InTechOpen uses cookies to offer you the best online experience. By continuing to use our site, you agree to our Privacy Policy.

Environmental Sciences » "Advanced Air Pollution", book edited by Farhad Nejadkoorki, ISBN 978-953-307-511-2, Published: August 17, 2011 under CC BY-NC-SA 3.0 license. © The Author(s).

Chapter 23

A New Air Quality Index for Cities

By Lígia T. Silva and José F. G. Mendes
DOI: 10.5772/16701

Article top


Normalisation functions
Figure 1. Normalisation functions
Dummy variables; a) Dummy variable function of PM10; b) Dummy variable function of CO; c) Dummy variable function of NO2 ; d) Dummy variable function of C6H6 ; e) Dummy variable function of O3
Figure 2. Dummy variables; a) Dummy variable function of PM10; b) Dummy variable function of CO; c) Dummy variable function of NO2 ; d) Dummy variable function of C6H6 ; e) Dummy variable function of O3
Automatic monitoring sites in Viana do Castelo
Figure 3. Automatic monitoring sites in Viana do Castelo
Temporal variation at A1
Figure 4. Temporal variation at A1
Temporal variation at A2
Figure 5. Temporal variation at A2
Temporal variation at A3
Figure 6. Temporal variation at A3
Air pollution maps, summer scenario
Figure 7. Air pollution maps, summer scenario
CityAIR, summer scenario
Figure 8. CityAIR, summer scenario

A New Air Quality Index for Cities

Lígia T. Silva1 and José F.G. Mendes

1. Introduction

Global population growth has led to increased populations living in urban areas. Often, this enhances stresses on space, ecosystems, infrastructures, facilities and personal lifestyles. Problems related to quality of life in cities are increasingly relevant, especially with regard to environmental issues.

Due to a generalised increase of mobility and road traffic in urban areas, the total emissions from road traffic have risen significantly, assuming the main responsibility for the disregard of air quality standards. In urban environment the typical anthropogenic sources are mainly the road traffic and, when existing, the industrial activity.

The quantitative evaluation of traffic air pollution levels is the basis on which air pollution control policies stand. The evaluation of air quality may be occasional or long-term. Occasional evaluation is useful in the context of information and alert systems for the population, working normally in real or almost-real time. Data is acquired through measurements made on an hourly or daily average basis and concentration episodes are evaluated and reported. When long-term data is considered, then we talk about long-term trend analysis, this kind of approach can be adequate for identifying the major emissionsource contributors to urban pollution (Butterwick et al. 1991).

In order to find an air quality index, the pollutant concentrations are combined through a classification scale anchored on the legal limits and, on the other side, on the impacts over human health. Typically these classification models consider only the worse pollutant, i.e. the one which concentration is higher given a certain scale. Two air quality evaluation models are referred, both working in real time: a Canadian and a Portuguese experience.

The objective of this chapter is to present a new air quality index, cityAIR, developed for urban contexts. The mathematical formulation of cityAIR stands on two logics: whenever at least one of the pollutants considered overcomes the legal limits for the concentration, this will be the only relevant one for the index calculation, and the value will be the minimum of the scale (zero or red); when there is no limit violation, then all the pollutants are considered for the overall air quality, which is calculated through a multi-criteria combination of the concentrations, where trade-off is allowed.

A case study is presented for Viana do Castelo, a mid-sized Portuguese city, in which cityAIR values were calculated in consideration of concentrations of CO, NO2, O3, C6H6 and PM10.

2.Urban air pollution

Urban air pollution became one of the main factors of degradation of the quality of life in cities. This problem tends to worsen due to the unbalanced development of urban spaces and the significant increase of mobility and road traffic. As a consequence, the total emissions from road traffic have risen significantly, assuming the main responsibility for the disregard of air quality standard (Butterwick, L. et al., 1991).

