Skip to main content
NCBI home page
EPA Author Manuscripts logoLink to EPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 15.
Published in final edited form as: Sci Total Environ. 2020 Feb 4;717:137136. doi: 10.1016/j.scitotenv.2020.137136

Enhancing the local air quality benefits of roadside green infrastructure using low-cost, impermeable, solid structures (LISS)

Khaled Hashad 1, Bo Yang 1, Richard W Baldauf 2,3, Parikshit Deshmukh 4, Vlad Isakov 2, K Max Zhang 1
PMCID: PMC7799493  NIHMSID: NIHMS1570866  PMID: 32062263

Abstract

Communities located in near-road environments face adverse health effects due to elevated exposures to traffic-related air pollution (TRAP). While the use of a combination of solid structures (i.e. sound walls) and vegetation barriers can be an effective TRAP mitigation tool, installing these barriers can also present challenges to local communities. Sound walls are costly, and building these structures often requires the involvement of federal, state, and local permitting agencies. In this paper, we proposed that the use of low-cost, impermeable, solid structures (LISS), e.g., an impermeable thin wooden, plastic or metal fence, combined with vegetation can provide an effective option for local communities to improve near-road air quality due to lower costs and easier implementation. We conducted Large Eddy Simulations (LES) for different potential design scenarios of LISS and vegetation barriers tested under various conditions. Our results indicate that (i) combining LISS and vegetation is more effective than either alone, (ii) combining a less dense vegetation and LISS can be as effective as a dense vegetation barrier, (iii) In certain scenarios, depending on wind speed and particle size, vegetation barriers alone might lead to elevated pollutant concentrations; however, combining LISS with vegetation can mitigate those negative impacts, (iv) placing LISS closer to the freeway and in front of the vegetation barrier enhances vertical dispersion of pollutants, and (v) increasing LISS height promotes pollutant concentration reduction. These design recommendations can be used by urban planners, developers, and local community leaders to evaluate and implement green infrastructure to mitigate TRAP.

Keywords: Air Pollution, CFD, Green Infrastructure, Urban Designs, Landscape Architecture, Community Engagement

Graphical Abstract

graphic file with name nihms-1570866-f0001.jpg

1. Introduction

People living, working, going to school, and traveling near large roadways are exposed to elevated traffic related air pollution (TRAP) (Baldauf, 2017). Roughly 45 million people in the US have been estimated to live within 100 m of a major transportation facility like highways (EPA, 2006), with many more people exposed to TRAP worldwide. Many studies have linked exposure to TRAP with a range of adverse health effects including cardiovascular and respiratory diseases, birth impairments, cancer, and premature mortality (HEI, 2010), putting millions of people at risk.

Previous studies have investigated the use of vegetation as roadside barriers to mitigate TRAP impacts. Vegetation foliage provides surfaces for pollutants to deposit on and be removed from the atmosphere. Furthermore, the presence of vegetation affects air flow, promoting mixing and vertical dispersion which further reduce pollutant concentrations. Multiple review articles describe the current state of the use of vegetation to mitigate TRAP in both modelling and field measurements (Janhall, 2015; Gallagher et al., 2015; Abhijith et al., 2017; Baldauf, 2017; Buccolieri et al., 2018; Tiwari et al., 2019). The effectiveness of roadside vegetation to reduce pollutant depends on multiple factors such as vegetation properties (e.g., density and shape), particle size and wind speed. Some field measurements have shown that vegetation can reduce downwind pollutant concentrations (Deshmukh et al., 2019; Lee et al., 2018; Al-Dabbous and Kumar, 2014; Abhijith and Kumar, 2019; Brantley et al., 2014; Lin et al., 2016). However, Hagler et al. (2012) found that the effectiveness of roadside vegetation to mitigate TRAP is inconclusive and attributed increased downwind pollutant concentrations in some locations to areas with thin vegetation, and tree spacing that allowed for gaps in the barrier. Deshmukh et al. (2019) later showed that highly porous vegetation resulted in negligible particle reduction and elevated gaseous concentrations downwind compared to air quality measurements in the clearing. Both those studies highlight that vegetation density will strongly influence the reduction or increase of pollutants by vegetation barriers. Brantley et al. (2014) and Fuller et al. (2017) showed that while vegetation reduced black carbon concentrations, there were no reductions for other particle types and larger particle sizes. In addition, both modeling (Steffens et al., 2012) and field studies (Lee et al. 2018; Ranasinghe et al., 2019) suggested that not only is reduction particle size dependent, but also reduction is dependent on wind speed. For example, Steffens et al. (2012) showed that the impact of wind speed on reduction of smaller particles (< 50 nm) is not as strong as larger particles (> 50 nm) due to strong deposition effects of the smaller particles, later confirmed by a field study (Lee et al., 2018). In addition, wind tunnel experiments (Heist et al., 2009), field measurements (Baldauf et al., 2016; Finn et al., 2010; Lee et al., 2018; Baldauf et al., 2008) and numerical simulations (Tong et al., 2016; Steffens et al., 2013, 2014) have shown that solid sound walls can also mitigate TRAP. Sound walls force the plume to go over the solid structure, which promotes vertical mixing and dispersion of pollutants.

Tong et al. (2016) compared six common roadside barrier scenarios and found that the combination of sound wall and vegetation barrier (SWVB) to be the most effective design option, consistent with field measurement results (Bowker et al., 2007; Lee et al., 2018). Combining the two structures promotes vertical mixing and enhances particle deposition at the same time. This design can be implemented in communities where sound walls are already in place, taking advantage of the existing roadway infrastructure. However, in the absence of existing sound walls, implementing the SWVB design is very challenging. First, the costs for constructing and maintaining sound walls can be high, on average ~$2.5 million per mile in the U.S. Our estimate was based on the statistics provided by FHWA (2018). Second, the existing sound walls are constructed primarily for mitigating noise pollution, not air pollution. Third, the construction of sound walls, usually within the highway right-of-way, is under the authority of the state or local department of transportation. Communities can provide input, but often have very limited control over the design and construction.

