DSpace at VNU: Stormwater quality management in rail transportation - Past, present and future

Science of the Total Environment 512–513 (2015) 353–363

Contents lists available at ScienceDirect

Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv

Review

Stormwater quality management in rail transportation — Past, present
and future
Phuong Tram Vo a, Huu Hao Ngo a,⁎, Wenshan Guo a, John L. Zhou a, Andrzej Listowski b, Bin Du c,
Qin Wei d, Xuan Thanh Bui e,f
a

Centre for Technology in Water and Wastewater, School of Civil and Environmental Engineering, University of Technology Sydney, Sydney, NSW 2007, Australia
Sydney Olympic Park Authority, 7 Figtree Drive, Sydney, NSW 2127, Australia
School of Resources and Environmental Sciences, University of Jinan, Jinan 250022, PR China
d
Key Laboratory of Chemical Sensing & Analysis in Universities of Shandong, School of Chemistry and Chemical Engineering, University of Jinan, Jinan 250022, PR China
e
Faculty of Environment, Ho Chi Minh City University of Technology, 268 Ly Thuong Kiet, District 10, Ho Chi Minh City, Viet Nam
f
Division of Environmental Engineering and Management, Ton Duc Thang University, District 7, Ho Chi Minh City, Viet Nam
b
c

H I G H L I G H T S
•
•
•

•

Stormwater management in the railway industry focused solely on drainage.
Stringent stormwater quality standards require urgent responses from the industry.
Railway transportation generates potential sources of pollutants for runoff.
Urban retrofitting provides opportunities for railway stormwater management.

a r t i c l e

i n f o

Article history:
Received 29 December 2014
Received in revised form 23 January 2015
Accepted 23 January 2015
Available online 29 January 2015
Editor: D. Barcelo
Keywords:
Stormwater quality
Railway industry
Stormwater treatment
Urban retrofit

a b s t r a c t
Railways currently play an important role in sustainable transportation systems, owing to their substantial
carrying capacity, environmental friendliness and land-saving advantages. Although total pollutant emissions
from railway systems are far less than that of automobile vehicles, the pollution from railway operations should
not be underestimated. To date, both scientific and practical papers dealing with stormwater management for rail
tracks have solely focused on its drainage function. Unlike roadway transport, the potential of stormwater pollution from railway operations is currently mishandled. There have been very few studies into the impact of its
operations on water quality. Hence, upon the realisation on the significance of nonpoint source pollution,

stormwater management priorities should have been re-evaluated. This paper provides an examination of past
and current practices of stormwater management in the railway industry, potential sources of stormwater
pollution, obstacles faced in stormwater management and concludes with strategies for future management
directions.
© 2015 Elsevier B.V. All rights reserved.

Contents
1.
2.
3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conventional approach to stormwater management in the
railway industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Stormwater quality management practices in the railway industry . . . . . . . . . .
3.1.
Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1.
Recognition of non-point source pollution from the transportation sector
3.1.2.
Contamination along railway tracks and stabling yards . . . . . . . .
3.2.
Potential sources of stormwater pollution in the railway industry . . . . . . .
3.2.1.
Wooden sleepers . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . .

354

.
.
.
.
.
.
.

354
355
355
355
355
355
355

.
.
.
.
.
.
.

.
.
.
.
.
.

.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.

.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.

.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.

.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.

.
.
.

⁎ Corresponding author at: School of Civil and Environmental Engineering, University of Technology, Sydney (UTS), P.O. Box 123, Broadway, NSW 2007, Australia.
E-mail addresses: , (H.H. Ngo).

/>0048-9697/© 2015 Elsevier B.V. All rights reserved.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.

.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.


354

P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

3.2.2.
Herbicides and pesticides . . . . . . . . . . . . . . . .
3.2.3.
Fuels, oils and lubricants . . . . . . . . . . . . . . . . .
3.2.4.
Wear and tear . . . . . . . . . . . . . . . . . . . . . .

3.2.5.
Embankments . . . . . . . . . . . . . . . . . . . . . .
3.2.6.
Human waste and littering . . . . . . . . . . . . . . . .
3.2.7.
Maintenance facilities . . . . . . . . . . . . . . . . . .
3.3.
Pollution routes . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.
Challenges in stormwater quality management in the railway industry
3.4.1.
Input data . . . . . . . . . . . . . . . . . . . . . . .
3.4.2.
Monitoring and modelling . . . . . . . . . . . . . . . .
3.4.3.
Treatment challenges . . . . . . . . . . . . . . . . . .
3.4.4.
Regulations, policies and standards . . . . . . . . . . . .
4.
Provisions for stormwater quality management in the railway industry . . .
4.1.
Source control . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.
Stormwater treatment and harvesting . . . . . . . . . . . . . . .
4.2.1.
Stormwater treatment . . . . . . . . . . . . . . . . . .
4.2.2.
Stormwater harvesting . . . . . . . . . . . . . . . . . .
4.3.
Urban retrofit . . . . . . . . . . . . . . . . . . . . . . . . . .

5.
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction
Among the many endeavours of society to promote a sustainable
transportation system, railway networks play a crucial part because of
their substantial carrying capacity. A rough statistic from the World
Bank (2014) showed that the combined length of the world's railway
lines increased dramatically by 40% from 1990 to 2012. Compared to
roadway transport, railway is considered more environmentally friendly in providing mass transporting services with less negative ecological
impact (Zimmerman, 2005). Nonetheless, the environmental benefits
from railway transportation over private vehicles are undeniable.
Hence, railway networks are likely to be upgraded in order to meet
greater transportation and environmental demands (Kamga and
Yazici, 2014; Zhiqun and Jiguang, 2011). Although emissions from
railway systems are far less than that of automobile vehicles, the
environmental pollution from railway operations should not be
underestimated. Frequently mentioned types of impact caused by
rail transportation include noise (Aasvang et al., 2007; Ali, 2005;
Trombetta Zannin and Bunn, 2014), vibration (Kouroussis et al., 2014;
Sanayei et al., 2013) and air pollution (Dincer and Elbir, 2007; Salma
et al., 2009). In contrast, there have been very few studies into the impact on water courses. This lack of interest does not imply that water
pollution from the railway industry is an insignificant issue. As
Osborne and Montague (2005) stated, “railway operations, both current
and in the past, have the potential to give rise to pollution, as water
drains from the railway into water courses”. Yet, to date, priorities in
water management for rail tracks still solely focus on its drainage
function. Hence, upon realising the significance of nonpoint source

pollution, stormwater management priorities should have been reevaluated. This paper will provide an examination of past and current
practices of stormwater management in the railway industry, potential
sources of stormwater pollution, management obstacles and future
directions.

2. Conventional approach to stormwater management in the
railway industry
Rail tracks and supporting systems attracted the most attention in
stormwater management plans for the railway industry as they were
the backbone of railway services. This heightened attention was due
to the negative impact of runoff on rail tracks directly threatening rail
safety.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

356
356
356

358
358
358
358
359
359
359
359
360
360
360
360
360
361
361
362
362
362

Based on the track support systems (or substructures), rail tracks are
divided into three categories: traditionally ballasted, modified ballasted
and ballastless. Configurations of these substructures were well presented in the works of Esveld (1997) and Teixeira et al. (2009). While
the latter types of rail tracks developed due to demands for highspeed trains and low maintenance frequency, ballasted railway tracks
have still been employed extensively, thanks to their enormous
economic advantages. A typical ballasted substructure comprises of a
top ballast layer (150–550 mm of single-sized rocks), a sub-ballast
layer (90–450 mm of well-graded crushed rock or a sandy gravel mixture) and an underlying subgrade layer (natural or amended soil).
Each layer performs different structural functions to ensure the durability and stability of a rail track. Precipitation falling on ballast quickly
drains to the sub-ballast layer and then runs into drainage systems.
The drainage system could be either a parallel pipework network or a

natural ditch, which is located along the sides of the embankment toe.
Similar mechanisms were found in depots or maintenance centres.
The influence of runoff from surrounding areas on the rail track areas
is often restricted to ensure the safety of the track bed.
The effect of runoff volume on rail tracks was investigated thoroughly, as the saturation of water in these layers can reduce the stiffness of
the track foundation (Australian Rail Track Corporation Ltd., 2006).
The flow hydraulic properties vary depending on the type and age of
the track bed. Drainage capacity of a track decreases over time, as
sediments accumulate in its body (Burkhardt et al., 2005).
Rushton and Ghataora (2014) observed that greater impact occurred
when water accumulated in the sub-ballast and subgrade layers, where
finer grains were predominant. Under the load of moving trains,
trapped water became pressurised, drawing clay or silt from the
subgrade upward to the ballast layer, known as the “clay pumping” phenomenon (Rushton and Ghataora, 2009). Together with the depositing
of dust and abrasive materials on the ballast surface, clay pumping can
cause ballast fouling (Indraratna et al., 2011). The fouled ballast further
degraded the drainage capacity of the track support system and led to
structural deformation. Due to its high risk of rail track structural
deformation, stormwater was a critical problem for rail operation.
Stormwater runoff had subsequently been perceived as a nuisance
that must be drained as quickly as possible.
For modified ballasted systems (with a bituminous or geotextile
layer working as the sub-ballast layer) and ballastless systems, the effects of stormwater on the foundation structure are less severe. EAPA
(2003) pointed out three main reasons for this improvement. Firstly,
an asphalt layer distributed train loadings more uniformly, hence


P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

eliminated the “clay pumping” phenomenon in the upper ballast layer.

