Environ. Sci. Technol. 2005, 39, 6730-6738
Development and Application of an Urban Tree Air Quality Score for Photochemical Pollution Episodes Using the Birmingham, United Kingdom, Area as a Case Study ROSSA G. DONOVAN,† HOPE E. STEWART,‡ SUSAN M. OWEN,§ A. ROBERT MACKENZIE,§ AND C . N I C H O L A S H E W I T T * ,§ Department of Environmental Science, Lancaster University, Lancaster, LA1 4YQ, United Kingdom, School of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, United Kingdom, and Environment Agency, Bristol, United Kingdom
An atmospheric chemistry model (CiTTyCAT) is used to quantify the effects of trees on urban air quality in scenarios of high photochemical pollution. The combined effects of both pollutant deposition to and emission of biogenic volatile organic compounds (BVOC) from the urban forest are considered, and the West Midlands, metropolitan area in the UK is used as a case study. While all trees can be beneficial to air quality in terms of the deposition of O3, NO2, CO, and HNO3, some trees have the potential to contribute to the formation of O3 due to the reaction of BVOC and NOx. A number of model scenarios are used to develop an urban tree air quality score (UTAQS) that ranks trees in order of their potential to improve air quality. Of the 30 species considered, pine, larch, and silver birch have the greatest potential to improve urban air quality, while oaks, willows, and poplars can worsen downwind air quality if planted in very large numbers. The UTAQS classification is designed with practitioners in mind, to help them achieve sustainable urban air quality. The UTAQS classification is applicable to all urban areas of the UK and other mid-latitude, temperate climate zones that have tree species common to those found in UK urban areas. The modeling approach used here is directly applicable to all areas of the world given the appropriate input data. It provides a tool that can help to achieve future sustainable urban air quality.
Introduction Trees are an integral part of the urban environment and have beneficial amenity, aesthetic, and physical effects. For example, they act as sound buffers (1) and as screening for privacy. They provide habitat for birds, insects, and mammals and benefit human health by reducing stress and promoting recovery from illness (2, 3). Urban trees modify the microclimate by humidifying the surrounding atmosphere, ame* Corresponding author phone: +44 1524 593931; fax: +44 1524 593985; e-mail:
[email protected]. † Lancaster University, now at the University of Birmingham. ‡ Lancaster University, now at the Environment Agency. § Lancaster University. 6730
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005
FIGURE 1. Location of West Midlands in the UK and the constitutive boroughs. Adapted from Ordnance Survey map data by permission of the Ordnance Survey Crown, copyright 2001. liorating the urban heat island effect (4), and providing shade and wind shelter that, in turn, can reduce building energy requirements (5, 6). This can itself result in reduced fossil fuel consumption (7). Trees also intercept rainfall, reducing erosion and flooding (8), sequester CO2, thus mitigating global warming (e.g., refs 9-11), and act as an enhanced deposition sink for gaseous and particulate pollution (12-14). Because of their role as a deposition sink, large-scale urban tree planting has been suggested as a means of improving urban air quality (15). Some trees emit volatile organic compounds (VOCs), the specific compounds and emission rates depending on tree species. Emitted VOCs, such as isoprene and monoterpenes, can contribute to the formation of secondary pollutants such as ozone (O3), peroxyacetyl nitrate (PAN), and secondary particulates following reaction with oxides of nitrogen in the presence of sunlight. The role that urban trees play in the formation of these secondary pollutants can be very significant (e.g., Atlanta, GA; ref 16). Because individual tree species may influence air quality, either in a positive sense by enhancing the deposition of pollutants or in a negative sense by contributing to secondary pollutant formation through the emission of reactive VOCs, it would be desirable to have a system that ranks tree species in the urban environment according to their effects on air quality. Here, we use an atmospheric chemistry model, information about species-specific VOC emission rates and pollutant deposition rates, and tree cover data to develop an urban tree air quality score (UTAQS). We used the West Midlands metropolitan area in the UK as a case study and ranked the 30 most common tree species found in this 900 km2 urban and semi-urban region, in terms of their effects on air quality. We examined changes in the mixing ratios of photochemical pollutants O3, NO2, HNO3, PAN, and NO due to changes in the emissions of biogenic VOC (BVOC) in scenarios with changing anthropogenic VOC (AVOC), NOx, temperature, and photosynthetic active radiation (PAR). Study Area. The West Midlands metropolitan area comprises the boroughs and cities of Birmingham, Coventry, Dudley, Sandwell, Solihull, Walsall, and Wolverhampton (Figure 1). The morphology of the area is very diverse, having substantial areas of emerging and mature urban woodlands; well-established residential areas with associated gardens, parks, and open spaces; large amounts of rural areas; major 10.1021/es050581y CCC: $30.25
2005 American Chemical Society Published on Web 07/27/2005
transport corridors; and numerous industrial and postindustrial sites. Many of these land-use types have a high potential for large-scale tree planting schemes, and much (if not all) of the region would benefit from an improvement in air quality.