As local communities strive to identify actionable strategies to address their local air pollution concerns, identifying cost-effective, practical, and passive design options to mitigate near-road air pollution have increased in interest. We posit that combining vegetation barriers with low-cost, impermeable, solid structures (LISS) can enhance the effectiveness of roadside green infrastructure on reducing near-road air pollutant concentrations for many community applications. For example, a thin impermeable fence, 2–3 m in height, can be installed next to a row of vegetation in a near-road environment at a fraction of the cost of a sound wall. The main objective of this paper is to explore different design scenarios for a combination of vegetation and LISS, evaluating their effectiveness in mitigating near-road air pollution and developing related design recommendations. We use “fence” in this paper as a generalized term to represent various types of LISS, which can be made of materials such as wood, glass, plastic, concrete or metal but share the same property of being impermeable to air flow.

While there are some review articles that provide valuable guidelines to design roadside vegetation barriers to mitigate TRAP (Baldauf, 2017; Abhijith et al., 2017), they are limited to only evaluating current existing roadside structures. To properly assess new roadside barrier designs, they should be tested under variable conditions. However, evaluating them using field measurements has its challenges like high cost and implementation time, and little control over environmental conditions which affect the effectiveness of those designs. Well-evaluated computational fluid dynamics (CFD) tools can be used to assess roadside barrier designs under various conditions as demonstrated in previous studies (Tong et al., 2016; Ghasemian et al., 2017; Santiago et al., 2019). As LISS is a new design concept, we conducted evaluations of our CFD tools and then perform multiple simulations to explore different design scenarios under variable conditions and provide design recommendations for a combination of vegetation and LISS.

This paper is organized as follows: Section 2 highlights the different design scenarios. Then, a description of our computational approach that includes the leaf area density (LAD) profile and the domain is presented. Next, we provide a brief description of the aerodynamic and deposition model and its evaluation. Section 3 includes an analysis and discussion of the different design scenarios and their impact on pollutant reduction. Finally, Section 4 discusses the limitations of this study, our design recommendations, and conclusion.

2. Design scenarios, simulations and evaluations

2.1. Design scenarios

Figure 1 and Table S1 summarize the different roadside barrier designs along with the tested variables. We developed design scenarios that could be easily and safely implemented by local communities in a cost-effective manner. Four variables were evaluated: leaf area density (LAD), wind speed, fence location, and fence height. The main wind speed used was 3 ms−1 (at a height of 10 m), and the maximum LAD used was 1.5 m−1. Unless otherwise specified, these values should be the assumed speed and LAD. Five main types of designs were tested in this study: no barrier (i.e., baseline), fence only, vegetation only, vegetation followed by fence, and fence followed by vegetation. All the designs tested were located in the unpaved roadside, approximately 1 m away from the paved outer highway shoulder, except for the fence only far from freeway design (Cases 3 and 4). In order to compare the “Vegetation followed by fence” designs (Cases 10–14) versus the “Fence only” designs, the fence in the “Fence only far from freeway” designs (Cases 3 and 4) was located 12 m further away, which corresponds to the width of the vegetation barrier.

Figure 1:

Figure 1:

Open in a new tab

Schematic of the design scenarios that were evaluated in this study. All the designs tested were located approximately 1 m away from the paved outer highway shoulder, except for the fence only far from freeway design (Cases 3 and 4) which was located 12 m further away.

The vegetation barrier in our design scenarios consists of two rows of trees to ensure that the vegetation barrier has no gaps. If only a single row is used and some trees die in the planting process or in the future as a result of disease or insect infection, it will create a gap in the barrier which might lead to undesired consequences. Since the planting space available in an urban environment is limited, we also simulated a case with a single row of vegetation only with LAD 1.5. The results of which were close to the results of two rows of vegetation with LAD 0.75 as illustrated in Figure S1 in the Supporting Information (SI). Therefore, only the results for the two rows of vegetation with LAD 0.75 is used to avoid redundancy. While we recommend using at least two rows of vegetation, if the planting space is limited a single row could be used as long as it is properly maintained. We simulated a tree of height 10 m since vegetation used in combination with sound walls tends to be taller than the sound wall height (~ 6 m). Therefore, the dimensions of the vegetation barrier, which consists of two rows of trees, were 10 m tall and 12 m wide.

2.2. Vegetation species and Leaf Area Density (LAD) profile

We chose to simulate coniferous trees because 1) this species does not lose leaves in the winter, 2) the LAD profile starts from the ground so pollutants cannot pass unobstructed under the tree canopy, and 3) they have a high leaf density, more effective in pollutant reduction (Tong et al., 2016; Steffens et al., 2012; Neft et al., 2016; Lin and Khlystov, 2012).

The drag force and dry deposition of vegetation are functions of LAD. The LAD of vegetation varies with height. We selected the LAD profile for coniferous vegetation derived by Lalic and Mihailovic (2004), which is illustrated in Figure S2, and we adjusted the maximum LAD to match the leaf area index (LAI) of a Norway Spruce (Pokorný and Stojnič, 2012; Gower and Norman, 1991), which led to a maximum LAD of 1.5. The same profile with a maximum LAD of 0.75 was used to study the impact of the fence with less dense vegetation and is discussed in Section 3.4.

2.3. Computational domain and boundary conditions

For all different design scenarios, a domain that reflects a generic highway setting was used. Figures 2a and 2b display a top and a side view of the computational domain, respectively. Two zones were created to represent two-way traffic on the highway. Each zone is 14 m wide and 3 m high, reflective of a four-lane bound. Hashad et al. (2017) showed that vehicle induced turbulence (VIT) can be accounted for by implementing a directional body force in the traffic zones. Hence, a positive and a negative body force were implemented in the traffic zones along the driving direction to account for VIT.

Figure 2:

Figure 2:

Open in a new tab

a) Top view of the domain; b) Side view of the domain; c) The plane over which the downwind concentration was evaluated which started right after the barrier at distance 0 m and extended for a 100 m downwind.

