Influence of low impact development construction on pollutant process of road-deposited sediments and associated heavy metals

Science of the Total Environment 613–614 (2018) 1130–1139

Contents lists available at ScienceDirect

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

Influence of low impact development construction on pollutant process of road-deposited sediments and associated heavy metals Yukun Ma a, Manli Gong a,b, Hongtao Zhao a,⁎, Xuyong Li a a b

State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, 18 Shuangqing Road, Beijing 100085, China College of Resources and Environment, University of Chinese Academy of Sciences, 19 Yuquan Road, Beijing 100049, China

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Excavation is the most critical LID construction stage for RDS amount. • LID construction leads to 5 to 50 times increase of heavy metal load in stormwater. • Influenced region: arterial road N collector road N access road N laneway. • TS in stormwater will increase by 36,694 t and 146,777 t by the year 2020 and 2030. • Suggestions are given to design LID construction regulations for pollution control.

a r t i c l e

i n f o

Article history: Received 30 June 2017 Received in revised form 13 September 2017 Accepted 17 September 2017 Available online 23 September 2017 Editor: D. Barcelo Keywords: LID construction Construction site pollution Road-deposited sediments Stormwater pollutant process Heavy metals

a b s t r a c t Intense Low-Impact Development (LID) construction in China could lead to increasingly severe stormwater and receiving water pollution due to the lack of appropriate regulation for mitigating pollution from LID construction. Samples of road-deposited sediments (RDS) were collected from 50 study sites at seven LID construction stages and four road hierarchies to analyze the pollution process and determine the size of the region influenced by LID construction. Six heavy metals were analyzed, and the RDS index model was adopted to estimate the potential heavy metal load washed off by stormwater runoff. Analysis of variance revealed that the excavation and gravel filling of rain gardens and excavation of porous pavements were critical LID construction stages that contributed the largest masses of RDS per unit area to road surfaces. Although the concentration of heavy metals at LID construction sites was lower than at sites without LID construction, the load of heavy metals washed off from LID construction was much higher. In addition, the sizes of regions influenced by accumulated RDS from LID construction descended in the following order: arterial road (600–775 m) N collector road (150–200 m) N access road (100–150 m) N laneway (20–30 m). According to the characteristics of LID construction at the study sites, the potential total solid loads in stormwater throughout China were estimated to reach 36,694 t by 2020 and 146,777 t by 2030. According to the results of analysis, several recommendations are provided for designing LID construction regulations to mitigate stormwater pollution. © 2017 Elsevier B.V. All rights reserved.

1. Introduction In China, sponge city construction has increased in response to the serious urban stormwater pollution issue, aiming to reduce runoff ⁎ Corresponding author. E-mail address: [email protected] (H. Zhao).

https://doi.org/10.1016/j.scitotenv.2017.09.174 0048-9697/© 2017 Elsevier B.V. All rights reserved.

volume and improve stormwater quality (Che et al., 2014). Low-Impact Development (LID) is an important part of the sponge city construction process, and it focuses on the source control of urban non-point pollution and reduction of runoff volume in small geographical areas (Coffman and France, 2002). In large geographical areas (for example, a watershed), other sponge city development measures, such as stormwater lakes and wetlands, are constructed to store and regulate

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

stormwater. The concept of LID has been well studied in other countries (Ahiablame et al., 2012; Baek et al., 2015; Liu et al., 2016), however, LID construction in China is still in its early stages and has only increased recently. According to the guidelines issued by the Chinese State Council (CSC, 2015), over 20% and 80% of urban areas should meet stormwater treatment targets through LID by 2020 and 2030, respectively. This will lead to rapid LID construction within the next decade. LID structures are primarily constructed in extant urban areas in China; therefore, construction activities could contribute large amounts of solids to road surfaces. Intense LID construction activities will affect the pollution process in terms of the build-up and wash-off of road-deposited sediments (RDS). Build-up is the accumulation of RDS on impervious surfaces during dry periods, while wash-off is the mobilization of RDS during rainfall events (Chiew et al., 1997). Pollutant build-up is influenced by anthropogenic factors, such as traffic, land use, street sweeping practices, and climate characteristics (Ball et al., 1998; Namdeo et al., 1999). The wash off process is mainly influenced by rainfall intensity; higher pollutant loads in stormwater runoff are often correlated with higher rainfall intensity (Egodawatta et al., 2007, Zhao et al., 2011). RDS are primary carriers of toxic chemical pollutants, such as heavy metals (Goonetilleke et al., 2005); therefore, pollutants generated from LID construction will enter urban stormwater and receiving water. The increased heavy metal loads in wash-off will pose risks to both human and ecological health, and limit urban stormwater recycling (Lu et al., 2014; Ma et al., 2016; Zhang et al., 2017). Thus, in-depth knowledge of the build-up and wash-off process of RDS at LID construction sites, and designing appropriate regulations to manage pollution due to LID construction in China, are critical. Although USEPA (2016) has issued guidelines to minimize environmental pollution from LID construction, they were not designed from a scientific perspective due to the lack of understanding about pollution from LID construction. The current regulations for controlling pollution from construction mainly cover urban buildings and road development (Fernández et al., 2009; Kong et al., 2015; Sillanpää and Koivusalo, 2015). However, the LID construction procedure differs significantly from other types. In addition, LID sites are often decentralized, and the surrounding population density is usually high. The differences between the construction of LID structures and other green infrastructures are described in Table 1. Consequently, the pollution process at LID sites could differ from that of other sites; therefore, it is unknown whether the current regulations are suitable for managing environmental pollution from LID construction. Considering that intense LID and urban receiving water pollution are both high in China, it is important to analyze the pollution process at LID construction sites and estimate pollutant loads that enter urban receiving water from LID construction. Numerous studies about pollution from construction have focused on water, air, and noise pollution (Ajayi et al., 2016; Hammad et al., 2016; Noor Ezlin, 2009; Wu et al., 2016). These studies have noted

