O

AM

ARa

b

c

d

a

ARR1AA

KADNOT

1

fuewIbr

h1

Urban Forestry & Urban Greening 22 (2017) 41–53

Contents lists available at ScienceDirect

Urban Forestry & Urban Greening

journa l h om epage: www.elsev ier .com/ locate /u fug

riginal article

ir quality affected by trees in real street canyons: The case ofarylebone neighbourhood in central London

ntoine P.R. Jeanjeana, Riccardo Buccolierib,∗, James Eddyc, Paul S. Monksd,oland J. Leigha

Department of Physics and Astronomy, University of Leicester, Leicester, UKDipartimento di Scienze e Tecnologie Biologiche ed Ambientali, University of Salento, S.P. 6 Lecce-Monteroni, 73100 Lecce, ItalyBluesky International Limited, Old Toy Factory, Jackson Street, Coalville, UKDepartment of Chemistry, University of Leicester, Leicester, UK

r t i c l e i n f o

rticle history:eceived 23 September 2016eceived in revised form8 November 2016ccepted 22 January 2017vailable online 28 January 2017

eywords:ir pollutionepositioneighbourhood scalepenFOAMrees

a b s t r a c t

This paper discusses the combined influence of building morphology and trees on air pollutant con-centrations in the Marylebone neighbourhood (central London). Computational Fluid Dynamics (CFD)simulations are performed with OpenFOAM using the k-ε model. Aerodynamic and deposition effects ofPlatanus acerifolia trees are considered. While aerodynamic effects are treated as typically done in theliterature, i.e. as a porous media, for the deposition an enhanced model with an additional sink term wasimplemented. CFD results are compared with UK AURN (Automatic Urban and Rural Network) stationconcentrations. Several meteorological conditions are analysed based on London City Airport weatherstation data, with attention to prevailing winds.

CFD simulations show that trees trap air pollution by up to about 7% at the Marylebone monitoringstation in the spring, autumn and summer seasons, suggesting that the aerodynamic effects are similarover the different leaf seasons. Aerodynamic effects are more important at lower wind speeds causinglittle turbulent dispersion. Deposition effects are found to be 4 times less important with reductions ofup to about 2%, with more deposition in summer due to a greater leaf area density. Furthermore, forwinds parallel to Marylebone Road, the aerodynamic effects decrease concentrations suggesting that insuch cases trees could be considered as a mitigation measures. This is different from perpendicular windsfor which trees exacerbate trapping, as found in previous studies. The analysis of concentration levels

obtained from CFD simulations across the whole street confirms a beneficial aerodynamic dispersiveeffect of trees of 0.7% in summer time for all wind directions averaged at a wind speed of 5 m/s (yearlyaverage wind speed observed in the area). Results highlight the need to account for both aerodynamicand dispersion effects of trees in CFD modelling to achieve a comprehensive evaluation and help cityplanners with a sustainable design of trees in urban environments.

ors. P

© 2017 The Auth

. Introduction

Many municipalities have shown a renewed interest in “urbanorestry” by incorporating green space and vegetation into therban environment. Urban greening usually refers to urban designlements such as trees and other plants in parks, sidewalks or else-here, employed for recreation or aesthetic improvement of a city.

n recent years, researchers have also been looking into potentialenefits of green space and vegetation, including lower energy use,educed air pollution (Gallagher et al., 2015; Gromke et al., 2016;

∗ Corresponding author.E-mail address: riccardo.buccolieri@unisalento.it (R. Buccolieri).

ttp://dx.doi.org/10.1016/j.ufug.2017.01.009618-8667/© 2017 The Authors. Published by Elsevier GmbH. This is an open access artic

ublished by Elsevier GmbH. This is an open access article under the CC BYlicense (http://creativecommons.org/licenses/by/4.0/).

Li et al., 2016), protection from harmful exposure to ultravioletrays, heat island mitigation, decreased storm water runoff, poten-tial reduced pavement maintenance (Roy et al., 2012; Maggiottoet al., 2014; Di Sabatino et al., 2015; Hsieh et al., 2016), improvedwellbeing of the urban population (White et al., 2013; Van denBerg et al., 2015) and reduced traffic noise levels (Kalansuriya et al.,2009). Although particle deposition on plant surfaces removes pol-lutants from the atmosphere, thus reducing their concentration,it also should be noted that trees themselves act as obstacles toairflow decreasing air exchange and leading to larger pollutantconcentrations.

Several experimental and modelling studies on the effects oftrees on urban air quality have been performed in the recent lit-erature (most of them have been collected in reviews by Janhall,

le under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

4 try &

2w(tow1dtswtAt2taB1naodedciwos1Jatbauutadcpti

totteaoidmhpw

smmdt

2 A.P.R. Jeanjean et al. / Urban Fores

015 and Gallagher et al., 2015). One of the pioneering experimentsas performed in the wind tunnel of the University of Karlsruhe

CODASC, 2008) where pollutant concentrations were measured athe leeward and windward of isolated symmetric street canyonsf several aspect ratios and approaching wind directions. Treesere found to increase wall-averaged concentrations up to about

00%. Results also showed that street-level concentrations cruciallyepend on the wind direction and street canyon aspect ratio ratherhan on tree crown porosity typically found in real scenarios andtand density. Several Computational Fluid Dynamics (CFD) modelsere then applied to simulate the CODASC case and parameteriza-

ions of aerodynamic (e.g. Buccolieri et al., 2011; Wania et al., 2012;morim et al., 2013; Gromke and Blocken, 2015a,b) and deposi-

ion effects of trees (e.g. Jeanjean et al., 2015, 2016; Santiago et al.,016; Selmi et al., 2016) were developed and employed in simula-ions of flow and pollutant dispersion within complex geometriesnd real scenarios. Among the most recent studies, Gromke andlocken (2015b) found low to moderate increases (up to about3%) of pollutant concentration at pedestrian level in generic urbaneighbourhood for various avenue-tree layouts (only consideringerodynamic effects), with pronounced locally restricted decreasesr increases (−87 to +1378%). Santiago et al. (2016) reportedecreased concentrations close to the ground up to 60% in sev-ral idealized arrays of different packing density depending on theeposition velocity, showing that the deposition effects are alsorucial in determining the final concentration levels. Real scenar-os were simulated by several authors such as Amorim et al. (2013)

ho investigated the aerodynamics effects of trees in selected areasf Lisbon and Aveiro (Portugal) for distinct relative wind directions,howing an average 12% increase of concentrations and an average6% decrease for oblique and parallel wind directions, respectively.eanjean et al. (2015, 2016) also found that trees are beneficial from

purely dynamic point of view, as they decreased concentration ofraffic emissions by 7% on average at pedestrian height in a neigh-ourhood in Leicester (UK). Recently, Gromke et al. (2016) showed

reduction up to 60% at pedestrian level in the presence of contin-ous hedgerows. These results show that the effects of trees andrban vegetation in general are strictly dependent on their interac-ion with geometry and meteorological conditions. Studies tend togree that aerodynamic effects of trees are more signification thaneposition (Vos et al., 2013; Jeanjean et al., 2016). Even thoughhallenges and strategies for urban green-space planning in com-act cities have been proposed (Haaland and van den Bosch, 2015),his topic needs to be further investigated before any action is takenn urban planning (Janhall, 2015).

