair 
Surface showing large areas of ravelling, bleeding, loss or polished aggregate. 15% to 20% Poor 
Surface is worn out, lots of wear and tear, typically the entire segment has pockets of ravelling, bleeding, loss or polished aggregate. >20% Very poor 

Source: adapted from BEAR Scotland (2010).

The RSCI rating is taken as the highest rating returned for any of the three categories in Table 1. Using the RSCI rating system, roads were rated and assigned a rating between 0 (excellent) and 5 (poor). In this study, roads were rated in two RSCI bandings: (1) between 0 and 3 and (2) 4 or 5. The rationale being that a road with an RSCI rating between 0 and 3 has a road surface with little evidence of fretting or fatigue on the surface. A road with an RSCI rating of 4 or 5 conversely shows signs of age, fatigue, and wear and tear on the surface and there is also evidence of localized loss of material. Comparing these RSCI ratings therefore permits analysis of the impact that road condition has on the generation and PSD grading of RDS.

RDS sample processing and particle fractionation

Upon arrival at the laboratory, samples were air-dried at room temperature to permit removal of materials greater than 20 mm. Samples were then oven-dried at 105 °C for 24 hours and then dry-sieved using British Standard stainless-steel sieves. Dry sieving was selected because this methodology produces reliable PSD results and is commonly used for RDS coarse particle analysis (Sartor & Boyd 1972).

The fractionated RDS was classified according to BS 1377-2:1990 (BSI 1990), e.g. silt and clay (<63 μm), fine sand (63–200 μm), medium sand (200–600 μm), coarse sand (600–2,000 μm), fine gravel (2,000–6,000 μm) and medium gravel (6,000–20,000 μm).

RESULTS AND DISCUSSION

RDS PSD grading envelope profiles

RDS PSD grading envelopes (% finer by mass) and corresponding mass percentage versus grain size fraction profiles for the 23 RDS samples are illustrated in Figure 3.

Figure 3

RDS PSD grading envelopes and corresponding mass percentages versus grain size fraction profiles.

Figure 3

RDS PSD grading envelopes and corresponding mass percentages versus grain size fraction profiles.

The findings reveal that RDS PSD gradation, and the percentage of each grain size, are highly variable across the 23 samples. However, despite being collected from nine trunk roads, with different trunk road catchments, PSD trends are similar, with a consistent distribution of particles ranging from clay and silt to medium gravel (Table 3).

Table 3

Distribution of particles

  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 3.8% 9.3% 7.4% 0% 270 
Max 16.5% 35.7% 58.5% 10.6% 1,100 
Mean 7.5% 15.7% 35% 2.8% 655 
Median 6.7% 13.7% 35.7% 3.2% 600 
  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 3.8% 9.3% 7.4% 0% 270 
Max 16.5% 35.7% 58.5% 10.6% 1,100 
Mean 7.5% 15.7% 35% 2.8% 655 
Median 6.7% 13.7% 35.7% 3.2% 600 

The results are consistent with those published by Walker & Wong (1999), with coarse sand and fine gravel particles recorded in all 23 samples and medium gravel particles recorded in 16 samples. Twelve of the 23 samples also had particles with a diameter >10,000 μm. The distribution of particles within the range 63 to 3,350 μm is also in line with Sartor & Boyd (1972), with 5 samples exceeding 70%, 10 exceeding 80% and 8 exceeding 90%. Across the 23 samples, the d50 range exceeds those recorded by Sansalone et al. (1998) and Maglionico & Pollicino (2004), but the median d50 is lower than the range recorded by Adachi & Tainosho (2005).

Analysis of RSCI rating data supports this conclusion, with wide variations being noted in the results between roads with RSCI rating between 0 and 3 and a rating of 4 or 5 (Figure 4).

Figure 4

(a) PSD grading envelopes (% finer by mass) for roads with an RSCI rating of between 0 and 3, and 4 or 5, (b) corresponding mean RDS PSD profiles, (c) corresponding mass percentage versus grain size fraction profiles.