1131

several specific regulations for mitigating these pollution types at construction sites, based on pollution processes (Ballesteros et al., 2010; Font et al., 2014; Ren et al., 2013; Trenouth, 2013; Zhang et al., 2013). However, studies discussing the influence of construction on RDS and stormwater pollution are limited. As RDS act as a main carrier of toxic chemical pollutants to the urban stormwater system (Li et al., 2016b; Sutherland and Tolosa, 2000), intense LID construction in China could become a primary nonpoint source for urban stormwater pollution. Pollutants in urban stormwater normally originate from traffic and land use activities (Ma et al., 2017b). Nevertheless, because of the specific characteristics of the LID procedure, existing knowledge of the pollution process is not sufficient to explain the build-up and wash-off of pollutants at LID construction sites. Therefore, considering the characteristics of LID construction and its influence on the RDS pollutant process is essential for mitigating stormwater pollution from LID construction sites. The primary aims of this study were to: (1) analyze the influence of LID construction on the build-up and wash-off of RDS and associated heavy metals; (2) investigate the regions of LID construction that influence RDS distribution; (3) estimate the potential solid loads washed-off in urban receiving water; (4) provide recommendations for designing regulations to mitigate LID construction pollution. 2. Methods and materials 2.1. Study sites This study was undertaken in Zhenjiang, Jiangsu, China. In 2014, Zhenjiang was selected as one of the first 16 sponge cities in China, and there are various case studies of LID construction in Zhenjiang that can provide a robust foundation for this study. To investigate the influence of LID construction characteristics on the RDS pollution process, different LID types at different construction stages were selected as an important criterion in study site selection. Rain gardens and porous pavements are among the most common LID structures; therefore, this study used them as examples for analyzing pollution resulting from their construction. The rain garden construction process consists of four stages, including excavation, gravel and soil filling, and planting; while porous pavement construction is conducted over three stages, including excavation, gravel filling, and paving. Considering the differences in pollutant build-up due to variations in LID construction at different sites, three roads alongside different LID structures were selected as study sites to represent one stage for each LID type. Thus, 21 study sites (3 sites × 7 construction stages) were identified for investigating the influence of LID construction on pollution. These sites were spread over seven residential districts (BHY, DWXY, JE, SMG, TYYP, XINLY, YLW) that were undergoing LID construction in two suburbs: Jingkou and Runzhou. In addition, three study sites were selected from two other residential districts (JS and XIANGLY) that were not

Table 1 Construction characteristics of LID and other green infrastructures within one residential district. Construction characteristics

Rain garden

Normal green space

Construction procedure Deepness Area Number Distribution Construction percentage

Four stages: excavation, gravel filling, soil filling, planting Deep: 1.2 m Small: 110 m2 Large: 30 Decentralized Low: 11.18%

Five stages: excavation, site preparation, soil improvement, fertilization, planting Shallow: b0.5 m Large: 1000 m2 Small: 10 Centralized High: 30%

Construction characteristics

Porous pavement

Normal pavement

Construction procedure Deepness Area Number Distribution Construction percentage

Three stages: excavation, gravel filling, paving Deep: 0.7 m Small: 150 m2 13 Decentralized Low: 6.45%

Three stages: gravel filling, land leveling, paving Shallow: 0.15 m Large: 10,000 m2 – Centralized Large: 40%

Note: all the data were collected to the study region.

1132

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

undergoing LID construction, and they were considered as background sites without LID. These two residential districts were close to the other seven to ensure that the background values were reliable for data analysis. The distribution of study sites is presented in Fig. 1, and the characteristics of each study site are presented in Fig. S1. Four roads surrounding the seven LID residential districts were selected to analyze the region influenced by LID construction (Fig. 1). It is important to note that the influenced region discussed in this paper was only attributed to extent of LID construction presented at the study sites. The extent of LID construction considered in this study referred to the area of the selected residential districts, LID construction percentage, LID construction area, and the area of individual LID structures. This information is presented in Table S1. It was assumed that the extent of LID construction at the study sites was representative of China, as the selected residential districts are typical in Zhenjiang, and Zhenjiang is a typical LID city. The road hierarchy represented four road types categorized by their usage, as follows: arterial road N collector road N access road N laneway. The classification of the four roads was based on the average daily traffic volume (27,000, 3000, 750, and 500 vehicles per day, respectively). The road site closest to the intense LID construction site was considered as the origin, followed by several sites along the roads. The distance between one site and the next was not constant and depended on the road condition. Thus, a total of 26 study sites were selected along the four roads. It is important to note that the road sites were of similar condition, in terms of road surface type, land use, and age. As the selected study sites were within the same region, the effects of climate characteristics, such as wind induced turbulence on the build-up of RDS, were not considered.