Within this context, the objectives of the present study arewofold. The first objective is to validate a CFD dispersion modelf NOx and PM2.5 (using the CFD code OpenFOAM) to account forhe aerodynamic effects of trees in combination with the deposi-ion effects for PM2.5. This allows a comprehensive evaluation of theffects of trees on pollutant dispersion. The second objective is topply the developed methodology to investigate the effects of treesn dispersion in a real scenario, i.e. in Marylebone neighbourhoodn central London. CFD results are compared with concentrationata from monitoring stations available from the UK AURN (Auto-atic Urban and Rural Network). Several meteorological conditions

ave been chosen based on data retrieved from the London City Air-ort weather station, paying particular attention to the prevalentind speeds and directions.

The paper is structured as follows. Section 2 describes the studyite and the cases investigated. Section 3 describes the general CFDodelling and the modelling of trees, with details on the develop-

ent of the deposition module. Section 4 presents the results and

iscusses the effects of trees on affecting road emission concentra-ions. Conclusions are given in Section 5.

Urban Greening 22 (2017) 41–53

2. The study site

2.1. Description of geometry and trees

Marylebone is an affluent inner-city area of central London (UK),located within the City of Westminster. It is characterised by majorstreets on a grid pattern such as Marylebone Road, one of the busi-est roads of central London, with smaller mews between the majorstreets. The area is characterized by a geometry typical of the archi-tecture of many European cities with several street canyons (DiSabatino et al., 2010). Marylebone Rd is characterised by a streetcanyon configuration with an aspect ratio (height over width) nearunity (Nikolova et al., 2016). It usually experiences high pollutionepisodes due to the passage of more than 80,000 vehicles per day onMarylebone Rd and regular traffic congestion (Crosby et al., 2014).This makes it one of the most polluted sites in the UK, with an aver-age NO2 concentration of 94 �g m−3 in 2014, according to the AURNmeasurements. Well above the European recommended thresh-old of 200 �g m−3, pollutant concentration thresholds are regularlyexceeded up to 35 times a year (Charron et al., 2007).

Roads, buildings and trees data were integrated to reconstruct a3-dimensional (3D) area around the study area. Roads and buildingsdata were taken by Ordnance Survey which is the UK governmen-tal mapping agency (OS, 2016). The National Tree MapTM (NTM)Crown Polygon produced by Bluesky International Ltd was used torepresent individual trees or closely grouped tree crowns (Bluesky,2016). Trees and bushes over 3 m in height were included in thedatabase. An overview of the study area can be seen in Fig. 1. TheNTMTM product provides a canopy top height but does not how-ever provide a canopy base height. Therefore, a canopy base heightof 1/3 of the canopy depth was assumed, as is commonly done incurrent literature (e.g. Gromke et al., 2008; Gromke and Blocken2015b).

2.2. Description of the cases investigated

Several cases have been simulated with the CFD code Open-FOAM (Table 1). Wind data for the year 2014 were retrieved fromthe London City Airport weather station (EGLC, available at https://www.wunderground.com), every 30 min with a wind directionaccuracy of 10◦. The station is located around 15 km west of themonitoring site. In 2014, the recorded average wind speed was4.3 m/s and the prevalent wind direction was South-West (Fig. 2).Specifically, 4 wind speeds and 15 wind directions were selected,i.e. every 30◦ in the range 270◦–180◦, and every 15◦ in the range180◦–270◦, the latter being the prevailing wind direction rangefound in the study area.

Leaf-free trees (winter, referred to as CB), trees with half-grownleaves (spring/autumn, referred to as CT1) and trees with fullygrown leaves (summer, referred to as CT2) were investigated foreach wind speed and direction. Scenarios CT1 and CT2 have beenmodelled with different porosities (see Subsection 3.3 and Table 3for further details). Overall, 4 wind speeds, 15 wind directions, 3 dif-ferent tree profiles and 2 pollutant species were simulated, givinga total of 360 individual simulations.

The year 2014 has been chosen as a reference year in this studyfor pollutant concentrations as it provides a recent annual baselineto investigate the interaction between trees and the atmosphere.Although excluded here, the investigation of this relationship overtime leaves room for further research.

2.3. Description of traffic data and pollutant concentration

analysis

Estimated Annual Average Daily Flows (AADF) from the Depart-ment for Transport (DfT, 2016) were used to estimate road

A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53 43

Fig. 1. Area of interest around the Marylebone monitoring site in London, UK. M stands for Marylebone, GP for Gloucester Place and BS for Baker Street. (a) GoogleEarthoverview (b) 3D model of the scene using roads and buildings from Ordnance Survey UK and tree data from Bluesky International Ltd.

Table 1Scenarios investigated with different types of trees, seasons and meteorological data simulated with the CFD code OpenFOAM.

Name Trees Season Wind speed (ms−1) Wind direction (◦)

CB (Case of Buildings only) Leaf-free winter 3579

0306090120150180195210225240255270300330

CT1 (Case of Trees 1) Half-grown leaves spring & autumnCT2 (Case of Trees 2) Fully grown leaves summer

44 A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53

Table 2Calculated NOx and PM2.5 emissions from Annual Average Daily Flows (AADF).

Marylebone (M) A41 − Gloucester Place (GP) A41 − Baker Street (BS)

AADF M1 =M2 =M3 =

788807882779528

GP1 =GP2 =

1353015627

BS1 =BS2 =

1381310583

NOx emission (mg/m-s) M1 =M2 =M3 =Average =

0.690.670.690.68

GP1 =GP2 = Average =

0.100.120.11

BS1 =BS2 = Average =

0.110.080.10

PM2.5 emission (mg/m-s) M1 =M2 =M3 =Average =

0.0310.0300.0310.031

GP1 =GP2 = Average =

0.0050.0060.006

BS1 =BS2 = Average =

0.0050.0040.005

Fig. 2. Wind rose plot showing the wind directions (◦) and wind speeds during theyear 2014 in London (data: London City Airport weather station).

Table 3Pressure loss coefficients of trees (�) of the modelled area across the seasons.

Season Spring & Autumn Summer Winter

ealf2fl

Ayotuw3T

RemlCd

upwind scheme.

Pressure loss coefficient of trees � (m−1) 0.26 0.4 0LAD (Leaf Area Density m2 m−3) 1.06 1.6 0

missions of nitrogen oxides (NOx) and particulate matter (PM2.5)round the monitoring site. These typical daily flows were trans-ated into road emissions using the Emissions Factors Toolkit (EFT)rom the Department for Environment, Food & Rural Affairs (DEFRA,016). Emissions were produced for the average London vehicleeet profile and are reported in Table 2.

To calculate pollutant concentrations in Marylebone Rd from theURN station (identification MY1), hourly measurements from theear 2014 were collected and distributed into classes dependingn wind speed, wind direction and seasonality. In order to matchhis hourly measure of pollution, only the hourly wind data wassed. Specifically, the averages reported for each wind directionere repeated for each wind speed (3; 5; 7 and 9 m/s) across the

modelled seasons (winter, spring & autumn and summer) (seeable 1).