The atmospheric pollutants are emitted from existent sources and, subsequently, transported and dispersed several times in the atmosphere before reaching receptors through wet deposition (rainout and washout by rain and snow) or dry deposition (particle adsorption). In an urban environment, typical anthropogenic sources are mainly the road traffic and, when existing, the industrial activity.Emissions from mobile sources contribute to primary and secondary air pollution that can threaten human health, damage ecosystems and influence climate (Sharma et al. 2010; Nagurney et al. 2010). Traffic patterns, vehicle characteristics, and street configurations have a cumulative effect on exhaust emissions (Pandian et al. 2009).

The combustion of hydrocarbon fuel in the air generates mainly carbon dioxide (CO2) and water (H2O). However, the combustion engines are not totally efficient, which means that the fuel is not totally burned. In this process the product of the combustion is more complex and could be constituted by hydrocarbons and other organic compounds as well as benzene (C6H6), carbon monoxide (CO) and particles (PM) that contain carbon and other pollutants. On the other hand, the combustion conditions - high pressures and temperatures - originate partial oxidation of the nitrogen present in the air and in the fuel, forming oxides of nitrogen (mainly nitric oxide and some nitrogen dioxides) conventionally designated by NOx.

Traffic-related air pollution levels can be evaluated by either directmeasurements or predictive models. The direct measurement method is only feasible for evaluating actual situations; predictive methods can be applied throughout theplanning process from the initial concept to the final detailed design of air pollutionabatement measures.However measurements provide essential information to validatethe predictive methods.

Numerous available dispersion models represent an important set of tools forsimulating air pollution scenarios. The model adopted for this research was developedby Cambridge Environmental Research Consultants (CERC) in the United Kingdom.

This model has been used by local authorities all over Europe for urban air qualityforecasting (Carruthers et al., 1997, 1998, 2003; Timmis et al, 2000; McHugh et al.,1997).It uses a parameterisation of boundary layer physics in terms ofboundary layer depth and Monin-Obukhov length, and it applies a skewed-Gaussianconcentration profile for convective meteorological conditions. For stable and neutralmeteorological conditions, the model assumes a Gaussian plume for the concentrationprofile distribution with reflection at the ground and in the inversion layer.

The dispersion model has a meteorological processor for input variables, which typically include day of the year, time of day, cloud cover, wind direction and speed and temperature. These variables are used to calculate model parameters such as boundary layer depth and Monin-Obukhov length. The model does not account for anthropogenic heat sources.

An additional and important feature that makes this dispersion model suitablefor modelling the urban environment is a chemistry scheme that facilitates thecalculation of chemical reactions between nitric oxide, nitrogen dioxide, ozone andvolatile organic compounds in the atmosphere.

2.1. Existing air quality evaluation models

The evaluation of air quality may be occasional or long-term. Occasional evaluation is useful in the context of information and alert systems for the population, working normally in real or almost-real time. Data is acquired through measurements made on an hourly or daily average basis and concentration episodes are evaluated and reported. When long-term data (6-month or yearly evaluations) is considered, than we talk about long-term trend analysis.

In the following subchapters two air quality evaluation models are referred, both working in real time: a Canadian and a Portuguese experience.

2.1.1. AQI,Canada

Integrated in a public information system of Vancouver, the FPCAP (Federal-Provincial Committee on Air Pollution) provides the information on pollution levels in form of an Air Quality Index (AQI). The AQI is based on measurements taken throughout the region of Greater Vancouver (Butterwick et al., 1991).

The AQI is expressed as a single value taking into consideration the concentrations of five major air pollutants (CO, NO2, O3, SO2, PM). The index is based on the pollutant with the highest concentration relative to Federal and Provincial air criteria. This pollutant is called the Index Pollutant. The values of the other four pollutants are then disregarded.

The numeric value of the Air Quality Index is correlated to a classification system. For each category of air quality, information is provided on the associated general health effects and recommended precautionary action. Table 1 summarizes this information.

AQIAir QualityGeneral Health EffectsCautionary Statements
0 – 25GoodNo measured effects are associatedNo precautions are necessary
26 – 50FairIs adequate protection against effects on general populationNo precautions are necessary
51 – 100PoorShort-term exposure may result in irritation or mild aggravation of symptoms in sensitive persons.Persons with heart or respiratory ailments should reduce physical action and outdoor activity
Over 100Very poorSignificant aggravation of persons with heart and lung disease. Many people may notice symptoms.Persons with respiratory and cardiovascular diseases should stay indoors and minimize physical activity.