2.2. RDS sample collection RDS samples were collected from the road surfaces at the selected study sites in December 2016. Sample collection was conducted after seven antecedent dry days, as RDS load maintains an almost constant value after this period (Egodawatta et al., 2009). The sampling plot at each study site was between the middle line and the curb, and was 1– 6 m2 to ensure that the collected samples were sufficient for laboratory analysis. RDS samples were collected via a dry vacuuming procedure using a vacuum cleaner (Philips FC9712/81). The vacuum cleaner had a high efficiency (99.998%) air filtration system to capture fine particles ranging from 0.1 μm to 0.3 μm in size. This sample collection procedure was efficient for collecting RDS samples in previous studies (Zhao et al., 2010). 2.3. Laboratory analysis The collected RDS samples were separated into eight particle size fractions: b 44 μm, 44–62 μm, 62–105 μm, 105–149 μm, 149–250 μm, 250–450 μm, 450–1000 μm, and 1000–2000 μm, using a Retsch AS200 jet sieving machine. This is because particle size is important in the pollution processes of RDS and associated heavy metals (Gunawardana et al., 2014). The masses of each particle size class were weighed for particle size distribution analysis. The RDS samples were tested for six heavy metals, including cadmium (Cd), chromium (Cr), copper (Cu), lead (Pb), nickel (Ni), and zinc (Zn). These heavy metals were analyzed as they are common and potentially toxic in RDS and urban stormwater (Ma et al., 2017a).

Fig. 1. (a) Distribution of two suburbs; (b) Study sites in Jingkou; (c) Study sites in Runzhou.

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

The extraction procedure was undertaken in accordance with the National Standards of the People's Republic of China GBT 17138-1997, GBT 17141-1997, GBT 17139-1997, HJ 491-2009 (PRCMEP, 1997a, PRCMEP, 1997b, PRCMEP, 1997c, PRCMEP, 2009). An aliquot sample (0.2500 ± 0.0005 g) of each particle size fraction was placed in a Teflon crucible with 10 mL HCL, and left overnight. The crucible was then heated in a hot block digester at 110 °C until 7 mL of the HCL had evaporated. After cooling, 5 mL HNO3, 5 mL HF, and 3 mL HCLO4 were added to the crucible. The crucible was covered with a lid and heated at 160 °C. After 1.5 h, the lid was removed, and heating of the crucible continued at 180 °C until the liquid had evaporated to 3 mL. The crucible was then covered again and heated until the black organic materials on the wall had disappeared. The lid was removed and the crucible was heated until the liquid had evaporated, leaving a colloidal material. It should be noted that the entire procedure was conducted in a fume hood. Finally, the crucible was removed from the heat and cooled. The extracted sample was diluted to 50 mL using 0.5% HNO3, and was kept in the refrigerator at 4 °C before analysis. PerkinElmer NexION 300X Inductively Coupled Plasma Mass Spectrometry (ICP-MS) was used to determine the concentrations of heavy metals, according to the US EPA method 200.8 (USEPA, 1994). A standard sample with similar concentrations to the samples was tested by ICP-MS after every 30 samples to ensure that testing was reliable. As part of the quality control and quality assurance procedures, two laboratory reagent blanks and two certified reference materials (CRM) were used in each batch of samples. Geochemical standard soil (GBW07423) was used as the CRM, and the percentage of recovery was between 85%–110%, which met the limit specified in the method used.

1133

2.5. Statistical data analysis A one-way analysis of variance (one-way ANOVA) test was conducted using SPSS 23. One-way ANOVA explores whether there is a statistical difference between the means of different sample groups (Tamhane, 1977). In this study, the amounts of RDS from three different sites at the same construction stage of each LID type were averaged for analysis. This was conducted to analyze the impact of LID construction stage on RDS load. If the difference between two sample groups at P b 0.05 is significant, then the means of the two groups are significantly different from each other. In this case, the two groups are labeled with different lowercase letters (i.e., a, b). If the means of the two groups were significantly differed at P b 0.01, the two groups were labeled with different uppercases (i.e. A, B). 3. Results 3.1. Influence of LID construction on RDS distribution The influence of LID construction activities on RDS distribution was analyzed, as solids are major pollutants in stormwater (Sartor et al., 1974). Fig. 2 compares the RDS mass per unit area at the different stages

2.4. RDS index model for the load The potential load of heavy metals washed-off by stormwater runoff was estimated using the RDS index model developed by Zhao et al. (2014). This model has been adopted in previous studies to calculate the ecological risk from heavy metals associated with different sized particles in stormwater (Liu et al., 2015), and to estimate the possible wash-off load of hydrocarbons in stormwater (Li et al., 2016a). This approach uses source and transport factors. The source factor refers to the build-up of pollutants associated with RDS, and determines the maximum amount of pollutants that could be washed-off in stormwater. The source factor includes the amount of RDS, and particle sizes, load, and species of pollutants in RDS. Transport factor is the mobility of RDS to stormwater. It can be expressed as the percentage of RDS mass in stormwater runoff compared to the initial mass of RDS on road surface. The mobility of RDS was developed according to particle sizes, regardless of RDS composition. The wash-off load could differ between different comparisons of RDS particle size. The RDS index model is described by Eq. (1). RDSindex ¼ M i  C ij  A  F wi

ð1Þ

where: RDSindex is the potential amount of pollutants in stormwater runoff [kg]; Mi is the RDS mass per unit area of particle size i [g/m2]; Cij is the build-up concentration of heavy metal j associated with particles of size i [mg/kg]; A is the road surface area [m2]; Fwi is the percentage of particles of size i washed-off in runoff; 17% for b 44 μm, 10% for 44–62 μm, 4.5% for 62–105 μm, 4.3% for 105–149 μm, 2.9% for 149–250 μm, 1.5% for 250–450 μm, and 1.0% for 450– 1000 μm at a rainfall intensity of 12.87 mm/h and rainfall duration of 1 h (Zhao et al., 2011); and i , j is the number of particle sizes and heavy metal species.