The vegetation barrier was modeled at 10 m in height and 12 m in width (representing two rows of vegetation as described in Section 2.1), spanning a length of 80 m along the driving direction. The dimensions of the domain were 180 m (along the driving direction) × 250 m (the span-wise direction) × 60 m (the vertical direction). The domain consisted of roughly 12 million cells with an average element size of 0.5 m in the vegetation zone. The mesh used was mostly uniform with finer cells present in the vegetation and traffic regions. Figure S3 depicts a side view of the mesh in our study. Results from a simulation with a finer grid with spacing of 0.42 m showed little sensitivity to grid size when compared to the original grid (Figure S4), indicating that the original grid captures most of the energy containing eddies responsible for the turbulent kinetic energy (TKE) in the domain. The vegetation barrier is the tallest object in the domain at a height (H) of 10 m. The height of the domain was 60 m, five times greater than the height (H) of the tallest object in the domain to ensure no unphysical flow acceleration or blocking effects (Tominaga et al., 2008). The guidelines recommended in Blocken (2015) were applied so a spacing of 5H upstream of the road and on the sides of the vegetation barrier, and a spacing of 15H downwind of the vegetation barrier was used (Figure 2a).

A neutral atmospheric boundary layer (ABL) velocity profile was specified at the inlet, pressure outflow conditions were specified at the outlet, symmetry boundary conditions were applied on top of the domain, and periodic boundary conditions were applied on the sides of the domain to account for traffic flow. In our simulations, we modelled the wind perpendicular to the barrier to account for the worst-case scenario where the community is expected to be exposed to the highest concentrations.

2.4. Coupled aerodynamic and deposition model

The coupled aerodynamic and deposition model used in this study is consistent with that reported in our previous work (Steffens et al., 2014; Tong et al., 2016). A brief description of the model is provided here, and more details can be found in Section S5 of the SI.

Vegetation consists of various structures like leaves, branches, and trunks that vary in size and shape. Different elements of the vegetation induce drag on the incoming flow and influence its turbulence characteristics. To explicitly model the impact of the various structures of vegetation on the flow would be computationally prohibitive; therefore, the vegetation was spatially averaged to produce average flow speed and turbulence statistics within the canopy (Wilson and Shaw, 1977), and the vegetation is modeled as a region of fluid only. The impacts of vegetation structures are accounted for by adding appropriate sink and source terms in the governing equations. Further details on the spatial averaging approach is highlighted in Steffens et al. (2012) and Tong et al. (2016).

The governing equations are shown in the SI section S5. For our LES simulations, the dynamic Smagorinsky model was implemented. To account for drag of the vegetation structures, a sink term was introduced to the momentum Equation S2. The drag term was modeled using the following equation (Shaw and Schumann, 1992),

S a(z)| (1)

which is proportional to the plant drag coefficient Cd (=0.3), a(z), the LAD as a function of height z, the total velocity of the flow |u|, and the velocity in the direction of interest ui.

Nine particle sizes were modeled, ranging between 15–253 nm. On-road measurements (Kittelson et al., 2004) and near road measurements (Hagler et al., 2012) indicate that this range is reflective of most exhaust particles from vehicle emissions. Since the particles are small, a one-way coupling between the fluid and particles was used; that is, the particles acted as tracers to the fluid. A scalar transport equation was used to model dispersion (Equation S4) and a sink term, Sd (Dp), was added to account for deposition, defined using the following equation:

(Dp)   (Dp)Np(Dp)˜a(z) (2)

where Vd(Dp) is the deposition velocity as a function of particle size Dp, adopted from the dry deposition model developed by Zhang et al. (2001), and Np is the average particle concentration for each particle size. Details on how to estimate deposition velocity are presented in the SI Section S5.3. The model did not account for any processes that could lead to particle transformation like chemical processes or coagulation. Steffens et al. (2012) showed that coagulation played a minor role, given the short residence time of the particles.

2.5. Integrated concentrations

In presenting our downwind concentration results, we integrated the concentrations calculated for heights between 0 and 2 m along a surface located at the center of the barrier and extended 100 m downwind (Figure 2c). A height of 2 m covers human breathing height, so a reduction in pollutant concentrations will be most beneficial within that region. A 100 m behind the barrier represents the region closest to the highway, and community residents living there are exposed to the highest TRAP concentrations which are necessary to reduce. Additionally, after a 100 m the differences in pollutant concentrations between most designs were not substantial as shown later. The origin (0 m) is located right behind the barrier as illustrated in Figure 2c. The concentrations have been normalized by the maximum concentration of the no-barrier case at location 0 m for each respective wind speed. The smallest and largest particle sizes (15 nm and 253 nm) are analyzed since the smallest particle has the greatest deposition while the biggest particle has the least deposition. This provided an insight into the impact that deposition plays in pollutant reduction. The downwind concentration of particle size 88 nm for selected cases is also shown in Figures S12a and S12b.

2.6. Model evaluation

Steffens et al. (2012) evaluated the coupled aerodynamic, deposition, and coagulation model using Reynolds averaged Navier-Stokes (RANS) against field measurements obtained in Chapel Hill, NC (Hagler et al., 2012). Tong et al. (2016) improved the model in Steffens et al. (2012) by adopting LES with a dynamic subgrid model for aerodynamics to better capture flow structures and species concentration and evaluated the model performance against the same field measurements, showing good agreement. This study used the same LES model as Tong et al. (2016).

For further validation, we used our model while matching the drag term to that specified in a study conducted in a maize canopy to study the dispersion of pollen particles (Pan et al., 2014). Figure 3 compares the mean velocity and mean Reynolds stress we obtained, and it shows that our model is in good agreement with the LES and field measurements reported by Pan et al. (2014).

Figure 3:

Figure 3:

Open in a new tab

Results from the LES model used in our study versus field measurement and LES results from Pan et al. (2014) (a) Normalized height versus normalized mean velocity; (b) Normalized height versus normalized Reynolds Stress.

3. Results and discussion

3.1. Overview of physical mechanisms

To facilitate explaining the mechanisms of pollutant transport for the different designs adopted in this study, Figure 4 shows the velocity, TKE and concentration contours for selected cases, organized as follows: the contours are listed vertically in order of increasing LAD, and horizontally the contours on the left are without fence, while those on the right are the same designs when combined with fence. The contours highlight a region of 120 m behind the different barriers. A brief explanation for the mechanisms of the designs without fence will be provided, then the impact of the added fence on those designs will be discussed.