Table 1.

Great Vancouver Air Quality Index.Source: (Butterwick et al., 1991)

2.1.2. QualAr, Portugal

The APA (Agência Portuguesa do Ambiente) of the Ministry of Environment of Portugal provides public information on pollution levels based on measurements taken through a pollution monitoring network. The information on pollutant levels is presented as an index called “Índice de QualidadedoAr” (QualAr) (APA, 2011). TheQualAr is based on 24–hour average concentrations, and therefore does not reflect short term peak levels.

The QualAr is expressed as a single value taking into consideration the concentrations of five major air pollutants (CO, NO2, O3, SO2, PM). The index is based on the pollutant with the highest concentration relative to the Portuguese annual limit values for the protection of human health. The values of the other four pollutants are then disregarded. The calculation of QualAr takes into account the following averages:

  • Nitrogen Dioxide (NO2) – hourly average

  • Sulphur Dioxide (SO2) – hourly average

  • Ozone (O3) – hourly average

  • Carbon Monoxide (CO) – 8-hour average

  • Suspended Particulates (PM10) – daily average

The air quality assumes the classification from Poor to Good according to a classification system summarized in the Table 2.

PollutantCO (μg/m3)NO2(μg/m3)O3(μg/m3)PM10(μg/m3)SO2(μg/m3)

Table 2.

Classification of QualAr for 2010.Source: (APA, 2010)

The classification of the air quality is based on the pollutant with the highest concentration relative to the Portuguese annual limit values for the protection of human health (Decreto-Lei 102/2010), [i.e. for an atmosphere with pollutants levels SO2 - 35 µg/m3 (very good), NO2 - 180 µg/m3 (fair); CO - 6000 µg/m3 (good), PM10 - 15 µg/m3 (very good) and O3 - 365 µg/m3 (very poor): Air Quality was Very Poor due to Ozone].

3. The cityAIR index

Both models presented above are approaches which prevent trade-off between pollutant concentrations because they are based on the pollutant with the highest concentration relative to the legal limits. For situations where the concentrations are below the legal limit, i.e. when there is no limit violation, a model integrating all the pollutants could offer a more complete evaluation of the air quality. Such a model requires that whenever at least one of the pollutants considered overcomes the legal limits for the concentration (or any other limit assumed for this purpose), this one will be the only relevant for the index calculation, and the value will be the minimum of the scale.

Amulticriteria air quality index is proposed, which allows for trade-off between pollutants whenever concentration values stay under the considered limits.

The cityAIR model proposed stands on the combination of long-term concentrations, which may result from past measurements or, differently, from mathematical simulation models providing in this case a prospective view of air quality.

When air pollution concentrations are computer-simulated for a city, the values for each point or area considered are compared to a standard (in this paper the legal limit).This comparisongenerates a dummy variable: zero if the standard is exceeded and one if it is not.

The cityAIR index results from the weighted linear combination of normalised concentration values, which are subjected to the product of the dummy variables (eqn 1).



w i is the relative weight of the pollutant i;

c i is the normalised concentration of the pollutant i;

v i is the dummy variable of the legal limit violation L i of pollutant i, defined as follows:

v i = 1 when c i L i

v i = 0 when c i >L i

The proposed model makes use of multi-criteria techniques for combining,aggregating and standardising pollutant concentration data.

3.1. Pollutants and weights

The selection of pollutants to be included in the cityAIR index may vary according to the type of sources or even the data availability. For the purpose of this paper we present the pollutants considered in the case study, which are typically result from road traffic:

CO: Carbon Monoxide

NO2: Nitrogen Dioxide

PM10: Particulate< 10 µm

C6H6: Benzene

O3: Ozone

Equal weights were considered, which means 0.2 for each of the five pollutants.