Fig. 2. (a) RDS mass per unit area during rain garden construction stages; (b) RDS mass per unit area during porous pavement construction stages. (Note: values expressed as mean ± standard deviation. Different lowercases and uppercases refer to mean difference is significant at 0.05 and 0.01 level (P b 0.05 and P b 0.01), respectively. RD-E, RD-GF, RDSF, RD-P = excavation, gravel filling, soil filling, planting of rain garden construction; PP-E, PP-GF, PP-P = excavation, gravel filling, paving of porous pavement construction).

1134

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

of rain garden and porous pavement construction. It is evident that the amount of RDS was much lower prior to LID construction than it was during construction, especially during the early stages. This indicates that LID construction activities lead to significant increases in RDS. As such, stormwater pollution during LID construction could be more serious as larger amounts of solids and associated chemical pollutants could enter stormwater runoff. From Fig. 2(a), it can also be seen that RDS mass per unit area during different rain garden construction stages decreased in the following order: excavation (337.79 ± 113.38 g/m2) N gravel filling (242.24 ± 44.62 g/m2) N soil filling (62.36 ± 14.05 g/m2) N planting (21.03 ± 9.29 g/m2). The ANOVA test found that RDS mass per unit area was not significantly different between excavation and gravel filling, while RDS mass from these stages significantly differed from the other stages and background value. However, there was no significant difference between those other two stages and the background value, suggesting that excavation and gravel filling contributed the highest mass of solids to road surfaces. This could be due to soil deposition onto road surfaces, and the transportation and stacking of materials in open areas at the LID construction sites. Fig. 2(b) compares RDS mass per unit area between the different stages of porous pavement construction. RDS mass per construction stage descended in the following order: excavation (350.68 ± 85.12 g/m2) N paving (128.83 ± 41.48 g/m2) N gravel filling (78.83 ± 23.46 g/m2). The significant differences between the three construction stages and the background value indicate that RDS mass during porous pavement construction was significantly higher than that prior to construction. Furthermore, the mean difference between RDS mass during excavation and the other two stages was significant at 0.01, indicating that the excavation stage released the largest amount of RDS during porous pavement construction. The high solid loads on road surfaces during paving could result from frequent brick cutting activities. The influence of LID construction on RDS particle size distribution was analyzed, as particle size is an important parameter in studying

pollutant process (Zhao et al., 2011). The RDS mass per unit area from sites at different LID construction stages, separated by particle size, are presented in Fig. S2 in the Supplementary Information. The particle size distribution pattern was similar at all construction stages. The mass per unit area of RDS b149 μm was lower than that of RDS N149 μm. Larger masses of RDS 250–450 μm were present compared to the other particle sizes. All particle sizes were the most abundant during the excavation and gravel filling of rain gardens, and excavation for porous pavement construction. This was consistent with changes in RDS mass at different LID construction stages (Fig. 2). 3.2. Influence of LID construction on heavy metal in build-up and wash-off 3.2.1. Build-up concentration of heavy metals The build-up concentration of each heavy metal associated with different sized RDS particles during LID construction was compared. Generally, larger amounts of heavy metals adsorbed to RDS b149 μm compared with RDS N 149 μm (Fig. 3). This finding was consistent with previous studies, which found that heavy metals tend to associate with fine particles (Aryal et al., 2010; Gunawardana et al., 2014). This suggests that RDS b149 μm could contribute to increased heavy metal concentrations in stormwater runoff. Although the mass per unit area of RDS b 149 μm was lower than that of RDS N 149 μm (Fig. S2), a larger amount of RDS b149 μm accumulated on road surfaces, which would need to be reduced during LID construction process due to their ability to readily adsorb heavy metals. In addition, Fig. 3 also shows that the variability of heavy metal concentration at different LID construction stages is not clear, which was supported by the results of the ANOVA. Cd variations are presented in Figs. S3 to S9 in the Supplementary Information, and similar outcomes were observed for the other heavy metals. Therefore, the different stages of LID construction slightly influence heavy metal concentration at different RDS particle sizes. Fig. 4 compares heavy metal build-up concentration during rain garden (RD) and porous pavement (PP) construction with background

Fig. 3. Heavy metal build-up concentrations on different sized particles during LID construction process. (Note: RD-E, RD-GF, RD-SF, RD-P = excavation, gravel filling, soil filling, planting of rain garden construction; PP-E, PP-GF, PP-P = excavation, gravel filling, paving of porous pavement construction).

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

1135

Fig. 4. Build-up concentrations of heavy metals at background sites and sites being constructed with rain garden and porous pavement. (Note: RD = rain garden; PP = porous pavement).

sites, where no LID construction had taken place. The concentration of heavy metals at background sites was much higher than that of the sites undergoing rain garden and porous pavement construction. This is because RDS at these sites primarily originate construction activities such as soil excavation (approximately 1 m depth), spillages of filter media for rain gardens, and the cutting of porous bricks. These sources contain lower heavy metal concentrations than normal RDS sources, such as traffic and other anthropogenic activities.