Figure 4:

Figure 4:

Open in a new tab

Contours at 3 ms−1 for the no barrier (Case 1), vegetation only LAD 0.75 (Case 7), vegetation only LAD 1.5 (Case 8), fence only far from highway (Case 3), vegetation LAD 0.75 followed by 3m fence (Case 10), and vegetation LAD 1.5 followed by 3m fence (case 11) (a) Velocity; (b) TKE; (c) Normalized concentration. The concentration has been normalized by the maximum concentration for the no barrier (Case 1). The contours highlight a region of 120 m behind the barrier as demonstrated by the scale above them. The dotted black lines highlight the concentration plateau region for the vegetation only cases where there is slow pollutant dispersion due to reduced velocity and TKE.

Cases 1, 7 and 8 describe the transition from no barrier to vegetation with increasing LAD. The drag force introduced by the vegetation slows the incoming air flow, which allows the plume to vertically disperse and diffuse. At the same time, since the vegetation barrier is porous, the plume also goes through it. Overall, cases with vegetation see 1) low velocity within the vegetation and accelerated flow around it (Figure 4a, Cases 7 and 8) compared to the no barrier case (Figure 4a, Case 1), 2) a region of reduced velocity (Figure 4a, Cases 7 and 8) and TKE (Figure 4b, Cases 7 and 8) right immediately behind the vegetation, referred to as “Low-TKE zone”, via short circuiting of turbulence by the vegetation (Poggi et al., 2004; Raupach and Thom, 1981; Finnigan, 2000), and 3) a “High-TKE zone”, after the Low-TKE zone, as a result of the turbulence produced from the shear at the vegetation top edge and in the wake. In addition, clearly shown on the velocity contours (Figure 4b), denser vegetation (i.e., LAD = 1.5) leads to more drag imposed on the flow, which creates a “Recirculation Zone” in the vegetation wake (Case 8) that was not observed for vegetation with LAD = 0.75 (Case 7). The presence of a recirculation zone behind dense vegetation has been reported in previous studies (Cassiani et al., 2008; Detto et al., 2008; Frank and Ruck, 2008). In terms of the concentration fields, pollutant dispersion is weak in the Low-TKE zone and the concentration roughly plateaus (Figure 4c, Cases 7 and 8, “Concentration plateau”). Denser vegetation leads to smaller plateau region due to the presence of a stronger High-TKE zone close to the barrier.

Compared to no-barrier case, the impermeable surface in the fence only case shifts the plume upwards and enhances vertical dispersion. Furthermore, the fence generates a recirculation region behind it (Figure 4a, Case 3, “Fence recirculation”). The recirculation along with the shear generated at the fence top edge produces TKE immediately behind the barrier as depicted in Figure 4b, Case 3 (“Fence TKE”), which disperses pollutants and generates a well-mixed region behind the fence. When the fence is combined with vegetation it promotes vertical dispersion of pollutants and produces TKE immediately behind the vegetation (Figure 4b, Cases 10 and 11, “Fence TKE”), which effectively mitigates the plateau region observed for the vegetation only cases. When the vegetation is less dense, the fence produces more turbulence since the velocity reaching the fence is higher due to less drag (Figure 4b, Cases 10) than the corresponding values for the denser vegetation case (Figure 4b, Cases 11).

The implications of the physical mechanisms described above are discussed the following sections in the context of roadside barrier designs.

3.2. Pollutant reduction evaluation for different designs

Evaluating the effectiveness of each design in terms of pollutant reduction is challenging as it depends on a number of many factors. For example, the reduction is both particle size- and locational dependent, as highlighted in Figures 5a and 5b showing the percent reduction for Case 8 (Vegetation only) compared to Case 1 (No barrier) at a distance of 30 and 70 m away from the barrier for particle sizes 15 and 253 nm, respectively. Enhanced deposition for 15 nm particles results in more reduction in concentrations than those of 253 nm particles. For both particle sizes, the relative reduction at 70 m downwind the barrier is greater than that at 30 m downwind. Furthermore, the reduction of any design will also depend on site-specific conditions such as wind speed, atmospheric stability, and traffic intensity. We tested designs at different wind speeds, and reduction values were highly influenced by the wind condition as discussed in Section 3.5.

Figure 5:

Figure 5:

Open in a new tab

Normalized concentration versus distance from the barrier for no barrier (Case 1), vegetation only (Case 8), and vegetation followed by 3m fence (Case 11) for (a) particle size 15 nm; (b) particle size 253 nm. The dashed black line on Figure 4a highlights the plateau region behind the barrier for the vegetation only (Case 8) where there is reduced pollutant dispersion. The dotted lines show the pollutant reduction value at 30 m and 70 m of the Vegetation only (Case 8) case compared to the No barrier (Case 1). Pollutant reduction depends on the particle size and location behind the barrier.

Table S2 in the SI provides the overall reduction of all the tested designs compared to the No barrier case integrated over a volume that spans horizontally from 0 to 100 m behind the barrier and vertically from 0 to 2 m above the ground, as illustrated in Figure 2. While the quantitative comparison is useful to evaluate different designs, it is important to note that those reduction values are valid for the scenarios described in this study.

3.3. Impact of adding a fence behind a vegetation barrier on pollutant concentration reduction

Figure 5 displays the downwind concentration behind the barrier for the no barrier (Case 1), vegetation only (Case 8) and vegetation followed by 3m fence (Case 11). The figure highlights that adding a fence behind the vegetation barrier can further reduce particle concentrations. Figure 6a shows the normalized vertical concentration at a distance 1H behind the barrier. As illustrated in the figure, the presence of the impermeable fence promotes vertical dispersion of the plume leading to substantial concentration reductions at ground level. Figure 6b displays the vertical profile of the TKE at a distance 1H behind the barrier for the three designs. The figure shows that the fence generates TKE immediately behind the barrier. The added TKE results in a well-mixed region behind the barrier (Figure 6a). Both the plume shift and turbulence production of the fence enhances the pollutant reductions from the vegetation barrier.

Figure 6:

Figure 6:

Open in a new tab

Vertical profile at 1H from the barrier for no barrier (Case 1), vegetation only (Case 8), and vegetation followed by 3m fence (Case 11) (a) 253 nm particle concentration; (b) TKE.

With a fence, the positive impacts of the vegetative barrier are still maintained. For vegetation only (Case 8) after the plateau region, there is a strong concentration reduction (Figure 5) from the high TKE generated behind the barrier as shown in Figure 4b, Case 8. Combining the fence with vegetation (Case 11), the vegetation recirculation is still maintained along with the high TKE region that occurs with it (Figures 4a and 4b, Case 11).