Secondly, a dense layer of asphalt moved water away quickly to protect
the top layer. Finally, the impermeable bituminous layer can act as a
barrier to block the upward movements of silt materials from the
subgrade (or foundation) layer.
Due to drainage being the focus of stormwater management
systems, only hydraulic profiles were considered in design. Collected
stormwater from the drainage systems were then discharged into natural water bodies, including rivers, streams, creeks or even drinking
water catchments, while its effects on the basin were completely
ignored. No quality consideration was found in any official technical
guidance for drainage systems in the railway industry (Australian Rail
Track Corporation Ltd., 2013; U.S. Army Corps of Engineers, 2004).
3. Stormwater quality management practices in the railway industry
3.1. Rationale
The above perception remained unchanged until recent years when
stormwater management objectives were re-assessed due to the
following reasons.
3.1.1. Recognition of non-point source pollution from the transportation
sector
An abundant number of papers have highlighted the existence of
pollution from diffuse sources over the last 40 years (Clark and Pitt,
2012). When most effluent discharges were moderately controlled,
stormwater then became one of the largest non-point pollution sources
contributing to the degradation of surface water resources (National
Research Council, 2008). Unlike effluent discharges with relatively stable characteristics, there is wide variation in stormwater quality and
quantity. Attempts to build roadway runoff profiles were accomplished
universally. While an insignificant concentration of biological oxygen
demand (BOD5), bacteria and nutrients are found in stormwater, it is a
substantial source of heavy metals and polycyclic aromatic hydrocarbons (PAHs) (Barbosa et al., 2012). These constituents were often
accompanied with finer particles (Barbosa et al., 2012; Kayhanian
et al., 2012).

To reduce the negative impacts of runoff, several countries set up
regulations for stormwater pollution control, such as the National Pollutant Discharge Elimination System Stormwater Program (1990) in
the US and the “Water Framework Directive” (2000) in the European
Union. A few states in the US even issued more stringent industrial
stormwater permits. As a result, the railway industry must comply
with these new requirements. An understanding of stormwater quality
from railway operations is a prerequisite for applying effective
pollution-reduction measures required to fulfil these tightening regulations (Kayhanian et al., 2012). Despite the practical needs to understand
stormwater quality profiles, available literature has shown little information pertaining to this issue. It imposed a large burden on the old
rigid engineering systems of the railway industry. As Dunning and
Weiner (2011) argued, stormwater management turns out to be one
of the largest struggles that the railway industry has to overcome in
the coming years.
3.1.2. Contamination along railway tracks and stabling yards
Railway is an important means of freight transport over long distance at reasonable cost. It is the preferred choice for transporting
crude oil. This, consequently, has caused the environmental risks associated with rail to increase. As evidence of this, more than 4350 m3 of oil
was released into the environment due to rail incidents in America in
2013, which was equal to 150% of past four decades put together
(Tate, 2014). Although clean-up activities could reduce the harmful effects to some extent, the accumulated oil in the soil could pollute
stormwater in much later years, as in the case of Osborn Yard in Louisville, Kentucky (Kurzanski et al., 2013). Signs of oil pollution appeared in

355

the stormwater run-off as a result of diesel spillage incidents that
occurred 20 years ago.
Apart from incident-related causes, daily railway operations were
also proven to affect the soil quality along rail tracks and supporting
infrastructures. Malawska and Wiłkomirski (2001) surveyed concentrations of nine metals (Co, Cd, Cr, Cu, Fe, Hg, Mo, Pb, Zn) and 14 priority
PAHs in soil samples taken from four railway locations – the siding,
the main track within the platform, the cleaning bay and the loading

ramp – at the Iława Głowna junction (Poland). Most of the substances
tested for, except Mo, were at substantially higher concentrations than
in the control sample. The largest concentrations of pollutants were
found near the platform and railway siding areas where trains spent
long periods of time at low speeds. Metals detected at high concentration were Fe (10,800–50,600 mg/kg dry weight), Zn (84–1244 mg/kg
dry weight), Pb (55–506 mg/kg dry weight) and Cu (24–115 mg/kg
dry weight). 13 years later, the authors executed comparative research
at the same locations. PAHs had significantly increased by 8–25 times,
which turned the soil from “slightly polluted” class to “polluted” and
“heavily polluted” classes, with reference to Polish and Dutch regulations (Wilkemirski et al., 2011). To a lesser extent, metal levels had
also magnified by 1.2 to 2 times. The investigation of heavy metals
along rail tracks in Qinghai–Tibet railway (Zn, Cd and Pb) and Suining
railway station (Pb and Cd) provided similar results (Chen et al., 2014;
Zhang et al., 2012).
Even though the movement of these pollutants from soil to water
environment has not been studied in great detail, soil pollution in rail
track areas could potentially result in stormwater contamination
(Burkhardt et al., 2008).
3.2. Potential sources of stormwater pollution in the railway industry
Emission sources from the railway industry can be divided into two
groups — those associated with daily operation (i.e. affected by the
frequency of trains) and those independent of rail traffic volume (i.e.
supporting infrastructures). The main sources of pollutants from daily
operation include: (1) wooden sleepers, (2) herbicides for vegetation
control, (3) fuelling and lubrication, (4) wear-and-tear processes and
corrosion-resistant poles, (5) embankment materials and (6) human
activities. The pollution from incidents such as oil spillage that were
briefly presented in the previous section is not the focus of this paper.
3.2.1. Wooden sleepers
The most significant source of organic compounds in railway runoff

comes from creosote-impregnated wooden sleepers. Creosote is a fusion mixture of more than 162 compounds including PAHs (69%), nitrogen heterocyclics (11%) and other aliphatic hydrocarbon (Utley, 2005).
It has been used as a fungicide to enhance the lifespan of wooden
sleepers (Brooks, 2001). Although the toxicity of creosote is low
(Chakraborty, 2001), it is identified as a potential carcinogen due to its
PAH components. In Switzerland, creosote-loaded sleepers accounted
for 43% of annual stock (Kohler et al., 2000). Nevertheless, the country
with the highest demand for wooden sleepers is the US. For instance,
the US required approximately 17 million new sleepers in the year
2008, 91% of which were wooden (AREMA, 2008). The main reasons
for the popularity of wooden sleepers come from their impressive ability in the dynamic attenuation of loadings, their light weight, their ease
to install and maintain, and most importantly, their economic viability.
The average loss rate of creosote in rail sleepers was about
210 mg/m2·day for 20–30 years in service, of which PAHs accounted
for 20 mg/m2·day (Kohler et al., 2000). Despite this, the asserted possibility of creosote leakage into water was considered debatable. Brooks
(2004), in his 18-month study to investigate the seepage of creosote
from railway sleepers to adjacent environments, suggested that leakage
of creosote via stormwater was negligible. The authors argued that creosote loss was accompanied mostly with vaporisation, weathering and
deposition in railway ballast. Thierfelder and Sandström (2008) also


356

P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

stated that creosote-impregnated wooden sleepers used for embankments would expose no risk to the water environment, albeit no
evidence was given for this conclusion.
On the other hand, Kang et al. (2005) explored the migration of 16
priority PAHs from impregnated wooden sleepers in fresh water
under different flow-rate regimes in one week. Seven lower molecular
weight PAHs (acenaphthene, anthracene, naphthalene, fluoranthene,

fluorene, phenanthrene and pyrene) were detected in the leakage
water in all cases. Chakraborty (2001) also studied three different
mechanisms of creosote loss – bleeding, leaching and vaporisation –
for eight light PAHs (the 7 previously mentioned PAHs and acenaphthylene). It was found that the main mechanism for PAHs loss was
leaching (more than 50%), rather than vaporisation and bleeding.
In addition, Becker et al. (2001) explored the leaching behaviour of
creosote in treated wood in three media — deionised water, buffered solution at pH 4.7 and humic mixture liquid. In their research, nitrogen
heterocyclics and several PAH compounds were leachable in all media.
Heterocyclic nitrogen substances (quinoline, isoquinoline, indole and
2-methyl-quinoline) were leaked with a higher rate than that of PAHs.
The highest leaching rate was quinoline with 1050 mg/kg of wood
after 24 h of submerging in water. The leaching rate for PAHs such as
naphthalene, dibenzofurane, phenanthrene and pyrene was much
lower, only 2–38 mg/kg of wood. The leachable quantities of PAHs
were minor compared to its extractable quantities (0.1–3.0%) by using
Soxhlet extraction method with toluene solvent.
It could be concluded that heavier PAHs tend to attach to organic
matters and sediments whereas lighter PAHs are able to dissolve with
a low concentration into stormwater, normally much lower than its
solubility (Table 1).
3.2.2. Herbicides and pesticides
Weed-growth on roadbeds or embankments is strictly controlled as
they may (1) impede a driver's ability to see signals, (2) impede staff
members working in rainy weather, (3) impede inspectors examining
track damage or (4) become fire hazards (Victorian Rail Industry Environmental Forum, 2007). Although different methods of weed control
have been considered, chemical herbicide spraying appears to be the
most economically feasible (Torstensson et al., 2005). Table 2 summarises several toxicological parameters of typical herbicides applied in
the railway industry.
Among these herbicides, diuron has been prohibited since the late
1990s in the railway industry because it is toxic and highly mobile. Its

strong mobility resulted in the destruction of a vast majority of pine
trees along the rail corridors in Sweden (Torstensson et al., 2002).
Table 1
Leaching rates for priority PAHs from impregnated wood in different types of water.
Experiment
conditions