(Fig. 5). This indicates that, although LID construction activities lead to reduced heavy metal build-up on road surfaces, these activities can increase heavy metal loading in stormwater runoff. This is because LID construction results in increased build-up of RDS on roads, which subsequently results in larger amounts of solids washed off into stormwater. Consequently, the increased solids can carry more heavy metals to stormwater, indicating that solids are the most serious pollutants from LID construction activities.

3.2.2. Wash-off load of heavy metals The potential load of heavy metals washed-off by stormwater runoff was estimated using Eq. (1). The RDSindex values for each heavy metal in stormwater generated from the LID and background sites are presented in Fig. 5. These RDSindex values were calculated as the load of heavy metals washed-off per unit area, regardless of total road surface area. Although the accumulation of heavy metals at LID construction sites was much lower than that at the background sites (Fig. 4), the RDSindex value of heavy metals at LID construction sites was 5 to 50 times higher

3.3. Influenced region of LID construction activities As LID construction activities are common in extant urban areas in China, it is important to analyze the region influenced by such activities to apply pollution control measures within an appropriate region. The region influenced was determined through comparing RDS mass per unit area along the road, starting from the most intense LID construction activities to un-influenced road surfaces. Fig. 6 provides the RDS mass per unit area along four different types of roads, according to the

1136

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

Fig. 5. Potential amount of heavy metals per unit area washed-off in stormwater. (Note: RD = rain garden; PP = porous pavement).

selected study sites. The horizontal axes in Fig. 6 represent the distance from the most intense LID construction activities (i.e., origin). The amount of RDS on road surfaces decreased in the following order: laneway (16.77–365.81 g/m2) N access road (15.37– 52.37 g/m2) N collector road (3.39–32.04 g/m2) N arterial road (3.04– 18.87 g/m2). This sequence demonstrates that pollutant load decreased with increasing road hierarchy. The higher pollutant loads on the laneway within the residential district might be attributed to the lower frequency of sweeping and proximity to the pollution source. With increasing sweeping frequency and distance from the pollutant source, the pollutant load of RDS decreases. However, RDS build-up is influenced by many other factors, as mentioned in the Introduction, which were not noted in the study. For example, traffic volume was lower on roads of lower hierarchy where RDS load was higher. Lower traffic volume leads to lower RDS loads (Barrett et al., 1998); therefore, traffic volume did not influence RDS load on the four road types. With increasing distance from the source, and becomes constant (dashed line in Fig. 6) at sites where RDS build-up was not influenced by LID construction. The largest region influenced by construction was along the arterial road at 600–775 m. This was followed by the collector road, which had an influenced region of 150–200 m, while the influenced region along the access road was 100–150 m. The region influenced by LID construction activities was the smallest along the

laneway at 20–30 m within the residential district. This suggests that the size of the region influenced by LID construction activities increases with increasing road hierarchy. This could be due to the increased traffic volume on the more heavily used roads, which can result in re-suspension of RDS. These findings suggest that control measures can be implemented according road hierarchy. For roads of lower hierarchy, sweeping frequency needs to be increased to reduce RDS loading. This is because street sweeping is a widespread technique that can remove all RDS and in turn reduce solid loading in stormwater (Vaze and Chiew, 2002). For roads of higher hierarchy, it is recommended that measures should be implemented to mitigate for the re-suspension of RDS, such as dampening of road surfaces and installation of fences. These control measures should be applied within areas according to the size of the region influenced by LID construction. 4. Discussion 4.1. Contribution of solids to stormwater from LID construction in China As solids were identified as the most serious pollutants in stormwater runoff during LID construction, the total solids (TS) loading that could be washed-off into the stormwater system was estimated to

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

1137

Fig. 6. RDS mass per unit area along the roads from the origin of the most intense LID construction activities.

develop mitigation strategies for stormwater pollution. This study estimated an increase of TS loading in stormwater due to LID construction in China as an example. This is because LID structures are currently being constructed in many Chinese areas for sponge city development. Increasingly intense LID construction could result in increasingly serious stormwater pollution; therefore, estimating the increased loading of TS entering stormwater runoff during LID construction is essential. A generalized catchment was simulated according to the study sites to estimate TS loading in stormwater runoff. The generalized catchment's characteristics (Fig. S10) and identification procedure are presented in the Supplementary Information. It was assumed that the generalized catchment was representative of a typical Chinese residential district. Section 3.1 found that RDS loading at LID construction sites was much higher than that at background sites, so intense LID construction activities were considered to be the most important pollution source during the development process. Consequently, the region influenced by construction discussed in this study was assumed to be the same for other cities. The generalized catchment was taken as a unit to estimate increased TS loading in stormwater runoff throughout China during LID construction. The information about LID construction within the generalized catchment is provided in Fig. S11 in the Supplementary Information, which suggests that the TS load in stormwater generated from the catchment would be 105.76 kg after one rainfall event (rainfall intensity: 12.87 mm/h, duration: 1 h). The calculation for this is presented in the Supplementary Information. The potential increased TS load in stormwater runoff throughout China by 2020 and 2030 were estimated based on the generalized catchment (Fig. S11). The estimations in Fig. S11 were from one rainfall event with an intensity of 12.87 mm/h and duration of 1 h. The methodology adopted in this study can be used for other conditions, and the result could differ with other rainfall characteristics. The current, total size of urban areas in China is 52,102.31 km2 (NBSPRC, 2016), and, according to the sponge city construction target, over 10,420.46 km2 and 41,681.85 km2 of the target catchment will be urbanized by 2020 and 2030, respectively. In this study, it was assumed that the target catchment will be constructed with LID, taking the generalized catchment as a unit. Consequently, the potential increased TS loading in