3.4. Combining less dense vegetation with a fence can have similar reduction to that of a denser vegetation only

Denser vegetation tends to perform better at pollutant reduction, which is discussed in Section S7 in the SI. At 3 ms−1, our results show that vegetation with LAD 1.5 resulted in 67 and 51 percent reduction for particle sizes 15 nm and 253 nm respectively, while vegetation with LAD 0.75 resulted in 42 and 29 percent reduction for particle sizes 15 nm and 253 nm respectively (Table S2). Sections S13 and 3.2 describe how those reduction values were obtained. While the use of denser vegetation is preferred, site-specific constraints like weather, water availability, land type, current existing vegetation, and cost will influence whether dense vegetation can be used. In addition, planting a vegetation only barrier with sufficient density can be very difficult due to the need for adequate spacing for future plant growth. Our results indicate that combining less dense vegetation with fence (Case 10) can have similar reduction in pollutant concentrations to that of a denser vegetation only barrier (Case 8) as illustrated in Figure 7. This can aid choosing roadside barrier designs based on site-specific constraints. If the use of dense vegetation is not feasible, combining less dense vegetation with a fence can be a sound alternative.

Figure 7:

Figure 7:

Open in a new tab

Normalized concentration versus distance from the barrier for the no barrier (Case 1), vegetation only LAD 0.75 (Case 7), vegetation only LAD 1.5 (Case 8), and vegetation LAD 0.75 followed by 3m fence (Case 10) (a) particle size 15 nm; (b) particle size 253 nm.

Adding a fence for vegetation with LAD 0.75 and 1.5 resulted in enhanced pollutant reductions. However, the fence recirculation and TKE generated for vegetation with LAD 0.75 followed by 3m fence (Case 10) is stronger in magnitude and more effective compared to a fence located behind vegetation LAD 1.5 (Figures 4a and 4b, Cases 10 and 11). Since the plateau region is longer for vegetation with LAD 0.75 (Figure 4c, Cases 10 and 11), adding a fence is advantageous because it results in enhanced reduction for a substantial distance behind the barrier.

3.5. Fence can mitigate the negative effects of vegetation: Impact of wind speed

Meteorological and local environmental conditions can influence the effectiveness of roadside barriers. Steffens et al. (2012) has shown that particles smaller than 50 nm are less sensitive to wind speed compared to particles greater than 50 nm. A recent field study also showed that at lower wind speeds (1–2 ms−1) vegetation only was effective at reducing particles less than 100 nm in diameter (up to 50% reduction) and little effect on particles greater than 100 nm and less than 250 nm in diameter (0–5% reduction) (Lee et al., 2018). Our results are consistent with those findings as smaller particles experienced reductions regardless of wind speed, whereas larger particles experienced an increase or no reduction in concentrations (Figure 8a). Larger particles experienced an increase at 1 ms−1 since the vegetation induced vertical dispersion was not a dominant factor for pollutant reduction. Furthermore, at low wind speeds, the TKE generated is lower in magnitude due to a weaker shear flow which contributes to less pollutant dispersion. Details on the flow structure and dispersion for vegetation and vegetation followed by fence cases at 1 ms−1 are provided in Section S8 of the SI.

Figure 8:

Figure 8:

Figure 8:

Open in a new tab

Normalized concentration versus distance from the barrier for no barrier (Case 2), vegetation only (Case 9), and vegetation followed by 3m fence (Case 12) for (a) particle size 15 nm; (b) particle size 253 nm; Particle size distribution at a distance of 15 m from the barrier for the no barrier, vegetation only and vegetation followed by 3m fence normalized by the no barrier concentration at 15 m c) Wind speed 3 ms−1 (Cases 1, 8, and 11) ( d) Wind speed 1 ms−1 (Cases 2, 9, and 12)

At 1 ms−1 (Case 9), the plateau region in the vegetation-only case, where little dispersion occurs due to reduced velocity and TKE, extends for roughly 20 m (Figure 8b). In this region, we can see an (undesirable) increase in pollutant concentration for large particles (Figures 8b) and (desirable) reductions for smaller particles (Figures 8a) compared to the no-barrier case. Adding a fence to the vegetative barrier for this condition (Case 12) can mitigate the concentration increase for the larger particles while maintaining the reduction of smaller particles (Figures 8a and 8b). This ensures that there are no harmful effects being caused by the vegetation barrier when the fence is present. Beyond the plateau region, the reduction for vegetation only and vegetation and fence cases are similar.

At low wind speeds, deposition can play an important role. Figure 8a shows that for particle size 15 nm, for vegetation and vegetation with fence designs there are substantial reduction compared to the no barrier case. At lower velocity, particles spend more time within the vegetation barrier which allows for more deposition. Figures 8c and 8d show the particle size distributions at a distance 15 m behind the barrier for two wind speeds, 3 and 1 ms−1. Compared to the condition at 3 ms−1, the concentration of smaller particles has a stronger rate of deposition at 1 ms−1 wind that results in a steeper slope closer to small particle size. Figure 8 further demonstrates that a combination of a vegetation and fence results in more reduction of pollutants and adding a fence to the vegetative barrier at 1 ms−1 ensures that there is almost no harm being caused by the vegetative barrier.

3.6. Other design considerations

3.6.1. Impact of fence location

Figure 9 shows the normalized concentration for fence only, far from highway (Case 3), fence only, close to highway (Case 5), vegetation followed by fence (Case 11), and fence followed by vegetation (Case 15). Figures 9a and 9b illustrate that for fence only more reduction is obtained when the fence is located closer to the highway due to earlier vertical dispersion of pollutants. In addition, having the fence closer to the highway also introduces the turbulence generated by the fence at an earlier location which aids pollutant dispersion. Figure S10b also demonstrates that at wind speed 1 ms−1 a fence only placed closer to the highway outperforms fence only placed further from the highway.

Figure 9:

Figure 9:

Open in a new tab

Normalized concentration contours for a) Fence only, far from highway (Case 3) b) Fence only, close to highway (Case 5), c) Vegetation followed by fence (Case 11), and d) Fence followed by vegetation (Case 15). Placing the fence close to the highway and adding vegetation to a fence increase pollutant dispersion.