3.2. Normalization of concentrations

Because of the different scales upon which concentrations are measured, it is necessary to standardize them before aggregation. The process of standardisation is essentially identical to that of fuzzification in fuzzy sets. Standardisation is intended to transform any scale into a normalised range (i.e. zero to one). In our case, the results express a membership grade that ranges from 0.0 to 1.0, representing a continuous spectrumfrom non-membership (bad air quality) to complete membership (very good air quality), on the basis of the criterion (pollutant concentration) being fuzzified.

For the standardization a sigmoidal function has been adopted (eqn 2).




Where x is the concentration value being normalized, and x a and x b are control points in the function. Figures 1a to 1e present this function graphically for each of the five pollutants. The control points adopted (a and b) are listed in Table 3.


Figure 1.

Normalisation functions

Control points of the sigmoidal functions were selected according to the following criteria: score = 0 for the concentration limit values considered in the Portuguese legislation for human health protection and score = 1 for the concentration guidance values recommended by the World Health Organisation (WHO, 2005)(NO2 and CO values represented a non-polluted atmosphere (Seinfeld, 1997). Table 3 presents the adopted values.

PollutantsScore = 0Score = 1Averaging period
CO[CO] "/ 10.0 mg/m3 [CO] ≤ 0.140 mg/m3 8 hours (rollingaverage) for calendaryear
PM[PM10] "/ 40.0 μg/m3 [PM10] ≤ 20.0 μg/m3Calendar year
NO2 [NO2] "/ 40.0 μg/m3 [NO2] ≤ 20.0 µg/m3 Calendar year
O3 [O3] "/ 110.0 μg/m3 [O3] ≤ 100.0 µg/m3 8-hour average for calendar year
C6H6 [C6H6] "/ 5.0 μg/m3 [C6H6] ≤ 1.0 µg/m3 Calendaryear

Table 3.

Control points of the fuzzy functions

3.3. Dummy variables

Dummy variables switch from zero to one at the concentration limits mentioned above (third column of Table 3). Figures 2a to 2e show a graphical view of the dummy variable functions.


Figure 2.

Dummy variables; a) Dummy variable function of PM10; b) Dummy variable function of CO; c) Dummy variable function of NO2 ; d) Dummy variable function of C6H6 ; e) Dummy variable function of O3

4. Case Study: Airquality index of one mid-sized city

A case study was undertaken to evaluate urban environmental quality in thePortuguese city of Viana do Castelo, which is located on the north-western seaside.

This mid-sized city has a population of around 36,000 in an overall area of 37 km2.The most notable source of noise and air pollution is a main road (Avenida 25 deAbril) that crosses the city and divides it into two parts.

Based on traffic data and the physical characteristics of the area, horizontal concentration maps were created for five main pollutants:CO, NO2, C6H6, PM10 and O3. A range of numerical models were used to produceresults.

The ADMS-Urban model was used for pollutant dispersion. TheHills model was used to calculate air flow and turbulence over complex terrain and toaccount for the effects of variable surface roughness (CERC, 2001). The COPERT4model (COPERT4), which is based on CORINAIR v.5 (CORINAIR, 2006), was usedto estimate traffic emissions.

4.1. Air pollution of VianadoCastelo

The sources characterization data,and considering that Viana doCastelo is a touristic seaside city, two traffic counting campaigns were carried out,one in winter time and another one in summer time, of which resulted the data for twoscenarios. Each campaign included most of the city streets and traffic was countedround-the-clock in a typical week day.

Main and secondary roads weremodelled explicitly, as were one pulp and paper mill located in the vicinity ofthe city.The factory was modelled as one point source that represents the stack.

One singleprofile was developed to represent the hourly variation of traffics flows on all theroads.A full survey, including topographic characteristics, surface roughness and thespecification of the emission sources, cross and longitudinal profiles (for canyonroads) was carried out for the whole city.

4.2. Validation

There is no direct technique for determining if a model is good or bad because model performance depends on so many factors. These are related with model input data, model set-up parameters and model algorithms. Besides model performance depends on the averaging time for the pollutant concentration, the pollutant itself and the monitoring sites locations. Much research has gone into prepare acceptable validations techniques. The usually used BOOT statistics approach derives from that of Hanna and Paine (Hanna & Paine, 1989) and employs a series of statistical measures comprising the mean, correlation, normal mean square error and fractional bias. The methodology adopted was based in BOOT statistical approach.