Materials and Methods Overview. For the purposes of this study, the West Midlands metropolitan area has been classified using 27 land-use attributes as input to principal components and cluster analysis, to yield eight different urban classes (17). A field survey of the urban forest of the area was conducted, based on the stratified random sampling of the eight land-use classes (32 300 trees surveyed). An estimate of the maximum proportion of each hectare that could realistically be planted with trees, called the future planting potential (FPP), was made for each hectare surveyed (18). Estimates of foliar biomass and leaf area were made for every tree recorded in the field survey. These were used in the estimation of BVOC emission potentials (EPs) for each urban class (18), which were found to be significantly higher than BVOC emission potential estimates (EPs) derived from a desk study for the same urban land classes (19). An urban-class-weighted average was estimated to derive BVOC emission potentials for use in the model. Predictions of typical tree populations for each urban land class were made by linear extrapolation of the survey data collected in the sample hectares to all the square kilometers of that class. These urban land-class predictions therefore maintained the species mix and typical tree dimensions recorded during the field survey (see ref 18 for details). This enabled estimates of foliar biomass, leaf area, and BVOC EPs to be made for each urban land class. A number of scenarios was then designed to model the effects of changes in the composition and size of the urban forest on air quality using an air quality model, and these model results were used to obtain the UTAQS rankings for different tree species. Model Description. The Cambridge tropospheric trajectory model of chemistry and transport (CiTTyCAT) is an air parcel model, described in detail by Evans et al. (20) and Emmerson et al. (21) and summarized here. It is developed from the work of Wild (22) and Wild et al. (23). The model considers the chemical composition of an air parcel subject to the following processes: photochemistry, emissions, deposition, and mixing from the free troposphere (20)
d[Ci] v i Ei vm ) P - L[Ci] - [Ci] + + ([Ci]ft - [Ci]) (1) dt h h h where [Ci] is the concentration of species i (cm-3), P and L[Ci] represent the photochemical production and loss rates of i (cm-3 s-1), vi is the deposition velocity of i (cm s-1), h is the height of the boundary layer (cm), Ei is the emission rate of i (cm-2 s-1), [Ci]ft is the concentration of species i in the free troposphere (cm-3), and vm is the mixing velocity (cm s-1). P and L are time-varying due to changes in solar input and temperature. Ei, vi, and h are time-varying if the air parcel follows a Lagrangian trajectory. In our study, Ei, vi, and h are constant because the model is run as a stationary box. The values for Ei and vi are discussed next; h is set at 1390 m, which is a typical daytime value for the mid-latitude continental boundary layer in midsummer. Also, we set vm ) 0 (i.e., we do not consider mixing from above). The chemistry module contains a detailed representation of Ox, HOx, and NOx chemistry and uses a parametrized hydrocarbon scheme that considers the degradation of 13 hydrocarbons (CH4, C2H6, C2H4, C2H2, C3H8, C3H6, C4H10, C5H12, C6H14, C6H6, C7H8, C5H8, isoprene, C10H16, and R-pinene) with updates from recent work. Degradation of isoprene is initiated by OH and O3; degradation of R-pinene is initiated
TABLE 1. July Day and (Night) Dry Deposition Velocities (cm s-1) for Different Land Use Types (26) O3 NO NO2 HNO3 CO NO3 N2O5 HNO4 H2O2 HCHO CH3CHO CH3COOH CH3COO2H PAN
water
woodland
grassland
built up (26)
0.03 (0.03) 0 (0) 0 (0) 0.8 (0.8) 0 (0) 0.02 (0.02) 1.0 (1.0) 1.0 (1.0) 1.0 (1.0) 1.0 (1.0) 0.02 (0.02) 0.25 (0.25) 0.36 (0.36) 0.01 (0.01)
1 (0.4) 0 (0) 0.4 (0) 4 (2) 0.05 (0.05) 0.83 (0.04) 4.0 (3.0) 4.0 (3.0) 1.25 (0.16) 1.0 (0.01) 0.31 (0) 0.83 (0.04) 0.71 (0.04) 0.53 (0.04)
0.8 (0.4) 0 (0) 0.4 (0) 2.5 (1.5) 0.05 (0.05) 0.63 (0.06) 2.5 (1.5) 2.5 (1.5) 1.25 (0.53) 0.71 (0.03) 0.26 (0) 0.63 (0.06) 0.53 (0.07) 0.42 (0.06)