The location of the fence with respect to the vegetation can play an important role in pollutant reduction. At 3 ms−1, the 3m fence followed by vegetation (Case 15, Figure 9d) had more pollutant reduction compared to vegetation followed by a 3m fence (Case 11, Figures 9c) (Figures S9a, S9b). Placing the fence before the vegetation barrier results in an earlier vertical dispersion of pollutants as shown in Figure 9. Detailed explanation of the flow structure for the 3m fence followed by vegetation is provided in Section S9 of the SI. Figure S10a also shows that at wind speed 1 ms−1 the 3m fence followed by vegetation (Case 16) results in further downwind pollutant reduction compared to vegetation followed by 3m fence (Case 12). Those results indicate that even at lower wind speed placing the fence closer to the highway and in front of the vegetation barrier is beneficial.

3.6.2. Impact of fence heights

For vegetation followed by fence, we tested three different fence heights, 2, 3, and 6 m. While a 6 m fence might not be feasible or cost effective for local communities, we included it to understand the impact of fence height on pollutant reduction, and to compare it to other more feasible options. At 3 ms−1, vegetation followed by a 6 m fence (Case 13) results in more reduction compared to vegetation followed by 2 or 3 m fences (Cases 14 and 11) (Figure S11a) since a taller fence promotes stronger vertical dispersion of pollutants as illustrated in Figure S11b. At 3 ms−1, the amount of total reduction over 100 m for particle size 15 nm was 89% for vegetation with 6m fence compared to 82% for vegetation followed by 2m fence. For particle size 253 nm, the reduction was 79% for vegetation followed by 6m fence compared to 67% for vegetation followed by 2m fence (Table S2). While there is some added reduction from increasing fence height, whether it is feasible for local communities to implement such design will depend on their resources and cost-benefit analysis. Additionally, the actual reduction will depend on the location as shown in Figure S11. Those reduction values will also change under different site conditions.

3.6.3. A combination of fence and vegetation performs better than fence only

Figure S5 demonstrates that combining a fence with vegetation, regardless of fence location, always outperforms the fence only cases in pollutant reduction. For all the other cases, regardless of wind speed or LAD, the combination always performed better than fence only. Adding vegetation to the fence promotes pollutant reduction through two different mechanisms. First, adding vegetation to a fence enhances vertical dispersion as shown in Figures 9c and 9d. Second, vegetation can further reduce smaller particles through deposition. Since we tested different combinations of fence only and vegetation with fence, we highlight some reduction values for certain cases, but we refer the reader to Table S2 if they are interested in all the cases. For example, at 3 ms−1, when the fence is located far from the highway, adding vegetation with LAD 1.5 to it increased the reduction of particle size 15 nm from 51% to 86% and reduction of particle size 253 nm from 51% to 72%. The reduction also depends on the location of fence, vegetation density, and wind speed as illustrated in Table S2.

3.6.4. Fence only versus vegetation only

The use of an impermeable fence only (Cases 3, 4, 5, and 6) can also be a useful tool to mitigate TRAP. Figure S13b shows that a 3m fence only, located close to the highway, can outperform the reduction of vegetation only for a particle size of 253 nm. For particle size 15 nm, initially fence only, close to highway, outperforms the vegetation only LAD 1.5 (Case 8). Then downwind the vegetation only has slightly more reduction due to deposition and TKE produced (Figure S13a). Since the differences in reduction between both designs was not substantial, whether fence or vegetation only is better, will depend on environmental conditions, particle size and location behind the barrier. Also, the results will be influenced by the fence and vegetation properties such as the heights of the barriers. Therefore, this comparison is meant to give some perspective about the pollutant reduction for only the designs explored in this study.

4. Conclusion and design implications

In this paper, we introduced the concept of low-cost, impermeable, solid structures (LISS) in green infrastructure designs. The effectiveness of vegetative roadside barriers is sensitive to site-specific conditions such as atmospheric stability, the presence of buildings, wind direction, and the traffic density. We demonstrated that a combination of LISS and vegetation performs better than either a fence or vegetation alone regardless of wind speed, LAD, or fence location. Combining the two enhances vertical dispersion; in addition, if vegetation is added to a fence, deposition also contributes to pollutant reduction.

Site specific constraints or current existing trees on site can limit the choice of vegetation used in the barrier. Our results indicate that a combination of a fence and less dense vegetation can achieve similar results to that of a dense vegetation barrier. Hence, if dense vegetation cannot be used, a less dense vegetation combined with LISS can be as effective. Also, the relative downwind reduction compared to vegetation only is substantial for less dense vegetation. This implies that adding LISS to a less dense vegetation barrier is beneficial.

The meteorological and highway conditions can dictate whether a vegetative barrier alone is effective, therefore, the use of a vegetative barrier alone should be site-specific. For vegetation barrier only at lower wind speed, whether reduction or increase of pollutants occurs, depends on the particle size. Adding a fence to the vegetative barrier ensures that this scenario does not lead to elevated pollutant concentrations caused by the vegetative barrier for all particle sizes.

The placing of the fence can also play an important role. If a fence only design is to be used it should be placed as close as possible to the highway. The closer the fence is located to the highway, the earlier the vertical dispersion will occur, resulting in enhanced downwind pollutant reduction. If a combination of vegetation and LISS is used, the fence should be placed before the vegetative barrier (closer to the highway) to achieve more pollutant reduction. Finally, an increased fence height results in more pollutant reduction.

Supplementary Material

Supplement1

Highlights.

  • We introduced the concept of low-cost, impermeable, solid structures (LISS).

  • A combination of LISS and vegetation is better than either alone in pollution mitigation.

  • Combining LISS with less dense vegetation is as effective as dense vegetation.

  • LISS can mitigate the elevated concentration vegetation alone might cause.

  • Increasing LISS height promotes pollutant concentration reduction.

Acknowledgements

The Cornell team acknowledges support from the National Science Foundation (NSF) through grant no. 1605407. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. The views expressed in this journal article are those of the authors and do not necessarily reflect the views or policies of the US Environmental Protection Agency.