Figure 4

(a) PSD grading envelopes (% finer by mass) for roads with an RSCI rating of between 0 and 3, and 4 or 5, (b) corresponding mean RDS PSD profiles, (c) corresponding mass percentage versus grain size fraction profiles.

Roads with an RSCI rating of 4 or 5 tend to have higher speed limits, higher traffic densities and more stopping and starting of vehicles at junctions, braking zones, etc. Together these factors contribute to enhanced degradation of the road surface because they induce friction between the vehicle tyre and the road surface, which gives rise to increased rates of abrasion of the road surface and degradation of vehicle components: brake linings, tyre wear etc. The results of this are reflected in the large percentage of coarse sand, fine gravel and medium gravel particles found in RDS PSD grading envelopes from roads with an RSCI rating of 4 or 5, compared to roads with an RSCI rating between 0 and 3 (Tables 4 and 5).

Table 4

Distribution of particles RSCI rating 0 and 3

  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 5.3% 9.3% 7.4% 0% 270 
Max 16.5% 35.7% 58.5% 9.8% 1,100 
Mean 8.7% 18.1% 30.3% 1.5% 571 
Median 8.3% 17.3% 29.9% 0.0% 500 
  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 5.3% 9.3% 7.4% 0% 270 
Max 16.5% 35.7% 58.5% 9.8% 1,100 
Mean 8.7% 18.1% 30.3% 1.5% 571 
Median 8.3% 17.3% 29.9% 0.0% 500 
Table 5

Distribution of particles RSCI rating 4 or 5

  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 3.8% 9.5% 31.0% 0% 520 
Max 8.5% 16.8% 52.9% 10.6% 1,100 
Mean 6.0% 12.5% 41.2% 4.4% 765 
Median 6.2% 12.1% 41.1% 4.3% 720 
  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 3.8% 9.5% 31.0% 0% 520 
Max 8.5% 16.8% 52.9% 10.6% 1,100 
Mean 6.0% 12.5% 41.2% 4.4% 765 
Median 6.2% 12.1% 41.1% 4.3% 720 

RDS PSD grading envelope profiles by trunk road

The greatest variability in grain size across the nine trunk roads was on the A90 (Table 6).

Table 6

Distribution of particles on the A90

  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 6.6% 10.9% 7.4% 0% 270 
Max 16.5% 35.7% 47.3% 9.8% 1,000 
Mean 9.3% 18.8% 28.4% 3.4% 556 
Median 7.8% 16% 29.9% 3.2% 500 
  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 6.6% 10.9% 7.4% 0% 270 
Max 16.5% 35.7% 47.3% 9.8% 1,000 
Mean 9.3% 18.8% 28.4% 3.4% 556 
Median 7.8% 16% 29.9% 3.2% 500 

On-site observation suggest that variations between the grain size characteristics reflect differences in local catchment characteristics and road dynamics on the A90. The five sample locations, for example, are spread over 63 km; therefore it is possible that hydraulic sorting of the RDS linked to topography, wind direction and strength, road gradient, condition of the road surface, etc. has dictated the rate, magnitude and distribution of particles dispersed across this 63 km stretch of trunk road. Different land uses and soil types are also reflected in the results. Variations in road dynamics are also reflected in the results. Three of the roads for example are asphaltic road surfaces and two are un-reinforced concrete surface slabs. RSCI ratings also vary from 1 to 3 and traffic flows range from 16,725 to 31,081 vehicles per hour. The A90(2) is also close to an off-slip and it is possible that the induced surface tension at the braking zone has skewed the grading envelope, with a higher percentage of coarse sand, fine gravel and medium gravel particles due to the increased rate of surface course abrasion. Three sections of the road also have a concrete central reserve safety barrier (CCRSB) and the ‘barrier’ effect that this safety feature creates to the movement of RDS may have skewed the actual grain size characteristics along these sections of the A90. This relates to the fact that observations made during on-site surveys qualitatively confirmed that a CCRSB, unlike a steel safety barrier, inhibits the natural movement of RDS, litter and vegetative detritus by the normal action of wind. As a result, detritus that would normally migrate to the adjoining landscape, or would be discharged to a central reserve drainage system, forms into large masses along the base of the CCRSB (Figure 5). It is hypothesized that wind vortices dictate the shape and mass of these formations.