Kang et al.
(2005)

+Medium
+Temperature
+Time

Fresh water Deionised
water
12–13 °C
–
7 days
120 h

Substance

Unit

Naphthalene
Acenaphthene
Fluorene
Phenanthrene
Fluoranthene

Pyrene
Quinoline
Isoquinoline
Indole
2-Methyl-quinoline
Dibenzofuran

Becker et al. (2001)
Buffer
solution
–
120 h

Humic
solution
–
120 h

μg/cm2·day mg/kg wood

mg/kg wood

mg/kg wood

N/A
N/A
N/A
0.2–0.5
N/A
N/A

–
–
–
–
–

108 ± 14
37 ± 1
30 ± 3
44 ± 5
6±2
4±2
1890 ± 180
205 ± 21
374 ± 33
150 ± 5
57 ± 3

105 ± 31
31 ± 1
31 ± 3
60 ± 27
27 ± 19
21 ± 16
1760 ± 170
354 ± 35
544 ± 66
254 ± 26
41 ± 1

77 ± 12
49 ± 6
42 ± 10
44 ± 2
22 ± 1
14 ± 2
2450 ± 180
427 ± 18
706 ± 23
354 ± 38
46 ± 6

Nevertheless, diuron has still been used for weed control in numerous
countries due to its long-lasting effectiveness. Glyphosate then emerged
as a safer alternative. Compared to other herbicides, glyphosate has
higher water solubility but lower toxicity (Schweinsberg et al., 1999).
In an investigation of pesticide application in the UK, Croll (1991)
discovered a disproportionate amount of triazine concentration in
surface water compare to the amount utilised in agriculture. Croll
suspected that a substantial part of this type of pesticide originated
from weed control for railway and roadway. A similar observation was
made by Skark et al. (2004). Indeed, the average application rate of
herbicides per area for railroads was claimed to be six times higher
than that being applied in agriculture (Schweinsberg et al., 1999).
Some papers discovered the existence of herbicides in surface water in
railway territories. The concentration of herbicides exceeded the drinking standard of 0.1 μg/L in the surface water near the railway lines
(Cooker, 1996; Schweinsberg et al., 1999). They accumulated in
the drainage ditch of a disused railway section at levels as high as
800 μg/L (Heather and Hollis, 1999).
Besides these contaminants, the arsenic level in soil along abandoned rail tracks in South Australia was measured to be within the

range of 17–1000 mg/kg, exceeding acceptable limits (5–40 mg/kg)
(Smith et al., 2006) as a consequence of the use of As-based herbicides.
3.2.3. Fuels, oils and lubricants
Leakages of petroleum products from fuel storage tanks, filling stations, locomotives and transformers are also frequent sources of water
pollution. Risks from oil leaks are directly proportional to the share of
diesel locomotives in the railway industry. The conversion from
diesel-powered trains to electric trains in railway networks is underway
worldwide, but has encountered various unfavourable hurdles. More
than two-thirds of locomotives in the railway industry are currently
powered by diesel (World Bank, 2007). The fraction of diesel locomotives is extremely high in regions where freight transportation over
long haulage distances is predominant, for instance North America
(99%), Latin America (97%) and Australia (95%) (World Bank, 2007).
This is because freight trains only need simple infrastructure and low
levels of electrification spanning over long sections. In contrast, passenger trains require higher levels of electrification because they run
through and connect multiple high-density metropolitan areas. Thus,
currently, railway electrification efforts are chiefly accelerated in populated metropolitan regions or in ambitiously developing countries such
as India and China (Juhasz et al., 2013).
Furthermore, oil and grease are also commonly used for lubricating
curves, gears and engines. Despite this, information relating to oil
leakage in railway operation is incredibly scarce. Only one Swedish
survey exists on this topic, presenting the oil leakage rates of various
transformers. The rates for large transformers, booster transformers
and auxiliary transformers are 10, 3 and 0.5 L/year, respectively
(Gustafsson et al., 2007). In a recent study on the stormwater runoff
profile from railway bridges, Gil and Im (2014) found the concentrations of oil and grease in a concrete road-bed and a gravel road-bed
were 0.20–2.90 and 0.61–6.70 mg/L. Oil leakages contain a high concentration of carcinogenic PAHs. Some organic compounds, even in small
amounts, can cause odour and aesthetic problems. Highly mobile hydrocarbon components pose higher risks to the receiving water bodies.
3.2.4. Wear and tear
The largest source of heavy metal emissions originated from friction
processes — rolling stock braking (73%), rail (21%), wheel (5%) and then

power line (1%) (Burkhardt et al., 2008). Both embankments and surrounding areas of the railroad were contaminated with metals
(Bukowiecki et al., 2007; Gustafsson et al., 2007). Iron accounted for
the highest portion of metal emission in the braking and abrasion processes, followed by Mn, Cr and trace amounts of Ni, Mo, V and Pb. Meanwhile, power line abrasion contributed the biggest quantity of Cu while
Zn was emitted from galvanised poles at a quantity of 140 g/pole/year


P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

357

Table 2
Representative herbicides used in railway industry.
Herbicide unit

LD50a (g/kg)

Phenoxy-carboxylic acids
2,4-D
0.4
2,4,5-T
MCPA

0.5
0.7

Dichlorprop

0.8

Triazines

Atrazine

2.0

Hexazinone
Simazine
Terbuthylazine
Propazine

1.7
N5.0
2.0

ADI (mg/kg)

RfD (mg/kg per day)

LC50 (mg/L)

DT50

Notes

0.3a
0.01f
0.03a
0.00015a
0.01f
0.03f

0.01a

100e

b7 dayse

–

0.01a
0.0005a
0.0044g
–

0.54–0.77h
232e

–
1–10 dayse

–
–

521e

21–25 dayse

–

0.0007a
0.005f

0.1f
0.005a
0.003f
0.02f

0.035a

176h

19–120 daysk

Germany: used until 1991a

0.035a
0.005a
0.02a

–
5h (rat)
–
2.04m

8–92 days k
27–216 days i
10–36 daysk
131 daysm

Germany: used until 1989a
–
–

–

Urea derivatives
Bromacil
Chlorotoluron
Diuron

5.2
N10.0
3.4

0.1f
0.0005a
0.00003a
0.007f

0.1g
0.002a
0.002g

–
–
–

12–46 daysk
–
12–48 monthsd

Isoproturon
Monuron

1.8
3.7

–
–

–
–

–
–

14–29.5 daysj
–

Germany: using from 1989a
–
Germany: used until 1996a
Holland: used until 1999
Sweden: used between 1974 and 1993b
UK: used until 2008
–
–

Miscellaneous
Amitrole
Dalapon
Picloram
Imazapyr

Glyphosate

N5.0
3.9
3.8
–
4.5

0.001l
–
0.07f
2.5f
0.3a

1.13g
–
0.2g
–
0.1a

0.439 (rat)l
–
26e
N100e
86e

50 daysl
–
30–90 days e
2–6 months d

2–5 months d
3–174 days e

Germany: used until 1989a
Germany: used until 1990a
Germany: used until 1989a
EU: used until 2004e
Germany: using since 1987a
Sweden: using since 1986c

Notes:
ADI: acceptable daily intake.
LD50: lethal dose for rat (oral).
RfD: reference dose for chronic oral exposure.
LC50: lethal concentration for fish.
DT50: disappearance time for 50% of substance.
a
Adapted from Schweinsberg et al. (1999).
b
Adapted from Torstensson et al. (2005).
c
Adapted from Gustafsson et al. (2007).
d
Adapted from Torstensson et al. (2005).
e
Adapted from Britt et al. (2003).
f
Adapted from Bending and Rodriguez-Cruz (2007).
g
Adapted from Department of Health — Office of Chemical Safety (2014).

h
Adapted from Dikshith and Diwan (2003).
k
Adapted from Directorate-general health and consumer protection (2001).
i
Adapted from Gunasekara et al. (2007).
j
Adapted from Hayes and Kruger (2014).
l
Adapted from Sarmah et al. (2009).
m
Adapted from University of Hertfordshire (2013).

due to data for the former group being both lacking and inadequate.
Thus, Table 3 was given as a rough guide to locate the ratio of railway runoff contaminants within the general railway runoff profile
and benchmark values. The benchmark values for stormwater pollutants proposed by the US. Environmental Protection Agency
(2009b) were used to represent the “level of concern” to receiving
water quality.