References

  1. Abhijith KV, Kumar P, 2019. Field investigations for evaluating green infrastructure effects on air quality in open-road conditions. Atmospheric Environment 201 (October 2018), 132–147. [Google Scholar]
  2. Abhijith KV, Kumar P, Gallagher J, McNabola A, Baldauf R, Pilla F, Broderick B, Di Sabatino S, Pulvirenti B, 2017. Air pollution abatement performances of green infrastructure in open road and built-up street canyon environments - A review. Atmospheric Environment 162, 71–86. [Google Scholar]
  3. Al-Dabbous AN, Kumar P, 2014. The influence of roadside vegetation barriers on air-borne nanoparticles and pedestrians exposure under varying wind conditions. Atmospheric Environment 90, 113–124. [Google Scholar]
  4. Amorim JH, Rodrigues V, Tavares R, Valente J, Borrego C, 2013. CFD modelling of the aerodynamic effect of trees on urban air pollution dispersion. Science of the Total Environment 461–462, 541–551. [DOI] [PubMed] [Google Scholar]
  5. Baldauf R, 2017. Roadside vegetation design characteristics that can improve local, near-road air quality. Transportation Research Part D: Transport and Environment 52, 354–361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Baldauf R, Thoma E, Khlystov A, Isakov V, Bowker G, Long T, Snow R, 2008. Impacts of noise barriers on near-road air quality. Atmospheric Environment 42 (32),7502–7507. [Google Scholar]
  7. Baldauf RW, Isakov V, Deshmukh P, Venkatram A, Yang B, Zhang KM, 2016. Influence of solid noise barriers on near-road and on-road air quality. Atmospheric Environment 129, 265–276. [Google Scholar]
  8. Blocken B, 2015. Computational Fluid Dynamics for urban physics: Importance, scales, possibilities, limitations and ten tips and tricks towards accurate and reliable simulations. Building and Environment 91, 219–245. [Google Scholar]
  9. Bowker GE, Baldauf R, Isakov V, Khlystov A, Petersen W, 2007. The effects of roadside structures on the transport and dispersion of ultrafine particles from highways. Atmospheric Environment 41 (37), 8128–8139. [Google Scholar]
  10. Brantley HL, Hagler GS, J. Deshmukh P, Baldauf RW, 2014. Field assessment of the effects of roadside vegetation on near-road black carbon and particulate matter. Science of the Total Environment 468–469, 120–129. 15 [DOI] [PubMed] [Google Scholar]
  11. Buccolieri R, Santiago J. l., Rivas E, Sanchez B, 2018. Review on urban tree modelling in CFD simulations: Aerodynamic, deposition and thermal effects. Urban Forestry & Urban Greening 31 (February), 212–220. [Google Scholar]
  12. Cassiani M, Katul GG, Albertson JD, 2008. The Effects of Canopy Leaf Area Index on Airflow Across Forest Edges: Large-eddy Simulation and Analytical Results. Boundary-Layer Meteorology 126 (3), 433–460. [Google Scholar]
  13. Deshmukh P, Isakov V, Venkatram A, Yang B, Zhang KM, Logan R, Baldauf R, 2019. The effects of roadside vegetation characteristics on local, near-road air quality. Air Quality, Atmosphere and Health 12 (3), 259–270. [PMC free article] [PubMed] [Google Scholar]
  14. Detto M, Katul GG, Siqueira M, Juang JY, Stoy P, 2008. The structure of turbulence near a tall forest edge: The backward-facing step flow analogy revisited. Ecological Applications 18 (6), 1420–1435. [DOI] [PubMed] [Google Scholar]
  15. EPA, 2016. Health Impacts of Near Roadway Air Pollution and Mitigation Strategies Different Types of Roadways.
  16. FHWA, 2018. Summary of Noise Barriers Constructed by December 31, 2016. URL https://www.fhwa.dot.gov/environment/noise/noise_barriers/inventory/ (accessed 01.12.19).
  17. Finn D, Clawson KL, Carter RG, Rich JD, Eckman RM, Perry SG, Isakov V, Heist DK, 2010. Tracer studies to characterize the effects of roadside noise barriers on near-road pollutant dispersion under varying atmospheric stability conditions. Atmospheric Environment 44 (2), 204–214. [Google Scholar]
  18. Finnigan J, 2000. Turbulence in plant canopies. Annual Review of Fluid Mechanics 32, 519–571. [Google Scholar]
  19. Frank C, Ruck B, 2008. Numerical study of the airflow over forest clearings. Forestry 81 (3), 259–277. [Google Scholar]
  20. Fuller CH, Carter DR, Hayat MJ, Baldauf R, Watts Hull R, 2017. Phenology of a vegetation barrier and resulting impacts on near-highway particle number and black carbon concentrations on a school campus. International Journal of Environmental Research and Public Health 14 (2). [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gallagher J, Baldauf R, Fuller CH, Kumar P, Gill LW, McNabola A, 2015. Passive methods for improving air quality in the built environment: A review of porous and solid barriers. Atmospheric Environment 120, 61–70. [Google Scholar]
  22. Ghasemian M, Amini S, Princevac M, 2017. The influence of roadside solid and vegetation barriers on near-road air quality. Atmospheric Environment. [Google Scholar]
  23. Gower S, Norman JM, 1991. Rapid Estimation of Leaf Area Index in Conifer and Broad-Leaf Plantations. Ecological Society of America 72 (5), 1896–1900. [Google Scholar]
  24. Hagler GSW, Lin MY, Khlystov A, Baldauf RW, Isakov V, Faircloth J, Jackson LE, 2012. Field investigation of roadside vegetative and structural barrier impact on near-road ultrafine particle concentrations under a variety of wind conditions. Science of the Total Environment 419, 7–15. [DOI] [PubMed] [Google Scholar]
  25. Hashad Khaled. “Comparing Different VIT Formulations on Near-Road Dispersion of Particulate and Gaseous Pollutants”. American Association for Aerosol Research Annual Conference. Rayleigh, NC: 2017. [Google Scholar]
  26. HEI, 2010. Traffic-related air pollution: a critical review of the literature on emissions, exposure, and health effects. Health Effects Institute, 1–386. [Google Scholar]