stormwater runoff during one rainfall event due to intense LID construction was estimated to be 36,694 t by 2020 and 146,777 t by 2030. The increased TS loading in stormwater due to LID construction is significant, according to current LID construction management practices. This could result in more serious pollution of urban receiving water due to the increased amounts of suspended solids and associated toxic chemical pollutants. Therefore, designing appropriate LID construction regulations to mitigate environmental pollution is essential. The obtained data discussed in this Section were extrapolated from the observed data in this study. Additionally, the estimation of the potential TS load in receiving water using a generalized catchment is an innovative methodology for estimating pollution of urban receiving water. 4.2. Practical recommendations for regulating LID construction Based on the findings of this study, recommendations are provided for designing LID construction regulations to mitigate stormwater pollution (Fig. 7). General recommendations are given for reducing solids at LID construction sites and the surrounding areas. Solids were identified as the most serious pollutants entering stormwater due to LID construction, so LID construction should be conducted during dry seasons with few rainfall events. Alternatively, road surface within the influenced region should be cleaned more often prior to rainfall events to reduce the amount of solids washed-off by stormwater runoff. In addition, RDS b 149 μm were found to carry higher concentrations of heavy metals; therefore effective road cleaning techniques should be adopted to remove RDS b149 μm. For example, a tandem operation of road cleaning methods could be helpful where larger particles are initially removed by a mechanical sweeper followed by a regenerative-air sweeper to remove finer particles (Amato et al., 2010). Recommendations for LID construction sites are also summarized in Fig. 7. It was found that excavation and gravel filling during rain garden construction, and excavation during porous pavement development, generated the largest amount of solids. Therefore, the frequency of road cleaning during these three construction stages should be increased, and the length of time for the three stages should be reduced. Construction of different LID structures within the same residential

1138

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

Fig. 7. Recommendations for LID construction regulation design.

district could be designed with interleaving construction stages to simultaneously mitigate for RDS. Although other stages generate fewer solids than these three stages, recommendations for them are also presented in Fig. 7. For example, a stack of construction materials in open areas was considered as a main source of solid build-up on road surfaces, and therefore materials at LID construction sites should be stored appropriately. Additionally, brick cutting during porous pavement development produces many solids, so improving technology for brick cutting could be considered. Recommendations are also provided for areas surrounding construction sites according to the analysis of influenced regions. The frequency of road cleaning should be increased, especially for laneways in residential districts. The area cleaned should be extended to 30 m along laneways, 150 m along access roads, 200 m along collector roads, and 775 m along arterial roads. To mitigate for the re-suspension of solids by vehicles, fence installation is recommended to protect the influenced region. 5. Conclusions This study analyzed the influence of rain garden and porous pavement construction on the build-up and wash-off of RDS. The outcomes from this study are expected to contribute to the understanding of how LID construction in China influences RDS pollution and stormwater pollution, due to intense LID construction without regulations for pollution control. The following important conclusions were derived from the study: • LID construction can significantly increase the amount of RDS on roads. RDS mass per unit area varied between different LID construction stages, and decreased in the following order for rain garden construction: excavation N gravel filling N soil filling N planting, and for porous pavement development: excavation N paving N gravel filling. Therefore, measures should be taken to remove RDS, especially during excavation and gravel filling of rain gardens and excavation in porous pavement development. • Although the amount of RDS b 149 μm on road surfaces during LID construction was less than that of RDS N149 μm, RDS b 149 μm

contains more heavy metals. This suggests that the removal of finer RDS particles should be enhanced during LID construction, due to their ability to adsorb heavy metals. • Heavy metal concentration at LID construction sites was less than that of the background sites without LID construction. However, wash-off of heavy metals generated from LID construction sites in stormwater was much higher than that of background sites as more RDS were washed off by runoff. • Solids are the most notable pollutants from LID construction contributing to stormwater, as the load of solids in stormwater determines heavy metal concentration. Increased solid loads in stormwater throughout China was estimated to reach 36,694 t by 2020 and 146,777 t by 2030. • The sizes of regions influenced by LID construction along four hierarchies of roads descended in the following order: arterial road (600– 775 m) N collector road (150–200 m) N access road (100–150 m) N laneway (20–30 m), while RDS loading decreased in the inverse order to this. Thus, appropriate measures should be taken to reduce RDS re-suspension on high hierarchy roads, and remove RDS on low hierarchy roads.