(Burkhardt et al., 2008). Most metals were bound with particles while
some were released in the dissolved phase (Zn, Cu and Ni). The distribution of metals depends on spatial and temporal scales which have not
been studied.
Three studies reported the existence of metals in railway runoff
(Gil and Im, 2014; Gill, 2012; Larsson, 2004). Apparently, it is difficult to compare between railway runoff and highway runoff quality

Table 3
Metal concentrations in runoff from highway and railway.
Component (μg/L)

Zn

Cu

Cd

Pb

Cr

Fe

Source of data

Railway
Railway bridge
Station
Stabling yard
Highway
Benchmark value

–
950
23–180
63–1784
117

25–270
46.3
25–92
5.5–11.7

63.6

0.015–3.1
1.3
b0.1
9.4–350.6
15.9

2–63
43
9.3–16
5.64–1860
81.6

–
17.3
2.9–5.3
0.056–16.6
–

–
–

Gil and Im (2014)
Gill (2012)
Larsson (2004)
Kayhanian et al. (2012)
USEPA (2009b)

Bold value: metals with high concentration in the railway industry.

334–89,000
1000


358

P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

Compared to highway runoff, the concentrations of Cd, Cr and Pb
were relatively low whereas Zn and Cu were typically high. The concentration of Cu in railway runoff was significantly greater than the amount
in roadway runoff, in some cases, up to 20 times higher. Iron content in
railway runoff has not been studied in existing research; however, its
content in the environment is expected to be much higher than other
metals (Wilkemirski et al., 2011).
The toxicity of these metals has been widely studied. Despite Cu, Mn
and Zn being less harmful than Pb and Cr, they are more soluble, and so
tend to have a greater impact on the water environment (Osborne and
Montague, 2005). While the toxicity of iron is low, it can affect water
colour and taste.
3.2.5. Embankments
3.2.5.1. Soil erosion. Erosion of rail embankments can result in a washing
out of sediments. These sediments themselves could be a source of pollution, depending on their particle size. Furthermore, heavy metals and
organic compounds tend to attach to particles. As a result, particles may
act as a medium for transporting pollutants into the water environment.
3.2.5.2. Substitute materials for ballasts. Steel furnace slags are often used
as a substitute for natural rocks in the ballast layer. Originating as a byproduct of the iron or steel processing industry, slag is a fused nonmetallic mixture which is rough surfaced and angular in external
shape. Steel slag is also highly resistant to physical and electrical forces.
These characteristics make steel slag a perfect candidate for railroad ballast. Most practitioners consider slag to be an inert and safe material.
Yet, Piatak et al. (in press) warned about the risks of slag usage. Most notably, the content of Al, As, Cd, Cr, Pb and Mn content in iron and steel

slag often exceeds the US EPA standards for residential and industrial
soil. Therefore, after interacting with air or water, derivative weathering
products from slag can release trace metal elements such as Cd, Cr and
Pb, especially under rainfall conditions. This is the case when the quality
of slag is not controlled.
3.2.6. Human waste and littering
In many developing countries, open carriage toilets are still being
utilised in train cars. Human excrement and garbage are discharged directly onto rail tracks and surrounding areas. This waste often contains
pathogens, nutrients and organic matter. They are deposited and then
accumulate in the environment without any treatment. Representative
cases can be found in many developing countries. For example, with
over 14 million people being transported via India's rail network every
day, it was estimated that about 3980 MT of human waste were disposed freely from 160,000 open toilets per day (Comptroller and
Auditor General of India, 2013). Moreover, railway corridors, which
are commonly viewed as vacant land, become an attractive environment for illegal discharges of sewage and domestic waste.
3.2.7. Maintenance facilities
Maintenance activities take place at all railway depots. Common
contaminants in runoff water from these areas are oil and grease, chlorinated and non-chlorinated solvents, phenols, antifreeze, detergents,
PAHs, sewage waste and several inorganic chemicals (Osborne and
Montague, 2005). These pollutants have resulted from different
processes in maintenance centres: metal processing, fuelling, repair of
machines and batteries, maintenance of rolling stocks, train cleaning
and so on. Apart from a study by Gill (2012), information on the runoff
quality from these areas is totally lacking.
3.3. Pollution routes
From the above analysis, the main pollutants in railway industries
are PAHs, herbicides and heavy metals. The possibility of runoff pollution is determined by numerous factors, such as precipitation regimes,

runoff flow dynamics, substance properties and its interaction with
surrounding soils (Fig. 1).

Rainfall is indeed a crucial factor in the environmental fate of
contaminants. It determines the movement of pollutants through rail
tracks and embankments. In general, ballast and subballast layers have
much higher permeability than the subgrade. Contaminants will be
transported downward and normally retained at the interface between
the subballast layer and the subgrade. At low rainfall intensity (under
15–20 mm), rain water may accumulate inside the track bed. The infiltration and evaporation rate might account for up to 75% of precipitation volume and sometimes does not generate the outflow of runoff
(Burkhardt et al., 2005). Under high intensity rainfall, contaminants
will be either washed out into drainage systems or infiltrated into adjacent soil. Consequently, the retention time of pollutants in rail tracks
fluctuates significantly, from half a day to three months from site to
site (Osborne and Montague, 2005). The first flush effect was detected
at locations that experienced great variation in wet and dry weather
conditions (Gil and Im, 2014), but in places where the rainfall regime
was more uniform, the first flush effect may not be observed (Lee
et al., 2007).
Secondly, the discharge of a contaminant into runoff is dependent
not only on its sources and characteristics, but also on its interactions
with the soil environment. The unsaturated soil near the track bed
could act either as a filter or a pathway to transport pollutants
(Burkhardt et al., 2005). Flow dynamics in soil is determined on one
hand by soil texture and structure, on the other hand, by soil water
content and tension. Two key mechanisms for the mass transfer of
heavy metals and organic compounds are degradation and adsorption/
desorption.
Heavy metals are neither biologically nor chemically degraded. They
will accumulate in the track bed or in the surrounding environment.
Heavy metals tend to attach to silt materials in natural soil. The adsorption of heavy metals on silt particles is influenced by pH. Alkaline soil
provides conditions advantageous for the adsorption of heavy metals,
whereas acidic soil is favourable for desorption. This typically increases
the toxicity of the metals and their complexes. For example, the low soil

pH (pH = 4.03–6.38) in Chengdu–Kunming railway (Sichuan, China)
showed the changes in Cd's mobile capacity (Liu et al., 2009). An interesting fact is that whereas most of the heavy metals in railway runoff are
toxic, Fe and Mn could act as absorbents for attaching heavy metals and

Emission sources

Substances
(Amounts, patterns,
characteristics)

Precipitation
(Intensity, frequency,
volume)

Runoff
(Track proϐile, drainage
system)

Mobility/ losses
(Soil, track proϐile):
Inϐiltration
Degradation
Sorption/ desorption

Runoff pollution
Fig. 1. The pollution pathway in rail track areas.
Modified from Burkhardt et al. (2008).


P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

anionic compounds such as glyphosate (Burkhardt et al., 2005). Nonetheless, their sorption capacities have not yet been investigated.
In contrast, organic compounds will degrade over time. Their degradation is characterised by disappearance time (DT50) which varies
greatly in relation to different types of environment. In the case of rail
tracks, the biodegradation of PAHs and herbicides is extremely low. As
Burkhardt et al. (2005) observed, the microbial biomass in a track area
is only one tenth of those in agricultural soil. This could be explained
by the coarse texture, low organic and nutrient contents of ballast and
embankment materials. Therefore, PAHs and herbicides used in railway
embankments usually had better mobility and prolonged persistence
(Cederlund et al., 2007). Many of these organic compounds are attached
to organic matters in soil. In favourable conditions, contaminants can
be reactivated and released slowly over long periods of time. The
stormwater runoff then acts as a pathway to transport the contaminants
from the soil into surface water bodies (Osborne and Montague, 2005).
3.4. Challenges in stormwater quality management in the railway industry
Huge challenges have been encountered from the implementation
of stormwater quality management in the railway industry. Apart
from common issues of stormwater management which were clearly
addressed by Langeveld et al. (2012) and Barbosa et al. (2012), the
railway industry experienced some particular struggles worth noting.
3.4.1. Input data
A lack of data is encountered as the most significant problem. In
reality, albeit the necessity of a thorough investigation on drainage
systems both quantitatively and qualitatively, it is a tough task for any
railway manager. Not only is railway runoff quality data frequently unavailable, but documentation for stormwater drainage systems has also
been inadequate (Singaraja et al., 2012). Historical modifications in
storm drains were not recorded properly. This is particularly problematic when it comes to predicting the movement of pollutants because they
may distribute and end up at various unknown receptors.
The co-mingled effects of runoff between railway land-use and surrounding land-uses are another troublesome issue, especially under

pressures of obtaining stormwater permits and abiding by surface
water standards. Some storm drain lines for the railway industry receive
both stormwater falling on its own assets and from other facilities in the
same watershed. Furthermore, cross connections between different
stormwater networks or between drainage and sewage systems make
the situation more challenging (Dunning, 2012).
3.4.2. Monitoring and modelling
Monitoring is a necessary step to determine baseline conditions,
levels of contamination, treatment methods and management measures. Key considerations for a stormwater monitoring program include