Figure 5

RDS mass build-up along CCRSB.

Figure 5

RDS mass build-up along CCRSB.

Results from the M9 established that RDS grading envelopes had a similar range of values at all five sample locations, suggesting that the RDS originates from a similar source(s) (Table 7).

Table 7

Distribution of particles on the M9

  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 4.4% 10.3% 31% 0% 520 
Max 6.4% 15.6% 41.8% 3.7% 730 
Mean 5.5% 12.2% 35.6% 2% 610 
Median 5.5% 11.8% 34.9% 2.7% 600 
  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm) 
Min 4.4% 10.3% 31% 0% 520 
Max 6.4% 15.6% 41.8% 3.7% 730 
Mean 5.5% 12.2% 35.6% 2% 610 
Median 5.5% 11.8% 34.9% 2.7% 600 

Similarity in the results is, in part, attributed to the fact that the five sample locations were located within a 20 km section of the M9 and all have similar land uses and soil types. In addition, all five locations had an asphaltic surface course and RSCI ratings of 4 or 5 were recorded at four of the sample locations. A steel central reserve safety barrier was also present at all five locations.

Results from the A68 established that the two RDS PSD grading envelopes are more heterogeneous in terms of particle size, with RDS from the A68(2) being predominantly coarser than those originating on the A68(1), the distribution of particles <100 μm being 9.8% on the A68(1) and 6% on the A68(2) and the distribution of particles <200 μm being 16.7% on the A68(1) and 9.3% on the A68(2). The distribution of particles >1,000 μm also recorded a large variation, with 39.1% on the A68(1) and 58.5% on the A68(2). Neither of the roads recorded particles >10,000 μm. The d50 ranged from 700 μm on the A68(1) to 1,100 μm on the A68(2). Differences were unexpected because the sample locations are only separated by 9 km of road, both have the same land use, soil types and RSCI rating, and both are single carriageway roads with asphaltic surface courses and low traffic flows ranging from 3,294 to 11,843 vehicles per hour. Although the results at first appeared difficult to explain, the on-site survey determined that differences were attributable to the characteristics of the road, surface course material and condition of the road surface. The A68(2), for example, is on a sharp bend and therefore the road surface is under completely different load conditions to that of the A68(1). The A68(2) also has a high friction surface course (HFSC) which is near the end of its service life.

HFSC treatments are typically used to reduce traffic accidents at high-risk locations on the trunk road network and tend to correspond to locations with a high traffic density or skidding risk. In the case of the A68(2), an HFSC had been applied on the approach to a tight bend and observation identified that the braking forces being generated on the bend had left large areas with little or no HFSC, with the HFSC detritus (resin and aggregate) migrating to the kerb-line. HFSC treatments use a high polished stone value aggregate; most commonly on the Scottish trunk road network this is calcined bauxite. The grading envelope for calcined bauxite is specified in BS EN933-1:1997 (BSI 1997) and this stipulates that 90% of the particles must fall within the range 600 to 3,350 μm.

The A68(2) had 57.7% of RDS particles within the range of 600 to 3,350 μm, compared to only 45.4% on the A68(1). This supports the conclusion that the A68(2) RDS PSD grading envelope has been skewed because of the HFSC. This was confirmed by a visual inspection of the A68(2) sieved RDS particles, which highlighted that a high percentage of the particles were derived from fragmentation of the calcined bauxite HFSC.

Trunk road RDS PSD grading envelopes

The 23 samples were separated into each of the nine constituent Scottish trunk roads included in this study and the RDS PSD profiles for these trunk roads are illustrated in Table 8.

