This may be the author’s version of a work that was submitted/acceptedfor publication in the following source:

Gbeddy, Gustav Kudjoe Seyram, Ayomi Lakmini, Ayomi Lakmini Ja-yarathne, Goonetilleke, Ashantha, Ayoko, Godwin, & Egodawatta,Prasanna(2018)Variability and uncertainty of particle build-up on urban road surfaces.Science of the Total Environment, 640 - 641, pp. 1432-1437.

This file was downloaded from: https://eprints.qut.edu.au/120273/

c© Consult author(s) regarding copyright matters

This work is covered by copyright. Unless the document is being made available under aCreative Commons Licence, you must assume that re-use is limited to personal use andthat permission from the copyright owner must be obtained for all other uses. If the docu-ment is available under a Creative Commons License (or other specified license) then referto the Licence for details of permitted re-use. It is a condition of access that users recog-nise and abide by the legal requirements associated with these rights. If you believe thatthis work infringes copyright please provide details by email to qut.copyright@qut.edu.au

License: Creative Commons: Attribution-Noncommercial-No DerivativeWorks 4.0

Notice: Please note that this document may not be the Version of Record(i.e. published version) of the work. Author manuscript versions (as Sub-mitted for peer review or as Accepted for publication after peer review) canbe identified by an absence of publisher branding and/or typeset appear-ance. If there is any doubt, please refer to the published source.

https://doi.org/10.1016/j.scitotenv.2018.05.384

https://eprints.qut.edu.au/view/person/Gbeddy,_Gustav_Kudjoe_Seyram.htmlhttps://eprints.qut.edu.au/view/person/Jayarathne,_Ayomi.htmlhttps://eprints.qut.edu.au/view/person/Jayarathne,_Ayomi.htmlhttps://eprints.qut.edu.au/view/person/Goonetilleke,_Ashantha.htmlhttps://eprints.qut.edu.au/view/person/Ayoko,_Godwin.htmlhttps://eprints.qut.edu.au/view/person/Egodawatta,_Prasanna.htmlhttps://eprints.qut.edu.au/view/person/Egodawatta,_Prasanna.htmlhttps://eprints.qut.edu.au/120273/https://doi.org/10.1016/j.scitotenv.2018.05.384

1

Variability and uncertainty of particle build-up on urban road

surfaces

Gustav Gbeddy, Ayomi Jayarathne, Ashantha Goonetilleke, Godwin A. Ayoko, Prasanna

Egodawatta

Science and Engineering Faculty, Queensland University of Technology (QUT), GPO Box

2434, Brisbane, 4001, Queensland, Australia

g.gbeddy@qut.edu.au; ayomi.jayarathne@qut.edu.au; a.goonetilleke@qut.edu.au;

g.ayoko@qut.edu.au; p.egodawatta@qut.edu.au

*Corresponding Author: Tel: 61 731384396; Fax: 61 7 31381170; Email:

p.egodawatta@qut.edu.au

2

Abstract

Particle build-up is a key stormwater pollutant process that is typically replicated using a power

function with increasing antecedent dry days. Though the use of power function is

recommended by a range of researchers, its applicability is demonstrated primarily for

residential roads. Particle build-up process is also subjected to significant variability due to

catchment heterogeneity and variability associated with source characteristics such as traffic

volumes and land use. Variability in build-up process and use of stereotype coefficients can

lead to significant model uncertainties. This study evaluates particle build-up characteristics on

urban road surfaces using an extensive field investigation program, giving specific priority to

industrial and commercial roads. Based on the outcomes, particle build-up process

characteristics and respective uncertainties were evaluated and compared for residential,

industrial and commercial road surfaces. The study primarily found that both industrial and

commercial land-uses generally manifested greater particle build-up loads compared with

residential land-uses. The study provides estimates for build-up coefficients for a range of land-

uses, including industrial and commercial with their potential uncertainties in build-up

predictions. This is a new addition to resources for stormwater quality modelling. Aside from

land use, the proximity of sites to major road networks was identified as a critical factor

influencing the variability and uncertainty of particle build-up. Variability of the fraction of

particles

3

1. Introduction

Stormwater pollution is of prime concern due to its significant negative effects on the

ecosystem and human health. A comprehensive understanding of stormwater pollutant

processes during the critical phase of pollutant build-up is crucial for curbing stormwater

pollution. This is due to the fact that subsequent pollutant wash-off is highly dependent on

preceding build-up process (Wijesiri et al., 2015). Ball et al. (1998) noted that the estimation of

accumulated pollutants available for transport during a storm event is a critical part of

stormwater quality modelling. Pollutant build-up refers to pollutants accumulation on

impervious catchment surfaces during dry weather periods (Liu et al., 2016). Significant spatial

variations have been observed in pollutant load and constituents of built-up (Deletic and Orr,