359

monitoring locations and safety, parameters and frequency, analytical
methods, precision and accuracy as well as cost-effectiveness (Fig. 2).
Stormwater monitoring is more challenging than effluent monitoring due to the irregular nature of runoff, along with a substantial
discrepancy of effluent throughout a rain event or time of discharge.
The number of events, the time and method of sampling during an
event may lead to a striking contrast in results (Lee et al., 2007). Common sampling methods are instantaneous grab sampling and flowweighted composite sampling. Although grab sampling is simple and
cheap, it may distort the result of runoff quality, depending on the
time of collecting samples (Lee et al., 2007). In contrast, flowweighted composite sampling gives a more precise outcome. However,
it is more complicated, costly and requires more training for practitioners. Selection of the monitoring method should be based on practical conditions. Nevertheless, grab sampling is still preferential in the
railway industry. To minimise the error in sampling, Leecaster et al.
(2002) suggested that the frequency for sampling be seven storm
events per year, 12 samples per event using volume-weighted ratio.
Further discussion on the stormwater monitoring program could be
seen in Lee et al. (2007). The sampling of track drainage systems is classified as a high-risk task. Safety issues may arise from either “adverse
climatic conditions” or the potential harm of being struck by running
trains. The deficiency of clearly designated stormwater drains makes
their monitoring even more troublesome (Copeland and Lefler, 2013).
With regards to the lack of monitoring data on stormwater quality, a

mathematical model is necessary for the prediction of its discharges
on the environment (Barbosa et al., 2012). Several sophisticated
stormwater quality management models have been incorporated with
the consideration of treatment methods (Elliott and Trowsdale, 2007).
Unfortunately, little information is available for movement kinetics of
pollutants through rail tracks. The application of available models for
modelling railway runoff, therefore, needs further study in order to
fine-tune and calibrate.

3.4.3. Treatment challenges
Besides uncertainties in stormwater quantity and quality, the
selection of treatment methods faces various challenges. First, space
constraints are encountered in many types of railway infrastructures.
Railway corridors are often narrow but they accommodate a wide
range of crucial infrastructures for daily operations such as power
lines, communication cables, signalling systems, accesses, barriers and
drainage systems. This is also the case for many stations and stabling
yards. The restricted availability of land in the areas results in the
limitation of treatment alternatives.
Maintenance frequency of treatment systems must follow the operational procedures of railways. In fact, maintenance is a key issue for rail
operations as anything within 3 m of a live track can only be accessed
during a possession. This is when a section of the track must be officially

Deϐine study type
Determine study scope
(Spatial boundaries, scale
and duration)
Consider sampling design
issues
Field

sampling
sites

Spatial
variablity

Frequency

Precision
and
accuracy

Measurement
parameters

Fig. 2. Framework for stormwater monitoring.
Adapted from ANZECC and ARMCANZ (2000).

Cost
effectiveness


360

P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

shut down for works. Because of this, maintenance requirements for
drainage should not be any more frequent than yearly.
3.4.4. Regulations, policies and standards
As stormwater quality management in the railway industry is a

newly emerging field, there is still debate as to whether railway runoff
should be integrated into a general urban stormwater management
scheme or be solely managed on its own. On one hand, the inclusion
of railway runoff into the general stormwater management program
can reduce the cost of management and treatment. On the other hand,
it can be argued that each industry has to eliminate its own pollution
and railway is not an exception. It is necessary to reduce the level of
metals and organic compounds which are specific in the industry.
Moreover, unlike other industrial sources, which have clear boundaries as well as outfalls for their assets, railway is mostly a line source
(except at stabling yards or depots) which runs over different territories
with various discharge regulations. Every territory has its own permits,
which poses a great challenge for a general control of stormwater quality in the industry, even though operational activities between different
regions are similar.
Bench-mark standards for stormwater management are also subjected to a number of critiques as they are ineffective in controlling
stormwater quality (National Research Council, 2008). Thus, a new
methodology of allocating pollutant loads for different sectors in a
catchment has recently been proposed to assist the enforcement of
bench-mark standards (Rogne, 2012). The new calculation method is
based on Total Maximum Daily Loads (TMDLs) for a catchment. The railway industry is accordingly required to employ an integrated system of
monitoring programs, stormwater control measures, appropriate
modelling and treatment methods to adhere with this new approach
(Schultz and Godlewski, 2012).
In short, the accountability of stormwater management programs is
often low. With ambiguous understandings of the nature of pollutants
in the railway environment, it is more challenging for rail companies
to appraise whether their runoff water meets the given standards or
not.
4. Provisions for stormwater quality management in the railway
industry
4.1. Source control

The pollution prevention measures are preferable in stormwater
quality management in the railway industry. To minimise the causes
of pollution is more cost-effective than treating its consequences downstream as simple changes could lead to long-term positive results. Common practices have been proposed to reduce pollutants at the source
and prevent contact between stormwater and potential contaminants
in the rail industry (US EPA, 2009a). However, these practices concentrated mainly on maintenance facilities and depots, rather than pollution along rail tracks, stabling yards and embankments. Therefore, the
following solutions are suggested for controlling potential sources of
stormwater pollution for these areas.
As wooden sleepers are of most concern in relation to PAHs sources,
they should be replaced by less harmful materials. Gustafsson et al.
(2007) investigated the replacement of wooden sleepers with concrete
sleepers. Through their experiments, concrete sleepers were proven to
be ecologically safe. The only additive in concrete that attracted
environmental concerns was sulphonate naphthalene. With a concentration of 1‰ in the concrete, it showed an insignificant leaching rate
(Gustafsson et al., 2007). Switching from wooden sleepers to concrete
sleepers and composite sleepers (a new type of sleeper) becomes
increasingly popular.
The application of herbicides in weed control practices should be
carefully planned with regards to types of weed, time and amount of
application, weed resistance capacity to herbicides and treatment

locations. The substitution of persistent and toxic herbicides for alternatives with quickly-degradable active ingredients is highly recommended. A new technique of applying herbicides in the railway is to use lowspeed swiping trains. These trains distribute herbicides directly onto
specific weed-ridden areas at speeds of 5–8 km/h rather than spraying
over entire areas, as is the present conventional method.
As Hansen and Clevenger (2005) argued, the disruption to natural
soil conditions along railway edges promoted the invasion of exotic
plant species. Therefore, an additional method for controlling weeds
that can be utilised is the re-plantation of indigenous flora (Victorian
Rail Industry Environmental Forum, 2007). The preservation of natural
vegetation on railway corridors also helps to reduce embankment erosion by increasing slope stabilities and eliminating water logging at
the track toes. In addition, mulches (organic and inorganic) could be

utilised to reduce the growth of weeds.
A large quantity of metal deposits originated from abrasion processes of brakes, rails, wheels and power lines. The magnitudes of deposits
caused by these abrasion processes vary according to the type of
materials involved; for instance, composite brakes and wheels emitted
the least amount of metals, compared to cast-iron and sintered iron.
The substitution of cast-iron and sintered iron brakes by composite
brakes could eliminate the emission of metals by approximately 90%
(Gustafsson et al., 2007).
As mentioned earlier, pollution along railway tracks is mainly
accompanied with fine fractions. It means that removing fine particles
could reduce high fraction of impurities. Therefore, increasing the frequency of ballast cleaning can reduce the accumulation of contaminants
onto track beds.
4.2. Stormwater treatment and harvesting
4.2.1. Stormwater treatment
Although source control measures are effective in reducing pollution
potentials, they alone cannot fulfil the requirements of stormwater discharge permits (Dunning and Weiner, 2011). In this case, a treatment
system is necessary. Unfortunately, the railway industry has not paid a
great attention to stormwater treatment. The judgement of selecting
treatment methods must be based on understandings of quantity and
quality characteristics of stormwater, treatment objectives, local conditions and possibilities for incorporating with educational or aesthetical
purposes (Barbosa et al., 2012). National Research Council (2008) and
Scholes et al. (2008) provided systematic comparisons of different alternatives for stormwater treatment.
It is obvious that no single treatment method would be effective for
removal of all pollutants. Learning from experiences of highway runoff
treatment, the conventional treatment chain often serves these
functions: (1) trapping litter and large objects, (2) detaining coarse sediments, (3) settling fine sediments, and (4) treatment of dissolved solids
and other contaminants. In the case of the railway industry, space
constraints and maintenance requirements are important factors in
the selection of treatment methods. Table 4 reviews different methods
in stormwater treatment and their applicability in the railway industry.