0.2 (0.2) 0 (0) 0 (0) 3 (2) 0 (0)
by OH, NO3, and O3 (24, 25). The photochemical contribution to PM10 is assumed to scale with nitric acid production. This is not intended to be a complete representation of PM10, nor of secondary aerosol production by sulfur and ammonia chemistry. It is also important to note that the model assumes instantaneous mixing in the boundary layer and is not intended to predict near-ground or curbside conditions. Because there is no diurnal change in the height of the boundary layer, concentrations of primary pollutants whose emissions tend to have an early morning peak will be underestimated in the model, and some of the nonlinearities in the daily NOx-VOC chemistry will not be captured. However, the VOC/NOx ratio achieved in the model is ∼5, which is typical of an urban environment. CiTTyCAT includes dry deposition for O3, NO, NO2, NO3, N2O5, HNO3, HNO4, H2O2, CO, HCHO, CH3CHO, CH3OOH, CH3COO2H, and PAN (Table 1), which are based on CIS landcover categories (Table 1). The CIS land-cover categories used do not include built-up land. In this study, deposition velocities for O3, NO, NO2, CO, and HNO3 were recalculated (0.4, 0.0, 0.12, 0.01, and 3.0 cm s-1, respectively) to reflect the integrated land cover (water, grass, forest, and buildings) for the West Midland region. Wet deposition is not simulated in the model. CiTTyCAT includes anthropogenic emissions for NOx, CO, and the hydrocarbons mentioned previously. Anthropogenic emissions in the UK are supplied at a resolution of 10 km × 10 km from the National Atmospheric Emissions Inventory (NAEI) (27) and are 197, 2290, and 127 tonne gridcell-1 year-1 for NOx (as NO2), CO, and total hydrocarbons, respectively. Emissions of CH4 (58 380 tonne gridcell-1 year-1) are supplied at a resolution of 50 km × 50 km by Core Inventory Air (CORINAIR) (28). Biogenic emissions for the West Midlands metropolitan area have been estimated at a resolution of 1 km × 1 km using the methodology described next and an urban class area-weighted average used in the model (1240 and 163 g km-2 h-1 for isoprene and monoterpenes (as R-pinene), respectively). Initial conditions are calculated from the Cambridge twodimensional model (29, 30) for O3, NOx, CH4, H2O2, CO, HCHO, CH3OOH, PAN, C2H6, and C3H8, with all other species initialized at zero concentration. For each run, there were constant anthropogenic emissions, as described previously. Biogenic emissions of isoprene varied diurnally according to light intensity and temperature, and R-pinene varied according to temperature alone (31). The model was run for 5 days until the chemistry stabilized. The model scenario, therefore, represents unusual but not unrealistic photochemistry in stagnant anticyclonic conditions. Such conditions are well-known to result in photochemical pollution episodes, to give, in general, worst-case air qualities in the VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6731
TABLE 2. Changes in Air Quality and Urban Tree Air Quality Scores for Different Sensitivity Model Runsa change in concentration (%)
c
scenario description
O3b
NO2c
HNO3d
NOc
PANc
S1: NOx emissions +50% S2: NOx emissions -50% S3: anthropogenic VOCs alone (no trees) S4: biogenic VOCs alone S5: July +50% biogenic VOC S6: July -50% biogenic VOC S7: urban stress EPs for bet pen and que rob (LC 4, 5, 6 only) S8: urban stress EPs for all species (LC 4, 5, 6 only) S9: July +50% deposition velocities S10: July -50% deposition velocities S11: July +20% foliar biomass S12: July -20% foliar biomass S13: Ta +2°C S14: Ta -2°C S15: PAR +200 µmol m-2 s-1 S16: PAR -200 µmol m-2 s-1
11.5 -18.5 -2.0
34.8 -39.6 -1.7
52.7 -50.3 7.7
33.0 -33.4 7.4
25.7 -34.1 -16.9
-9.0 1.4 -1.7 -0.3 -4.6 -4.5 7.8 0.1 -0.1 0.9 -0.8 0.4 -0.8
-6.5 0.6 -0.7 -0.4 -20.9 0.7 -1.3 0.3 -0.3 0.6 -0.5 0.6 -1.1
11.0 -3.5 3.6 -0.1 -36.6 -2.9 7.1 -1.6 1.7 -1.7 1.4 -0.4 0.6
27.6 -3.7 5.1 1.1 8.5 9.0 -12.5 -0.6 0.7 -2.4 2.5 -0.5 1.6
-70.4 10.4 -7.6 0.3 6.7 -1.4 3.9 5.1 -1.9 7.0 -2.6 4.2 -2.9
a CS output values: [O ] ) 66 ppb; [NO ] ) 0.9 ppb; [HNO ] ) 4.6 ppb; [NO] ) 0.1 ppb; and [PAN] ) 1.3 ppb. 3 2 3 Maximum 1 h mean on fifth day. d 24 h running mean on fifth day.