2005; Wijesiri et al., 2016) due to differences in influential factors such as land use and traffic

volumes (Helmreich et al., 2010). The build-up of particulate pollutants is generally replicated

as a power function with an asymptotic pattern with increasing antecedent dry days (Ball et al.,

1998; Goonetilleke et al., 2014). However, Wijesiri et al. (2015) have shown that variability of

build-up processes for particles with varying characteristics can affect the accuracy of power

function’s build-up prediction as well as the extent of associated uncertainties. In this regard,

particles with varying characteristics such as size are expected to exhibit different behaviour

and build-up patterns during an antecedent dry period.

Uncertainty is commonly used to describe the inaccuracies associated with any predictive

model with respect to natural processes (Rabinovich, 2005). Assessment of uncertainty in

stormwater quality modelling is an emerging field in urban stormwater quality modelling. Thus

the influence of different land use and pollutant source characteristics on the replicating

capacity of the power function has not been adequately addressed. Considering the very

dynamic nature across different land uses (Wijesiri et al., 2016) further knowledge is required

on the variability and uncertainty of pollutant build-up on urban road surfaces as most of the

4

existing studies are primarily focussed on residential land use. This will facilitate the holistic

appreciation of these concepts in stormwater quality modelling thereby aiding the formulation

of up-to-date stormwater remediation strategies. The development of an effective stormwater

management strategy is highly dependent on the availability of comprehensive knowledge

(Hvitved-Jacobsen et al., 2010) and thus the need for further investigations on pollutant build-

up variability and uncertainty.

The research under discussion seeks to examine the influence of varied land use types and

source characteristics on particle build-up variability and uncertainty using the most applicable

build-up replication equation for pollutants on urban road surfaces. Secondly, the role of

particle size on build-up pollutant variability across a varied set of land uses will be assessed.

This is highly crucial since the sorption of toxic pollutants in the stormwater environment is

highly influenced by particle size distribution. Finally, recommendations will be proposed

based on the research findings in order to enhance sustainable management of stormwater

quality during the critical phase of pollutant build-up.

2. Materials and methods

2.1 Study area and site description

Gold Coast region of Queensland, Australia was selected as the study area. Gold Coast is the

sixth largest and one of the rapidly growing cities in Australia. Due to the diversity of land uses

it offers, Gold Coast can be regarded as a truly representative urban setting to carry out an in-

depth study on particulate pollutant build-up on urban road surfaces. In this regard, five road

sites were selected for particle build-up investigations in the field. These sites entail two

industrial roads located within Nerang suburb, and two residential roads and one commercial

road located within Benowa suburb. Each site has varying traffic characteristics such as daily

traffic volumes as shown in Table 1 and proximity to major road network other than primary

5

variability in land-use. The locations of the study sites are represented in Fig.S1 in the

Supplementary Information.

2.2 Sampling and Laboratory testing

Build-up samples were collected from half-width of each road site for 1, 4, 7 and 11 antecedent

dry days (ADDs). Half-width was considered due to non-uniform distribution of particulate

build-up across road surfaces (Sartor et al., 1974). A portable dry and wet vacuuming system

(Delonghi Aqualand Model) was used for sample collection. The sampling efficiency of the

vacuum system was assessed prior to field sampling. During the assessment, a sandy loam

sample with particle sizes ranging from 0.45 to 3000 µm was collected with 92% efficiency. In

this study, particles less than 0.45µm are considered as soluble fractions. A total of twenty (20)

build-up samples were collected from all the study sites at the end of the sampling period.

Further details on similar build-up sampling can be found in Gunawardana et al. (2012) and

Jayarathne et al. (2017).