Table 8

Distribution of particles across the Scottish trunk road network

  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm)
 
Min Max Mean Median 
A7 9.2% 21.2% 24.2% 0.0% 400 570 448 410 
A68 7.9% 13% 48.8% 0.0% 700 1,100 900 900 
A90 9.3% 18.8% 28.4% 3.4% 270 1,000 556 500 
A912 6.2% 13.7% 40.3% 10.6% 710 710 710 710 
M8 7.7% 15.7% 35.3% 2.9% 460 880 653 620 
M9 5.5% 12.2% 35.6% 2% 520 730 610 600 
M80 7.1% 12.4% 46.1% 5.1% 900 900 900 900 
M90 3.8% 9.5% 52.9% 7.5% 1,100 1,100 1,100 1,100 
M876 6.1% 11% 45.2% 4.9% 980 980 980 980 
  <100 μm <200 μm >1,000 μm >10,000 μm d50 (μm)
 
Min Max Mean Median 
A7 9.2% 21.2% 24.2% 0.0% 400 570 448 410 
A68 7.9% 13% 48.8% 0.0% 700 1,100 900 900 
A90 9.3% 18.8% 28.4% 3.4% 270 1,000 556 500 
A912 6.2% 13.7% 40.3% 10.6% 710 710 710 710 
M8 7.7% 15.7% 35.3% 2.9% 460 880 653 620 
M9 5.5% 12.2% 35.6% 2% 520 730 610 600 
M80 7.1% 12.4% 46.1% 5.1% 900 900 900 900 
M90 3.8% 9.5% 52.9% 7.5% 1,100 1,100 1,100 1,100 
M876 6.1% 11% 45.2% 4.9% 980 980 980 980 

Five of the nine Scottish trunk road RDS PSDs (A7, A90, M8, M9 and M90) peaked at a particle diameter of 425 μm. Of the remaining four trunk roads, the A912 and M876 PSD peaked at a particle diameter of 600 μm, the M80 PSD peaked at a particle diameter of 1,180 μm and the A68 PSD peaked at particle diameters of 2,120 and 3,350 μm.

The 23 samples were also separated into their constituent trunk roads and the data was synthesized and RDS classified according to BS 1377-2:1990 (BSI 1990) for each of the nine Scottish trunk road network roads included in this study (Table 9).

Table 9

RDS classified according to BS 1377-2:1990 (BSI 1990).

Trunk road Silt and clay (%) Fine sand (%) Medium sand (%) Coarse sand (%) Fine gravel (%) Medium gravel (%) 
A7 3.5 17.7 41.4 25.9 11.5 0.0 
A68 2.4 10.6 25.5 20.7 31.7 2.9 
A90 3.7 15.1 41.4 20.4 11.9 7.5 
A912 2.5 11.2 32.0 25.4 12.5 16.4 
M8 3.2 12.5 35.9 26.8 14.4 15.9 
M9 2.2 10.0 37.7 25.3 17.5 7.3 
M80 2.9 9.5 32.4 21.7 21.2 12.2 
M90 1.2 8.4 29.8 18.8 22.8 19.1 
M876 2.4 8.6 26.7 30.6 18.8 13.0 
Mean 2.7 11.5 33.6 24.0 18.0 10.5 
Median 2.5 10.6 32.4 25.3 17.5 12.2 
Trunk road Silt and clay (%) Fine sand (%) Medium sand (%) Coarse sand (%) Fine gravel (%) Medium gravel (%) 
A7 3.5 17.7 41.4 25.9 11.5 0.0 
A68 2.4 10.6 25.5 20.7 31.7 2.9 
A90 3.7 15.1 41.4 20.4 11.9 7.5 
A912 2.5 11.2 32.0 25.4 12.5 16.4 
M8 3.2 12.5 35.9 26.8 14.4 15.9 
M9 2.2 10.0 37.7 25.3 17.5 7.3 
M80 2.9 9.5 32.4 21.7 21.2 12.2 
M90 1.2 8.4 29.8 18.8 22.8 19.1 
M876 2.4 8.6 26.7 30.6 18.8 13.0 
Mean 2.7 11.5 33.6 24.0 18.0 10.5 
Median 2.5 10.6 32.4 25.3 17.5 12.2 

All nine trunk road RDS PSD grading envelopes are consistent with those published by Walker & Wong (1999). Coarse sand and fine gravel particles are recorded on all nine trunk roads and medium gravel particles are recorded on eight of the nine trunk roads. The distribution of particles is also consistent with research by Regenmorter et al. (2002) and Sansalone & Tribouillard (1999) who identified that individual RDS particles accumulating on road surfaces range in size from 1.0 μm to >10,000 μm. The results obtained in this study identify that medium and coarse sand particles account for the majority of RDS accumulating on road surfaces.