According to the analysis results, detailed recommendations were provided for each construction stage and the surrounding influenced areas. Due to a lack of knowledge about the influence of other construction activities on RDS and stormwater pollution, pollution from LID construction was not compared with other types of construction in this study. This could be conducted in future studies by collecting samples at other construction sites. In addition, because of the limited data, the estimation of increased TS loads in stormwater could be different in other regions in China. The methodology adopted is recommended for use to improve estimation and mitigation of stormwater pollution. Acknowledgements We are grateful to Nian She and Jiang Zhao for helpful suggestions, to Changliang Zou, Qian Jiang and Wenyan He for their help collecting

Y. Ma et al. / Science of the Total Environment 613–614 (2018) 1130–1139

road-deposited sediments. This research was supported by the National Natural Science Foundation of China (No. 41701586, 41471417, 41771529). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2017.09.174. References Ahiablame, L.M., Engel, B.A., Chaubey, I., 2012. Effectiveness of low impact development practices: literature review and suggestions for future research. Water Air Soil Pollut. 223 (7), 4253–4273. Ajayi, S.O., Oyedele, L.O., Bilal, M., Akinade, O.O., Alaka, H.A., Owolabi, H.A., 2016. Critical management practices influencing on-site waste minimization in construction projects. Waste Manag. 330–339 págs. Amato, F., Querol, X., Johansson, C., Nagl, C., Alastuey, A., 2010. A review on the effectiveness of street sweeping, washing and dust suppressants as urban PM control methods. Sci. Total Environ. 408 (16), 3070–3084. Aryal, R., Vigneswaran, S., Kandasamy, J., Naidu, R., 2010. Urban stormwater quality and treatment. Korean J. Chem. Eng. 27 (5), 1343–1359. Baek, S.-S., Choi, D.-H., Jung, J.-W., Lee, H.-J., Lee, H., Yoon, K.-S., Cho, K.H., 2015. Optimizing low impact development (LID) for stormwater runoff treatment in urban area, Korea: experimental and modeling approach. Water Res. 86, 122–131. Ball, J., Jenks, R., Aubourg, D., 1998. An assessment of the availability of pollutant constituents on road surfaces. Sci. Total Environ. 209 (2), 243–254. Ballesteros, M.J., Fernández, M.D., Quintana, S., Ballesteros, J.A., González, I., 2010. Noise emission evolution on construction sites. Measurement for controlling and assessing its impact on the people and on the environment. Build. Environ. 45 (3), 711–717. Barrett, M.E., Irish Jr., L.B., Malina Jr., J.F., Charbeneau, R.J., 1998. Characterization of highway runoff in Austin, Texas, area. J. Environ. Eng. 124 (2), 131–137. Che, W., Zhao, Y., Yang, Z., Li, J., Shi, M., 2014. Integral stormwater management master plan and design in an ecological community. J. Environ. Sci. 26 (9), 1818–1823. Chiew, F.H.S., Mudgway, L.B., McMahon, T.A., 1997. Urban Stormwater Pollution. Clayton, CRC for Catchment Hydrology. Coffman, L.S., France, R.L., 2002. Low-impact Development: An Alternative Stormwater Management Technology. CSC, 2015. Guideline to Promote Building Sponge Cities. Retrieved 25 May, 2017, 2017, from. http://english.gov.cn/policies/latest_releases/2015/10/16/content_ 281475212984264.htm. Egodawatta, P., Thomas, E., Goonetilleke, A., 2007. Mathematical interpretation of pollutant wash-off from urban road surfaces using simulated rainfall. Water Res. 41 (13), 3025–3031. Egodawatta, P., Thomas, E., Goonetilleke, A., 2009. Understanding the physical processes of pollutant build-up and wash-off on roof surfaces. Sci. Total Environ. 407 (6), 1834–1841. Fernández, M.D., Quintana, S., Chavarría, N., Ballesteros, J.A., 2009. Noise exposure of workers of the construction sector. Appl. Acoust. 70 (5), 753–760. Font, A., Baker, T., Mudway, I.S., Purdie, E., Dunster, C., Fuller, G.W., 2014. Degradation in urban air quality from construction activity and increased traffic arising from a road widening scheme. Sci. Total Environ. 497–498, 123–132. Goonetilleke, A., Thomas, E., Ginn, S., Gilbert, D., 2005. Understanding the role of land use in urban stormwater quality management. J. Environ. Manag. 74 (1), 31–42. Gunawardana, C., Egodawatta, P., Goonetilleke, A., 2014. Role of particle size and composition in metal adsorption by solids deposited on urban road surfaces. Environ. Pollut. 184, 44–53. Hammad, A.W.A., Akbarnezhad, A., Rey, D., 2016. A multi-objective mixed integer nonlinear programming model for construction site layout planning to minimise noise pollution and transport costs. Autom. Constr. 61, 73–85. Kong, S., Li, X., Li, Q., Yin, Y., Li, L., Chen, K., Liu, D., Yuan, L., Pang, X., 2015. Subway construction activity influence on polycyclic aromatic hydrocarbons in fine particles: comparison with a background mountainous site. Atmos. Res. 161–162, 82–92. Li, Y., Song, N., Yu, Y., Yang, Z., Shen, Z., 2016a. Characteristics of PAHs in street dust of Beijing and the annual wash-off load using an improved load calculation method. Sci. Total Environ. 581. Li, Y., Yu, Y., Yang, Z., Shen, Z., Wang, X., Cai, Y., 2016b. A comparison of metal distribution in surface dust and soil among super city, town, and rural area. Environ. Sci. Pollut. Res. 23 (8), 7849–7860.