To compare model outputs with monitored data for Maryleboned, an urban background concentration was added to the mod-lled data for each case investigated. The closest urban backgroundonitoring station in central London away from Marylebone Rd is

ocated in Russell’s Square (DEFRA monitoring site identification:LL2). Similar to what was done for AURN station concentrationata, hourly measurements from the year 2014 were distributed

Fig. 3. Example of the wind rose plot showing the method used to average hourlyNOx data over wind directions, here corresponding to the urban background pollu-tion measured in Russell Square for the winter season at a wind speed of 5 m/s.

into classes depending on wind speeds and wind directions asshown on Fig. 3.

The final modelled concentration for each case was thenobtained as follows:

Model (WS, WD, season) = (Background (WS, WD, season)

+road contribution (WS, WD)) (1)

where road contribution is the concentration obtained from modelsimulations and remains the same across the seasons, whereasbackground and model data were variable across the seasons. It isworth noting that the annual average daily traffic flow assumptionused in this study doesn’t take into account all the temporal vari-ations across the change of traffic during day and night, weekdaysand weekend and so on, as the spread of traffic is averaged for atypical day.

3. CFD modelling

3.1. Flow modelling set-up

Simulations were performed using the OpenFOAM (Open FieldOperation and Manipulation) open source software platform (freelyavailable at http://www.openfoam.com). Wind flow calculationswere performed under the steady-state simpleFOAM solver forincompressible, isothermal and turbulent flows. This steady-statesolver is based on the Reynolds-Averaged Navier–Stokes (RANS)with the standard k–ε closure model (Launder and Spalding,1974). The governing equations were discretized with second order

The steady RANS approach with the standard k–ε model include,among other, the underestimation of the size of separation andrecirculation regions on the roof and the side faces of a building,

A.P.R. Jeanjean et al. / Urban Forestry &

Fa

awsattpla2taastBetFtStsps(tm2fyhttcu

tidfd

ig. 4. Modelled area of interest inside the CFD OpenFOAM software. Coordinatesre in British National Grid (UK coordinate system expressed in metres).

s well as the underestimation of turbulence kinetic energy in theake. To overtake such limitations, several RANS turbulence clo-

ures have been proposed in the literature for the study of flownd pollutant dispersion in idealized and real scenarios. However,here are not still guidelines on which RANS model performs bet-er, thus a validation study is always compulsory to evaluate theerformance case by case. On the other hand, Large Eddy Simu-

ations (LES) perform better in predicting turbulence than RANSpproaches (see recent reviews by Di Sabatino et al., 2013; Blocken,015; Lateb et al., 2016). There are however still challenges toheir applications. These include difficulty in specifying appropri-te time-dependent inlet and wall boundary conditions, as wells longer computational times. In spite of their limitations, RANSimulations have shown their reliability in reproducing the spa-ial distribution of mean velocity and concentration fields (e.g.uccolieri et al., 2015; Hang et al., 2015; Lateb et al., 2016; Santiagot al., 2016). In the present paper, the starting point is to employhe same CFD methodology (using the standard k-ε model in Open-OAM) used in previous studies for similar geometries to evaluatehe influence of trees on final concentration levels of pollutants.pecifically the use of OpenFOAM-RANS k-ε to assess air quality inhe study area is supported by the validation performed in a recenttudies (Jeanjean et al., 2015; Vranckx et al., 2015), where flow andollutant dispersion results for an idealized street canyon wereuccessfully validated against the CODASC wind tunnel databaseCODASC, 2008). A grid sensitivity analysis was performed againsthe wind tunnel data to ensure that the grid resolution used by the

odel was fine enough to provide stable results (Jeanjean et al.,015). The focus here is on the development of a deposition moduleor the parameterization of the deposition effects which have notet been fully evaluated in the literature. Present CFD simulationsave been further validated against monitored data (see Subsec-ion 4.1), showing that this study can constitute a starting point forhe improvement and the development of a modelling tool for theomprehensive assessment of tree-atmosphere interaction in therban environments.

Best practice guidelines were followed to build the computa-ional domain (Franke et al., 2007). The maximum reported heightn the domain is a building height (H) of 63 m. The computational

omain was built with its boundaries placed more than 15H awayrom the modelled area (Fig. 4). The top of the computationalomain was set to 570 m, which corresponds to 8H. A maximum

Urban Greening 22 (2017) 41–53 45

expansion ratio between two consecutive cells was kept below 1.3.With an average building height of 12 m across the modelled area,the blocking ratio was kept below 1% inclination and is thereforebelow the 3% recommended threshold.

A hexahedral mesh of more than 4 million cells was used. A highmesh resolution of 0.5 m in the vertical direction close to the bottomof the computational domain was chosen (<1 m in order to ensureproper flow modelling at pedestrian height as advised by Blocken(2015)). A cell size of 1.25 m along the X and Y axis was applied forthe buildings, trees and roads. This resolution allows more than 10cells to be present across the main street canyon to ensure properflow modelling (Fig. 5).

The boundary conditions were chosen to reflect an atmosphericboundary layer. Single inlet and outlet conditions were used forNorthern, Eastern, Southern and Western winds using the 4 sidesof the outer domain as shown in Fig. 4. For the other wind direc-tions, 2 sides of the domain were defined as inlets and two asoutlets, to model the change in wind direction. The mean velocityboundary flow and the turbulent dissipation were set up to followa logarithmic law using the “atmBoundaryLayerInletVelocity” and“atmBoundaryLayerInletEpsilon” utilities in OpenFOAM such that:

U = u∗�

ln(

z + z0

z0

)(2)

ε = u3∗� · z

(1 − z

ı

)(3)

where U is the fluid velocity (m/s), u∗ is the frictional velocity (m/s),� is the Karman’s constant, z is the vertical co-ordinate (m), ı is theboundary layer depth (m) and z0 is the surface roughness (m).

As recommended by the Cost Action 732 best practice guidelines(Franke et al., 2007), the top of the domain was set as a symmetryplane. The wind atmospheric boundary layer was set to reach thewind speed at a height of 10 m to match with the wind measure-ments height. A surface roughness of z0 = 0.10 m was set for theground, which corresponds to sparse, large obstacles (WMO, 2008).This surface roughness is set to account for the presence of parkedcar, bus stop and so on across the modelled scene.

For the wind flow calculation, a residual convergence of 10−5

was used for all field variables and of 10−4 for the pressure. Forthe scalar transport simulation (pollutant dispersion), a residualconvergence of 10−6 was reached. The simulation time per uniquewind condition was 3 days on average on a single core running witha RAM of 8 GB.

3.2. Gaseous pollutant modelling set-up

The transport equation “scalarTransportFoam” of OpenFoamwas used and modified to take into account the turbulent diffusivityas follows:

∂C

∂t+ � (U · C) = �2 ((D + K) · C) (4)

where C is the transported scalar, U is the wind velocity, D is thediffusion coefficient (m2/s) and K is the eddy diffusion coefficient(m2/s). The eddy diffusion coefficient can be expressed as: K = �t /Sct

where �t is the eddy viscosity or turbulent viscosity (m2/s) andSct is the turbulent Schmidt number. A Sct value of 0.5 was used,which gave the best model agreement when compared to windtunnel data (Jeanjean et al., 2015). The zones for the road cells wereselected up to 1.5 m height for pollution emission.