Constructed wetlands and detention basins, two common methods
in urban stormwater treatment, would rarely be used due to limited
space availability at railway areas. Initial sizing of wetlands should be
based on pollution reduction targets, but in reality, other factors such
as topography may mean that excessive land area is required. Sedimentation basins as a part of a wetland may have design particle sizes dictated by the catchment management authority. The trapping of fine clay
and silt particles makes basin size prohibitively expensive.
Simple methods such as buffer strips and grass swales are helpful in
the removal of several pollutants. The removal of contaminants by buffer strips is dependent on slope, length, runoff velocity, topography and
vegetation type. Buffer strips are suitable for removing coarse sediments
and some finer particles such as heavy metals, PAHs and nutrients.
Stagge et al. (2012) reported high removal efficiency for grass swales


P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

361

Table 4
Comparison of different stormwater treatment methods.
Modified from NSW Environment Protection Agency (1997).
Targeted pollutants

Scale of
catchment

Space
constraint

Environmental
and community

amenity

Operational and maintenance requirement

Litter and gross pollutants
Coarse sediments
Litters, coarse sediments
Oil, coarse sediments

b1 ha
8–20 ha
8–20 ha
N1 ha

Low
Low
Low
Low

Low
Low
Low
Low

Simple maintenance
Simple maintenance
Simple maintenance
Simple

b1 ha

Moderate

Moderate–high

b2 ha

Moderate

Moderate–high

Simple maintenance
Slope of the strip b 5%
Maximum flow depth = 12 mm
Simple maintenance

Sand filters

Litter and gross pollutants, coarse sediments,
suspended solids (SS), total phosphorus (T-P),
total nitrogen (T-N), bacteria
SS, T-N, T-P, organic matters, oil and grease,
bacteria
SS, T-N, T-P, bacteria

1–6 ha

Low

Low

Infiltration trenches and basins

SS, T-N, T-P, organic matters, bacteria

b6 ha

High

Moderate–high

Extended detention basins

Coarse sediments, SS, bacteria

N6 ha

High

Moderate

Tertiary treatment
Biofilters

SS, T-N, T-P, organic matters, bacteria

N/A

Low

Moderate

Constructed wetlands

Coarse sediments, SS, T-N, bacteria

N6 ha

High

High

Types of
treatment
Primary treatment
Litter pits, baskets and racks
Sediment traps
Gross pollutant traps
Oil/grit separators
Secondary treatment
Buffer (filter) strips

Grass swales

in treating highway runoff (50–60% of sediment, 46–81% of Zn, 27–75%
of Cu and 41–72% of Cd). Analysis of data on the International
Stormwater BMP Database roughly supports the figures by Wong et al.
(2000) with the exception of generally lower rates for TSS removal in
swales. Grass swales were very effective in removing zinc from runoff.
Stagge et al. (2012) emphasised that in the cases of physical limitation

as in railway corridors, the application of grass swales (about 200 m
length) can significantly improve the effluent water quality. However,
finer sediments deposited during smaller flows may be remobilised
during larger events. Vegetation types must be carefully selected to
prevent harmful effects on track foundation.
Infiltration systems often achieve moderate levels of pollutant removal due to the close contact between the runoff and substrate surface
during the infiltration of the runoff through the media (Scholes et al.,
2008), but they can have high failure rates. Cleaning time is an essential
factor in the design of these systems.
Wong et al. (2013) advocated the use of biofiltration systems
with a submerged zone for urban stormwater treatment. The
biofilters showed high removal efficiencies of TSS (N90%), pathogens
(1–3 log for Clostridium perfringens, Escherichia coli, and F-RNA
coliphages), heavy metals (N 90% for Cu, Zn, Cd and Pb), PAHs
(N 80%), oil and grease (N95%) and glyphosate (N80%) (Bratieres
et al., 2008; Li et al., 2012; Lim et al., 2015; Zhang et al., 2014). However, the biofilters were not effective in removing the triazine herbicides. This was due to the short hydraulic retention time (3–5 h) of
the biofilters (Zhang et al., 2014), which was insufficient for biodegradation of these herbicides. Thus, there is a tendency to seek out
novel filtration media to improve the capacity of removing pathogens, arsenic and micropollutants as herbicides.
For stabling yards, drainage systems have to move water away from
the track formation quickly, denying the possibility of retention for filtering or sedimentation. However, some stabling yards run in parallel
to an access road, which may allow for the possibility of long narrow
options such as grass swales, bioretention pits, and underground infiltration trenches. For most depot sites, parking lots and stations, which
are mostly impervious surfaces, filtration tanks could be located underneath. In addition, bioretention could be reserved for landscaped garden
beds in these areas.

Moderate maintenance due to sediment build-up
May require pre-treatment
Mostly suitable for sandy loam to loam soil
type with the infiltration rate of 13–25 mm/h
Simple maintenance

May require pre-treatment

Moderate to complex
May require pre-treatment
Moderate to complex

4.2.2. Stormwater harvesting
To fulfil the goals of a sustainable transportation system, the railway
industry aims to investigate opportunities for harvesting stormwater
(Transportation for NSW, 2013). Compared to wastewater reuse,
stormwater harvesting receives less public objections. While most
stations would not be able to incorporate any water storage under platforms due to structural elements and amount of services present, many
would have space on the platforms for above-ground water tanks to
capture rainwater from the station roofs. In addition, a spacious area
under stabling yards becomes an attractive opportunity for storing
stormwater. Australia is a pioneering country in this area, having
constructed the largest underground stormwater storage system in a
railway area. Being constructed at Auburn station (Sydney) in 2011,
the underground structure stores and treats more than 11,000 m3 of
stormwater.
4.3. Urban retrofit
Urban retrofit is an innovative planning and design approach that
considers the resilience of urban water that is aimed at developed or
brownfield areas. It fosters the incorporation of stormwater into the
urban landscape for environmental, social and economic benefits.
Railway corridors are a promising candidate for urban designers to
look for opportunities to incorporate natural elements and transform
“vacant” space into liveable space. As Penone et al. (2012) discussed,
railway could bear an ecologically functional connectivity in the
fragmented urban context, especially when it runs across densely populated areas. An example is the conversion of rail spaces to the green

recreational belt “Green Rail Track” in Amersfoort (Utrecht, Holland)
from 2011 to 2013 (Hoofwijk et al., 2013). Therefore, the wellplanned application of vegetated treatment methods along the tracks
serves multiple purposes — reducing pollutant flux, providing structural
connectivity for plant communities and integration of greenery into the
grey infrastructure. These outcomes look promising, not only in regards
to its corridors, but also its green track application. Green tracks for light
railways have been very common in the UK, Germany, Netherlands and
France. In the study of Tapia Silva et al. (2006), they assessed the ability
of green tracks in reducing runoff volume by infiltration and


362

P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363

evaporation. Nevertheless, the development of ballastless rail tracks
provides a good opportunity for incorporation of green turf on the surface of concrete slabs.
5. Conclusion
To the best of our knowledge, this is the first paper to review
stormwater management practices in the railway industry. This paper
points out the changes in perception of stormwater management
systems, from a drainage system (in the past) to a quality control (in
the present) and a resource in urban areas (in the future). To date, the
contamination of stormwater in railway areas has not been properly
studied. From the limited literature available, this paper tries to analyse
potential sources of pollutants and their pollution pathways. However,
stormwater pollution and management levels vary from country to
country. Some pioneering countries, such as the US, the UK and
Australia, have already issued relevant regulations and guidelines
for controlling pollution from the railway industry. This paper also

addresses the managerial challenges and provides provisions for future
management of stormwater quality in the railway industry.
From this study, we have found that there are many gaps in this field
that are open to further research:
– To survey stormwater quality from different assets of railway infrastructures such as rail tracks and embankments, stations, stabling
yards and depots;
– To model the transport behaviours and mass balance of pollutants
through various types of track bed and embankments;
– To explore the environmental fates of contaminants (PAHs, herbicides and heavy metals) under real railway conditions;
– To investigate particle size distributions along rail tracks and their
effects on the selection of treatment methods for stormwater;
– To study different methods for treating stormwater in the railway
industry.

Acknowledgements
The authors acknowledge the Sustainable Water Program — Wastewater treatment and Reuse Technologies, the Centre for Technology in
Water and Wastewater (CTWW), the School of Civil and Environmental
Engineering, the University of Technology, Sydney (UTS) and UTS —
Vietnam International Education Development Scholarship.
References
Aasvang, G.M., Engdahl, B., Rothschild, K., 2007. Annoyance and self-reported sleep disturbances due to structurally radiated noise from railway tunnels. Appl. Acoust. 68,
970–981.
Ali, S.A., 2005. Railway noise levels, annoyance and countermeasures in Assiut, Egypt.
Appl. Acoust. 66, 105–113.
ANZECC, ARMCANZ, 2000. Australian Guidelines for Water Quality Monitoring and
Reporting. Australian Water Association, Artarmon.
AREMA, 2008. M/W budgets to climb in 2008. Rail Tracks and Structures 104. SimmonsBoardman Publishing Company, New York.
Australian Rail Track Corporation Ltd., 2006. Track Drainage — Inspection and Maintenance. Australian Rail Track Corporation Ltd., Australia.
Australian Rail Track Corporation Ltd., 2013. Track Drainage — Design and Construction.
Australian Rail Track Corporation Ltd., Australia.