summertime in the UK. The model conditions have been chosen here to maximize the modeled effects that urban tree canopies may have on air quality, so that a ranking can be achieved for urban planning and landscape practitioners. Biogenic Emission Potentials. The BVOC emissions for the model were generated in stages. The 900 km2 metropolitan area was classified into eight urban land classes (17). Twentytwo 1 km squares were randomly selected across the eight urban land classes. In each square, three 4 hectare blocks were randomly selected, and every tree >2 m and hedge >1 m was recorded. Trunk height (TH), crown height (CH), crown spread (CSp), and diameter at breast height (DBH) were recorded. In excess of 32 000 trees, shrubs, and hedges across 260 hectares were recorded by a professional arboriculturalist (RGD) between July 1999 and August 2000 (18). The method of ref 32 was used to calculate tree species leaf area from CSp and CH. Species-specific leaf weights were used to convert leaf areas to species-specific foliar biomass. Foliar biomass and leaf areas were used to generate typical urban land-class biomass and leaf area estimates (18). A database of biogenic VOC emission potentials (33) was used to assign isoprene and monoterpene EPs to each sampled tree species. Where there were no reported EPs for a species, a taxonomic value was assigned using the method of ref 34. Where there were no taxonomic values available, default values were assigned (0.1 µg g-1 dw h-1 for isoprene and 0.01 µg g-1 dw h-1 for monoterpenes). All emission potentials are estimated for standard conditions of temperature and PAR (30 °C and 1000 µmol m-2 s-1, respectively). Using the estimates of foliar biomass (18), land class EPs standardized to 30 °C and 1000 µmol m-2 s-1 were generated. The error associated with extrapolating branch scale emission rates to canopy scale emission potentials has been estimated to be a factor of 3-5 (31, 35). BVOC emissions were measured from two stressed and two natural location urban specimens of Quercus robur and Betula pendula using a dynamic branch enclosure and analytical methods previously described in detail (36-38). Here, the term stressed is used in a generic way, to denote trees in an urban environment that could be suffering from water stress (e.g., paved locations), high pollution doses (e.g., in curbside locations), and extreme shading, etc. Emission rates were normalized using the method of ref 39. Isoprene emissions from stressed trees were reduced by a factor of 0.9, as compared to normal trees, and total monoterpene emissions increased by a factor of 2.2 in stressed trees. These 6732
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005
b
8 h running mean on fifth day.
factors were used to modify estimated emissions from all stressed trees, and while they are based on limited measurements, they are only used here to generate what if scenarios rather than to draw quantitative conclusions about the effect of urban stress on BVOC emissions.