The build-up samples were tested for total suspended solids using similar gravimetric methods

as reported by Wijesiri et al. (2015). The particle size distribution (PSD) was determined using

Malvern Mastersizer 3000 instrument, which is capable of measuring particle sizes ranging

from 10nm to 3.5mm based on a laser diffraction technique (Malvern Instruments Ltd, 2015).

6

3. Results and discussion

3.1 Mathematical replication of pollutant build-up

The mathematical replication of the observed particle build-up data conforms very well to the

power function as shown in Fig. 1. The power function, M (t) = atb adopted in this study is

indicated in the Supplementary Information as Equation (S1). In order to characterize particle

build-up, the build-up coefficients, ‘a’ and ‘b’ in the power function were estimated such that

the root mean standard error (RMSE) associated with the observed and predicted build-up is

minimal (Billo, 2007; Egodawatta and Goonetilleke, 2006). The build-up coefficients and their

uncertainties in predicting pollutant build-up are presented in Table 1. The uncertainties

accompanying the build-up coefficients were assessed in the form of prediction and confidence

intervals as illustrated in Equations (S2) and (S3) respectively in the Supplementary

Information. The prediction interval (PI) indicates that there is a 95 percent probability of

predicted pollutant loads falling within this interval whilst the confidence interval (CI) shows

that there is a 95 percent probability that the true best-fit line for the measurement will occur

within the confidence limits (Verschuuren, 2013). The resultant PI and CI plots associated with

industrial and commercial, and residential land uses are shown in Fig.1 and Table 1.

Industrial and commercial land uses have similar observed particle build-up patterns and were

therefore combined during the uncertainty assessment. All the observed pollutant loads fall

within the prediction interval bands for all land uses under study; an indication of excellent

fitting of the prediction line where the majority of the observed loads also lie within the 95%

confidence interval as indicated by Fig. 1a and 1b. The anticipated particle build-up load across

various land uses are shown in Fig. 1c including the outcomes of a research conducted by

Egodawatta & Goonetilleke, 2006.

7

Estimated build-up coefficients ‘a’ and ‘b’ indicate that the variability of particle build-up

across various land uses is highly influenced by the magnitude of these parameters (see Table

1). The ‘a’ mostly determines the rate and magnitude of the particle load whilst ‘b’, the

exponent parameter normally dictates the asymptotic net particle built-up pattern as indicated

by Fig. S2. In this regard, industrial and commercial land uses in this study have higher particle

accumulation rate (1-8 g/m2/day) than residential areas (1-4 g/m

2/day). This is further evident

in the large uncertainty band of the ‘a’ estimated for industrial and commercial areas as shown

in Table 1 and Fig. 1a. Particle build-up load, therefore, correlates positively with ‘a’ and

further indicates the potential particle generation capacity of a particular land use. It was also

determined from the data simulation (see Fig. S2) that ‘b’ values greater than 0.20 will result in

particle build-up pattern reaching an asymptotic value at longer antecedent dry days (ADD). On

the other hand, those land-uses with b

8

(a)

(b) (c)

Fig. 1: Pollutant build-up and associated uncertainties for different land-uses: (a) industrial and

commercial; (b) residential; and (c) projected build-up patterns (Residential areas (ii)

and (iii) were studied by Egodawatta and Goonetilleke (2006)).

0

2

4

6

8

10

12

14

16

18

20

0 5 10 15

Par

ticl

e b

uild

-up

load

(g/

m2 )

Antecedent dry days

Observed loadPrediction linePrediction interval bandEgodawatta & GoonetillekeEgodawatta & GoonetillekeConfidence interval band

0

2

4

6

8

10

12

14

16

18

20

0 10 20

Par

ticl

e b

uild

-up

load

(g/

m2 )

Antecedent dry days

Industrial&Commercial

Residential(i)

Residential(ii)

Residential(iii)

0

2

4

6

8

10

12

14

16

18

20

0 5 10 15

Par

ticl

e b

uild

-up

load

(g/

m2)

Antecedent dry days

Observed load

Prediction line

Prediction interval band

Egodawatta & Goonetilleke

Egodawatta & Goonetilleke

Confidence interval band

9

Table 1: Study site characteristics with estimated build-up coefficients and uncertainties