Implication of RDS for the operational lifecycle of HFDs

Observations made during on-site surveys qualitatively confirmed that RDS derived from agricultural farmland and/or roads that have poor or very poor service conditions can overload the filter drain graded stone and promote acceleration of a cake-layer on the surface of the filter drain (Figure 6(a)). It was also observed that high intensity rainfall events and the aerodynamic drag of large vehicles convey large RDS particles towards the edge of the road (Figure 6(b)). These two processes also fragment the RDS formations that build up along the base of CCRSBs. When these events take place next to HFDs, the road runoff flushed into the filter drain (rich in large particles) overloads the surface of the filter drain, resulting in an exaggerated horizontal flow regime, coupled with a greatly reduced vertical component. Once the maximum capacity for the filter drain to retain RDS particles is reached, subsequently deposited particles form a cake-layer on the filter drain graded stone surface (Figure 6(c)). Long term, this cake-layer impedes the management of surface water runoff from the road and can cause ponding on the road and filter drain (Figure 6(d)). On a high-speed trunk road network, ponding can be particularly dangerous and lead to life-threatening driving conditions because it reduces a road's surface friction, which in turn increases stopping distances and can induce aquaplaning. Moreover, spray from rainwater being thrown up by vehicle tyres can reduce visibility, which can lead to delays in reacting to events on the road. In response vehicles must slow down, which leads to travel disruption through longer journey times.

Figure 6

(a) Deposition of eroded soil particles from surrounding farmland migrating into HFDs, (b) RDS, rich in large particles, migrating towards the edge of the road, (c) completely clogged HFDs running parallel with CCRSB, (d) ponding on trunk roads and HFDs.

Figure 6

(a) Deposition of eroded soil particles from surrounding farmland migrating into HFDs, (b) RDS, rich in large particles, migrating towards the edge of the road, (c) completely clogged HFDs running parallel with CCRSB, (d) ponding on trunk roads and HFDs.

HFDs servicing roads generating RDS rich in large particles or high loads should therefore be accounted for when considering clogging and the operational lifecycle of HFDs. Large particles migrating from the road to the filter drain, for example, will increase the likelihood of ‘bridging’ of the void spaces within the filter drain graded stone. Bridging inhibits the downward migration of RDS particles through the filter drain graded stone matrix, which in turn increases the rate of clogging, and the formation of a cake-layer, at the surface of the filter drain. As clogging intensifies, the free void space below the clogged surface layer will become redundant.

ACCURACY OF RESULTS

Due to ‘non-point origin’ and random occurrence of RDS at any given location on the Scottish trunk road network, there are some important points to note when interpreting the grading envelopes depicted in this paper. Particles recorded in the clay and silt range, for example, generally represents a small fraction of the entire RDS particle fraction, the mean being 2.9% for all 23 sites. However, this may not be truly representative of a Scottish trunk road network RDS PSD grading envelope because mechanical sweeping is not a regular occurrence; therefore it is likely that over an extended timeframe the accumulating RDS has become enriched with coarser particles through re-suspension and loss of clay and silt particles. This relates to the fact that short intense rainfall events, following long dry periods, wash off a considerable percentage of the finer fraction of the RDS load, leaving behind an RDS particle fraction containing a high percentage of large particles. It is also hypothesized that dislodgement processes associated with wind dispersion and air turbulence caused by high traffic densities, high-speed vehicle movement and a high percentage of heavy goods vehicles have probably contributed to re-suspension and loss to the surrounding environment of clay and silt particles. This is supported by Abu-Allaban et al. (2003) who have shown that a heavy goods vehicle contributes eight times more re-suspended RDS than a small goods vehicle. If this is the case, then the observed RDS PSD grading envelopes in this study may only partially reflect the characteristics of actual Scottish trunk road network RDS PSD grading envelopes.