Model Runs and Scenarios A number of scenarios was run using the CiTTyCAT model to assess the impact that changes to the various inputs would have on air quality. The control scenarios (CS: the sensitivity model run control, CU: the UTAQS model run control, and CF: the future scenario model run control) (against which all other model runs are compared) consider the tree populations defined below with corresponding BVOC emission potential values and representative AVOC emission potentials, deposition potentials, PAR, and temperature values for the conurbation for July 19-24, 1996. The model runs were categorized into three types: sensitivity, UTAQS, and future. In the sensitivity runs, all parameters of the control scenario (CS) (i.e., foliar biomass, biogenic emission potentials, deposition potentials, temperature, PAR, urban stress effect) were altered to assess the sensitivity of the model to these factors (Table 2). In the UTAQS runs, the species composition of the urban forest was altered to assess the effects of different tree species on air quality and to develop the UTAQS classification (Table 3). The future runs were used to investigate the impacts of possible future changes to the urban forest and environmental conditions on air quality within the model domain (Table 4). Sensitivity Runs. The sensitivity scenarios presented in Table 2 (numbered S1-S16) were used to determine the integrity of the model. The output of each sensitivity run was compared with the control (CS), which considers the effects the current West Midlands urban forest has on air quality for current temperatures Ta. Peak foliar biomass was increased (S11) and decreased (S12) by 20%, and BVOC emission potentials were increased (S5) and decreased (S6) by 50% to assess the sensitivity of the model to changes in the tree population and the proportion of low and high BVOC emitting species. The effect of urban stress on trees was modeled by altering isoprene emissions by a factor of 0.9 and total monoterpene emissions by a factor of 2.2. S7 altered emissions for all English oak and silver birch trees in urban classes 4-6, and S8 altered emissions for all trees in urban classes 4-6. The effect of temperature
TABLE 3. Modeled Changes in Air Quality and Urban Tree Air Quality Scores for the Large-Scale Planting of Different Tree Speciesa UTAQS option change in concentration (%)
scenario description: existing urban forest + 20% of the following:
O3b
U1: English oak (Quercus robur L.) U2: White willow (Salix alba) U3: Crack willow (Salix fragilis) U4: aspen (Populus tremula) U5: Sessile oak (Quercus petraea) U6: Red oak (Quercus rubra) U7: Goat willow (Salix caprea) U8: lilac (Syringa vulgaris) U9: sycamore (Acer pseudoplatanus) U10: Mountain ash (Sorbus aucuparia) U11: apple (Malus spp.) U12: elder (Sambucus nigra) U13: Common lime (Tilia x europea) U14: holly (Ilex aquifolium) U15: Italian alder (Alnus cordata) U16: hazel (Corylus avellana) U17: Leyland cypress (x Cupressocyparis leylandii) U18: Grey alder (Alnus incana) U19: cherry (Prunus avium) U20: English elm (Ulmus procera) U21: ash (Fraxinus excelsior) U22: Field maple (Acer campestre) CU: existing composition U23: alder (Alnus glutinosa) U24: laurel (Prunus laurocerasus) U25: Lawson cypress (Chamaecyparis lawsoniana) U26: hawthorn (Crataegus monogyna) U27: Norway maple (Acer platanoides) U28: Silver birch (Betula pendula) U29: larch (Larix decidua) U30: Austrian, Corsican, Maritime pine (Pinus nigra)
2.9 2.4 2.3 1.9 1.2 1.0 0.8 -0.1 0.0 -0.1 -0.1 -0.2 -0.1 -0.2 -0.2 -0.2 -0.2 -0.2 -0.2 -0.2 -0.2 -0.3 -0.3 -0.3 -0.2 -0.2 -0.2 -0.3 -0.3 -0.4 -0.8
NO2c
HNO3d
2.7 -5.0 2.5 -3.3 2.5 -3.1 1.4 -4.5 1.5 -1.6 1.2 -1.4 0.9 -1.4 0.0 -0.3 0.3 -0.4 0.1 -0.2 -0.1 -0.5 0.0 -0.2 0.2 -0.4 0.0 -0.4 0.0 -0.2 0.1 -0.3 0.0 -0.4 0.0 -0.2 0.0 -0.3 0.0 -0.3 0.0 -0.2 0.0 -0.2 0.0 -0.2 0.0 -0.2 -0.2 -0.8 -0.2 -0.9 -0.4 -1.2 -0.2 -0.8 -0.5 -1.4 -2.2 -5.2 -4.9 -10. 8
HNO3 O3 + NO2
NO
PAN
O3
NO2
-8.0 -6.6 -6.4 -5.3 -3.4 -2.9 -2.2 0.1 -0.2 0.1 0.4 0.5 0.0 0.5 0.5 0.5 0.5 0.6 0.6 0.6 0.6 0.7 0.7 0.5 0.6 0.5 0.5 0.5 0.5 0.5 1.9