Land and road use characteristics Estimated daily traffic

volume*

Power function build-up

coefficient

a (95% PI) b (95% PI)

Urban industrial (NSS &NHC) &

commercial (BST), roads with through

traffic and close proximity to major road

500, 3500, 3000 for NSS,

NHC & BST respectively

6.74

(2.46 ̶ 11.03)

0.26

(0.18 ̶ 0.53)

Urban residential (BVH & BMT ) with

high population density (townhouse and

duplex housing), roads located in close

proximity to major road

1537, 750 for BVH & BMT

respectively

3.62

(1.03 ̶ 6.37)

0.43

(0.28 ̶ 0.86)

Urban residential (G ) with high

population density (townhouse and

duplex housing), roads located far away

from a major road

2.90

0.16

Urban residential (L & P ) with moderate

population density (single detached

housing), with no through traffic located

far from a major road

1.65 0.16

Notes: PI is predictive interval; Build-up coefficients for G, L & P were adapted from

Egodawatta and Goonetilleke (2006); and * data for daily traffic volume was acquired from

Mummullage et al. (2016).

The proximity of study sites to major road networks has been identified as a critical factor in

particle build-up load on roads. This emphasizes the important role of surrounding pollutant

10

emitting sources as noted by Goonetilleke et al. (2014). In this context, urban residential areas

with single detached housing and low population density located far away from major road

networks attain faster asymptotic particle build-up. This is evident in the build-up patterns

reported by Egodawatta and Goonetilleke (2006) compared to those in the current study.

Furthermore, residential areas in this study (Residential (i)) are most likely to exhibit delayed

asymptotic net particle load compared to industrial and commercial land uses as evident in the

build-up patterns of Fig. 1c. This observation further collaborates with the higher uncertainty

margins in ‘b’ estimate for residential land-use. A previous study conducted by Ball et al.

(1998) in a suburban residential area with similar characteristics as Residential area (i) in this

research obtained comparable empirical ‘a’ and ‘b’ values of 3.77 and 0.57 respectively. This,

therefore, underscores the spatial variation of particle build-up coefficients with respect to

changes in residential land use characteristics. The differences in particle build-up coefficients

for different sites may be an indication of the disparities in site location, pollutant generation

capacity, pollutant composition and the influence of natural and artificial redistribution factors

such as wind and vehicular movement across various land-uses. The influence of these factors

on ‘a’ and ‘b’ is however, interlinked thereby presenting a unique challenge during the generic

modelling of particle build-up (Goonetilleke et al., 2014). The need to adopt the most

appropriate stormwater management system for various land-uses must, therefore, be based on

an in-depth investigation.

The values determined for ‘a’ and ‘b’ as shown in Table 1 will enable stormwater quality

modelling possible for catchments with varied land uses. This will facilitate the accurate

prediction of particle build-up behaviour with reference to a particular land-use type. It will

further assist in estimating the most probable ADD to carry out cost and time effective street

sweeping practices with respect to different land-uses. In the context of this study, the

maximum ADD of 11 days employed is relatively inadequate to clearly observe and also

11

determine the most applicable asymptotic values for each land-use. However, with the help of a

detailed examination of the particle size distribution pattern in the observed particle build-up

load, a viable ADD can be projected for an effective protection of stormwater quality.

3.2 Role of particle size distribution in particle build-up variability

The assessment of particle size distribution (PSD) of particle built-up load is highly essential.

This is due to the significant influence of particle size on particle mobility and their associated

high pollutant concentrations (Deletic and Orr, 2005). It also influences the health risk posed by

particles to humans and other animals as finer particles are highly capable of reaching the

alveoli of the respiratory systems. Moreover, Wijesiri et al. (2015) noted that the behavioural

variability of size-fractionated particulate build-up influences pollutant build-up process

variability. As a result, the assessment of build-up variability across particle size ranges for

different land-uses is absolutely vital to develop management strategies for stormwater

pollution.