3.3. Modelling the effects of trees

Both aerodynamic and deposition effects of trees have beenmodelled. In the case of aerodynamic effects, trees were treatedas a porous media by adding a pressure loss coefficient to the cells

46 A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53

F to carry out the CFD simulations. A maximum resolution of 1.25 m was used across the Xa

ovSb

�

wLs

1oi“TbFb2C1mctgmscu

wP

�

w(L

Table 4Literature values of deposition velocities of PM2.5 on vegetation.

Pollutant species Deposition velocityon vegetation(cm s−1)

(Author, year)

PM2.5 0.02 (lower end)0.64 (intermediate)30 (higher end)

Peters and Eiden(1992)Pugh et al. (2012)White and Turner

ig. 5. (a) View from the top and (b) View from the street canyon of the mesh usednd Y axis and 50 cm along the Z axis.

ccupied by the tree canopy. This is the same approach used by pre-ious studies (Jeanjean et al., 2015, 2016; Vranckx et al., 2015; Diabatino et al., 2015). The pressure loss coefficient � (m−1) inducedy trees is expressed as:

= Cd · LAD (5)

here Cd is the drag coefficient induced by trees and LAD is theeaf Area Index (m2 m−3). With the assumption of a homogeneouspread of tree species across South East England and London,

it can be estimated that London has 80.3% deciduous trees and9.7% coniferous trees (Forestry-Commission, 2013). Only decidu-us trees were considered in this study as they are predominantn London. The Platanus acerifolia (Aiton) Willd, commonly calledLondon plane”, is the species mainly present in Marylebone Rd.he average LAD through the canopy can be approximated toe between 1.06 and 2.18 m2 m−3 (Lalic and Mihailovic, 2004).rom previous studies, the drag coefficient can be estimated to beetween 0.1 ≤ Cd ≤ 0.3 for most types of vegetation (Katul et al.,004). For the summer season, an average drag coefficient ofd = 0.25 for the tree canopy was used with an average LAD of.6 m2 m−3, which is consistent with that estimated via ceptometereasurements by Di Sabatino et al. (2015). The final pressure loss

oefficient � was equal to 0.4 m−1. With a LAD of 0 m2 m−3 in win-er, the drag of the trees was neglected for this season. With therowth and fall of leaves in spring and autumn respectively, theodel used the lower end of the LAD value of 1.06 m2 m−3 with a

imilar drag coefficient of Cd = 0.25 leading to a pressure loss coeffi-ient of 0.26 m−1. Table 3 summarises the pressure loss coefficientssed across all 4 seasons.

As for tree deposition, the OpenFOAM CFD model was enhancedith an additional sink term taking into account the deposition of

M2.5 on trees as follows:

C = C0 · LAD · Vd (6)

here �C is the change in particle concentration via depositiong m−3), C0 is the initial particle concentration (g m−3), LAD is theeaf Area Density previously estimated at 1.6 m2 m−3 for London

(1970)

and Vd is the deposition velocity (m/s) as summarised in Table 4.As shown in the table, the range of dry deposition velocities in theliterature is very wide. Deposition velocities are highly dependenton the vegetation species as well as the various particle diameters(size distribution) constituting PM2.5. Although an average deposi-tion velocity can be challenging to estimate, an intermediate valueof 0.64 cm/s for PM2.5 was used for modelling the deposition of thispollutant on trees as used by Pugh et al. (2012).

4. Results and discussion

In this section, modelled results are first compared with mon-itored data to validate CFD simulations (Subsection 4.1). A similarmethodology has been applied by Jeanjean et al. (2015) and val-idated against the CODASC database (2008). Once validated, dataobtained from both simulations and measures are used to evaluatethe effects of trees on concentration levels of NOx and PM2.5 at themonitoring site in the study area. Particular attention is paid to therelative contribution of aerodynamic and deposition effects for allthe wind directions and speeds summarised in Table 1 (Subsections4.2 and 4.3). Finally CFD simulations are used to provide a compre-hensive evaluation of concentration levels and of the effects of trees

over the whole study area, which is not possible for data monitoredat a single point (Subsections 4.4).

A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53 47

1 and

4M

4

Mwagtfwirsi(

Mams

ta

Fig. 6. Wind rose plots comparing modelled NOx and PM2.5 for the cases CB, CT

.1. Comparison between simulations and data monitored in thearylebone site

.1.1. Statistical analysis of model performanceWind roses comparing monitored and modelled data for the

arylebone site are shown in Fig. 6 for a wind speed of 5 m/s,hich is close to the yearly average wind speed observed in the

rea (4.3 m/s, see Subsection 2.2). The figure shows that the shapeenerated by measured pollutant concentrations is consistent withhat reproduced by the model. Quantitatively, it can be noted thator the wind direction equal to 225◦ in the absence of trees (CB inintertime), the modelled data is out of range. This is not observed

n the presence of trees (CT1 and CT2). A reason for this shift in theesults could be the fact that trees are not modelled for the wintereason. In reality however, tree trunks and branches are still presentn winter suggesting that a model parameterisation for winter treeswithout leaves) is needed for the CFD model.

An underestimation of NOx concentrations for winds parallel toarylebone Rd street canyon, i.e. Easterly and Westerly winds, is in

greement with findings from previous studies suggesting that CFDodels usually underestimate concentrations for winds parallel to

treet canyons with trees (Gromke and Blocken 2015a).As for PM2.5, the model overestimated concentrations by more

han 20% (see Table 5), which corresponds to the average modelccuracy (Jeanjean et al., 2015). A reason for this overestimation

CT2 against monitored data by Marylebone AURN site at a wind speed of 5 m/s.

could be the assignment of too large values for either the road emis-sions or the calculated urban background. It is worth noting thatthe urban background has a greater influence on PM2.5 concentra-tions of about 67%, with Marylebone Rd emissions contributing to33% of total PM2.5 (Charron and Harrison, 2005). Another possibilitycould be the fact that wet deposition was not modelled, rain actingas a skin of PM2.5 (Nowak et al., 2013) which would decrease themodelled concentrations.

To assess the model performance, several standard metrics havebeen calculated; namely the normalized mean square error (NMSE),the fraction of predictions within a factor of two for observations(FAC2) and the fractional bias (FB). According to COST Action 732 (DiSabatino et al., 2011), the recommended criteria are: NMSE ≤ 1.5;FAC2 ≥ 0.5; −0.3 ≤ FB ≤ 0.3. Results of the statistical analysis arepresented in Fig. 7. It can be noted that NMSE and FAC2 are withinthe acceptable range for most cases. FB is out of range for a few mod-elled cases, especially for high wind speeds for NOx and low windspeeds for PM2.5. The FB indicator gives information about over-estimation (negative values) or underestimation (positive values).As pointed out previously, this indicator shows that NOx values areunderestimated by the model and PM2.5 values are overestimated.

It is worth emphasising that the buildings database employed inthis study has an accuracy of 20 cm–50 cm, but does not includeroof shapes. Most of the features of buildings were thus accuratelyreconstructed, with the exception of roof shape which could intro-

48 A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53

Table 5Seasonal average bias for modelled concentrations of NOx and PM2.5 compared to Marylebone monitoring station measurements.