20.6 16.6 16.3 14.8 9.8 8.3 7.1 2.0 3.3 2.2 2.1 1.9 2.8 2.0 1.9 2.1 2.0 1.9 1.9 1.9 1.9 1.9 1.9 1.9 2.1 2.2 2.3 2.1 2.3 3.6 5.1
-3.85 -3.17 -3.06 -2.53 -1.53 -1.35 -0.99 0.09 0.06 0.18 0.12 0.20 0.17 0.21 0.25 0.27 0.23 0.28 0.27 0.29 0.32 0.34 0.34 0.40 0.29 0.26 0.24 0.37 0.33 0.50 1.11
-1.58 × 10-02 -1.47 × 10-02 -1.49 × 10-02 -8.03 × 10-03 -9.02 × 10-03 -6.71 × 10-03 -4.97 × 10-03 9.33 × 10-05 -1.93 × 10-03 -5.18 × 10-04 4.28 × 10-04 -1.67 × 10-04 -1.32 × 10-03 1.85 × 10-04 -2.12 × 10-04 -3.33 × 10-04 2.38 × 10-04 -2.32 × 10-04 8.17 × 10-05 -1.32 × 10-04 -2.22 × 10-04 -2.27 × 10-04 -2.27 × 10-04 -2.65 × 10-04 1.26 × 10-03 1.44 × 10-03 2.24 × 10-03 1.24 × 10-03 2.74 × 10-03 1.28 × 10-02 2.84 × 10-02
1.22 0.81 0.76 1.10 0.39 0.36 0.34 0.07 0.11 0.06 0.12 0.05 0.09 0.09 0.05 0.06 0.11 0.06 0.07 0.07 0.05 0.05 0.05 0.06 0.19 0.22 0.28 0.19 0.34 1.28 2.65
-3.87 -3.19 -3.07 -2.54 -1.54 -1.36 -1.00 0.09 0.06 0.18 0.12 0.20 0.17 0.21 0.25 0.27 0.23 0.28 0.27 0.29 0.32 0.34 0.34 0.40 0.29 0.26 0.24 0.37 0.33 0.51 1.14
O3 + NO2 + HNO3 -2.65 -2.38 -2.31 -1.44 -1.15 -1.00 -0.65 0.16 0.17 0.24 0.24 0.24 0.26 0.29 0.30 0.33 0.34 0.34 0.34 0.36 0.38 0.39 0.39 0.46 0.48 0.48 0.53 0.57 0.67 1.79 3.79
a Runs are listed in rank order of multi-UTAQS option (O + NO + HNO ). CU output values: [O ] ) 66 ppb; [NO ] ) 0.9 ppb; [HNO ] ) 12 µg 3 2 3 3 2 3 m-3; [NO] ) 0.1 ppb; and [PAN] ) 1.3 ppb. b 8 h running mean on fifth day. c Maximum 1 h mean on fifth day. d 24 h running mean on fifth day.
TABLE 4. Modeled Changes in Air Quality and Urban Tree Air Quality Scores for Different Future Scenariosa UTAQS value scenario description
O3b
change in concentration (%) NO2c HNO3d NO PAN
F1: as now + low UTAQS species (aged 50) in FPP areas for Ta + 2°C 6.2 6.0 -12.0 F2: as now + low UTAQS species (aged 10) in FPP areas 5.0 5.7 -10.5 F3: as now + low UTAQS species (aged 50) in FPP areas for Ta 4.4 4.7 -8.4 F4: as now + existing species (aged 50) in FPP areas for Ta +2 °C 3.9 3.1 -11.3 F5: as now + 25% FPP low UTAQS species aged 50 and Ta +2 °C 2.6 2.3 -4.4 F6: as now + 25% FPP existing species aged 50 and Ta +2 °C 1.9 1.4 -4.1 F7: as now + 25% FPP low UTAQS species aged 10 1.5 1.7 -2.7 F8: as now + existing species (aged 50) in FPP areas for Ta 2.4 2.1 -7.9 F9: as now + 25% FPP high UTAQS species aged 50 and Ta +2 °C 0.5 -0.6 -4.5 F10: Ta +2 °C 0.9 0.6 -1.7 F11: as now + existing species (aged 10) in FPP areas 2.1 2.0 -8.1 F12: PAR +200 µmol m-2 s-1 0.4 0.6 -0.4 F13: as now + 25% FPP existing species aged 50 0.7 0.6 -2.0 F14: as now + 25% FPP existing species aged 10 0.6 0.6 -2.1 CF: existing tree numbers + 20% of TC of existing composition -0.3 0.0 -0.2 F15: PAR -200 µmol m-2 s-1 -0.8 -1.1 0.6 F16: as now + 25% FPP high UTAQS species aged 10 -0.6 -1.4 -3.3 F17: as now + high UTAQS species (aged 50) in FPP areas for Ta +2 °C -1.0 -4.6 -12.9 F18: as now + high UTAQS species (aged 50) in FPP areas for Ta -1.7 -4.2 -9.6 F19: as now + high UTAQS species (aged 10) in FPP areas -2.6 -5.7 -13.0 F20: NOx emissions -50% -18.5 -39.6 -50.3 F21: July -50% anthropogenic VOC and NOx with Ta +2 °C -20.2 -40.5 -49.2 F22: July -50% anthropogenic VOC and NOx -20.8 -40.9 -47.7 c
-17.8 -15.0 -12.7 -11.9 -6.9 -5.1 -4.4 -7.6 -1.5 -2.4 -7.0 -0.5 -2.1 -1.9 0.7 1.6 1.2 1.3 3.3 4.9 -33.4 -28.7 -27.1
a CF output values: [O ] ) 66 ppb; [NO ] ) 0.9 ppb; [HNO ] ) 12 µg m-3; [NO] ) 0.1 ppb; and [PAN] ) 1.3 ppb. 3 2 3 Maximum 1 h mean on fifth day. d 24 h running mean on fifth day.