For the validation process it was guaranteed the same meteorological conditions, the same geographical base and the same reading points (coordinates x, y, z). Pollutant concentrations are predicted for each hour of the monitoring period. For hours with inadequate met data predictions are not made and the corresponding measured values are neglected.

The following simplifications were assumed:

The same flow and composition of traffic and the same traffic daily profile in both periods (measurement and modelling period);

  • The validation process was developed at two levels:

  • averaging the data in order to obtain daily concentrations profiles, both for monitored and predicted data;

  • for each monitoring site comparison of the averaged daily concentrations profiles by the BOOT statistical methodology.

  • The pollutant used in the validating process was CO, a primary and typical road traffic pollutant.

Pollutants were measured at three monitoring sites in the city (Fig. 3) during the monitoring periods shown in Table 4.

Monitoring sitesMonitoring periods
A10h00 19.Jan. to 24h00 21.Jan
A20h00 23.Jan. to 24h00 25.Jan
A30h00 30.Jan. to 24h00 1.Feb

Table 4.

Monitoring periods

Point A1 is located at the Largo JoãoTomás da Costa, next to the River Garden. This site is particularly influenced by the road traffic that circulates near the garden.

Point A2 is located in the Campo do Castelo, a large square where some outdoor activities take place.

The third point, A3, is located in Rio Lima's Street, at the South edge of the City’s Park and close to highway A28.


Figure 3.

Automatic monitoring sites in Viana do Castelo

The statistics calculated include Average, Standard Deviation, Normalised Mean Square Error (NMSE), FractionalBias (FB) and the FAC2. The data format was hour by hour for the measured concentrations and predicted concentration. A perfect model would have FAC2=1.0, NMSE=0.0 and FB=0.0.

The Table 5 present statistics of comparisons between measured concentrations and the ADMS-Urban calculations. Statistics have been calculated based on hourly comparisons for each site (A1, A2 and A3) and for overall statistics (Ov.S.).

Figures 4 to 6 compare the predicted (ADMS-Urban calculations) and observed (measured) average daily concentrations of CO at the monitoring sites A1, A2 and A3.

Monitoring SitesAverageStandard deviationFAC2 (objective 1)NMSE (objective 0)FB (Objective 0)

Table 5.

Monitored and predicted CO concentrations (μg/m3)

The comparison of the output of ADMS-Urban with pollutant concentrations measured in the control points has confirmed the generally good performance of the model.

The variations of the mean concentrations along the day, shown in Figures 4 to 6, reveal a quite fair agreement between predicted and measured values.


Figure 4.

Temporal variation at A1


Figure 5.

Temporal variation at A2


Figure 6.

Temporal variation at A3

4.3. Air pollutant maps

Horizontal concentration maps were created using ADMS-Urban model.These maps represent the average atmospheric pollution situation in one year. Thefollowing calculation parameters were adopted:

  • Grid spacing:variable grid spacing (less than 10 meters);

  • Height of the map: 1.20 m;

  • Meteorological conditions:Data gathered at the automatic monitoring sites forone year (hourly);

  • Monin-Obukhov length: 30 m ;

  • Surface roughness: 0.5 m;

  • Emissions inventory: database prepared for Viana do Castelo including roadsources and industrial sources;

  • Background file: annual average background concentration of NO2, CO, PM10and O3 at background monitoring sites (Silva, 2008);

  • Output: hourly average CO [mg/m3], NO2[μg/m3],PM10[μg/m3], C6H6,O3[μg/m3];

  • Average speed: variable.

This article presents results for the summer scenario, the most critical (Figures 7a to 7e).


Figure 7.

Air pollution maps, summer scenario

4.4. CityAIRof VianadoCastelo

The combination of the concentration maps, according to eqn.(1), results in anoverall air qualitymap (Fig. 8).


Figure 8.