Barbosa, A.E., Fernandes, J.N., David, L.M., 2012. Key issues for sustainable urban
stormwater management. Water Res. 46, 6787–6798.
Becker, L., Matuschek, G., Lenoir, D., Kettrup, A., 2001. Leaching behaviour of wood treated
with creosote. Chemosphere 42, 301–308.
Bending, G.D., Rodriguez-Cruz, M.S., 2007. Microbial aspects of the interaction between
soil depth and biodegradation of the herbicide isoproturon. Chemosphere 66,
664–671.
Bratieres, K., Fletcher, T.D., Deletic, A., Zinger, Y., 2008. Nutrient and sediment removal by
stormwater biofilters: a large-scale design optimisation study. Water Res. 42,
3930–3940.
Britt, C., Mole, A., Kirkham, F., Terry, A., 2003. The Herbicide Handbook: Guidance on the
Use of Herbicides on Nature Conservation Sites. English Nature, Yorkshire, the UK.

Brooks, K.M., 2001. The Environmental Risks Associated with the Use of Pressure Treated
Wood in Railway Rights-of-way. Railway Tie Association, Fayetteville, GA.
Brooks, K.M., 2004. Polycyclic Aromatic Hydrocarbon Mitigation from Creosote-treated
Railway Ties into Ballast and Adjacent Wetlands. United States Department of Agriculture, Madison, WI.
Bukowiecki, N., Gehrig, R., Hill, M., Lienemann, P., Zwicky, C.N., Buchmann, B., et al., 2007.
Iron, manganese and copper emitted by cargo and passenger trains in Zürich
(Switzerland): size-segregated mass concentrations in ambient air. Atmos. Environ.
41, 878–889.
Burkhardt, M., Rossi, L., Chevre, N., Boller, M., Steidle, L., Abrecht, J., et al., 2005.
Gewässerschutz an Bahnanlagen - Emittierte Stoffe im Normalbetrieb der SBB
sowie Grundlagen zu deren Umweltverhalten. Eawag, Dubendorf.
Burkhardt, M., Rossi, L., Boller, M., 2008. Diffuse release of environmental hazards by railways. Desalination 226, 106–113.
Cederlund, H., Börjesson, E., Önneby, K., Stenström, J., 2007. Metabolic and cometabolic
degradation of herbicides in the fine material of railway ballast. Soil Biol. Biochem.
39, 473–484.
Chakraborty, A., 2001. Investigation of the Loss of Creosote Components From Railroad
Ties. Graduate Department of Chemical Engineering and Applied Chemistry. Master

of Applied Science (University of Toronto, Toronto, Canada).
Chen, Z., Ai, Y., Fang, C., Wang, K., Li, W., Liu, S., et al., 2014. Distribution and
phytoavailability of heavy metal chemical fractions in artificial soil on rock cut slopes
alongside railways. J. Hazard. Mater. 273, 165–173.
Clark, S.E., Pitt, R., 2012. Targeting treatment technologies to address specific stormwater
pollutants and numeric discharge limits. Water Res. 46, 6715–6730.
Comptroller and Auditor General of India, 2013. Performance Audit of Environment Management in Indian Railways. International Centre for Environmental Audit and Sustainable Development, India.
Cooker, J., 1996. Monitoring of Pesticides by the NRA in Connection With the Spraying of
Railway Lines to Control Weed Growth. Environment Agency Report, Australia.
Copeland, E., Lefler, D., 2013. Negotiations and challenges to revise NPDES permit limit for
copper and ammonia. 2013 Railroad Environmental Conference. University of Illinois,
Urbana-Champaign, Illinois, p. 9.
Croll, B.T., 1991. Pesticides in surface waters and groundwaters. Water Environ. J. 5,
389–395.
Department of Health — Office of Chemical Safety, 2014. Acceptable Daily Intakes for Agricultural and Veterinary Chemicals. The Office of Chemical Safety (Office of Health
Protection, Department of Health), Canberra.
Dikshith, T.S.S., Diwan, P.V., 2003. Industrial Guide to Chemical and Drug Safety. A John
Wiley & Sons, Inc., Publication, New Jersey.
Dincer, F., Elbir, T., 2007. Estimating national exhaust emissions from railway vehicles in
Turkey. Sci. Total Environ. 374, 127–134.
Directorate-general health & consumer protection, 2001. Review Report for the Active
Substance Amitrole (Unit E1 — Legislation Relating to Crop Products and Animal Nutrition). European Commission, Belgium.
Dunning, R., 2012. All of your storm drain maps are wrong. 2012 Railroad Environmental
Conference. University of Illinois, Urbana-Champaign, Illinois, p. 9.
Dunning, R., Weiner, J.L., 2011. Hope is not an approved BMP — lessons in Railroad
Stormwater Management from the West Coast. 2011 Railroad Environmental Conference. University of Illinois, Urbana-Champaign, Illinois, p. 28.
Elliott, A.H., Trowsdale, S.A., 2007. A review of models for low impact urban stormwater
drainage. Environ. Model Softw. 22, 394–405.
Esveld, C., 1997. Low-maintenance ballastless track structures. Rail Eng. Int. Ed. 3, 13–16.
European Asphalt Pavement Association (EAPA), 2003. Asphalt in Railway Track.

Breukelen, European Asphalt Pavement Association, The Netherlands.
Gil, K., Im, J., 2014. Analysis of non-point source characteristics of heavy metals and
oil and grease at railway bridge area with various land uses. Desalin. Water Treat.
1–7.
Gill, G., 2012. Development of Stormwater Quality Profiles and Treatment Strategies for
Selected Railway Infrastructure. School of Civil and Environmental Engineering, Faculty of Engineering and IT. Bachelor of Engineering (University of Technology, Sydney, UTS).
Gunasekara, A.S., Troiano, J., Goh, K.S., Tjeerdema, R.S., 2007. Chemistry and Fate of Simazine vol. 189. Springer, The US.
Gustafsson, M., Blomqvist, G., Håkansson, K., Lindeberg, J., Nilsson-Påledal, S., 2007.
Järnvägens föroreningar — källor, spridning och åtgärder: En litteraturstudie (Railway Pollution — Sources, Dispersion and Measure: A Literature Review). VTI (Statens
väg- och transportforskningsinstitut), Sweden.
Hansen, M.J., Clevenger, A.P., 2005. The influence of disturbance and habitat on the presence of non-native plant species along transport corridors. Biol. Conserv. 125,
249–259.
Hayes, A.W., Kruger, C.L., 2014. Hayes' Principles and Methods of Toxicology. Sixth
Edition. CRC Press, London.
Heather, A.I.J., Hollis, J.M., 1999. Agency E. Losses of Six Herbicides From a Disused
Railway Formation. Environment Agency.
Hoofwijk, H., Stobbelaar, D.J., Gestel, Dv., 2013. Een Groen Spoor door Amersfoort
Wageningen UR Science Shop Amersfoort.
Indraratna, B., Salim, W., Rujikiatkamjorn, C., 2011. Chapter 8 — track drainage and use of
geotextiles. In: Rujikiatkamjorn, C. (Ed.), Advanced Rail Geotechnology — Ballasted
Track. CRC Press, London, UK, pp. 203–217.
Juhasz, M., Princz-Jakovics, T., Vörös, T., 2013. What are the real effects of railway electrification in Hungary? European Transport Conference 2013 40. AET, Franfurt,
Germany.
Kamga, C., Yazici, M.A., 2014. Achieving environmental sustainability beyond technological improvements: potential role of high-speed rail in the United States of America.
Transp. Res. Part D: Transp. Environ. 31, 148–164.


P.T. Vo et al. / Science of the Total Environment 512–513 (2015) 353–363
Kang, S.-M., Morrell, J.J., Simonsen, J., Lebow, S., 2005. Creosote movement from treated
wood immersed in fresh water. For. Prod. J. 55, 42–46.

Kayhanian, M., Fruchtman, B.D., Gulliver, J.S., Montanaro, C., Ranieri, E., Wuertz, S., 2012.
Review of highway runoff characteristics: comparative analysis and universal implications. Water Res. 46, 6609–6624.
Kohler, M., Künniger, T., Schmid, P., Gujer, E., Crockett, R., Wolfensberger, M., 2000. Inventory
and emission factors of creosote, polycyclic aromatic hydrocarbons (PAH), and phenols
from railroad ties treated with creosote. Environ. Sci. Technol. 34, 4766–4772.
Kouroussis, G., Pauwels, N., Brux, P., Conti, C., Verlinden, O., 2014. A numerical analysis of
the influence of tram characteristics and rail profile on railway traffic ground-borne
noise and vibration in the Brussels Region. Sci. Total Environ. 482–483, 452–460.
Kurzanski, P., Custer, B., McFarland, T., 2013. The story of a drainage ditch: from iron staining to PCBs to Mother Nature. 2013 Railroad Environmental Conference. University of
Illinois, Urbana-Champaign, Illinois.
Langeveld, J.G., Liefting, H.J., Boogaard, F.C., 2012. Uncertainties of stormwater characteristics and removal rates of stormwater treatment facilities: implications for
stormwater handling. Water Res. 46, 6868–6880.
Larsson, M., 2004. Toxicity Analysis by Luminous Bacteria — Investigation of Surface Water
and Groundwater From Gallivare Yard. Division of waste technology — Department
of Civil Engineering (Master of Engineering. Lulea University of Technology, Sweden).
Lee, H., Swamikannu, X., Radulescu, D., Kim, S.-j., Stenstrom, M.K., 2007. Design of
stormwater monitoring programs. Water Res. 41, 4186–4196.
Leecaster, M.K., Schiff, K., Tiefenthaler, L.L., 2002. Assessment of efficient sampling designs
for urban stormwater monitoring. Water Res. 36, 1556–1564.
Li, Y.L., Deletic, A., Alcazar, L., Bratieres, K., Fletcher, T.D., McCarthy, D.T., 2012. Removal of
Clostridium perfringens, Escherichia coli and F-RNA coliphages by stormwater
biofilters. Ecol. Eng. 49, 137–145.
Lim, H.S., Lim, W., Hu, J.Y., Ziegler, A., Ong, S.L., 2015. Comparison of filter media materials
for heavy metal removal from urban stormwater runoff using biofiltration systems.
J. Environ. Manag. 147, 24–33.
Liu, H., Chen, L.-P., Ai, Y.-W., Yang, X., Yu, Y.-H., Zuo, Y.-B., et al., 2009. Heavy metal contamination in soil alongside mountain railway in Sichuan, China. Environ. Monit. Assess. 152, 25–33.
Malawska, M., Wiłkomirski, B., 2001. An analysis of soil and plant (Taraxacum officinale)
contamination with heavy metals and polycyclic aromatic hydrocarbons (PAHs) in
the area of the railway junction Iława Główna, Poland. Water Air Soil Pollut. 127,
339–349.