  27. Heist DK, Perry SG, Brixey LA, 2009. A wind tunnel study of the effect of roadway configurations on the dispersion of traffic-related pollution. Atmospheric Environment 43 (32), 5101–5111. [Google Scholar]
  28. Janhall S, 2015. Review on urban vegetation and particle air pollution - Deposition and dispersion. Atmospheric Environment 105, 130–137.16 [Google Scholar]
  29. Kittelson DB, Watts WF, Johnson JP, 2004. Nanoparticle emissions on Minnesota highways. Atmospheric Environment 38 (1), 9–19. [Google Scholar]
  30. Lalic B, Mihailovic DT, 2004. An Empirical Relation Describing Leaf-Area Density inside the Forest for Environmental Modeling. Journal of Applied Meteorology 43 (4), 641–645. [Google Scholar]
  31. Lee ES, Ranasinghe DR, Ahangar FE, Amini S, Mara S, Choi W, Paulson S, Zhu Y, 2018. Field evaluation of vegetation and noise barriers for mitigation of near-freeway air pollution under variable wind conditions. Atmospheric Environment 175 (June 2017), 92–99. [Google Scholar]
  32. Lin MY, Hagler G, Baldauf R, Isakov V, Lin HY, Khlystov A, 2016. The effects of vegetation barriers on near-road ultrafine particle number and carbon monoxide concentrations. Science of the Total Environment 553, 372–379. [DOI] [PubMed] [Google Scholar]
  33. Lin MY, Khlystov A, 2012. Investigation of ultrafine particle deposition to vegetation branches in a wind tunnel. Aerosol Science and Technology 46 (4), 465–472. [Google Scholar]
  34. Neft I, Scungio M, Culver N, Singh S, 2016. Simulations of aerosol filtration by vegetation: Validation of existing models with available lab data and application to near-roadway scenario. Aerosol Science and Technology 50 (9), 937–946. [Google Scholar]
  35. Ning Z, Hudda N, Daher N, Kam W, Herner J, Kozawa K, Mara S, Sioutas C, 2010. Impact of roadside noise barriers on particle size distributions and pollutants concentrations near freeways. Atmospheric Environment 44 (26), 3118–3127. [Google Scholar]
  36. Pan Y, Chamecki M, Isard SA, 2014. Large-eddy simulation of turbulence and particle dispersion inside the canopy roughness sublayer. Journal of Fluid Mechanics, 499–534. [Google Scholar]
  37. Petroff A, Zhang L, 2010. Development and validation of a size-resolved particle dry deposition scheme for application in aerosol transport models. Geoscientific Model Development 3 (2), 753–769. [Google Scholar]
  38. Poggi D, Porporato A, Ridolfi L, Albertson JD, Katul GG, 2004. The effect of vegetation density on canopy sub-layer turbulence. Boundary-Layer Meteorology 111 (3), 565–587. [Google Scholar]
  39. Pokorný R, Stojnič S, 2012. Leaf area index of Norway spruce stands in relation to age and defoliation. Beskydy 5 (2), 173–180. [Google Scholar]
  40. Ranasinghe D, Lee ES, Zhu Y, Frausto-Vicencio I, Choi W, Sun W, Mara S, Seibt U, Paulson SE, 2019. Effectiveness of vegetation and sound wall-vegetation combination barriers on pollution dispersion from freeways under early morning conditions. Science of the Total Environment 658, 1549–1558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Raupach MR, Thom AS, 1981. Turbulence in and above Plant Canopies. Annual Review of Fluid Mechanics 13 (1), 97–129. [Google Scholar]
  42. Santiago JL, Buccolieri R, Rivas E, Calvete-Sogo H, Sanchez B, Martilli A, Alonso R, Elustondo D, Santamaría JM, Martin F, 2019. CFD modelling of vegetation barrier effects on the reduction of traffic-related pollutant concentration in an avenue of Pamplona, Spain. Sustainable Cities and Society 48 (April), 101559. [Google Scholar]
  43. Shaw RHR, Schumann U, 1992. Large-eddy simulation of turbulent flow above and within a forest. Boundary-Layer Meteorology 61 (1), 47–64. [Google Scholar]
  44. Steffens JT, Heist DK, Perry SG, Isakov V, Baldauf RW, Zhang KM, 2014. Effects of roadway configurations on near-road air quality and the implications on roadway designs.
  45. Steffens JT, Heist DK, Perry SG, Zhang KM, 2013. Modeling the effects of a solid barrier on pollutant dispersion under various atmospheric stability conditions. Atmospheric Environment 69, 76–85. 17 [Google Scholar]
  46. Steffens JT, Wang YJ, Zhang KM, 2012. Exploration of effects of a vegetation barrier on particle size distributions in a near-road environment. Atmospheric Environment 50, 120–128. [Google Scholar]
  47. Tiwari A, Kumar P, Baldauf R, Zhang KM, Pilla F, Di Sabatino S, Brattich E, Pulvirenti B, 2019. Considerations for evaluating green infrastructure impacts in microscale and macroscale air pollution dispersion models. Science of the Total Environment 672, 410–426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Tominaga Y, Mochida A, Yoshie R, Kataoka H, Nozu T, Yoshikawa M, Shirasawa T, 2008. AIJ guidelines for practical applications of CFD to pedestrian wind environment around buildings. Journal of Wind Engineering and Industrial Aerodynamics 96 (10–11), 1749–1761. [Google Scholar]
  49. Tong Z, Baldauf RW, Isakov V, Deshmukh P, Max Zhang K, 2016. Roadside vegetation barrier designs to mitigate near-road air pollution impacts. Science of the Total Environment 541, 920–927. [DOI] [PubMed] [Google Scholar]
  50. Wilson NR, Shaw RH, 1977. A Higher Order Closure Model or Canopy Flow. Journal of Applied Meteorology 16, 1197–1205. [Google Scholar]
  51. Zhang L, Gong S, Padro J, Barrie L, 2001. A size-segregated particle dry deposition scheme for an atmospheric aerosol module. Atmospheric Environment 35 (3), 549–560. [Google Scholar]
  52. Zhu Y, Hinds WC, Kim S, Sioutas C, 2002. Concentration and size distribution of ultrafine particles near a major highway. Journal of the Air and Waste Management Association 52 (9), 1032–1042 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement1
NIHMS1570866-supplement-Supplement1.pdf (10MB, pdf)

RESOURCES