CityAIR, summer scenario

Model results were overlain with a population GIS layer to estimate theaffected population. Table 6 presents a synthesis of the areas and populations affectedby air pollution in the city.

hab%m2 %
= 0690.2%263320.2%
[0 ; 0.35[00.0%00.0%
[0.35 ; 0.65[90.0%32960.0%
[0.65; 0.85[2047771.7%515248447.3%
[0.85; 1.0]800228.0%571176852.4%

Table 6.

Areas and populations affected by air pollution

4.5. Analysis

The cityAIR index was developed to quantify city air quality. It depends on theconcentrations of the five major urban pollutants: CO, NO2, O3, C6H6 and PM10.When none of the concentrations of these pollutants exceed the legal limit, cityAIR iscalculated by combining the concentrations of different species. The cityAIR index iszero for areas with at least one pollutant concentration above the limit. Each area wasalso described as a binary data set, whereby a value of one signified that allconcentrations were below the limit and a value of zero signified a thresholdviolation.

In combination with a Geographic Information System platform, the modeland technologies used in this study proved to be useful for evaluating urbanenvironmental quality, allowingthe calculation of the areas andpopulations affected by air pollution in the city.

Although the inventory of air pollutant emission sources in the city of Vianado Castelo included one integrated pulp and paper mill, the road traffic sources madethe greatest contribution to air pollution in the city. A dispersion model was used tocalculate air pollution in a continuous space of urban pollutant concentrations in thecity. Horizontal maps were created for major air pollutants in urban areas (CO2, C6H6, CO, NO2, PM10 and O3). The results demonstrated that the highest concentrations ofprimary air pollutants were in areas adjacent to the major roads. The obtained resultsare in agreement with field measurements and expected values. The highest pollutantconcentrations were found in areas with a greater traffic flow or on roads with channelcharacteristics. Additionally, the weather conditions in the summer scenario wererelatively unfavourable for dispersion and natural pollutant removal.

The cityAIR index defines air quality within a range from zero (poor airquality) to one (good air quality). Applying this model to the city of Viana do Casteloproved very useful for comparing the concentrations of major air pollutant species totheir standards. This model generates a summary index of air quality that is easy tounderstand and intuitive for the general public.

Of the species studied, only NO2 was found to be above legal limits. Airpollution maps reveal that the concentrations of PM10, NO2, CO, CO2 and C6H6 arehigher in areas that are adjacent to high-traffic roads running through and around thecity.

Because ozone is a secondary pollutant, its highest concentrations are notfound near emission sources. Horizontal maps of O3 show that the maximum ozoneconcentrations do not exceed the legal limit. However, in outlying areas, ozoneconcentrations reach the threshold for vegetation protection (65 μg/m3).

The distribution of the calculated cityAIR index over Viana do Castelorevealed that air quality is globally acceptable in this city (Figure 8). Nonetheless, thedearth of small zones on this map may be problematic. Small zones, including Av. 25de Abril (the ramp to the bridge), the roundabouts of the hospital and football fieldand the access to the main road (IC1), have high levels of NO2 over the legal limit.

Based on an analysis of the Table 6, we can conclude that only 0.2% of thepopulation (69 pop.) in summer is exposed to a cityAIR index of zero, during whichtime 71.7% and 28.0% of the population benefit a cityAIR index of the populationbenefit a cityAIR index ranging between [0.65 ; 0.85[ and above 0.85, respectively.

5. Conclusion

The urban air index used in Viana do Castelo, cityAIR,aggregates data for the air quality of a city and presents results in thecontext of standardised legal limits for air pollution.

Based on cityAIR, several priority areas for future mitigation andmonitoring are proposed: Avenida 25 de Abril, Avenida Gaspar de Castro, access to west IC1 and access to north and south EN13.

When the results of this analysis were presented to the Municipality of Vianado CastelocityAIR index proved to be easily understood and quite intuitive.For the problematic areas, a Monitoring & Mitigation Plan is being prepared.

In order to estimate pollutant emissions to the atmosphere from vehiculartraffic, which is essential for evaluating both local and global air quality, emissionfactors were obtained for vehicles in Viana do Castelo and combined with thecalculated vehicle circulation in that city. Thus, it is now generally possible toestimate the contribution of car traffic to climate change due to CO2 emissions.