Visual observation of the sieved RDS particles also qualitatively confirmed that particles >1,000 μm were generally derived from deterioration of the road surface and were non-cohesive and granular in nature. Most were coated in bitumen, road paint or expansion jointing compounds and some of these particles were observed to have adhered (aggregated) to particles of a similar or larger size. As such RDS PSD grading envelopes containing particles derived from deterioration of the road surface are likely to be somewhat skewed, with envelopes with a larger percentage of particles in the coarse sand, fine gravel and medium gravel ranges. However, it is hypothesized that particle aggregation derived from deterioration of the road surface has occurred in most, if not all, of the RDS PSD research studies referenced in this paper. On that basis, one can conclude that a direct comparison can be made of the RDS PSD grading envelopes in this research and those referenced in this paper.

The results may also be skewed somewhat by the fact that several of the sampling sites had potholes containing RDS that could not be collected as the RDS was lodged firmly within the pothole and this was inaccessible with a dustpan and brush. However, visual observations at the time of collecting RDS samples showed that the majority of the RDS at these sites was concentrated at the side of the road and therefore the RDS PSD grading curves were likely to be typical of the site's road catchment.

CONCLUSIONS

RDS PSD grading envelopes measured as part of this study are comparable to those reported in previous studies, with a consistent distribution of particles ranging from clay and silt to medium gravel. However, the results identify that it is important to recognize that grading envelopes are essentially instantaneous values and, given the scale of a strategic trunk road network, assuming a single RDS PSD profile at any given location is unlikely to be representative of a trunk road, or trunk road catchment profile. This is based on the premise that RDS is heterogeneous in composition and local catchment variables and factors influencing hydraulic sorting will dictate the rate, magnitude and distribution of particles dispersed across a strategic trunk road network.

Strategic HFDs operate under very variable catchment environments and this study has shown that assuming a single operational lifecycle profile for a filter drain servicing a trunk road or trunk road network is unlikely to be representative across a whole trunk road or trunk road catchment.

Results from this study can also be used to develop trunk road asset management strategies to minimize the risk that roads generating RDS rich in large particles or high loads pose to the operational lifecycle of HFDs. Existing trunk road asset management inspection regimes, for example, could be modified to include an assessment of RDS build-up at known ‘hot-spots’ to determine whether or not road sweeping was required. If sweeping was required, proactive intervention to remove the RDS would ensure that the HFD's operational lifecycle could be extended.

ACKNOWLEDGEMENTS

This study was supported in part by Transport Scotland and BEAR Scotland. The writers are grateful for their continuous support.