The built-up particles generally ranged from 0.46 ̶ 3080 µm. In order to fully comprehend

particles behaviour during build-up on road surfaces across varied land-uses, particle size

distribution data for built-up particles were fractionated into three size ranges of 0.45-75µm,

75-150µm and 150-3500µm as specified in Table S1. Fig. 2 illustrates the variations of build-

up particles in the form of percentage by volume for different particle size ranges against the

ADDs for each study site. The highly dynamic nature of particle build-up due to the influence

of re-distribution forces is clearly evident in Fig.2. First of all, particles >150 µm generally

show an increasing relationship with ADD across all land-uses. This observation agrees with an

earlier study conducted by Wijesiri et al. (2015). As they noted, re-distribution forces such as

natural wind and vehicular movement (Egodawatta and Goonetilleke, 2006; Namdeo et al.,

1999) results in re-suspension and relocation of finer particles thereby enabling coarser

particles to accumulate with increasing ADD.

12

(a) (b)

(c) (d)

(e)

Fig. 2: Behaviour pattern of particle size ranges for different land-uses: (a) & (b) residential

BVH & BMT; (c) commercial BST; and (d) & (e) industrial NHC & NSS respectively

(Legend in (a) is valid for (b) - (e)).

0

20

40

60

80

100

0 5 10

% b

y vo

lum

e P

SD

Antecedent dry days

0.45 - 75 75 - 150150-3500

0

20

40

60

80

100

0 5 10

% b

y vo

lum

e P

SD

Antecedent dry days

0

20

40

60

80

100

0 5 10

% b

y vo

lum

e P

SD

Antecedent dry days

0

20

40

60

80

100

0 5 10

% b

y vo

lum

e P

SD

Antecedent dry days

0

20

40

60

80

100

0 5 10

% b

y vo

lum

e P

SD

Antecedent dry days

13

The behaviour pattern of particles within 75-150 µm size range shows partial variations for

different land-uses. Some of the sites such as a and e in Fig. 2 exhibit gentle decline of particles

in this size range with ADD whilst sites b, c and d in Fig. 2 show gradual increase of particles

with ADD. Finally, from Fig. 2 (a) and (c), particles within 0.45-75 µm range exhibit distinct

characteristics where the % by volume decreases significantly with ADD. However, in terms of

Fig. 2 (b) and (d), this sharp decrease reaches a minimum point around the 7th

ADD before

ascending with higher ADDs. This particle behaviour can be attributed to the attainment of

build-up equilibrium where the rate of 0.45-75 µm particles generation and loss becomes equal

around the 7th

ADD at industrial site of Hilldon Court and residential site of Mediterranean

Drive. However, after the 7th

ADD more particles are produced by the sources than lost due to

redistribution factors at these sites.

This study partly affirms the conclusion made by Wijesiri et al. (2015) that particle

14

(Kayhanian et al., 2012). In this regard, fine particles could pose an enormous challenge to the

effective performance of commonly deployed stormwater particle removal processes such as

street sweeping, filtration and sedimentation. In addition, this presents a unique health

implication to humans, other animals and plants since particles

15

4. Conclusions

This research shows the existence of significant variation and uncertainty in particle build-up

across residential, commercial and industrial land-uses. Industrial and commercial land-uses

generally manifested greater build-up of particles compared with residential land-uses. The

study provides estimates of coefficients in particle build-up replicating equation for a range of

land-uses, including industrial and commercial with their potential uncertainties in particle

build-up predictions. This is a new addition to resources for stormwater quality modelling and

will greatly help in enhancing the accuracy and reliability of modelling outcomes.

Estimated built-up coefficients for the power function indicated that industrial and commercial

land-uses accumulate more particles (1-8 g/m2/day) compared to residential areas (1-4

g/m2/day) which are evident in the large difference of the estimated coefficients compared to

residential areas. Commercial and industrial land-uses must, therefore, receive greater attention

during the deployment of stormwater mitigation measures coupled with the fact that higher

loads of deleterious stormwater pollutants such as PAHs and heavy metals are normally

associated with these areas. The behaviour and variability of particles between 0.45-3500 µm

are influenced by either particles in size ranges of >150µm and 75µm and

16

Acknowledgement

The authors will like to acknowledge the Queensland University of Technology (QUT) for

extending the postgraduate research award to Gustav Gbeddy to undertake this study. Further

appreciation goes to the Central Analytical Research Facility (CARF) under the Institute of

Future Environments, QUT where the data employed in this paper were acquired. Access to

CARF was facilitated by generous funding from the Science and Engineering Faculty, QUT.