NOx PM2.5

�g/m3 % �g/m3 %

Winter: −75.7 −24.6 3.8 24.5Spring andAutumn:With treedeposition:

−68.3N/A

−21.5N/A

4.74.4

27.520.3

Summer:With treedeposition:

−43.6N/A

−16.1N/A

5.65.3

37.025.4

ns wh

d(ii

ti

4

octqtu

bit

Fig. 7. Statistical analysis of modelled pollutant concentratio

uce significant changes in the wind flow in street canyon studiesYazid et al., 2014). The analysis of statistical data suggests that FBs highly dependent on wind speed; higher wind speeds employedn the CFD simulations tend to exacerbate this issue.

Figs. 6 and 7 show a satisfactory model performance in terms ofhe essential features of the mean concentrations and support thenvestigation of the effects of trees on air quality in the study area.

.1.2. Model seasonal uncertaintyThe seasonal model uncertainty for simulated concentrations

f NOx and PM2.5 is reported in Table 5. It can be seen that NOx

oncentrations are at least underestimated by 20% when comparedo measurements. This underestimation is a common feature of airuality dispersion model in the UK (Williams et al., 2011), one ofhe main reason being caused by diesel vehicles emissions beingnderestimated (e.g. Oldenkamp et al., 2016).

PM2.5 concentrations are, however, overestimated by the modely at least 20%. This could be due to the choice of deposition veloc-

ty, which could further enhance a sink in the leaf seasons andherefore reduce modelled PM2.5 concentrations. It can be seen

en compared to the monitored data for the Marylebone site.

that even the winter case (no leaves), modelled concentrations areoverestimated which suggests that other mechanisms are causingthe model overestimation. Another reason for this overestimationcould be that wet deposition has not been modelled in this studywith rain being a sink of particulate matters. Rainy weather islikely to impact modelled concentrations of PM2.5, even though thiscould be counterbalanced by the resuspension which would leadto higher concentrations.

4.2. The influence of trees under different wind directions

To investigate the aerodynamic and deposition effects of treesunder different wind directions, the following section focusespurely on traffic emissions of PM2.5. As discussed in the previoussubsection, the analysis of wind roses and the statistical analysisshowed that the wind direction of 225◦ was not reproduced cor-

rectly by the model, therefore this wind direction was removedfrom the analysis below.

Table 6 shows the influence of wind direction on PM2.5 concen-trations emitted by traffic for a wind speed (WS) of 5 m/s, without

A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53 49

Table 6Influence of wind direction on PM2.5 concentrations emitted by traffic at the Marylebone monitoring station, for a wind speed of 5 m/s, without urban background concentra-tions. Total percentages (%) are calculated in comparison with seasonal average concentrations of PM2.5, corresponding to the wind speed of 5 m/s. The 225◦ wind directionwas not included in the calculation as the results were not satisfactory when compared to measurements. Highlighted numbers correspond to wind directions which resultin a beneficial reduction in air pollution because of the aerodynamic dispersion of trees.

Aerodynamic dispe rsion of trees (μg/m3) Tree deposi�on (μg/m3)

WD(°) WD probabili ty (%)

Spring and Autumn Summer

Spring and Autumn Summer

0 11.5 0.8 0.3 -0.2 -0.230 6.2 1.6 1.7 -0.4 -0.560 6.0 -1.0 -0.8 -0.2 -0.390 10.6 1.8 2.1 0.0 0.0120 5.1 0.3 0.8 0.0 0.0150 5.4 6.3 0.4 -0.3 0.0180 4.4 7.6 11.7 -0.5 -0.7195 6.9 2.6 3.8 -0.4 -0.7210 6.1 6.6 4.8 -0.3 -0.4240 7.0 -3.1 -2.9 -0.3 -0.4255 13.7 -3.2 -2.7 -0.4 -0.5270 5.5 1.6 2.1 -0.2 -0.3300 6.6 2.4 2.5 -0.3 -0.4330 5.0 0.1 0.1 0.0 0.0

Total (μg/m3): 1.2 1.2 -0.2 -0.3

uao

T

tbetdtamm

2acas2

4

diAt3mchsc

f

Total (%): 7.0 7.5 -1.4

rban background concentrations. In the table total percentages (%)re calculated in comparison with seasonal average concentrationsf PM2.5 as follows:

otal (%) = 1 − [Mes.seasonal average] (WS) − [Mod.seasonal average] (WS)[Mes.seasonal average] (WS)

(7)

The table shows that trees are found to trap air pollution athe Marylebone monitoring site by 7.0% in spring and autumn andy 7.5% in summer. These increases in PM2.5 concentrations arequivalent to an increase of about 1.2 �g/m3, which shows thathe aerodynamic effects are fairly similar over the leaf seasons. Theeposition effects are found to be around 4 times less importanthan the aerodynamic effects, with reductions of 1.4% in spring andutumn and 2.1% in summer. More deposition is found over sum-er than in spring and autumn, as a greater leaf area density offeredore surfaces for deposition (Steffens et al., 2012).It should be noted that for wind directions of 60◦, 240◦ and

55◦, which are parallel to Marylebone Rd street canyon, theerodynamic effects of trees decrease street concentrations. Thisonclusion supports previous results in real scenarios which report

12% increase in concentration for winds perpendicular to thetreet canyon and a 16% decrease for parallel winds (Amorim et al.,013).

.3. The influence of trees under different wind speeds

Table 7 shows the effects of trees on PM2.5 concentrations forifferent wind speeds. Results show that the effectiveness of trees

n altering street concentrations is greater at lower wind speeds.t a wind speed of 3 m/s, trees increase road emission concentra-

ions by 16.7% in summer, whilst providing a beneficial decrease of.4% via deposition. The dispersive effect of trees is subsequentlyuch lower at a wind speed of 9 m/s, with road emission con-

entrations increasing by 3.7%. Dispersion and deposition do notowever decrease at the same rate across the different wind speeds,

uggesting that the effects of trees are not linear with wind speedhanges.

The ability of trees to increase turbulent dispersion accountsor their beneficial aerodynamic effects (Jeanjean et al., 2015).

-2.1

However, little turbulent dispersion occurs at lower wind speedsbecause trees act as obstacles to wind flow, explaining the greaterwind-trapping ability of trees at lower wind speeds. This findingis in agreement with a previous experimental study (Di Sabatinoet al., 2015). An explanation for greater deposition effects occur-ring at lower wind speeds could be due to the fact that more timeis left for the suspended particles to deposit on leaves. In addi-tion, less dispersion occurs at lower wind speeds, which increasesconcentrations of pollution therefore leading to greater depositionflux.

4.4. Discussion

4.4.1. Effects of trees at the Marylebone measurement stationThe analysis of CFD results and data monitored at the Maryle-

bone measurement station shows that trees could be considered asa mitigation measure only in the case of streets that are parallel tothe prevailing winds. However, trees also exacerbate trapping forwind directions perpendicular to the street canyon orientation andtree planting would not improve air pollution in this situation. Thissupports previous findings which have shown an increase in streetcanyon pollution concentrations for perpendicular winds (see forexample Gromke et al., 2008; Buccolieri et al., 2011; Wania et al.,2012) and demonstrates that local meteorology should be takeninto account for tree planting policies.

In the CFD model employed in this study, no effect of trees hasbeen taken into account when the LAD was supposed to be zero (CBcase in winter), i.e. branches and trunks have been neglected. Hereresults suggest that the remains of trunks and branches or possiblythe presence of pine trees which are evergreen seem to play a rolein winter. In fact, for a particular wind direction (225◦) a significantdiscrepancy between modelled and monitored date was found forthe CB case (buildings only) and not for the other cases with trees(CT1 and CT2).