National Research Council, 2008. Urban Stormwater Management in the United States.
The National Academies Press, Washington, D.C.
NSW Environment Protection Authority, 1997. Managing Urban Stormwater: Treatment
Techniques. NSW Environment Protection Authority, Sydney.
Osborne, M., Montague, K., 2005. The Potential for Water Pollution from Railways. CIRIA
(Construction Industry Research and Information Association), London, UK.
Penone, C., Machon, N., Julliard, R., Le Viol, I., 2012. Do railway edges provide functional
connectivity for plant communities in an urban context? Biol. Conserv. 148, 126–133.
Piatak, N.M., Parsons, M.B., Seal Ii, R.R., 2015. Characteristics and environmental aspects of
slag: a review. Appl. Geochem. (in press, Corrected Proof (available online 30/04/
2014)).
Rogne, B., 2012. Stormwater survival guide: preparation strategies for stormwater regulations in flux. 2012 Railroad Environmental Conference. University of Illinois, UrbanaChampaign, Illinois, p. 50.
Rushton, K.R., Ghataora, G., 2009. Understanding and modelling drainage of railway ballast. Proc. Inst. Civ. Eng. Transp. 162, 227–236.
Rushton, K.R., Ghataora, G., 2014. Design for efficient drainage of railway track foundations. Proc. Inst. Civ. Eng. Transp. 167, 3–14.
Salma, I., Pósfai, M., Kovács, K., Kuzmann, E., Homonnay, Z., Posta, J., 2009. Properties and
sources of individual particles and some chemical species in the aerosol of a metropolitan underground railway station. Atmos. Environ. 43, 3460–3466.
Sanayei, M., Maurya, P., Moore, J.A., 2013. Measurement of building foundation and
ground-borne vibrations due to surface trains and subways. Eng. Struct. 53, 102–111.
Sarmah, A.K., Close, M.E., Mason, N.W.H., 2009. Dissipation and sorption of six commonly
used pesticides in two contrasting soils of New Zealand. J. Environ. Sci. Health B 44,
325–336.
Scholes, L., Revitt, D.M., Ellis, J.B., 2008. A systematic approach for the comparative assessment of stormwater pollutant removal potentials. J. Environ. Manag. 88, 467–478.
Schultz, D., Godlewski, M., 2012. Railway stormwater permits: all 46 industrial
stormwater permits are not equally created. 2012 Railroad Environmental Conference. University of Illinois, Urbana-Champaign, Illinois, p. 21.

363

Schweinsberg, F., Abke, W., Rieth, K., Rohmann, U., Zullei-Seibert, N., 1999. Herbicide use
on railway tracks for safety reasons in Germany? Toxicol. Lett. 107, 201–205.
Singaraja, D., Gayman, T., Shepard, R., 2012. Spill and contaminant release from storm

water infrastructure through a comprehensive investigation approach. 2012 Railroad
Environmental Conference. University of Illinois, Urbana-Champaign, Illinois, p. 30.
Skark, C., Zullei-Seibert, N., Willme, U., Gatzemann, U., Schlett, C., 2004. Contribution of
non-agricultural pesticides to pesticide load in surface water. Pest Manag. Sci. 60,
525–530.
Smith, E., Smith, J., Naidu, R., 2006. Distribution and nature of arsenic along former railway corridors of South Australia. Sci. Total Environ. 363, 175–182.
Stagge, J.H., Davis, A.P., Jamil, E., Kim, H., 2012. Performance of grass swales for improving
water quality from highway runoff. Water Res. 46, 6731–6742.
Tapia Silva, F.O., Wehrmann, A., Henze, H.-J., Model, N., 2006. Ability of plant-based surface technology to improve urban water cycle and mesoclimate. Urban For. Urban
Green. 4, 145–158.
Tate, C., 2014. More Oil Spilled From Trains in 2013 Than in Previous 4 Decades, Federal
Data Show. McClatchyDC — Watching Washington and the world (McClatchyDC,
Washington DC).
Teixeira, P.F., Ferreira, P.A., Pita, A.L., Casas, C., Bachiller, A., 2009. The use of bituminous
subballast on future high-speed lines in Spain: structural design and economical impact. IJ. Int. J. Railw. 2, 1–7.
Thierfelder, T., Sandström, E., 2008. The creosote content of used railway crossties as compared with European stipulations for hazardous waste. Sci. Total Environ. 402,
106–112.
Torstensson, L., Cederlund, H., Borjesson, E., Stenstrom, J., 2002. Environmental problems
with the use of diuron on Swedish railways. Pestic. Outlook 13, 108–111.
Torstensson, L., Börjesson, E., Stenström, J., 2005. Efficacy and fate of glyphosate on Swedish railway embankments. Pest Manag. Sci. 61, 881–886.
Transportation for NSW, 2013. NSW Sustainable Design Guidelines Version 3.0 Transportation for NSW. NSW.
Trombetta Zannin, P., Bunn, F., 2014. Noise annoyance through railway traffic — a case
study. J. Environ. Health Sci. Eng. 12, 1–12.
U.S. Army Corps of Engineers, 2004. Unified Facilities Criteria — Railroad Design and Rehabilitation. U.S. Army Corps of Engineers, Washington, DC (USA).
University of Hertfordshire, 2013. The Pesticide Properties DataBase (PPDB) (2006–2013).
Agriculture & Environment Research Unit (AERU), University of Hertfordshire, UK.
US EPA, 2009a. Industrial Stormwater Fact Sheet Series — Sector P: Motor Freight Transportation Facilities, Passenger Transportation Facilities, Petroleum Bulk Oil Stations
and Terminals, Rail Transportation Facilities, and United States Postal Service Transportation Facilities. US EPA, USA.
US EPA, 2009b. Industrial Stormwater Monitoring and Sampling Guide — Final Draft. US
EPA, USA.

Utley, W.S., 2005. Creosote. In: Wexler, P. (Ed.), Encyclopedia of Toxicology, Second Edition Elsevier, New York, pp. 677–678.
Victorian Rail Industry Environmental Forum, 2007. Vegetation Management Guidelines
for Rail Corridors. Victorian Rail Industry Environmental Forum, Victoria, Australia.
Wilkemirski, B., Sudnik-Wojcikowska, B., Galera, H., Wierzbicka, M., Malawska, M., 2011.
Railway transportation as a serious source of organic and inorganic pollution. Water
Air Soil Pollut. 218, 333–345.
Wong, T., Breen, P., Lloyd, S., 2000. Water Sensitive Road Design – Design Options for Improving Stormwater Quality of Road Runoff, Technical Report 00/1. Cooperative Research Centre for Catchment Hydrology.
Wong, T.H.F., Allen, R., Brown, R.R., Deletić, A., Gangadharan, L., Gernjak, W., et al., 2013.
Blueprint2013 — Stormwater Management in a Water Sensitive City. In: Wong,
T.H.F. (Ed.), Cooperative Research Center for Water Sensitive Cities, Melbourne,
Australia.
World Bank, 2007. Railway Database 2007. World Bank, Washington DC.
World Bank, 2014. Railway Lines. World Bank, Washington DC (the US).
Zhang, H., Wang, Z., Zhang, Y., Hu, Z., 2012. The effects of the Qinghai–Tibet railway on
heavy metals enrichment in soils. Sci. Total Environ. 439, 240–248.
Zhang, K., Randelovic, A., Page, D., McCarthy, D.T., Deletic, A., 2014. The validation of
stormwater biofilters for micropollutant removal using in situ challenge tests. Ecol.
Eng. 67, 1–10.
Zhiqun, S., Jiguang, L., 2011. Enlarging Chinese railway scale in Chinese transportation system is the key of energy conservation. Energy Procedia 5, 824–828.
Zimmerman, R., 2005. Mass transit infrastructure and urban health. J. Urban Health 82,
21–32.