The model for air pollutant dispersion is consistent with field measurements.The model calculates concentrations of PM10, a component of secondary emissionthat originates under typical city conditions. The developed database of secondaryparticle concentrations in the city facilitates simulations of PM10 concentrations.

The cityAIRindex developed here is transparent, simple and easy tounderstand. Indicator weights and dimensions used in the cityAIR calculationdepend on the control points defined during the normalisation process. A variety ofoptions can be used during the cityAIR calculation to focus index results ondifferent dimensions and indicators of overall urban environmental quality.


This work was partially supported by Department of Civil Engineering, University ofMinho and the Municipality of Viana do Castelo.The authors would like to thank theanonymous reviewers for their valuable comments and suggestions to improve thequality of the paper.


1 - APA 2010 on line:
2 - L. Butterwick, Roy. e. Harrison, Merritt, 1991 Handbook for Urban AirImprovement. Commission of the EuropeanCommunities.
3 - D. J. Carruthers, H. A. Edmunds, M. Bennett, P. T. Woods, M. J. T. Milton, R. Robison, B. Y. Underwoon, C. J. Franklin, 1997 Validation of ADMS dispersionmodel and assessment of its performance relative to R-19 and ISC usingarchivedLIDAR data. Int. J. Environment and Pollution, 8 264-278.
4 - D. J. Carruthers, H. A. Edmunds, A. E. Lester, C. A. Mc Hugh, Singles, R. , 1998 Use and Validation of ADMS-Urban in ContrastingUrban and IndustrialLocations. Int. J. Environment and Pollution, 14 364-374.
5 - D. J. Carruthers, J. W. Blair, K. L. Johnson, 2003 Validation and SensivityStudy ofthe ADMS-Urban, Cambridge EnvironmentalResearch Consultants (2003) TR-0191.
6 - CERC 2001 ADMS-Urban User Guide, Version 1.6.
7 - COPERT4, Acessed 2008 Availablefrom:
8 - CORINAIR 2006 Emission InventoryGuidebook, Version 5.0, September.
9 - Diário da República 2010 Dec. Lei nº102/10, 23 de Set. 2010. I Série-A, N.º 186, Lisboa.
10 - Hanna, Steven R., Robert J. Paine. 1989.Hybrid Plume Dispersion Model (HPDM) Development and Evaluation. J. Appl. Meteor., 28, 206-224.
11 - C. A. Mc Hugh, D. J. Carruthers, H. A. Edmunds, 1997 ADMS-Urban: an air qualitymanagement system for traffic, domestic and industrial pollution, InternationalJournal of Environment and Pollution, 666 675 .
12 - A. Nagurney, Q. Qiang, Nagurney, 2010 2010. Environmental Impact Assessment ofTransportation Networks withDegradable Links in an Era of Climate Change.International Journal of Sustainable Transportation, 4: 3, 154- 171.
13 - S. Pandian, S. Gokhale, A. Ghoshal, 2009 2009. Evaluatingeffects of traffic and vehiclecharacteristics on vehicularemissionsneartraffic intersections. TransportationResearch Part D 14 180 196 .
14 - J. H. Seinfeld, S. Padis, 1997 1997. AtmosphericChemistry and PhysicsFrom AirPollution To Climate Change, John Wiley Sons.
15 - A. R. Sharma, S. K. Kharol, K. V. S. Badarinath, 2010 Influence of vehiculartraffic onurban air quality- A case study of Hyderabad, India. Transp.Research Part D15 154 159 .
16 - L. T. Silva, 2008 Avaliação da QualidadeAmbiental Urbana. PhDThesis,Universidade do Minho, Braga, Portugal.
17 - R. Timmis, S. Wilkinson, D. Carruthers, Mc Hugh, 2000 2000. Recentstudies tovalidate and compare atmospheric dispersion models for regulatorypurposes in theU.K., International Journal of Environment and Pollution 14 431 442 .
18 - WHO 2005 Air quality guidelines global update 2005 meeting report. Report on aworking group meeting. WHO, Bonn, Germany.