4.4.2. Effects of trees on pedestrians in Marylebone RdTo investigate the effects of trees on pedestrians, CFD concentra-

tions were sampled across the whole of Marylebone Rd on a regular

50 A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53

Table 7Influence of wind speed on aerodynamic dispersion and deposition of trees on PM2.5 concentrations emitted by traffic at the Marylebone monitoring site.

Aerodynamic dispersion of trees Tree deposition

WS (m/s) Spring and Autumn Summer Spring and Autumn Summer

3 (�g/m3) 2.2 2.8 −0.4 −0.6(%) 10.8 16.7 −2.1 −3.4

5 (�g/m3) 1.2 1.2 −0.2 −0.3(%) 7.0 7.5 −1.4 −2.1

7 (�g/m3) 0.3 0.2 −0.1 −0.2(%) 1.8 1.8 −0.9 −1.5

9 (�g/m3) 0.2 0.4 −0.1 −0.1(%) 1.7 3.7 −0.9 −1.3

F g winw s of PM

2ifv

tbislTttimif

ig. 8. Street effects of trees in Marylebone Rd at pedestrian height over prevailinith trees in summer (CT2). (c) Aerodynamic effects of trees in summer (%). (d) Los

m grid at a pedestrian height of 1.5 m (see concentrations of PM2.5n Fig. 8). The aerodynamic effects of trees are found to be similaror both PM2.5 and NOx, the only difference being in terms of thealue of the emissions emitted by them.

Street concentrations of PM2.5 can be seen in Fig. 8a withouthe effect of trees and in Fig. 8b with trees, the concentrationseing averaged for Fig. 8a,b across prevailing winds (see prevail-

ng winds probability in Table 6). The introduction of trees in thetreet slightly changes the distribution of concentrations, particu-arly of hotspots, which subsequently appear to be more spread out.he aerodynamic effects of trees (Fig. 8c) is quite heterogeneous inhat trees can either decrease or increase concentrations at pedes-rian height. The pollution-trapping ability of trees is particularly

mportant around the monitoring site. A reason for this increase

ight be the high number of trees towards the East of the mon-toring site which act as a barrier to the prevailing winds comingrom the South West. This has the effect of reducing wind veloci-

d directions. (a) PM2.5 concentrations without trees (CB). (b) PM2.5 concentrations2.5 via deposition on trees in summer (%).

ties and therefore increasing pollution concentrations. Changes inPM2.5 via deposition on trees (Fig. 8d) are less important than theaerodynamic effects in terms of magnitude.

When considering the effects of trees across the whole street,the aerodynamic dispersion of trees is actually found to decreasestreet pollution concentrations by 1.1% in spring and autumnand 0.7% in summer (Table 8). In addition, the loss of PM2.5 viadeposition on trees is greater than the aerodynamic effects whenconsidering the whole street, with 2.9% loss in spring and autumnand 4.6% in summer. In most of the wind directions, trees arefound to increase concentrations. In the special case of Maryle-bone Rd, where prevailing winds are parallel to the street canyon,more weight is given to favorable wind directions, as illustrated

by Table 8. The resultant weighted effect of trees also happens tobe beneficial for pedestrians. This suggests that results found fora single measurement point location (Table 7) are not necessar-

A.P.R. Jeanjean et al. / Urban Forestry & Urban Greening 22 (2017) 41–53 51

Table 8Effect of trees on average street concentrations of PM2.5 emitted by traffic, for a wind speed of 5 m/s, without urban background concentrations. Total percentages (%) arecalculated in comparison with seasonal average concentrations of PM2.5 corresponding to the wind speed of 5 m/s. The 225◦ wind direction was not included in the calculationas the results were not satisfactory when compared to measurements. Highlighted numbers correspond to wind directions for which trees aerodynamic dispersion is beneficial.

Aerodynamic dispersion of tree s (μg/m3) Tree dep osi�on (μg/m3)

WD(°) WD probabili ty (%)

Spring and Autumn Summer

Spring and Autumn Summer

0 11.5 3.5 3.9 -0.4 -0.530 6.2 1.9 2.1 -0.4 -0.660 6.0 -6.0 -6.7 -0.7 -1.090 10.6 -3.9 -4.8 -0.6 -0.8120 5.1 2.0 1.3 -0.3 -0.4150 5.4 -3.5 3.2 -0.2 -0.3180 4.4 3.4 3.2 -0.2 -0.3195 6.9 3.3 3.0 -0.2 -0.3210 6.1 3.1 2.8 -0.3 -0.4240 7.0 -1.8 -2.7 -0.8 -1.1255 13.7 -4.2 -4.8 -0.9 -1.2270 5.5 -2.8 -3.3 -0.6 -0.8300 6.6 4.2 4.3 -0.6 -0.9330 5.0 3.6 4.3 -0.4 -0.6

ie

4

taHepisewwosiicpiJpls0ti

pwisscGtt

Total (μg/m3): -0.2 -0.1 -0.5Total (%): -1.1 -0.7 -2.9

ly representative of the overall effects of trees and can lead torroneous conclusions.

.4.3. Links with previous studiesIn this paper it is found that in the case of perpendicular winds,

rees act as obstacle to airflow leading to increased concentrationss found in many previous studies mentioned in the Introduction.owever, for parallel wind directions to a street canyon, the pres-nce of trees leads to a reduction in pollutant concentrations. Somerevious studies reported the opposite, with trees leading to an

ncrease in pollutant concentrations for parallel wind directions,uch as in the CODASC wind tunnel experiment realised by Gromket al. (2008) in an idealized street canyon. When compared to thisind tunnel experiment, the CFD model used in this study agreesith this increase (Jeanjean et al., 2015), suggesting the suitability

f the employed model for evaluating flow and pollutant disper-ion in idealized cases. In real street canyon scenarios, such as thosenvestigated here, the extra turbulent mixing due to the surround-ng obstacles and other non-uniformities is taken into account andould greater influence the final concentration levels. This is sup-orted by the comparison with monitored data performed here and

n previous studies in complex geometries (Amorim et al., 2013;eanjean et al., 2015, 2016; Santiago et al., 2016). Further, the highorosity used for trees in the CODASC experiment, with a pressure

oss coefficient of 250 m−1 at 1:150 model scale (1.67 m−1 in fullcale, see Gromke et al., 2016) which compared to 0.26 m−1 and.4 m−1 used here could be a reason for an increase in concentra-ions. This may suggest that greater porosities would lead to anncreased trapping effect of trees.

Recently Gromke and Blocken (2015b) reported an increase inollutant concentrations for parallel winds. Trees were modelledith a pressure loss coefficient of 0.2 m−1, similar to the ones used

n this study, however their modelled scenarios were for deepertreet canyons with a height over width aspect ratio 1.7. This coulduggest that the trapping effect of trees is increased in deeper street

anyons. Idealised buildings with square blocks were also used inromke and Blocken (2015b) work, which further confirms that

he buildings geometry or density could possibly impact as wellhe effects of trees in the case of parallel wind directions.

-0.7-4.6

These discussions suggest that the mechanism for aerodynamicconcentration reduction by including trees for parallel wind direc-tions leaves room for further investigation.