and PAR changes on model sensitivity was modeled by increasing and decreasing the temperature by 2 °C (S13 and S14, respectively) and PAR by 200 µmol m-2 s-1 (S15 and S16,
b
0.48 0.42 0.34 0.35 0.18 0.14 0.12 0.24 0.08 0.07 0.23 0.04 0.07 0.07 0.02 -0.03 0.03 0.10 0.04 0.05 -0.34 -0.47 -0.51
UTAQS (O3 + NO2 + PM10) -5.25 -3.97 -3.69 -2.41 -2.32 -1.42 -1.36 -1.18 -0.92 -0.83 -0.76 -0.50 -0.42 -0.30 0.39 0.91 1.63 4.57 4.60 6.59 36.81 38.78 39.22
8 h running mean on fifth day.
respectively, to simulate changes in cloudiness). Scenarios S1 and S2 increased and decreased emissions of NOx by 50%, and S9 and S10 increased and decreased deposition potentials VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6733
(using a peak summer Vd value) by 50%. The model was also run with AVOC emissions only (S3) and with BVOC emissions only (S4). UTAQS Species Runs. Thirty UTAQS scenarios (numbered U1-U30 in Table 3) were used to assess the relative contribution that the 30 most dominant tree species make to air quality in the area. Each scenario maintains the existing urban forest but adds an additional 20% of the total number of trees. The extra trees are of one species (i.e., the tree species for which the UTAQS is being calculated). The output of each run is compared with that of the control scenario (CU), which maintains the existing urban forest and adds 20% of the tree count for the existing species mix. In each scenario, the age mix of that species within the current urban forest is maintained for the extra trees. Although we treat all hedge and tree canopies as functionally equivalent for our purposes, it should be remembered that they are, of course, very different in other terms (e.g., habitat, shade provision, aesthetic value, etc.). The model was run for typical July temperatures (Ta). Future Scenario Runs. Twenty-two future scenarios (F1-F22 shown in Table 4) were used to assess the effects of changes to the urban forest and environment on air quality. Each model run output was compared with that of the control (CF), which maintains the existing urban forest and expands it by 20% (by number), keeping the same species and age mix modeled for Ta. Emissions of anthropogenic VOC and NOx are likely to reduce in the future as a result of legislation and technology. However, global warming is likely to alter air temperatures, which has implications for BVOC emissions and atmospheric chemistry. Therefore, the effect of 50% less anthropogenic VOC and NOx emissions on air quality was modeled for July temperatures (F22) and potential future temperatures (Ta + 2 °C) (F21). The effects of large-scale tree planting on air quality were modeled for various scenarios. These maintain the species composition, tree count, and tree size of the present West Midlands urban forest with the addition of trees planted of different ages and areas. This assumes that the composition of the present urban forest will remain unchanged in the future, and while this may not be realistic, it is assumed that the loss of trees due to death and removal will be compensated for by the planting of new trees. Where the planting areas in the scenarios are planted with 30 regionally dominant species, numbers may vary as each species has a different mean projected crown area (PCA). The projected crown area describes the area of ground that the tree’s canopy would shade if the sun were directly overhead, while the crown spread is twice the mean radius of the tree’s canopy. The following relationship has been assumed for the purposes of this study: PCA ) π(0.5CS)2. The maximum number of trees is planted in the area assuming no overlap of PCA. A first set of urban greening scenarios assumes that the existing urban forest remains intact and that 10% of the landclass FPP (future planting potential) area is planted with trees. In three scenarios, 10-year-old trees are planted in the potential planting area using (i) the same species mix as the existing urban forest (F11); (ii) low ranking UTAQS species (F2); and (iii) high ranking UTAQS species (F19) for current July temperatures (Ta). Similar scenarios are used but with 50-year-old trees for Ta (F8, F3, and F18) and Ta + 2 °C (F4, F1, and F17). A second set of urban greening scenarios assumes that the existing forest remains intact and that 25% of the land class FPP area is planted with extra trees. In three scenarios, 10-year-old trees are planted in the potential planting area using (i) the same species mix as the existing urban forest (F14); (ii) low ranking UTAQS species (F7); and (iii) high ranking UTAQS species for Ta (F16). Similar scenarios are 6734
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005