Finally, the significant role of the Ghana Atomic Energy Commission (GAEC) is highly

recognized for granting study leave to Gustav Gbeddy in order to embark upon this study.

References

Ball, J.E., Jenks, R., Aubourg, D., 1998. An assessment of the availability of pollutant

constituents on road surfaces. The Science of the Total Environment 209, 243-254.

Billo, E.J., 2007. Excel for Scientists and Engineers Numerical Methods. John Wiley & Sons,

Inc., New Jersey.

Deletic, A., Orr, D.W., 2005. Pollution Buildup on Road Surfaces. Journal of Environmental

Engineering 131, 49-59.

Egodawatta, P., Goonetilleke, A., 2006. Characteristics of pollution build-up on residential road

surfaces. Paper presented at the 7th International conference on hydroscience and engineering

(ICHE 2006) (Eds: Piasecki M, Wang S.S.Y, Holz K.P, Kawahara M, Gonzalez A, Beran B)

Philadelphia, USA.

Goonetilleke, A., Yigitcanlar, T., Ayoko, G.A., Egodawatta, P., 2014. Sustainable Urban Water

Environment; Climate, Pollution and Adaptation. Edward Elgar, Cheltenham.

Grottker, M., 1987. Runoff quality from a street with medium traffic loading. Science of the

Total Environment 59, 457-466.

Gunawardana, C., Goonetilleke, A., Egodawatta, P., Dawes, L., Kokot, S., 2012. Source

characterisation of road dust based on chemical and mineralogical composition. Chemosphere

87, 163-170.

Helmreich, B., Hilliges, R., Schriewer, A., Horn, H., 2010. Runoff pollutants of a highly

trafficked urban road--correlation analysis and seasonal influences. Chemosphere 80, 991-997.

Hvitved-Jacobsen, T., Vollertsen, J., Nielsen, A.H., 2010. Urban and highway stormwater

pollution: Concepts and engineering. CRC Press, Taylor and Francis Group, New York.

17

Jayarathne, A., Egodawatta, P., Ayoko, G.A., Goonetilleke, A., 2017. Geochemical phase and

particle size relationships of metals in urban road dust. Environ Pollut 230, 218-226.

Kayhanian, M., McKenzie, E.R., Leatherbarrow, J.E., Young, T.M., 2012. Characteristics of

road sediment fractionated particles captured from paved surfaces, surface run-off and

detention basins. Sci Total Environ 439, 172-186.

Liu, A., Miguntanna, N.S., Miguntanna, N.P., Egodawatta, P., Goonetilleke, A., 2016.

Differentiating Between Pollutants Build-Up on Roads and Roofs: Significance of Roofs as a

Stormwater Pollutant Source. CLEAN - Soil, Air, Water 44, 538-543.

Malvern Instruments Ltd, M., 2015. Mastersizer 3000 User manual, MAN0474-07-EN-00.

Malvern Instruments Ltd, Worcestershire.

Mummullage, S., Egodawatta, P., Ayoko, G.A., Goonetilleke, A., 2016. Sources of

hydrocarbons in urban road dust: Identification, quantification and prediction. Environ Pollut

216, 80-85.

Namdeo, A.K., Colls, J.J., Baker, C.J., 1999. Dispersion and re-suspension of fine and coarse

particulates in an urban street canyon. The Science of the Total Environment 235, 3-13.

Rabinovich, S.G., 2005. Measurement Errors and Uncertainties Theory and Practice, Third

Edition. Springer Science and Media, Inc, New York.

Sartor, J.D., Boyd, G.B., Agardy, F.J., 1974. Water pollution aspects of street surface

contaminants. Journal of Water Pollution Control Federation 46 458–467.

Verschuuren, G.M., 2013. Excel 2013 for scientists. Revised and Expanded 3rd edition. Holy

Macro! Books Uniontown.

Wijesiri, B., Egodawatta, P., McGree, J., Goonetilleke, A., 2015. Process variability of

pollutant build-up on urban road surfaces. Sci Total Environ 518-519, 434-440.

Wijesiri, B., Egodawatta, P., McGree, J., Goonetilleke, A., 2016. Assessing uncertainty in

pollutant build-up and wash-off processes. Environ Pollut 212, 48-56.