5. Conclusions

Dispersion of traffic emissions (NOx and PM2.5) in a real neigh-bourhood were investigated by means of monitored data and CFDsimulations. Analyses were performed to clarify the relative con-tribution of both aerodynamic and deposition effects of trees onpollutant concentrations at pedestrian level for several wind speedsand directions. Several conclusions were achieved from this study:

- the aerodynamic effects of trees prevailed over the depositioneffects for all the cases investigated at the Marylebone monitoringsite. As a consequence, the worst effects of trees with respect toair quality was found for lower wind speeds, since the turbulentmixing was inhibited;

- this study confirms previous findings that the street air quality isaltered by trees, with increases of 7% for typical meteorologicalconditions at the monitoring site, and an additional benefit of 2%reduction of PM2.5 via deposition;

- while perpendicular winds lead to larger pollution concentrationsin street canyons in the presence of trees, the effects of trees underparallel winds have been shown to be beneficial for air quality;

- when considering the average effects of trees on pedestrians, abeneficial reduction of 0.7% due to aerodynamic effects was foundin summer, with an additional 4.6% reduction via deposition loss.This shows the importance of the local meteorology, as both winddirection and wind speed distribution have a critical impact onthe effects of trees within the whole street canyon.

It should be noted that, as discussed in the Results section,some assumptions have been made in this study and could haveaffected modelled results. We in particular recall here that yearlytraffic emissions were averaged to get a typical day and thus they

were same across the seasons; background concentrations wereprovided by the closest station located in a square and thus theymay not represent realistic values. Further, deposition velocity wasconsidered the same across the wind speed investigated and par-

5 try &

tmbl

gtsom

A

FNNiff

R

A

B

B

B

B

C

C

C

C

D

D

D

D

D

D

F

F

G

2 A.P.R. Jeanjean et al. / Urban Fores

iculate resuspension was neglected which could lead to higherodelled values, even though this may have been counterbalanced

y the neglecting of wet deposition which is likely to decrease PM2.5evels.

The above findings leave room to further studies and sug-est that there is a crucial need for research to provide effectiveree-planting policy advice for urban planners. This could lead toubstantial air quality improvements depending on the interactionf trees with local meteorological conditions and building arrange-ents.

cknowledgements

This research used the ALICE High Performance Computingacility at the University of Leicester and was supported by theERC knowledge exchange fellowship for R. Leigh under grantE/L002930/1. We would like to thank OpenFOAM Ltd for provid-

ng the platform software used in this study and Ordnance Surveyor provide the building data. We express our thanks to Glose Diasor valuable discussions.

eferences

morim, J.H., Rodrigues, V., Tavares, R., Valente, J., Borrego, C., 2013. CFD modellingof the aerodynamic effect of trees on urban air pollution dispersion. Sci. TotalEnviron. 461, 541–551.

locken, B., 2015. Computational fluid dynamics for urban physics: importance,scales, possibilities, limitations and ten tips and tricks towards accurate andreliable simulations. Build. Environ. 91, 219–245.

luesky, 2016. UK National Tree Map (NTM), internet database http://www.bluesky-world.com/national-tree-map, (Accessed 18 November 2016).

uccolieri, R., Salim, S.M., Leo, L.S., Di Sabatino, S., Chan, A., Ielpo, P., Gromke, C.,2011. Analysis of local scale tree–atmosphere interaction on pollutantconcentration in idealized street canyons and application to a real urbanjunction. Atmos. Environ. 45, 1702–1713.

uccolieri, R., Salizzoni, P., Soulhac, L., Garbero, V., Di Sabatino, S., 2015. Thebreathability of compact cities. Urban Clim. 13, 73–93.

ODASC, 2008. Concentration Data of Street Canyon, internet database http://www.windforschung.de/CODASC.htm, (Accessed 18 November 2016).

harron, A., Harrison, R.M., 2005. Fine (PM2. 5) and coarse (PM2. 5–10) particulatematter on a heavily trafficked London highway: sources and processes.Environ. Sci. Technol. 39, 7768–7776.

harron, A., Harrison, R.M., Quincey, P., 2007. What are the sources and conditionsresponsible for exceedences of the 24 h PM 10 limit value (50 �gm-3) at aheavily trafficked London site? Atmos. Environ. 41, 1960–1975.

rosby, C.J., Fullen, M.A., Booth, C.A., Searle, D.E., 2014. A dynamic approach tourban road deposited sediment pollution monitoring (Marylebone Road,London, UK). J. Appl. Geophys. 105, 10–20.

EFRA, 2016. UK Department for Environment, Food & Rural Affairs EmissionsFactors Toolkit (EFT), internet database http://laqm.defra.gov.uk/review-and-assessment/tools/emissions-factors-toolkit.html (Accessed 11 November2016).

fT, 2016. UK Department for Transport Traffic Counts, internet database http://www.dft.gov.uk/traffic-counts/cp.php, (Accessed 11 November 2016).

i Sabatino, S., Leo, L.S., Cataldo, R., Ratti, C., Britter, R.E., 2010. Construction ofdigital elevation models for a southern European city and a comparativemorphological analysis with respect to Northern European and NorthAmerican cities. J. Appl. Meteorol. Climatol. 49, 1377–1396.

i Sabatino, S., Buccolieri, R., et al., 2011. COST 732 in practice: the MUST modelevaluation exercise. Int. J. Environ. Pollut. 44, 403–418.

i Sabatino, S., Buccolieri, R., Salizzoni, P., 2013. Recent advancements innumerical modelling of flow and dispersion in urban areas: a short review. Int.J. Environ. Pollut. 52, 172–191.

i Sabatino, S., Buccolieri, R., Pappaccogli, G., Leo, L.S., 2015. The effects of trees onmicrometeorology in a real street canyon: consequences for local air quality.Int. J. Environ. Pollut. 58, 100–111.

orestry-Commission, 2013. National Forest Inventory: Standing Timber Volumefor Coniferous Trees in Britain, http://www.forestry.gov.uk/pdf/FCNFI111.pdf/$FILE/FCNFI111.pdf, (Accessed 11 September 2016).

ranke, J., Schlünzen, H., Carissimo, B., 2007. Best practice guideline for the CFDsimulation of flows in the urban environment. COST Action 732 – qualityassurance and improvement of microscale meteorological models. Distributed

by University of Hamburg (Germany), Meteorological Institute, ISBN:3-00-018312-4.

allagher, J., Baldauf, R., Fuller, C.H., Kumar, P., Gill, L.W., McNabola, A., 2015.Passive methods for improving air quality in the built environment: a reviewof porous and solid barriers. Atmos. Environ. 120, 61–70.

Urban Greening 22 (2017) 41–53

Gromke, C., Blocken, B., 2015a. Influence of avenue-trees on air quality at the urbanneighborhood scale. Part I: quality assurance studies and turbulent Schmidtnumber analysis for RANS CFD simulations. Environ. Pollut. 196, 214–223.

Gromke, C., Blocken, B., 2015b. Influence of avenue-trees on air quality at theurban neighborhood scale. Part II: Traffic pollutant concentrations atpedestrian level. Environ. Pollut. 196, 176–184.

Gromke, C., Buccolieri, R., Di Sabatino, S., Ruck, B., 2008. Dispersion study in astreet canyon with tree planting by means of wind tunnel and numericalinvestigations–evaluation of CFD data with experimental data. Atmos. Environ.42, 8640–8650.