1139

Liu, A., Liu, L., Li, D., Guan, Y., 2015. Characterizing heavy metal build-up on urban road surfaces: implication for stormwater reuse. Sci. Total Environ. 515–516, 20–29. Liu, A., Guan, Y., Egodawatta, P., Goonetilleke, A., 2016. Selecting rainfall events for effective Water Sensitive Urban Design: a case study in Gold Coast City, Australia. Ecol. Eng. 92, 67–72. Lu, X., Wu, X., Wang, Y., Chen, H., Gao, P., Fu, Y., 2014. Risk assessment of toxic metals in street dust from a medium-sized industrial city of China. Ecotoxicol. Environ. Saf. 106 (106), 154. Ma, Y., Egodawatta, P., McGree, J., Liu, A., Goonetilleke, A., 2016. Human health risk assessment of heavy metals in urban stormwater. Sci. Total Environ. 557-558, 764–772. Ma, Y., Liu, A., Egodawatta, P., McGree, J., Goonetilleke, A., 2017a. Assessment and management of human health risk from toxic metals and polycyclic aromatic hydrocarbons in urban stormwater arising from anthropogenic activities and traffic congestion. Sci. Total Environ. 579, 202–211. Ma, Y., Liu, A., Egodawatta, P., McGree, J., Goonetilleke, A., 2017b. Quantitative assessment of human health risk posed by polycyclic aromatic hydrocarbons in urban road dust. Sci. Total Environ. 575, 895–904. Namdeo, A., Colls, J., Baker, C., 1999. Dispersion and re-suspension of fine and coarse particulates in an urban street canyon. Sci. Total Environ. 235 (1), 3–13. NBSPRC, 2016. China Statistical Yearbook. Retrieved 5 June, 2017, 2017, from. http:// www.stats.gov.cn/tjsj/ndsj/2016/indexeh.htm. Noor Ezlin, A.B., 2009. Rapid prototyping expert systems for minimising river pollution during highway construction activities. J. Appl. Sci. 9 (12), 2280–2286. PRCMEP, 1997a. Soil Quality-determination of Copper, Zinc-flame Atomic Absorption Spectrophotometry Beijing. Ministry of Environmental Protection of the People's Republic of China. PRCMEP, 1997b. Soil Quality-determination of Lead, Cadmium-Graphite Furnace Atomic Absorption Spectrophotometry. Ministry of Environmental Protection of the People's Republic of China, Beijing. PRCMEP, 1997c. Soil Quality-determination of Nickel-flame Atomic Absorption Spectrophotometry Beijing. Protection of the People's Republic of China, Ministry of Environmental. PRCMEP, 2009. Soil Quality-determination of Total Chromium-flame Atomic Absorption Spectrophotometry Beijing. Protection of the People's Republic of China, Ministry of Environmental. Ren, J.M., Ran, X.M., Jiang, C.R., 2013. Prediction and prevention of water and soil erosionin qilinsi hydro-electric power station project construction. Appl. Mech. Mater. 298, 2107–2111. Sartor, J.D., Boyd, G.B., Agardy, F.J., 1974. Water pollution aspects of street surface contaminants. Water Pollut. Control Fed. 46 (3), 458–467. Sillanpää, N., Koivusalo, H., 2015. Stormwater quality during residential construction activities: influential variables. Hydrol. Process. 29 (19), 4238–4251. Sutherland, R.A., Tolosa, C.A., 2000. Multi-element analysis of road-deposited sediment in an urban drainage basin, Honolulu, Hawaii. Environ. Pollut. 110 (3), 483–495. Tamhane, A.C., 1977. Multiple comparisons in model i one-way anova with unequal variances. Commun. Stat. Theory Methods 6 (1), 15–32. Trenouth, W., 2013. Better management of construction sites to protect inland waters. Inland Waters 3 (2), 167–178. USEPA, 1994. Metod 200.8: Trace Elements in Waters and Wastes By Inductively Coupled Plasma-mass Spectrometry. US Environmental Protection Agency, Washington, DC. USEPA, 2016. Low Impact Development Construction, Inspection & Maintenance Guide. US Environmental Protection Agency, Washington, DC. Vaze, J., Chiew, F.H.S., 2002. Experimental study of pollutant accumulation on an urban road surface. Urban Water 4 (4), 379–389. Wu, Z., Zhang, X., Wu, M., 2016. Mitigating construction dust pollution: state of the art and the way forward. J. Clean. Prod. 112 (Part 2), 1658–1666. Zhang, X., Shen, L., Zhang, L., 2013. Life cycle assessment of the air emissions during building construction process: a case study in Hong Kong. Renew. Sust. Energ. Rev. 17, 160–169. Zhang, J., Hua, P., Krebs, P., 2017. Influences of land use and antecedent dry-weather period on pollution level and ecological risk of heavy metals in road-deposited sediment. Environ. Pollut. 228, 158. Zhao, H., Li, X., Wang, X., Tian, D., 2010. Grain size distribution of road-deposited sediment and its contribution to heavy metal pollution in urban runoff in Beijing, China. J. Hazard. Mater. 183 (1–3), 203–210. Zhao, H., Li, X., Wang, X., 2011. Heavy metal contents of road-deposited sediment along the urban-rural gradient around Beijing and its potential contribution to runoff pollution. Environ. Sci. Technol. 45 (17), 7120–7127. Zhao, H., Zhao, J., Yin, C., Li, X., 2014. Index models to evaluate the potential metal pollution contribution from washoff of road-deposited sediment. Water Res. 59, 71–79.