REFERENCES

REFERENCES
Abu-Allaban
M.
,
Gillies
J. A.
,
Gertler
A. W.
,
Clayton
R.
&
Proffitt
D.
2003
Tailpipe, resuspended road dust, and brake wear emission factors from on-road vehicles
.
Atmospheric Environment
37
,
5283
5293
.
Adachi
K.
&
Tainosho
Y.
2005
Single particle characterization of size-fractionated road sediments
.
Applied Geochemistry
20
,
849
859
.
Ball
J. E.
,
Jenks
R.
&
Aubourg
D.
1998
An assessment of the availability of pollutant constituents on road surfaces
.
Science of the Total Environment
209
,
243
254
.
BEAR Scotland
2010
Road Surface Condition Index Rating System
.
BEAR Scotland, Perth
,
Scotland, UK
.
Bian
B.
&
Zhu
W.
2009
Particle size distribution and pollutants in road-deposited sediments in different areas of Zhenjiang, China
.
Environmental/Geochemistry and Health
31
,
511
5520
.
Bruen
M.
,
Johnston
P.
,
Quinn
M. K.
,
Desta
M.
,
Higgins
N.
,
Bradley
C.
&
Burns
S.
2006
Impact Assessment of Highway Drainage on Surface Water Quality
.
Centre for Water Resources Research, University College Dublin
,
Dublin, Ireland
.
BSI (British Standards Institute)
1990
BS 1377-2:1990 Methods of Test for Soils for Civil Engineering Purposes. Classification Tests
.
BSI
,
London, UK
.
BSI (British Standards Institute)
1997
BS EN 933-1:1997 Tests for Geometrical Properties of Aggregates. Determination of Particle Size Distribution. Sieving Method
.
BSI
,
London, UK
.
Charters
F. J.
,
Cochrane
T. A.
&
O'Sullivan
A. D.
2016
Untreated runoff quality from roof and road surfaces in a low intensity rainfall climate
.
Science of the Total Environment
550
,
265
272
.
Ellis
J. B.
&
Revitt
D. M.
1982
Incidence of heavy metals in street surface sediments: solubility and grain size studies
.
Water Air Soil Pollution
17
,
87
100
.
Grigoratos
T.
&
Martini
G.
2015
Break wear particle emissions: a review
.
Environmental Science Pollution Research
22
,
2491
2504
.
Gunawardana
C.
,
Egodawatta
P.
&
Goonetilleke
A.
2014
Role of particle size and composition in metal adsorption by solids deposited on urban road surfaces
.
Environmental Pollution
184
,
44
53
.
Lau
S. L.
&
Stenstrom
M. K.
2001
Best Management Practices to Reduce Pollution from Stormwater in Highly Urbanized Areas
.
WEFTec
,
Chicago, Illinois
,
US
.
Li
H.
&
Davis
A.
2008
Urban particle capture in bioretention media. II: theory and model development
.
Journal of Environmental Engineering
134
(
6
),
419
432
.
Li
Y.
,
Lau
S.-L.
,
Kayhanian
M.
&
Stenstrom
M. K.
2005
Particle size distribution in highway runoff
.
Journal of Environmental Engineering
131
(
9
),
1267
.
Loganathan
P.
,
Vigneswaran
S.
&
Kandasamy
J.
2013
Road-deposited sediment pollutants: a critical review of their characteristics, source apportionment, and management
.
Critical Reviews in Environmental Science and Technology
43
(
13
),
1315
1348
.
Maglionico
M.
&
Pollicino
F.
2004
Experimental analysis of the build-up of pollutants on an urban road surface
.
NOVATECH
,
Lyon, France
.
Regenmorter
L. C.
,
Kayhanian
M.
,
Chappell
R. W.
,
Burgessor
T. E.
&
Tsay
K.
2002
Particles and the associated pollutant concentrations in highway runoff in Lake Tahoe, California
. In
StormCon 2002
,
August 12–15
,
San Marco Island, Florida
.
Sansalone
J. J.
&
Tribouillard
T.
1999
Variation in characteristics of abraded roadway particles as a function of particle size
.
Transportation Research Record
1690
,
153
163
.
Sansalone
J. J.
,
Koran
J. M.
,
Smithson
J. A.
&
Buchberger
S. G.
1997
Characterization of solids transported from an urban roadway surface
.
WEFTEC '97
.
Chicago
,
Illinois, USA
.
Sansalone
J. J.
,
Koran
J. M.
,
Smithson
J. A.
&
Buchberger
S. G.
1998
Physical characteristics of urban roadway solids transported during rain events
.
Journal of Environmental Engineering
124
(
5
),
427
440
.
Sartor
J. D.
&
Boyd
G. B.
1972
Water Pollution Aspects of Street Surface Contaminants
.
EPA-R2-72-081, US Environmental Protection Agency, Washington, DC, USA
.
Stylianides
T.
,
Frost
M. W.
,
Fleming
P. R.
,
El-Jaber
A.
,
Mageean
M.
,
Huetson
A.
&
Klimczak
T.
2015
Highway filter drains: precursors for maintenance management
.
Infrastructure Asset Management
2
(
4
),
159
172
.
Transport Scotland
2016
Road Asset Management Plan for Scottish Trunk Roads
.
Walker
T.
&
Wong
T. H. F.
1999
Effectiveness of Street Sweeping for Stormwater Pollution Control
.
Technical Report 99/8
.
Cooperative Research Centre for Catchment Hydrology
.