Gromke, C., Jamarkattel, N., Ruck, B., 2016. Influence of roadside hedgerows on airquality in urban street canyons. Atmos. Environ. 139, 75–86.

Haaland, C., van den Bosch, C.K., 2015. Challenges and strategies for urbangreen-space planning in cities undergoing densification: a review. Urban For.Urban Green. 14, 760–771.

Hang, J., Wang, Q., Chen, X., Sandberg, M., Zhu, W., Buccolieri, R., Di Sabatino, S.,2015. City breathability in medium density urban-like geometries evaluatedthrough the pollutant transport rate and the net escape velocity. Build.Environ. 94, 166–182.

Hsieh, C.-M., Jan, Zhang, L., 2016. A simplified assessment of how tree allocation,wind environment, and shading affect human comfort. Urban For. UrbanGreen. 18, 126–137.

Janhall, S., 2015. Review on urban vegetation and particle air pollution −depositionand dispersion. Atmos. Environ. 105, 130–137.

Jeanjean, A.P.R., Hinchliffe, G., McMullan, W.A., Monks, P.S., Leigh, R.J., 2015. A CFDstudy on the effectiveness of trees to disperse road traffic emissions at a cityscale. Atmos. Environ. 120, 1–14.

Jeanjean, A.P.R., Monks, P.S., Leigh, R.J., 2016. Modelling the effectiveness of urbantrees and grass on PM2.5 reduction via dispersion and deposition at a cityscale. Atmos. Environ. 147, 1–10.

Kalansuriya, C.M., Pannila, A.S., Sonnadara, D.U.J., 2009. Effect of roadsidevegetation on the reduction of traffic noise levels. In: Proceedings of theTechnical Sessions of the Institute of Physics, Sri Lanka, pp. 1–6.

Katul, G., Mahrt, L., Poggi, D., Sanz, C., 2004. One- and two-equation models forcanopy turbulence. Bound Layer Meteor. 113, 81–109.

Lalic, B., Mihailovic, D.T., 2004. An empirical relation describing leaf-area densityinside the forest for environmental modelling. J. Appl. Meteorol. 43, 641–645.

Lateb, M., Meroney, R.N., Yataghene, M., Fellouah, H., Saleh, F., Boufadel, M.C., 2016.On the use of numerical modelling for near-field pollutant dispersion in urbanenvironments. A review. Environ. Pollut. 208, 271–283.

Launder, B.E., Spalding, D.B., 1974. The numerical computation of turbulent flows.Comput. Methods Appl. Mech. Eng. 3, 269–289.

Li, X.-B., Lu, Q.-C., Lu, S.I., He, H.D., Peng, Z.R., Gao, Y., Wang, Z.-Y., 2016. Theimpacts of roadside vegetation barriers on the dispersion of gaseous trafficpollution in urban street canyons. Urban For. Urban Green. 17, 80–91.

Maggiotto, G., Buccolieri, R., Santo, M.A., Leo, L.S., Di Sabatino, S., 2014. Validationof temperature-perturbation and CFD-based modelling for the prediction ofthe thermal urban environment: the Lecce (IT) case study. Environ. Modell.Softw. 60, 69–83.

Nikolova, I., MacKenzie, A.R., Cai, X., Alam, M.S., Harrison, R.M., 2016. Modellingcomponent evaporation and composition change of traffic-induced ultrafineparticles during travel from street canyon to urban background. FaradayDiscuss. 189, 529–546.

Nowak, D.J., Hirabayashi, S., Bodine, A., Hoehn, R., 2013. Modeled PM 2.5 removalby trees in ten US cities and associated health effects. Environ. Pollut. 178,395–402.

OS, 2016. Ordnance Survey: Britain Mapping Agency, internet database https://www.ordnancesurvey.co.uk/, (Accessed 18 November 2016).

Oldenkamp, R., van Zelm, R., Huijbregts, M.A., 2016. Valuing the human healthdamage caused by the fraud of Volkswagen. Environ. Pollut. 212, 121–127.

Peters, K., Eiden, R., 1992. Modelling the dry deposition velocity of aerosol particlesto a spruce forest. Atmos. Environ. 26, 2555–2564.

Pugh, T.A.M., MacKenzie, A.R., Whyatt, J.D., Hewitt, C.N., 2012. Effectiveness ofgreen infrastructure for improvement of air quality in urban street canyons.Environ. Sci. Technol. 46, 7692–7699.

Roy, S., Byrne, J., Pickering, C., 2012. A systematic quantitative review of urban treebenefits, costs, and assessment methods across cities in different climaticzones. Urban For. Urban Green. 11, 351–363.

Santiago, J.-L., Martilli, A., Martin, F., 2016. On dry deposition modelling ofatmospheric pollutants on vegetation at the microscale: application to theimpact of street vegetation on air quality. accepted. Bound . Layer Meteorol.http://dx.doi.org/10.1007/s10546-016-0210-5.

Selmi, W., Weber, C., Rivière, E., Blond, N., Mehdi, L., Nowak, D., 2016. Air pollutionremoval by trees in public green spaces in Strasbourg city, France. Urban For.Urban Green. 17, 192–201.

Steffens, J.T., Wang, Y.J., Zhang, K.M., 2012. Exploration of effects of a vegetationbarrier on particle size distributions in a near-road environment. Atmos.Environ. 50, 120–128.

Van den Berg, M., Wendel-Vos, W., van Poppel, M., Kemper, H., van Mechelen, W.,Maas, J., 2015. Health benefits of green spaces in the living environment: asystematic review of epidemiological studies. Urban For. Urban Green. 14,

806–816.

Vos, P.E., Maiheu, B., Vankerkom, J., Janssen, S., 2013. Improving local air quality incities: to tree or not to tree? Environ. Pollut. 183, 113–122.

try &

V

WW

W

W

modelling in Defra. A report by the Air Quality Modeling Review SteeringGroup.

Yazid, A.W.M., Sidik, N.A.C., Salim, S.M., Saqr, K.M., 2014. A review on the flowstructure and pollutant dispersion in urban street canyons for urban planningstrategies. Simulation, 0037549714528046.

A.P.R. Jeanjean et al. / Urban Fores

ranckx, S., Vos, P., Maiheu, B., Janssen, S., 2015. Impact of trees on pollutantdispersion in street canyons: a numerical study of the annual average effects inAntwerp, Belgium. Sci. Total Environ. 532, 474–483.

MO, 2008. Guide to Meteorological Instruments and Methods of Observation.ania, A., Bruse, M., Blond, N., Weber, C., 2012. Analysing the influence of different

street vegetation on traffic-induced particle dispersion using microscalesimulations. J. Environ. Manage. 94, 91–101.

hite, E.J., Turner, F., 1970. A method of estimating income of nutrients in a catchof airborne particles by a woodland canopy. J. Appl. Ecol. 7, 441–461.

hite, M.P., Alcock, I., Wheeler, B.W., Depledge, M.H., 2013. Would you be happierliving in a greener urban area? A fixed-effects analysis of panel data. Psychol.Sci. 24, 920–928.s

Urban Greening 22 (2017) 41–53 53

Williams, M., Barrowcliffe, R., Laxen, D., Monks, P., 2011. Review of Air Quality