used but with 50-year-old trees for Ta + 2 °C (F6, F5, and F9, respectively). In addition, the planting area is planted with 50-year-old trees using the existing species mix and modeled for Ta (F13). Assessing Changes in Air Quality. The outputs of all model runs were compared with the outputs of the relevant control scenario (CS, CU, or CF). When using a photochemical box model to classify the effects of trees on urban air quality, changes to VOC emission potentials and the deposition potentials of O3, NO, NO2, HNO3, and CO may not have a simple additive effect on model outputs because several components of the model behave nonlinearly. We chose to quantify the effect of trees on urban air quality by focusing on those chemical species for which there is a noticeable change in concentration following a change in the urban forest and for which there are air quality standards (AQS) in the UK. O3 and NO2 have air quality standards of 50 ppb (8 h running mean) and 150 ppb (1 h running mean), respectively. HNO3 does not have an AQS but is a component of secondary inorganic aerosol (SIA) and so, although CiTTyCAT does not model SIA, changes in HNO3 are regarded as being indicative of the changes that could be expected in the photochemical contribution to SIA. Changes in the mass concentrations of HNO3 are compared between each sensitivity run and the control with the AQS for PM10 (50 µg m-3). We define a number of different UTAQS equations to assess the change in air quality for each model run. The UTAQS equations investigate the effect on each of the pollutants in turn and on combinations of the pollutants to illustrate the importance of each component
( ) ( ) ( ) ∆O3 AQSO3
(2)
UTAQS(NO2) ) - 100
∆NO2 AQSNO2
(3)
UTAQS(PM10) ) - 100
∆HNO3 AQSPM10
(4)
UTAQS(O3) ) - 100
UTAQS(O3 + NO2) ) - 100
(
∆NO2 ∆O3 + AQSO3 AQSNO2
)
UTAQS(O3 + NO2 + PM10) ) ∆NO2 ∆HNO3 ∆O3 + + - 100 AQSO3 AQSNO2 AQSPM10
(
)
(5)
(6)
where ∆O3 is the difference between the peak 8 h running mean modeled concentration of O3 on the fifth day of the UTAQS model run and that of the control (CU); ∆NO2 is the difference between the peak 1 h concentration of NO2 on the fifth day of the UTAQS model run and that of the control (CU); ∆HNO3 is the difference between the modeled 24 h running mean of HNO3 on the fifth day of the UTAQS model run (in µg m-3) and the control (CU); and AQSO3, AQSNO2, and AQSPM10 are the air quality standards for O3 (50 ppb), NO2 (150 ppb), and PM10 (50 µg m-3), respectively. The values are multiplied by -100 to give a positive value for an improvement in air quality when compared to the control tree population (CU) and to scale UTAQS scores to be between -10 and +10.
Results and Discussion Model results, for the sensitivity study control run (run CS), are shown in Figure 2. The accumulation of secondary pollutants (O3, HNO3, and PAN) through the model run is evident. NO and NO2 show pronounced diurnal cycles,
FIGURE 2. Modeled mixing ratios of O3, NO2, NO, HNO3, and PAN over 5 days for the control scenario CS. reflecting diurnal patterns in partitioning between NO, NO2, NO3, and N2O5. The emissions and box-height used in our simulations produce NO2 concentrations of up to 0.9 ppb. Concentrations of NO2 of 6 ppb were recorded for Birmingham in July 1996, which, although higher than the model output, indicate the tendency for emissions and atmospheric chemistry at that time to yield relatively low concentrations of NO2. The discrepancy between the modeled NO2 concentrations for the control scenario and actual measurements is due to the high (1390 m) and constant boundary layer height used in these CiTTyCAT model runs. Similarly, the unusually high concentrations of HNO3 relative to NO2 are explained by the conditions of the model run and the fact that wet deposition is not accounted for, which in reality would remove HNO3. The following discussion uses the relative differences of the model runs as compared to the relevant control scenario (CS, CU, or CF), and the percentages shown in Tables 3 and 4 are actual percentage changes to the control run values (i.e., +10% ) control value + 10% ) control value and -10% ) control value - 10% ) control value). Sensitivity Runs. Changing NOx emissions has the greatest effect on the concentrations of O3, NO2, HNO3, PAN, and NO. Increased concentrations of O3 (+12%) and NO2 (+35%) in the elevated NOx scenario (S1) (Table 2) indicate that the chemistry of NOx is limited. However, the chemistry is reasonably close to the compensation point between NOxand VOC-limited chemistry since the reduction of VOC input (e.g., S3, S4, and S6) decreases the O3 concentration and increasing VOC input (S5) increases the O3 concentration. Altering BVOC emissions has a much smaller effect than altering NOx emissions on the concentrations of all five target pollutants. Increasing BVOC emissions (S5) results in an increase (0-10%) in the production of O3, NO2, and PAN but a decrease (
